| 1 | use crate::{ |
| 2 | core_arch::{simd::*, x86::*}, |
| 3 | intrinsics::simd::*, |
| 4 | ptr, |
| 5 | }; |
| 6 | |
| 7 | #[cfg (test)] |
| 8 | use stdarch_test::assert_instr; |
| 9 | |
| 10 | /// Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst. |
| 11 | /// |
| 12 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_abs_epi16&expand=30) |
| 13 | #[inline ] |
| 14 | #[target_feature (enable = "avx512bw" )] |
| 15 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 16 | #[cfg_attr (test, assert_instr(vpabsw))] |
| 17 | pub fn _mm512_abs_epi16(a: __m512i) -> __m512i { |
| 18 | unsafe { |
| 19 | let a: i16x32 = a.as_i16x32(); |
| 20 | let cmp: i16x32 = simd_gt(x:a, y:i16x32::ZERO); |
| 21 | transmute(src:simd_select(mask:cmp, if_true:a, if_false:simd_neg(a))) |
| 22 | } |
| 23 | } |
| 24 | |
| 25 | /// Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 26 | /// |
| 27 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_abs_epi16&expand=31) |
| 28 | #[inline ] |
| 29 | #[target_feature (enable = "avx512bw" )] |
| 30 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 31 | #[cfg_attr (test, assert_instr(vpabsw))] |
| 32 | pub fn _mm512_mask_abs_epi16(src: __m512i, k: __mmask32, a: __m512i) -> __m512i { |
| 33 | unsafe { |
| 34 | let abs: i16x32 = _mm512_abs_epi16(a).as_i16x32(); |
| 35 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:src.as_i16x32())) |
| 36 | } |
| 37 | } |
| 38 | |
| 39 | /// Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 40 | /// |
| 41 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_abs_epi16&expand=32) |
| 42 | #[inline ] |
| 43 | #[target_feature (enable = "avx512bw" )] |
| 44 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 45 | #[cfg_attr (test, assert_instr(vpabsw))] |
| 46 | pub fn _mm512_maskz_abs_epi16(k: __mmask32, a: __m512i) -> __m512i { |
| 47 | unsafe { |
| 48 | let abs: i16x32 = _mm512_abs_epi16(a).as_i16x32(); |
| 49 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:i16x32::ZERO)) |
| 50 | } |
| 51 | } |
| 52 | |
| 53 | /// Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 54 | /// |
| 55 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_abs_epi16&expand=28) |
| 56 | #[inline ] |
| 57 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 58 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 59 | #[cfg_attr (test, assert_instr(vpabsw))] |
| 60 | pub fn _mm256_mask_abs_epi16(src: __m256i, k: __mmask16, a: __m256i) -> __m256i { |
| 61 | unsafe { |
| 62 | let abs: i16x16 = _mm256_abs_epi16(a).as_i16x16(); |
| 63 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:src.as_i16x16())) |
| 64 | } |
| 65 | } |
| 66 | |
| 67 | /// Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 68 | /// |
| 69 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_abs_epi16&expand=29) |
| 70 | #[inline ] |
| 71 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 72 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 73 | #[cfg_attr (test, assert_instr(vpabsw))] |
| 74 | pub fn _mm256_maskz_abs_epi16(k: __mmask16, a: __m256i) -> __m256i { |
| 75 | unsafe { |
| 76 | let abs: i16x16 = _mm256_abs_epi16(a).as_i16x16(); |
| 77 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:i16x16::ZERO)) |
| 78 | } |
| 79 | } |
| 80 | |
| 81 | /// Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 82 | /// |
| 83 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_abs_epi16&expand=25) |
| 84 | #[inline ] |
| 85 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 86 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 87 | #[cfg_attr (test, assert_instr(vpabsw))] |
| 88 | pub fn _mm_mask_abs_epi16(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 89 | unsafe { |
| 90 | let abs: i16x8 = _mm_abs_epi16(a).as_i16x8(); |
| 91 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:src.as_i16x8())) |
| 92 | } |
| 93 | } |
| 94 | |
| 95 | /// Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 96 | /// |
| 97 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_abs_epi16&expand=26) |
| 98 | #[inline ] |
| 99 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 100 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 101 | #[cfg_attr (test, assert_instr(vpabsw))] |
| 102 | pub fn _mm_maskz_abs_epi16(k: __mmask8, a: __m128i) -> __m128i { |
| 103 | unsafe { |
| 104 | let abs: i16x8 = _mm_abs_epi16(a).as_i16x8(); |
| 105 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:i16x8::ZERO)) |
| 106 | } |
| 107 | } |
| 108 | |
| 109 | /// Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst. |
| 110 | /// |
| 111 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_abs_epi8&expand=57) |
| 112 | #[inline ] |
| 113 | #[target_feature (enable = "avx512bw" )] |
| 114 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 115 | #[cfg_attr (test, assert_instr(vpabsb))] |
| 116 | pub fn _mm512_abs_epi8(a: __m512i) -> __m512i { |
| 117 | unsafe { |
| 118 | let a: i8x64 = a.as_i8x64(); |
| 119 | let cmp: i8x64 = simd_gt(x:a, y:i8x64::ZERO); |
| 120 | transmute(src:simd_select(mask:cmp, if_true:a, if_false:simd_neg(a))) |
| 121 | } |
| 122 | } |
| 123 | |
| 124 | /// Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 125 | /// |
| 126 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_abs_epi8&expand=58) |
| 127 | #[inline ] |
| 128 | #[target_feature (enable = "avx512bw" )] |
| 129 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 130 | #[cfg_attr (test, assert_instr(vpabsb))] |
| 131 | pub fn _mm512_mask_abs_epi8(src: __m512i, k: __mmask64, a: __m512i) -> __m512i { |
| 132 | unsafe { |
| 133 | let abs: i8x64 = _mm512_abs_epi8(a).as_i8x64(); |
| 134 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:src.as_i8x64())) |
| 135 | } |
| 136 | } |
| 137 | |
| 138 | /// Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 139 | /// |
| 140 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_abs_epi8&expand=59) |
| 141 | #[inline ] |
| 142 | #[target_feature (enable = "avx512bw" )] |
| 143 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 144 | #[cfg_attr (test, assert_instr(vpabsb))] |
| 145 | pub fn _mm512_maskz_abs_epi8(k: __mmask64, a: __m512i) -> __m512i { |
| 146 | unsafe { |
| 147 | let abs: i8x64 = _mm512_abs_epi8(a).as_i8x64(); |
| 148 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:i8x64::ZERO)) |
| 149 | } |
| 150 | } |
| 151 | |
| 152 | /// Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 153 | /// |
| 154 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_abs_epi8&expand=55) |
| 155 | #[inline ] |
| 156 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 157 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 158 | #[cfg_attr (test, assert_instr(vpabsb))] |
| 159 | pub fn _mm256_mask_abs_epi8(src: __m256i, k: __mmask32, a: __m256i) -> __m256i { |
| 160 | unsafe { |
| 161 | let abs: i8x32 = _mm256_abs_epi8(a).as_i8x32(); |
| 162 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:src.as_i8x32())) |
| 163 | } |
| 164 | } |
| 165 | |
| 166 | /// Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 167 | /// |
| 168 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_abs_epi8&expand=56) |
| 169 | #[inline ] |
| 170 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 171 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 172 | #[cfg_attr (test, assert_instr(vpabsb))] |
| 173 | pub fn _mm256_maskz_abs_epi8(k: __mmask32, a: __m256i) -> __m256i { |
| 174 | unsafe { |
| 175 | let abs: i8x32 = _mm256_abs_epi8(a).as_i8x32(); |
| 176 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:i8x32::ZERO)) |
| 177 | } |
| 178 | } |
| 179 | |
| 180 | /// Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set) |
| 181 | /// |
| 182 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_abs_epi8&expand=52) |
| 183 | #[inline ] |
| 184 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 185 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 186 | #[cfg_attr (test, assert_instr(vpabsb))] |
| 187 | pub fn _mm_mask_abs_epi8(src: __m128i, k: __mmask16, a: __m128i) -> __m128i { |
| 188 | unsafe { |
| 189 | let abs: i8x16 = _mm_abs_epi8(a).as_i8x16(); |
| 190 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:src.as_i8x16())) |
| 191 | } |
| 192 | } |
| 193 | |
| 194 | /// Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 195 | /// |
| 196 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_abs_epi8&expand=53) |
| 197 | #[inline ] |
| 198 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 199 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 200 | #[cfg_attr (test, assert_instr(vpabsb))] |
| 201 | pub fn _mm_maskz_abs_epi8(k: __mmask16, a: __m128i) -> __m128i { |
| 202 | unsafe { |
| 203 | let abs: i8x16 = _mm_abs_epi8(a).as_i8x16(); |
| 204 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:i8x16::ZERO)) |
| 205 | } |
| 206 | } |
| 207 | |
| 208 | /// Add packed 16-bit integers in a and b, and store the results in dst. |
| 209 | /// |
| 210 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_add_epi16&expand=91) |
| 211 | #[inline ] |
| 212 | #[target_feature (enable = "avx512bw" )] |
| 213 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 214 | #[cfg_attr (test, assert_instr(vpaddw))] |
| 215 | pub fn _mm512_add_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 216 | unsafe { transmute(src:simd_add(x:a.as_i16x32(), y:b.as_i16x32())) } |
| 217 | } |
| 218 | |
| 219 | /// Add packed 16-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 220 | /// |
| 221 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_add_epi16&expand=92) |
| 222 | #[inline ] |
| 223 | #[target_feature (enable = "avx512bw" )] |
| 224 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 225 | #[cfg_attr (test, assert_instr(vpaddw))] |
| 226 | pub fn _mm512_mask_add_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 227 | unsafe { |
| 228 | let add: i16x32 = _mm512_add_epi16(a, b).as_i16x32(); |
| 229 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i16x32())) |
| 230 | } |
| 231 | } |
| 232 | |
| 233 | /// Add packed 16-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 234 | /// |
| 235 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_add_epi16&expand=93) |
| 236 | #[inline ] |
| 237 | #[target_feature (enable = "avx512bw" )] |
| 238 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 239 | #[cfg_attr (test, assert_instr(vpaddw))] |
| 240 | pub fn _mm512_maskz_add_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 241 | unsafe { |
| 242 | let add: i16x32 = _mm512_add_epi16(a, b).as_i16x32(); |
| 243 | transmute(src:simd_select_bitmask(m:k, yes:add, no:i16x32::ZERO)) |
| 244 | } |
| 245 | } |
| 246 | |
| 247 | /// Add packed 16-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 248 | /// |
| 249 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_add_epi16&expand=89) |
| 250 | #[inline ] |
| 251 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 252 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 253 | #[cfg_attr (test, assert_instr(vpaddw))] |
| 254 | pub fn _mm256_mask_add_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 255 | unsafe { |
| 256 | let add: i16x16 = _mm256_add_epi16(a, b).as_i16x16(); |
| 257 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i16x16())) |
| 258 | } |
| 259 | } |
| 260 | |
| 261 | /// Add packed 16-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 262 | /// |
| 263 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_add_epi16&expand=90) |
| 264 | #[inline ] |
| 265 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 266 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 267 | #[cfg_attr (test, assert_instr(vpaddw))] |
| 268 | pub fn _mm256_maskz_add_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 269 | unsafe { |
| 270 | let add: i16x16 = _mm256_add_epi16(a, b).as_i16x16(); |
| 271 | transmute(src:simd_select_bitmask(m:k, yes:add, no:i16x16::ZERO)) |
| 272 | } |
| 273 | } |
| 274 | |
| 275 | /// Add packed 16-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 276 | /// |
| 277 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_add_epi16&expand=86) |
| 278 | #[inline ] |
| 279 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 280 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 281 | #[cfg_attr (test, assert_instr(vpaddw))] |
| 282 | pub fn _mm_mask_add_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 283 | unsafe { |
| 284 | let add: i16x8 = _mm_add_epi16(a, b).as_i16x8(); |
| 285 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i16x8())) |
| 286 | } |
| 287 | } |
| 288 | |
| 289 | /// Add packed 16-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 290 | /// |
| 291 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_add_epi16&expand=87) |
| 292 | #[inline ] |
| 293 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 294 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 295 | #[cfg_attr (test, assert_instr(vpaddw))] |
| 296 | pub fn _mm_maskz_add_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 297 | unsafe { |
| 298 | let add: i16x8 = _mm_add_epi16(a, b).as_i16x8(); |
| 299 | transmute(src:simd_select_bitmask(m:k, yes:add, no:i16x8::ZERO)) |
| 300 | } |
| 301 | } |
| 302 | |
| 303 | /// Add packed 8-bit integers in a and b, and store the results in dst. |
| 304 | /// |
| 305 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_add_epi8&expand=118) |
| 306 | #[inline ] |
| 307 | #[target_feature (enable = "avx512bw" )] |
| 308 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 309 | #[cfg_attr (test, assert_instr(vpaddb))] |
| 310 | pub fn _mm512_add_epi8(a: __m512i, b: __m512i) -> __m512i { |
| 311 | unsafe { transmute(src:simd_add(x:a.as_i8x64(), y:b.as_i8x64())) } |
| 312 | } |
| 313 | |
| 314 | /// Add packed 8-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 315 | /// |
| 316 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_add_epi8&expand=119) |
| 317 | #[inline ] |
| 318 | #[target_feature (enable = "avx512bw" )] |
| 319 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 320 | #[cfg_attr (test, assert_instr(vpaddb))] |
| 321 | pub fn _mm512_mask_add_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 322 | unsafe { |
| 323 | let add: i8x64 = _mm512_add_epi8(a, b).as_i8x64(); |
| 324 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i8x64())) |
| 325 | } |
| 326 | } |
| 327 | |
| 328 | /// Add packed 8-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 329 | /// |
| 330 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_add_epi8&expand=120) |
| 331 | #[inline ] |
| 332 | #[target_feature (enable = "avx512bw" )] |
| 333 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 334 | #[cfg_attr (test, assert_instr(vpaddb))] |
| 335 | pub fn _mm512_maskz_add_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 336 | unsafe { |
| 337 | let add: i8x64 = _mm512_add_epi8(a, b).as_i8x64(); |
| 338 | transmute(src:simd_select_bitmask(m:k, yes:add, no:i8x64::ZERO)) |
| 339 | } |
| 340 | } |
| 341 | |
| 342 | /// Add packed 8-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 343 | /// |
| 344 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_add_epi8&expand=116) |
| 345 | #[inline ] |
| 346 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 347 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 348 | #[cfg_attr (test, assert_instr(vpaddb))] |
| 349 | pub fn _mm256_mask_add_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 350 | unsafe { |
| 351 | let add: i8x32 = _mm256_add_epi8(a, b).as_i8x32(); |
| 352 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i8x32())) |
| 353 | } |
| 354 | } |
| 355 | |
| 356 | /// Add packed 8-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 357 | /// |
| 358 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_add_epi8&expand=117) |
| 359 | #[inline ] |
| 360 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 361 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 362 | #[cfg_attr (test, assert_instr(vpaddb))] |
| 363 | pub fn _mm256_maskz_add_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 364 | unsafe { |
| 365 | let add: i8x32 = _mm256_add_epi8(a, b).as_i8x32(); |
| 366 | transmute(src:simd_select_bitmask(m:k, yes:add, no:i8x32::ZERO)) |
| 367 | } |
| 368 | } |
| 369 | |
| 370 | /// Add packed 8-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 371 | /// |
| 372 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_add_epi8&expand=113) |
| 373 | #[inline ] |
| 374 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 375 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 376 | #[cfg_attr (test, assert_instr(vpaddb))] |
| 377 | pub fn _mm_mask_add_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 378 | unsafe { |
| 379 | let add: i8x16 = _mm_add_epi8(a, b).as_i8x16(); |
| 380 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i8x16())) |
| 381 | } |
| 382 | } |
| 383 | |
| 384 | /// Add packed 8-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 385 | /// |
| 386 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_add_epi8&expand=114) |
| 387 | #[inline ] |
| 388 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 389 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 390 | #[cfg_attr (test, assert_instr(vpaddb))] |
| 391 | pub fn _mm_maskz_add_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 392 | unsafe { |
| 393 | let add: i8x16 = _mm_add_epi8(a, b).as_i8x16(); |
| 394 | transmute(src:simd_select_bitmask(m:k, yes:add, no:i8x16::ZERO)) |
| 395 | } |
| 396 | } |
| 397 | |
| 398 | /// Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst. |
| 399 | /// |
| 400 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_adds_epu16&expand=197) |
| 401 | #[inline ] |
| 402 | #[target_feature (enable = "avx512bw" )] |
| 403 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 404 | #[cfg_attr (test, assert_instr(vpaddusw))] |
| 405 | pub fn _mm512_adds_epu16(a: __m512i, b: __m512i) -> __m512i { |
| 406 | unsafe { transmute(src:simd_saturating_add(x:a.as_u16x32(), y:b.as_u16x32())) } |
| 407 | } |
| 408 | |
| 409 | /// Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 410 | /// |
| 411 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_adds_epu16&expand=198) |
| 412 | #[inline ] |
| 413 | #[target_feature (enable = "avx512bw" )] |
| 414 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 415 | #[cfg_attr (test, assert_instr(vpaddusw))] |
| 416 | pub fn _mm512_mask_adds_epu16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 417 | unsafe { |
| 418 | let add: u16x32 = _mm512_adds_epu16(a, b).as_u16x32(); |
| 419 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_u16x32())) |
| 420 | } |
| 421 | } |
| 422 | |
| 423 | /// Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 424 | /// |
| 425 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_adds_epu16&expand=199) |
| 426 | #[inline ] |
| 427 | #[target_feature (enable = "avx512bw" )] |
| 428 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 429 | #[cfg_attr (test, assert_instr(vpaddusw))] |
| 430 | pub fn _mm512_maskz_adds_epu16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 431 | unsafe { |
| 432 | let add: u16x32 = _mm512_adds_epu16(a, b).as_u16x32(); |
| 433 | transmute(src:simd_select_bitmask(m:k, yes:add, no:u16x32::ZERO)) |
| 434 | } |
| 435 | } |
| 436 | |
| 437 | /// Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 438 | /// |
| 439 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_adds_epu16&expand=195) |
| 440 | #[inline ] |
| 441 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 442 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 443 | #[cfg_attr (test, assert_instr(vpaddusw))] |
| 444 | pub fn _mm256_mask_adds_epu16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 445 | unsafe { |
| 446 | let add: u16x16 = _mm256_adds_epu16(a, b).as_u16x16(); |
| 447 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_u16x16())) |
| 448 | } |
| 449 | } |
| 450 | |
| 451 | /// Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 452 | /// |
| 453 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_adds_epu16&expand=196) |
| 454 | #[inline ] |
| 455 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 456 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 457 | #[cfg_attr (test, assert_instr(vpaddusw))] |
| 458 | pub fn _mm256_maskz_adds_epu16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 459 | unsafe { |
| 460 | let add: u16x16 = _mm256_adds_epu16(a, b).as_u16x16(); |
| 461 | transmute(src:simd_select_bitmask(m:k, yes:add, no:u16x16::ZERO)) |
| 462 | } |
| 463 | } |
| 464 | |
| 465 | /// Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 466 | /// |
| 467 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_adds_epu16&expand=192) |
| 468 | #[inline ] |
| 469 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 470 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 471 | #[cfg_attr (test, assert_instr(vpaddusw))] |
| 472 | pub fn _mm_mask_adds_epu16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 473 | unsafe { |
| 474 | let add: u16x8 = _mm_adds_epu16(a, b).as_u16x8(); |
| 475 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_u16x8())) |
| 476 | } |
| 477 | } |
| 478 | |
| 479 | /// Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 480 | /// |
| 481 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_adds_epu16&expand=193) |
| 482 | #[inline ] |
| 483 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 484 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 485 | #[cfg_attr (test, assert_instr(vpaddusw))] |
| 486 | pub fn _mm_maskz_adds_epu16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 487 | unsafe { |
| 488 | let add: u16x8 = _mm_adds_epu16(a, b).as_u16x8(); |
| 489 | transmute(src:simd_select_bitmask(m:k, yes:add, no:u16x8::ZERO)) |
| 490 | } |
| 491 | } |
| 492 | |
| 493 | /// Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst. |
| 494 | /// |
| 495 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_adds_epu8&expand=206) |
| 496 | #[inline ] |
| 497 | #[target_feature (enable = "avx512bw" )] |
| 498 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 499 | #[cfg_attr (test, assert_instr(vpaddusb))] |
| 500 | pub fn _mm512_adds_epu8(a: __m512i, b: __m512i) -> __m512i { |
| 501 | unsafe { transmute(src:simd_saturating_add(x:a.as_u8x64(), y:b.as_u8x64())) } |
| 502 | } |
| 503 | |
| 504 | /// Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 505 | /// |
| 506 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_adds_epu8&expand=207) |
| 507 | #[inline ] |
| 508 | #[target_feature (enable = "avx512bw" )] |
| 509 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 510 | #[cfg_attr (test, assert_instr(vpaddusb))] |
| 511 | pub fn _mm512_mask_adds_epu8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 512 | unsafe { |
| 513 | let add: u8x64 = _mm512_adds_epu8(a, b).as_u8x64(); |
| 514 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_u8x64())) |
| 515 | } |
| 516 | } |
| 517 | |
| 518 | /// Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 519 | /// |
| 520 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_adds_epu8&expand=208) |
| 521 | #[inline ] |
| 522 | #[target_feature (enable = "avx512bw" )] |
| 523 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 524 | #[cfg_attr (test, assert_instr(vpaddusb))] |
| 525 | pub fn _mm512_maskz_adds_epu8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 526 | unsafe { |
| 527 | let add: u8x64 = _mm512_adds_epu8(a, b).as_u8x64(); |
| 528 | transmute(src:simd_select_bitmask(m:k, yes:add, no:u8x64::ZERO)) |
| 529 | } |
| 530 | } |
| 531 | |
| 532 | /// Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 533 | /// |
| 534 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_adds_epu8&expand=204) |
| 535 | #[inline ] |
| 536 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 537 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 538 | #[cfg_attr (test, assert_instr(vpaddusb))] |
| 539 | pub fn _mm256_mask_adds_epu8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 540 | unsafe { |
| 541 | let add: u8x32 = _mm256_adds_epu8(a, b).as_u8x32(); |
| 542 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_u8x32())) |
| 543 | } |
| 544 | } |
| 545 | |
| 546 | /// Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 547 | /// |
| 548 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_adds_epu8&expand=205) |
| 549 | #[inline ] |
| 550 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 551 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 552 | #[cfg_attr (test, assert_instr(vpaddusb))] |
| 553 | pub fn _mm256_maskz_adds_epu8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 554 | unsafe { |
| 555 | let add: u8x32 = _mm256_adds_epu8(a, b).as_u8x32(); |
| 556 | transmute(src:simd_select_bitmask(m:k, yes:add, no:u8x32::ZERO)) |
| 557 | } |
| 558 | } |
| 559 | |
| 560 | /// Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 561 | /// |
| 562 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_adds_epu8&expand=201) |
| 563 | #[inline ] |
| 564 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 565 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 566 | #[cfg_attr (test, assert_instr(vpaddusb))] |
| 567 | pub fn _mm_mask_adds_epu8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 568 | unsafe { |
| 569 | let add: u8x16 = _mm_adds_epu8(a, b).as_u8x16(); |
| 570 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_u8x16())) |
| 571 | } |
| 572 | } |
| 573 | |
| 574 | /// Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 575 | /// |
| 576 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_adds_epu8&expand=202) |
| 577 | #[inline ] |
| 578 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 579 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 580 | #[cfg_attr (test, assert_instr(vpaddusb))] |
| 581 | pub fn _mm_maskz_adds_epu8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 582 | unsafe { |
| 583 | let add: u8x16 = _mm_adds_epu8(a, b).as_u8x16(); |
| 584 | transmute(src:simd_select_bitmask(m:k, yes:add, no:u8x16::ZERO)) |
| 585 | } |
| 586 | } |
| 587 | |
| 588 | /// Add packed signed 16-bit integers in a and b using saturation, and store the results in dst. |
| 589 | /// |
| 590 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_adds_epi16&expand=179) |
| 591 | #[inline ] |
| 592 | #[target_feature (enable = "avx512bw" )] |
| 593 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 594 | #[cfg_attr (test, assert_instr(vpaddsw))] |
| 595 | pub fn _mm512_adds_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 596 | unsafe { transmute(src:simd_saturating_add(x:a.as_i16x32(), y:b.as_i16x32())) } |
| 597 | } |
| 598 | |
| 599 | /// Add packed signed 16-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 600 | /// |
| 601 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_adds_epi16&expand=180) |
| 602 | #[inline ] |
| 603 | #[target_feature (enable = "avx512bw" )] |
| 604 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 605 | #[cfg_attr (test, assert_instr(vpaddsw))] |
| 606 | pub fn _mm512_mask_adds_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 607 | unsafe { |
| 608 | let add: i16x32 = _mm512_adds_epi16(a, b).as_i16x32(); |
| 609 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i16x32())) |
| 610 | } |
| 611 | } |
| 612 | |
| 613 | /// Add packed signed 16-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 614 | /// |
| 615 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_adds_epi16&expand=181) |
| 616 | #[inline ] |
| 617 | #[target_feature (enable = "avx512bw" )] |
| 618 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 619 | #[cfg_attr (test, assert_instr(vpaddsw))] |
| 620 | pub fn _mm512_maskz_adds_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 621 | unsafe { |
| 622 | let add: i16x32 = _mm512_adds_epi16(a, b).as_i16x32(); |
| 623 | transmute(src:simd_select_bitmask(m:k, yes:add, no:i16x32::ZERO)) |
| 624 | } |
| 625 | } |
| 626 | |
| 627 | /// Add packed signed 16-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 628 | /// |
| 629 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_adds_epi16&expand=177) |
| 630 | #[inline ] |
| 631 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 632 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 633 | #[cfg_attr (test, assert_instr(vpaddsw))] |
| 634 | pub fn _mm256_mask_adds_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 635 | unsafe { |
| 636 | let add: i16x16 = _mm256_adds_epi16(a, b).as_i16x16(); |
| 637 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i16x16())) |
| 638 | } |
| 639 | } |
| 640 | |
| 641 | /// Add packed signed 16-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 642 | /// |
| 643 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_adds_epi16&expand=178) |
| 644 | #[inline ] |
| 645 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 646 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 647 | #[cfg_attr (test, assert_instr(vpaddsw))] |
| 648 | pub fn _mm256_maskz_adds_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 649 | unsafe { |
| 650 | let add: i16x16 = _mm256_adds_epi16(a, b).as_i16x16(); |
| 651 | transmute(src:simd_select_bitmask(m:k, yes:add, no:i16x16::ZERO)) |
| 652 | } |
| 653 | } |
| 654 | |
| 655 | /// Add packed signed 16-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 656 | /// |
| 657 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_adds_epi16&expand=174) |
| 658 | #[inline ] |
| 659 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 660 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 661 | #[cfg_attr (test, assert_instr(vpaddsw))] |
| 662 | pub fn _mm_mask_adds_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 663 | unsafe { |
| 664 | let add: i16x8 = _mm_adds_epi16(a, b).as_i16x8(); |
| 665 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i16x8())) |
| 666 | } |
| 667 | } |
| 668 | |
| 669 | /// Add packed signed 16-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 670 | /// |
| 671 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_adds_epi16&expand=175) |
| 672 | #[inline ] |
| 673 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 674 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 675 | #[cfg_attr (test, assert_instr(vpaddsw))] |
| 676 | pub fn _mm_maskz_adds_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 677 | unsafe { |
| 678 | let add: i16x8 = _mm_adds_epi16(a, b).as_i16x8(); |
| 679 | transmute(src:simd_select_bitmask(m:k, yes:add, no:i16x8::ZERO)) |
| 680 | } |
| 681 | } |
| 682 | |
| 683 | /// Add packed signed 8-bit integers in a and b using saturation, and store the results in dst. |
| 684 | /// |
| 685 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_adds_epi8&expand=188) |
| 686 | #[inline ] |
| 687 | #[target_feature (enable = "avx512bw" )] |
| 688 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 689 | #[cfg_attr (test, assert_instr(vpaddsb))] |
| 690 | pub fn _mm512_adds_epi8(a: __m512i, b: __m512i) -> __m512i { |
| 691 | unsafe { transmute(src:simd_saturating_add(x:a.as_i8x64(), y:b.as_i8x64())) } |
| 692 | } |
| 693 | |
| 694 | /// Add packed signed 8-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 695 | /// |
| 696 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_adds_epi8&expand=189) |
| 697 | #[inline ] |
| 698 | #[target_feature (enable = "avx512bw" )] |
| 699 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 700 | #[cfg_attr (test, assert_instr(vpaddsb))] |
| 701 | pub fn _mm512_mask_adds_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 702 | unsafe { |
| 703 | let add: i8x64 = _mm512_adds_epi8(a, b).as_i8x64(); |
| 704 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i8x64())) |
| 705 | } |
| 706 | } |
| 707 | |
| 708 | /// Add packed signed 8-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 709 | /// |
| 710 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_adds_epi8&expand=190) |
| 711 | #[inline ] |
| 712 | #[target_feature (enable = "avx512bw" )] |
| 713 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 714 | #[cfg_attr (test, assert_instr(vpaddsb))] |
| 715 | pub fn _mm512_maskz_adds_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 716 | unsafe { |
| 717 | let add: i8x64 = _mm512_adds_epi8(a, b).as_i8x64(); |
| 718 | transmute(src:simd_select_bitmask(m:k, yes:add, no:i8x64::ZERO)) |
| 719 | } |
| 720 | } |
| 721 | |
| 722 | /// Add packed signed 8-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 723 | /// |
| 724 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_adds_epi8&expand=186) |
| 725 | #[inline ] |
| 726 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 727 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 728 | #[cfg_attr (test, assert_instr(vpaddsb))] |
| 729 | pub fn _mm256_mask_adds_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 730 | unsafe { |
| 731 | let add: i8x32 = _mm256_adds_epi8(a, b).as_i8x32(); |
| 732 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i8x32())) |
| 733 | } |
| 734 | } |
| 735 | |
| 736 | /// Add packed signed 8-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 737 | /// |
| 738 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_adds_epi8&expand=187) |
| 739 | #[inline ] |
| 740 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 741 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 742 | #[cfg_attr (test, assert_instr(vpaddsb))] |
| 743 | pub fn _mm256_maskz_adds_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 744 | unsafe { |
| 745 | let add: i8x32 = _mm256_adds_epi8(a, b).as_i8x32(); |
| 746 | transmute(src:simd_select_bitmask(m:k, yes:add, no:i8x32::ZERO)) |
| 747 | } |
| 748 | } |
| 749 | |
| 750 | /// Add packed signed 8-bit integers in a and b using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 751 | /// |
| 752 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_adds_epi8&expand=183) |
| 753 | #[inline ] |
| 754 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 755 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 756 | #[cfg_attr (test, assert_instr(vpaddsb))] |
| 757 | pub fn _mm_mask_adds_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 758 | unsafe { |
| 759 | let add: i8x16 = _mm_adds_epi8(a, b).as_i8x16(); |
| 760 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i8x16())) |
| 761 | } |
| 762 | } |
| 763 | |
| 764 | /// Add packed signed 8-bit integers in a and b using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 765 | /// |
| 766 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_adds_epi8&expand=184) |
| 767 | #[inline ] |
| 768 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 769 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 770 | #[cfg_attr (test, assert_instr(vpaddsb))] |
| 771 | pub fn _mm_maskz_adds_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 772 | unsafe { |
| 773 | let add: i8x16 = _mm_adds_epi8(a, b).as_i8x16(); |
| 774 | transmute(src:simd_select_bitmask(m:k, yes:add, no:i8x16::ZERO)) |
| 775 | } |
| 776 | } |
| 777 | |
| 778 | /// Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst. |
| 779 | /// |
| 780 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_sub_epi16&expand=5685) |
| 781 | #[inline ] |
| 782 | #[target_feature (enable = "avx512bw" )] |
| 783 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 784 | #[cfg_attr (test, assert_instr(vpsubw))] |
| 785 | pub fn _mm512_sub_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 786 | unsafe { transmute(src:simd_sub(lhs:a.as_i16x32(), rhs:b.as_i16x32())) } |
| 787 | } |
| 788 | |
| 789 | /// Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 790 | /// |
| 791 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_sub_epi16&expand=5683) |
| 792 | #[inline ] |
| 793 | #[target_feature (enable = "avx512bw" )] |
| 794 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 795 | #[cfg_attr (test, assert_instr(vpsubw))] |
| 796 | pub fn _mm512_mask_sub_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 797 | unsafe { |
| 798 | let sub: i16x32 = _mm512_sub_epi16(a, b).as_i16x32(); |
| 799 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i16x32())) |
| 800 | } |
| 801 | } |
| 802 | |
| 803 | /// Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 804 | /// |
| 805 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_sub_epi16&expand=5684) |
| 806 | #[inline ] |
| 807 | #[target_feature (enable = "avx512bw" )] |
| 808 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 809 | #[cfg_attr (test, assert_instr(vpsubw))] |
| 810 | pub fn _mm512_maskz_sub_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 811 | unsafe { |
| 812 | let sub: i16x32 = _mm512_sub_epi16(a, b).as_i16x32(); |
| 813 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:i16x32::ZERO)) |
| 814 | } |
| 815 | } |
| 816 | |
| 817 | /// Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 818 | /// |
| 819 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_sub_epi16&expand=5680) |
| 820 | #[inline ] |
| 821 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 822 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 823 | #[cfg_attr (test, assert_instr(vpsubw))] |
| 824 | pub fn _mm256_mask_sub_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 825 | unsafe { |
| 826 | let sub: i16x16 = _mm256_sub_epi16(a, b).as_i16x16(); |
| 827 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i16x16())) |
| 828 | } |
| 829 | } |
| 830 | |
| 831 | /// Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 832 | /// |
| 833 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_sub_epi16&expand=5681) |
| 834 | #[inline ] |
| 835 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 836 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 837 | #[cfg_attr (test, assert_instr(vpsubw))] |
| 838 | pub fn _mm256_maskz_sub_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 839 | unsafe { |
| 840 | let sub: i16x16 = _mm256_sub_epi16(a, b).as_i16x16(); |
| 841 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:i16x16::ZERO)) |
| 842 | } |
| 843 | } |
| 844 | |
| 845 | /// Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 846 | /// |
| 847 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_sub_epi16&expand=5677) |
| 848 | #[inline ] |
| 849 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 850 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 851 | #[cfg_attr (test, assert_instr(vpsubw))] |
| 852 | pub fn _mm_mask_sub_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 853 | unsafe { |
| 854 | let sub: i16x8 = _mm_sub_epi16(a, b).as_i16x8(); |
| 855 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i16x8())) |
| 856 | } |
| 857 | } |
| 858 | |
| 859 | /// Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 860 | /// |
| 861 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_sub_epi16&expand=5678) |
| 862 | #[inline ] |
| 863 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 864 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 865 | #[cfg_attr (test, assert_instr(vpsubw))] |
| 866 | pub fn _mm_maskz_sub_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 867 | unsafe { |
| 868 | let sub: i16x8 = _mm_sub_epi16(a, b).as_i16x8(); |
| 869 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:i16x8::ZERO)) |
| 870 | } |
| 871 | } |
| 872 | |
| 873 | /// Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst. |
| 874 | /// |
| 875 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_sub_epi8&expand=5712) |
| 876 | #[inline ] |
| 877 | #[target_feature (enable = "avx512bw" )] |
| 878 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 879 | #[cfg_attr (test, assert_instr(vpsubb))] |
| 880 | pub fn _mm512_sub_epi8(a: __m512i, b: __m512i) -> __m512i { |
| 881 | unsafe { transmute(src:simd_sub(lhs:a.as_i8x64(), rhs:b.as_i8x64())) } |
| 882 | } |
| 883 | |
| 884 | /// Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 885 | /// |
| 886 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_sub_epi8&expand=5710) |
| 887 | #[inline ] |
| 888 | #[target_feature (enable = "avx512bw" )] |
| 889 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 890 | #[cfg_attr (test, assert_instr(vpsubb))] |
| 891 | pub fn _mm512_mask_sub_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 892 | unsafe { |
| 893 | let sub: i8x64 = _mm512_sub_epi8(a, b).as_i8x64(); |
| 894 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i8x64())) |
| 895 | } |
| 896 | } |
| 897 | |
| 898 | /// Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 899 | /// |
| 900 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_sub_epi8&expand=5711) |
| 901 | #[inline ] |
| 902 | #[target_feature (enable = "avx512bw" )] |
| 903 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 904 | #[cfg_attr (test, assert_instr(vpsubb))] |
| 905 | pub fn _mm512_maskz_sub_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 906 | unsafe { |
| 907 | let sub: i8x64 = _mm512_sub_epi8(a, b).as_i8x64(); |
| 908 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:i8x64::ZERO)) |
| 909 | } |
| 910 | } |
| 911 | |
| 912 | /// Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 913 | /// |
| 914 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_sub_epi8&expand=5707) |
| 915 | #[inline ] |
| 916 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 917 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 918 | #[cfg_attr (test, assert_instr(vpsubb))] |
| 919 | pub fn _mm256_mask_sub_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 920 | unsafe { |
| 921 | let sub: i8x32 = _mm256_sub_epi8(a, b).as_i8x32(); |
| 922 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i8x32())) |
| 923 | } |
| 924 | } |
| 925 | |
| 926 | /// Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 927 | /// |
| 928 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_sub_epi8&expand=5708) |
| 929 | #[inline ] |
| 930 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 931 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 932 | #[cfg_attr (test, assert_instr(vpsubb))] |
| 933 | pub fn _mm256_maskz_sub_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 934 | unsafe { |
| 935 | let sub: i8x32 = _mm256_sub_epi8(a, b).as_i8x32(); |
| 936 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:i8x32::ZERO)) |
| 937 | } |
| 938 | } |
| 939 | |
| 940 | /// Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 941 | /// |
| 942 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_sub_epi8&expand=5704) |
| 943 | #[inline ] |
| 944 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 945 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 946 | #[cfg_attr (test, assert_instr(vpsubb))] |
| 947 | pub fn _mm_mask_sub_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 948 | unsafe { |
| 949 | let sub: i8x16 = _mm_sub_epi8(a, b).as_i8x16(); |
| 950 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i8x16())) |
| 951 | } |
| 952 | } |
| 953 | |
| 954 | /// Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 955 | /// |
| 956 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_sub_epi8&expand=5705) |
| 957 | #[inline ] |
| 958 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 959 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 960 | #[cfg_attr (test, assert_instr(vpsubb))] |
| 961 | pub fn _mm_maskz_sub_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 962 | unsafe { |
| 963 | let sub: i8x16 = _mm_sub_epi8(a, b).as_i8x16(); |
| 964 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:i8x16::ZERO)) |
| 965 | } |
| 966 | } |
| 967 | |
| 968 | /// Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst. |
| 969 | /// |
| 970 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_subs_epu16&expand=5793) |
| 971 | #[inline ] |
| 972 | #[target_feature (enable = "avx512bw" )] |
| 973 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 974 | #[cfg_attr (test, assert_instr(vpsubusw))] |
| 975 | pub fn _mm512_subs_epu16(a: __m512i, b: __m512i) -> __m512i { |
| 976 | unsafe { transmute(src:simd_saturating_sub(lhs:a.as_u16x32(), rhs:b.as_u16x32())) } |
| 977 | } |
| 978 | |
| 979 | /// Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 980 | /// |
| 981 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_subs_epu16&expand=5791) |
| 982 | #[inline ] |
| 983 | #[target_feature (enable = "avx512bw" )] |
| 984 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 985 | #[cfg_attr (test, assert_instr(vpsubusw))] |
| 986 | pub fn _mm512_mask_subs_epu16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 987 | unsafe { |
| 988 | let sub: u16x32 = _mm512_subs_epu16(a, b).as_u16x32(); |
| 989 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_u16x32())) |
| 990 | } |
| 991 | } |
| 992 | |
| 993 | /// Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 994 | /// |
| 995 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_subs_epu16&expand=5792) |
| 996 | #[inline ] |
| 997 | #[target_feature (enable = "avx512bw" )] |
| 998 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 999 | #[cfg_attr (test, assert_instr(vpsubusw))] |
| 1000 | pub fn _mm512_maskz_subs_epu16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1001 | unsafe { |
| 1002 | let sub: u16x32 = _mm512_subs_epu16(a, b).as_u16x32(); |
| 1003 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:u16x32::ZERO)) |
| 1004 | } |
| 1005 | } |
| 1006 | |
| 1007 | /// Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1008 | /// |
| 1009 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_subs_epu16&expand=5788) |
| 1010 | #[inline ] |
| 1011 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1012 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1013 | #[cfg_attr (test, assert_instr(vpsubusw))] |
| 1014 | pub fn _mm256_mask_subs_epu16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1015 | unsafe { |
| 1016 | let sub: u16x16 = _mm256_subs_epu16(a, b).as_u16x16(); |
| 1017 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_u16x16())) |
| 1018 | } |
| 1019 | } |
| 1020 | |
| 1021 | /// Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1022 | /// |
| 1023 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_subs_epu16&expand=5789) |
| 1024 | #[inline ] |
| 1025 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1026 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1027 | #[cfg_attr (test, assert_instr(vpsubusw))] |
| 1028 | pub fn _mm256_maskz_subs_epu16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1029 | unsafe { |
| 1030 | let sub: u16x16 = _mm256_subs_epu16(a, b).as_u16x16(); |
| 1031 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:u16x16::ZERO)) |
| 1032 | } |
| 1033 | } |
| 1034 | |
| 1035 | /// Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1036 | /// |
| 1037 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_subs_epu16&expand=5785) |
| 1038 | #[inline ] |
| 1039 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1040 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1041 | #[cfg_attr (test, assert_instr(vpsubusw))] |
| 1042 | pub fn _mm_mask_subs_epu16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1043 | unsafe { |
| 1044 | let sub: u16x8 = _mm_subs_epu16(a, b).as_u16x8(); |
| 1045 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_u16x8())) |
| 1046 | } |
| 1047 | } |
| 1048 | |
| 1049 | /// Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1050 | /// |
| 1051 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_subs_epu16&expand=5786) |
| 1052 | #[inline ] |
| 1053 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1054 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1055 | #[cfg_attr (test, assert_instr(vpsubusw))] |
| 1056 | pub fn _mm_maskz_subs_epu16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1057 | unsafe { |
| 1058 | let sub: u16x8 = _mm_subs_epu16(a, b).as_u16x8(); |
| 1059 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:u16x8::ZERO)) |
| 1060 | } |
| 1061 | } |
| 1062 | |
| 1063 | /// Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst. |
| 1064 | /// |
| 1065 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_subs_epu8&expand=5802) |
| 1066 | #[inline ] |
| 1067 | #[target_feature (enable = "avx512bw" )] |
| 1068 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1069 | #[cfg_attr (test, assert_instr(vpsubusb))] |
| 1070 | pub fn _mm512_subs_epu8(a: __m512i, b: __m512i) -> __m512i { |
| 1071 | unsafe { transmute(src:simd_saturating_sub(lhs:a.as_u8x64(), rhs:b.as_u8x64())) } |
| 1072 | } |
| 1073 | |
| 1074 | /// Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1075 | /// |
| 1076 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_subs_epu8&expand=5800) |
| 1077 | #[inline ] |
| 1078 | #[target_feature (enable = "avx512bw" )] |
| 1079 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1080 | #[cfg_attr (test, assert_instr(vpsubusb))] |
| 1081 | pub fn _mm512_mask_subs_epu8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 1082 | unsafe { |
| 1083 | let sub: u8x64 = _mm512_subs_epu8(a, b).as_u8x64(); |
| 1084 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_u8x64())) |
| 1085 | } |
| 1086 | } |
| 1087 | |
| 1088 | /// Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1089 | /// |
| 1090 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_subs_epu8&expand=5801) |
| 1091 | #[inline ] |
| 1092 | #[target_feature (enable = "avx512bw" )] |
| 1093 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1094 | #[cfg_attr (test, assert_instr(vpsubusb))] |
| 1095 | pub fn _mm512_maskz_subs_epu8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 1096 | unsafe { |
| 1097 | let sub: u8x64 = _mm512_subs_epu8(a, b).as_u8x64(); |
| 1098 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:u8x64::ZERO)) |
| 1099 | } |
| 1100 | } |
| 1101 | |
| 1102 | /// Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1103 | /// |
| 1104 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_subs_epu8&expand=5797) |
| 1105 | #[inline ] |
| 1106 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1107 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1108 | #[cfg_attr (test, assert_instr(vpsubusb))] |
| 1109 | pub fn _mm256_mask_subs_epu8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 1110 | unsafe { |
| 1111 | let sub: u8x32 = _mm256_subs_epu8(a, b).as_u8x32(); |
| 1112 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_u8x32())) |
| 1113 | } |
| 1114 | } |
| 1115 | |
| 1116 | /// Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1117 | /// |
| 1118 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_subs_epu8&expand=5798) |
| 1119 | #[inline ] |
| 1120 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1121 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1122 | #[cfg_attr (test, assert_instr(vpsubusb))] |
| 1123 | pub fn _mm256_maskz_subs_epu8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 1124 | unsafe { |
| 1125 | let sub: u8x32 = _mm256_subs_epu8(a, b).as_u8x32(); |
| 1126 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:u8x32::ZERO)) |
| 1127 | } |
| 1128 | } |
| 1129 | |
| 1130 | /// Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1131 | /// |
| 1132 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_subs_epu8&expand=5794) |
| 1133 | #[inline ] |
| 1134 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1135 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1136 | #[cfg_attr (test, assert_instr(vpsubusb))] |
| 1137 | pub fn _mm_mask_subs_epu8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 1138 | unsafe { |
| 1139 | let sub: u8x16 = _mm_subs_epu8(a, b).as_u8x16(); |
| 1140 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_u8x16())) |
| 1141 | } |
| 1142 | } |
| 1143 | |
| 1144 | /// Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1145 | /// |
| 1146 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_subs_epu8&expand=5795) |
| 1147 | #[inline ] |
| 1148 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1149 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1150 | #[cfg_attr (test, assert_instr(vpsubusb))] |
| 1151 | pub fn _mm_maskz_subs_epu8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 1152 | unsafe { |
| 1153 | let sub: u8x16 = _mm_subs_epu8(a, b).as_u8x16(); |
| 1154 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:u8x16::ZERO)) |
| 1155 | } |
| 1156 | } |
| 1157 | |
| 1158 | /// Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst. |
| 1159 | /// |
| 1160 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_subs_epi16&expand=5775) |
| 1161 | #[inline ] |
| 1162 | #[target_feature (enable = "avx512bw" )] |
| 1163 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1164 | #[cfg_attr (test, assert_instr(vpsubsw))] |
| 1165 | pub fn _mm512_subs_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 1166 | unsafe { transmute(src:simd_saturating_sub(lhs:a.as_i16x32(), rhs:b.as_i16x32())) } |
| 1167 | } |
| 1168 | |
| 1169 | /// Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1170 | /// |
| 1171 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_subs_epi16&expand=5773) |
| 1172 | #[inline ] |
| 1173 | #[target_feature (enable = "avx512bw" )] |
| 1174 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1175 | #[cfg_attr (test, assert_instr(vpsubsw))] |
| 1176 | pub fn _mm512_mask_subs_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1177 | unsafe { |
| 1178 | let sub: i16x32 = _mm512_subs_epi16(a, b).as_i16x32(); |
| 1179 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i16x32())) |
| 1180 | } |
| 1181 | } |
| 1182 | |
| 1183 | /// Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1184 | /// |
| 1185 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_subs_epi16&expand=5774) |
| 1186 | #[inline ] |
| 1187 | #[target_feature (enable = "avx512bw" )] |
| 1188 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1189 | #[cfg_attr (test, assert_instr(vpsubsw))] |
| 1190 | pub fn _mm512_maskz_subs_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1191 | unsafe { |
| 1192 | let sub: i16x32 = _mm512_subs_epi16(a, b).as_i16x32(); |
| 1193 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:i16x32::ZERO)) |
| 1194 | } |
| 1195 | } |
| 1196 | |
| 1197 | /// Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1198 | /// |
| 1199 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_subs_epi16&expand=5770) |
| 1200 | #[inline ] |
| 1201 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1202 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1203 | #[cfg_attr (test, assert_instr(vpsubsw))] |
| 1204 | pub fn _mm256_mask_subs_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1205 | unsafe { |
| 1206 | let sub: i16x16 = _mm256_subs_epi16(a, b).as_i16x16(); |
| 1207 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i16x16())) |
| 1208 | } |
| 1209 | } |
| 1210 | |
| 1211 | /// Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1212 | /// |
| 1213 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_subs_epi16&expand=5771) |
| 1214 | #[inline ] |
| 1215 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1216 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1217 | #[cfg_attr (test, assert_instr(vpsubsw))] |
| 1218 | pub fn _mm256_maskz_subs_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1219 | unsafe { |
| 1220 | let sub: i16x16 = _mm256_subs_epi16(a, b).as_i16x16(); |
| 1221 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:i16x16::ZERO)) |
| 1222 | } |
| 1223 | } |
| 1224 | |
| 1225 | /// Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1226 | /// |
| 1227 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_subs_epi16&expand=5767) |
| 1228 | #[inline ] |
| 1229 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1230 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1231 | #[cfg_attr (test, assert_instr(vpsubsw))] |
| 1232 | pub fn _mm_mask_subs_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1233 | unsafe { |
| 1234 | let sub: i16x8 = _mm_subs_epi16(a, b).as_i16x8(); |
| 1235 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i16x8())) |
| 1236 | } |
| 1237 | } |
| 1238 | |
| 1239 | /// Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1240 | /// |
| 1241 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_subs_epi16&expand=5768) |
| 1242 | #[inline ] |
| 1243 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1244 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1245 | #[cfg_attr (test, assert_instr(vpsubsw))] |
| 1246 | pub fn _mm_maskz_subs_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1247 | unsafe { |
| 1248 | let sub: i16x8 = _mm_subs_epi16(a, b).as_i16x8(); |
| 1249 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:i16x8::ZERO)) |
| 1250 | } |
| 1251 | } |
| 1252 | |
| 1253 | /// Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst. |
| 1254 | /// |
| 1255 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_subs_epi8&expand=5784) |
| 1256 | #[inline ] |
| 1257 | #[target_feature (enable = "avx512bw" )] |
| 1258 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1259 | #[cfg_attr (test, assert_instr(vpsubsb))] |
| 1260 | pub fn _mm512_subs_epi8(a: __m512i, b: __m512i) -> __m512i { |
| 1261 | unsafe { transmute(src:simd_saturating_sub(lhs:a.as_i8x64(), rhs:b.as_i8x64())) } |
| 1262 | } |
| 1263 | |
| 1264 | /// Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1265 | /// |
| 1266 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_subs_epi8&expand=5782) |
| 1267 | #[inline ] |
| 1268 | #[target_feature (enable = "avx512bw" )] |
| 1269 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1270 | #[cfg_attr (test, assert_instr(vpsubsb))] |
| 1271 | pub fn _mm512_mask_subs_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 1272 | unsafe { |
| 1273 | let sub: i8x64 = _mm512_subs_epi8(a, b).as_i8x64(); |
| 1274 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i8x64())) |
| 1275 | } |
| 1276 | } |
| 1277 | |
| 1278 | /// Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1279 | /// |
| 1280 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_subs_epi8&expand=5783) |
| 1281 | #[inline ] |
| 1282 | #[target_feature (enable = "avx512bw" )] |
| 1283 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1284 | #[cfg_attr (test, assert_instr(vpsubsb))] |
| 1285 | pub fn _mm512_maskz_subs_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 1286 | unsafe { |
| 1287 | let sub: i8x64 = _mm512_subs_epi8(a, b).as_i8x64(); |
| 1288 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:i8x64::ZERO)) |
| 1289 | } |
| 1290 | } |
| 1291 | |
| 1292 | /// Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1293 | /// |
| 1294 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_subs_epi8&expand=5779) |
| 1295 | #[inline ] |
| 1296 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1297 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1298 | #[cfg_attr (test, assert_instr(vpsubsb))] |
| 1299 | pub fn _mm256_mask_subs_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 1300 | unsafe { |
| 1301 | let sub: i8x32 = _mm256_subs_epi8(a, b).as_i8x32(); |
| 1302 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i8x32())) |
| 1303 | } |
| 1304 | } |
| 1305 | |
| 1306 | /// Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1307 | /// |
| 1308 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_subs_epi8&expand=5780) |
| 1309 | #[inline ] |
| 1310 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1311 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1312 | #[cfg_attr (test, assert_instr(vpsubsb))] |
| 1313 | pub fn _mm256_maskz_subs_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 1314 | unsafe { |
| 1315 | let sub: i8x32 = _mm256_subs_epi8(a, b).as_i8x32(); |
| 1316 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:i8x32::ZERO)) |
| 1317 | } |
| 1318 | } |
| 1319 | |
| 1320 | /// Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1321 | /// |
| 1322 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_subs_epi8&expand=5776) |
| 1323 | #[inline ] |
| 1324 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1325 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1326 | #[cfg_attr (test, assert_instr(vpsubsb))] |
| 1327 | pub fn _mm_mask_subs_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 1328 | unsafe { |
| 1329 | let sub: i8x16 = _mm_subs_epi8(a, b).as_i8x16(); |
| 1330 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i8x16())) |
| 1331 | } |
| 1332 | } |
| 1333 | |
| 1334 | /// Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1335 | /// |
| 1336 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_subs_epi8&expand=5777) |
| 1337 | #[inline ] |
| 1338 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1339 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1340 | #[cfg_attr (test, assert_instr(vpsubsb))] |
| 1341 | pub fn _mm_maskz_subs_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 1342 | unsafe { |
| 1343 | let sub: i8x16 = _mm_subs_epi8(a, b).as_i8x16(); |
| 1344 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:i8x16::ZERO)) |
| 1345 | } |
| 1346 | } |
| 1347 | |
| 1348 | /// Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst. |
| 1349 | /// |
| 1350 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mulhi_epu16&expand=3973) |
| 1351 | #[inline ] |
| 1352 | #[target_feature (enable = "avx512bw" )] |
| 1353 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1354 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
| 1355 | pub fn _mm512_mulhi_epu16(a: __m512i, b: __m512i) -> __m512i { |
| 1356 | unsafe { |
| 1357 | let a: u32x32 = simd_cast::<_, u32x32>(a.as_u16x32()); |
| 1358 | let b: u32x32 = simd_cast::<_, u32x32>(b.as_u16x32()); |
| 1359 | let r: u32x32 = simd_shr(lhs:simd_mul(a, b), rhs:u32x32::splat(16)); |
| 1360 | transmute(src:simd_cast::<u32x32, u16x32>(r)) |
| 1361 | } |
| 1362 | } |
| 1363 | |
| 1364 | /// Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1365 | /// |
| 1366 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_mulhi_epu16&expand=3971) |
| 1367 | #[inline ] |
| 1368 | #[target_feature (enable = "avx512bw" )] |
| 1369 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1370 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
| 1371 | pub fn _mm512_mask_mulhi_epu16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1372 | unsafe { |
| 1373 | let mul: u16x32 = _mm512_mulhi_epu16(a, b).as_u16x32(); |
| 1374 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_u16x32())) |
| 1375 | } |
| 1376 | } |
| 1377 | |
| 1378 | /// Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1379 | /// |
| 1380 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_mulhi_epu16&expand=3972) |
| 1381 | #[inline ] |
| 1382 | #[target_feature (enable = "avx512bw" )] |
| 1383 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1384 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
| 1385 | pub fn _mm512_maskz_mulhi_epu16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1386 | unsafe { |
| 1387 | let mul: u16x32 = _mm512_mulhi_epu16(a, b).as_u16x32(); |
| 1388 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:u16x32::ZERO)) |
| 1389 | } |
| 1390 | } |
| 1391 | |
| 1392 | /// Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1393 | /// |
| 1394 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_mulhi_epu16&expand=3968) |
| 1395 | #[inline ] |
| 1396 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1397 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1398 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
| 1399 | pub fn _mm256_mask_mulhi_epu16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1400 | unsafe { |
| 1401 | let mul: u16x16 = _mm256_mulhi_epu16(a, b).as_u16x16(); |
| 1402 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_u16x16())) |
| 1403 | } |
| 1404 | } |
| 1405 | |
| 1406 | /// Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1407 | /// |
| 1408 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_mulhi_epu16&expand=3969) |
| 1409 | #[inline ] |
| 1410 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1411 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1412 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
| 1413 | pub fn _mm256_maskz_mulhi_epu16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1414 | unsafe { |
| 1415 | let mul: u16x16 = _mm256_mulhi_epu16(a, b).as_u16x16(); |
| 1416 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:u16x16::ZERO)) |
| 1417 | } |
| 1418 | } |
| 1419 | |
| 1420 | /// Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1421 | /// |
| 1422 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_mulhi_epu16&expand=3965) |
| 1423 | #[inline ] |
| 1424 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1425 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1426 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
| 1427 | pub fn _mm_mask_mulhi_epu16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1428 | unsafe { |
| 1429 | let mul: u16x8 = _mm_mulhi_epu16(a, b).as_u16x8(); |
| 1430 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_u16x8())) |
| 1431 | } |
| 1432 | } |
| 1433 | |
| 1434 | /// Multiply the packed unsigned 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1435 | /// |
| 1436 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_mulhi_epu16&expand=3966) |
| 1437 | #[inline ] |
| 1438 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1439 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1440 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
| 1441 | pub fn _mm_maskz_mulhi_epu16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1442 | unsafe { |
| 1443 | let mul: u16x8 = _mm_mulhi_epu16(a, b).as_u16x8(); |
| 1444 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:u16x8::ZERO)) |
| 1445 | } |
| 1446 | } |
| 1447 | |
| 1448 | /// Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst. |
| 1449 | /// |
| 1450 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mulhi_epi16&expand=3962) |
| 1451 | #[inline ] |
| 1452 | #[target_feature (enable = "avx512bw" )] |
| 1453 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1454 | #[cfg_attr (test, assert_instr(vpmulhw))] |
| 1455 | pub fn _mm512_mulhi_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 1456 | unsafe { |
| 1457 | let a: i32x32 = simd_cast::<_, i32x32>(a.as_i16x32()); |
| 1458 | let b: i32x32 = simd_cast::<_, i32x32>(b.as_i16x32()); |
| 1459 | let r: i32x32 = simd_shr(lhs:simd_mul(a, b), rhs:i32x32::splat(16)); |
| 1460 | transmute(src:simd_cast::<i32x32, i16x32>(r)) |
| 1461 | } |
| 1462 | } |
| 1463 | |
| 1464 | /// Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1465 | /// |
| 1466 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_mulhi_epi16&expand=3960) |
| 1467 | #[inline ] |
| 1468 | #[target_feature (enable = "avx512bw" )] |
| 1469 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1470 | #[cfg_attr (test, assert_instr(vpmulhw))] |
| 1471 | pub fn _mm512_mask_mulhi_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1472 | unsafe { |
| 1473 | let mul: i16x32 = _mm512_mulhi_epi16(a, b).as_i16x32(); |
| 1474 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x32())) |
| 1475 | } |
| 1476 | } |
| 1477 | |
| 1478 | /// Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1479 | /// |
| 1480 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_mulhi_epi16&expand=3961) |
| 1481 | #[inline ] |
| 1482 | #[target_feature (enable = "avx512bw" )] |
| 1483 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1484 | #[cfg_attr (test, assert_instr(vpmulhw))] |
| 1485 | pub fn _mm512_maskz_mulhi_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1486 | unsafe { |
| 1487 | let mul: i16x32 = _mm512_mulhi_epi16(a, b).as_i16x32(); |
| 1488 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:i16x32::ZERO)) |
| 1489 | } |
| 1490 | } |
| 1491 | |
| 1492 | /// Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1493 | /// |
| 1494 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_mulhi_epi16&expand=3957) |
| 1495 | #[inline ] |
| 1496 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1497 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1498 | #[cfg_attr (test, assert_instr(vpmulhw))] |
| 1499 | pub fn _mm256_mask_mulhi_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1500 | unsafe { |
| 1501 | let mul: i16x16 = _mm256_mulhi_epi16(a, b).as_i16x16(); |
| 1502 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x16())) |
| 1503 | } |
| 1504 | } |
| 1505 | |
| 1506 | /// Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1507 | /// |
| 1508 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_mulhi_epi16&expand=3958) |
| 1509 | #[inline ] |
| 1510 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1511 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1512 | #[cfg_attr (test, assert_instr(vpmulhw))] |
| 1513 | pub fn _mm256_maskz_mulhi_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1514 | unsafe { |
| 1515 | let mul: i16x16 = _mm256_mulhi_epi16(a, b).as_i16x16(); |
| 1516 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:i16x16::ZERO)) |
| 1517 | } |
| 1518 | } |
| 1519 | |
| 1520 | /// Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1521 | /// |
| 1522 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_mulhi_epi16&expand=3954) |
| 1523 | #[inline ] |
| 1524 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1525 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1526 | #[cfg_attr (test, assert_instr(vpmulhw))] |
| 1527 | pub fn _mm_mask_mulhi_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1528 | unsafe { |
| 1529 | let mul: i16x8 = _mm_mulhi_epi16(a, b).as_i16x8(); |
| 1530 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x8())) |
| 1531 | } |
| 1532 | } |
| 1533 | |
| 1534 | /// Multiply the packed signed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the high 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1535 | /// |
| 1536 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_mulhi_epi16&expand=3955) |
| 1537 | #[inline ] |
| 1538 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1539 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1540 | #[cfg_attr (test, assert_instr(vpmulhw))] |
| 1541 | pub fn _mm_maskz_mulhi_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1542 | unsafe { |
| 1543 | let mul: i16x8 = _mm_mulhi_epi16(a, b).as_i16x8(); |
| 1544 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:i16x8::ZERO)) |
| 1545 | } |
| 1546 | } |
| 1547 | |
| 1548 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits \[16:1\] to dst. |
| 1549 | /// |
| 1550 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mulhrs_epi16&expand=3986) |
| 1551 | #[inline ] |
| 1552 | #[target_feature (enable = "avx512bw" )] |
| 1553 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1554 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
| 1555 | pub fn _mm512_mulhrs_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 1556 | unsafe { transmute(src:vpmulhrsw(a.as_i16x32(), b.as_i16x32())) } |
| 1557 | } |
| 1558 | |
| 1559 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits \[16:1\] to dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1560 | /// |
| 1561 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_mulhrs_epi16&expand=3984) |
| 1562 | #[inline ] |
| 1563 | #[target_feature (enable = "avx512bw" )] |
| 1564 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1565 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
| 1566 | pub fn _mm512_mask_mulhrs_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1567 | unsafe { |
| 1568 | let mul: i16x32 = _mm512_mulhrs_epi16(a, b).as_i16x32(); |
| 1569 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x32())) |
| 1570 | } |
| 1571 | } |
| 1572 | |
| 1573 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits \[16:1\] to dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1574 | /// |
| 1575 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_mulhrs_epi16&expand=3985) |
| 1576 | #[inline ] |
| 1577 | #[target_feature (enable = "avx512bw" )] |
| 1578 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1579 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
| 1580 | pub fn _mm512_maskz_mulhrs_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1581 | unsafe { |
| 1582 | let mul: i16x32 = _mm512_mulhrs_epi16(a, b).as_i16x32(); |
| 1583 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:i16x32::ZERO)) |
| 1584 | } |
| 1585 | } |
| 1586 | |
| 1587 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits \[16:1\] to dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1588 | /// |
| 1589 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_mulhrs_epi16&expand=3981) |
| 1590 | #[inline ] |
| 1591 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1592 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1593 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
| 1594 | pub fn _mm256_mask_mulhrs_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1595 | unsafe { |
| 1596 | let mul: i16x16 = _mm256_mulhrs_epi16(a, b).as_i16x16(); |
| 1597 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x16())) |
| 1598 | } |
| 1599 | } |
| 1600 | |
| 1601 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits \[16:1\] to dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1602 | /// |
| 1603 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_mulhrs_epi16&expand=3982) |
| 1604 | #[inline ] |
| 1605 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1606 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1607 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
| 1608 | pub fn _mm256_maskz_mulhrs_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1609 | unsafe { |
| 1610 | let mul: i16x16 = _mm256_mulhrs_epi16(a, b).as_i16x16(); |
| 1611 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:i16x16::ZERO)) |
| 1612 | } |
| 1613 | } |
| 1614 | |
| 1615 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits \[16:1\] to dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1616 | /// |
| 1617 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_mulhrs_epi16&expand=3978) |
| 1618 | #[inline ] |
| 1619 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1620 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1621 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
| 1622 | pub fn _mm_mask_mulhrs_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1623 | unsafe { |
| 1624 | let mul: i16x8 = _mm_mulhrs_epi16(a, b).as_i16x8(); |
| 1625 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x8())) |
| 1626 | } |
| 1627 | } |
| 1628 | |
| 1629 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Truncate each intermediate integer to the 18 most significant bits, round by adding 1, and store bits \[16:1\] to dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1630 | /// |
| 1631 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_mulhrs_epi16&expand=3979) |
| 1632 | #[inline ] |
| 1633 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1634 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1635 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
| 1636 | pub fn _mm_maskz_mulhrs_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1637 | unsafe { |
| 1638 | let mul: i16x8 = _mm_mulhrs_epi16(a, b).as_i16x8(); |
| 1639 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:i16x8::ZERO)) |
| 1640 | } |
| 1641 | } |
| 1642 | |
| 1643 | /// Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst. |
| 1644 | /// |
| 1645 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mullo_epi16&expand=3996) |
| 1646 | #[inline ] |
| 1647 | #[target_feature (enable = "avx512bw" )] |
| 1648 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1649 | #[cfg_attr (test, assert_instr(vpmullw))] |
| 1650 | pub fn _mm512_mullo_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 1651 | unsafe { transmute(src:simd_mul(x:a.as_i16x32(), y:b.as_i16x32())) } |
| 1652 | } |
| 1653 | |
| 1654 | /// Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1655 | /// |
| 1656 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_mullo_epi16&expand=3994) |
| 1657 | #[inline ] |
| 1658 | #[target_feature (enable = "avx512bw" )] |
| 1659 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1660 | #[cfg_attr (test, assert_instr(vpmullw))] |
| 1661 | pub fn _mm512_mask_mullo_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1662 | unsafe { |
| 1663 | let mul: i16x32 = _mm512_mullo_epi16(a, b).as_i16x32(); |
| 1664 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x32())) |
| 1665 | } |
| 1666 | } |
| 1667 | |
| 1668 | /// Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1669 | /// |
| 1670 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_mullo_epi16&expand=3995) |
| 1671 | #[inline ] |
| 1672 | #[target_feature (enable = "avx512bw" )] |
| 1673 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1674 | #[cfg_attr (test, assert_instr(vpmullw))] |
| 1675 | pub fn _mm512_maskz_mullo_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1676 | unsafe { |
| 1677 | let mul: i16x32 = _mm512_mullo_epi16(a, b).as_i16x32(); |
| 1678 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:i16x32::ZERO)) |
| 1679 | } |
| 1680 | } |
| 1681 | |
| 1682 | /// Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1683 | /// |
| 1684 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_mullo_epi16&expand=3991) |
| 1685 | #[inline ] |
| 1686 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1687 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1688 | #[cfg_attr (test, assert_instr(vpmullw))] |
| 1689 | pub fn _mm256_mask_mullo_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1690 | unsafe { |
| 1691 | let mul: i16x16 = _mm256_mullo_epi16(a, b).as_i16x16(); |
| 1692 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x16())) |
| 1693 | } |
| 1694 | } |
| 1695 | |
| 1696 | /// Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1697 | /// |
| 1698 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_mullo_epi16&expand=3992) |
| 1699 | #[inline ] |
| 1700 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1701 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1702 | #[cfg_attr (test, assert_instr(vpmullw))] |
| 1703 | pub fn _mm256_maskz_mullo_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1704 | unsafe { |
| 1705 | let mul: i16x16 = _mm256_mullo_epi16(a, b).as_i16x16(); |
| 1706 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:i16x16::ZERO)) |
| 1707 | } |
| 1708 | } |
| 1709 | |
| 1710 | /// Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1711 | /// |
| 1712 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_mullo_epi16&expand=3988) |
| 1713 | #[inline ] |
| 1714 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1715 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1716 | #[cfg_attr (test, assert_instr(vpmullw))] |
| 1717 | pub fn _mm_mask_mullo_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1718 | unsafe { |
| 1719 | let mul: i16x8 = _mm_mullo_epi16(a, b).as_i16x8(); |
| 1720 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x8())) |
| 1721 | } |
| 1722 | } |
| 1723 | |
| 1724 | /// Multiply the packed 16-bit integers in a and b, producing intermediate 32-bit integers, and store the low 16 bits of the intermediate integers in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1725 | /// |
| 1726 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_mullo_epi16&expand=3989) |
| 1727 | #[inline ] |
| 1728 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1729 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1730 | #[cfg_attr (test, assert_instr(vpmullw))] |
| 1731 | pub fn _mm_maskz_mullo_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1732 | unsafe { |
| 1733 | let mul: i16x8 = _mm_mullo_epi16(a, b).as_i16x8(); |
| 1734 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:i16x8::ZERO)) |
| 1735 | } |
| 1736 | } |
| 1737 | |
| 1738 | /// Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst. |
| 1739 | /// |
| 1740 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_max_epu16&expand=3609) |
| 1741 | #[inline ] |
| 1742 | #[target_feature (enable = "avx512bw" )] |
| 1743 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1744 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
| 1745 | pub fn _mm512_max_epu16(a: __m512i, b: __m512i) -> __m512i { |
| 1746 | unsafe { |
| 1747 | let a: u16x32 = a.as_u16x32(); |
| 1748 | let b: u16x32 = b.as_u16x32(); |
| 1749 | transmute(src:simd_select::<i16x32, _>(mask:simd_gt(a, b), if_true:a, if_false:b)) |
| 1750 | } |
| 1751 | } |
| 1752 | |
| 1753 | /// Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1754 | /// |
| 1755 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_max_epu16&expand=3607) |
| 1756 | #[inline ] |
| 1757 | #[target_feature (enable = "avx512bw" )] |
| 1758 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1759 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
| 1760 | pub fn _mm512_mask_max_epu16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1761 | unsafe { |
| 1762 | let max: u16x32 = _mm512_max_epu16(a, b).as_u16x32(); |
| 1763 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_u16x32())) |
| 1764 | } |
| 1765 | } |
| 1766 | |
| 1767 | /// Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1768 | /// |
| 1769 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_max_epu16&expand=3608) |
| 1770 | #[inline ] |
| 1771 | #[target_feature (enable = "avx512bw" )] |
| 1772 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1773 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
| 1774 | pub fn _mm512_maskz_max_epu16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1775 | unsafe { |
| 1776 | let max: u16x32 = _mm512_max_epu16(a, b).as_u16x32(); |
| 1777 | transmute(src:simd_select_bitmask(m:k, yes:max, no:u16x32::ZERO)) |
| 1778 | } |
| 1779 | } |
| 1780 | |
| 1781 | /// Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1782 | /// |
| 1783 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_max_epu16&expand=3604) |
| 1784 | #[inline ] |
| 1785 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1786 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1787 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
| 1788 | pub fn _mm256_mask_max_epu16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1789 | unsafe { |
| 1790 | let max: u16x16 = _mm256_max_epu16(a, b).as_u16x16(); |
| 1791 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_u16x16())) |
| 1792 | } |
| 1793 | } |
| 1794 | |
| 1795 | /// Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1796 | /// |
| 1797 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_max_epu16&expand=3605) |
| 1798 | #[inline ] |
| 1799 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1800 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1801 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
| 1802 | pub fn _mm256_maskz_max_epu16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1803 | unsafe { |
| 1804 | let max: u16x16 = _mm256_max_epu16(a, b).as_u16x16(); |
| 1805 | transmute(src:simd_select_bitmask(m:k, yes:max, no:u16x16::ZERO)) |
| 1806 | } |
| 1807 | } |
| 1808 | |
| 1809 | /// Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1810 | /// |
| 1811 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_max_epu16&expand=3601) |
| 1812 | #[inline ] |
| 1813 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1814 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1815 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
| 1816 | pub fn _mm_mask_max_epu16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1817 | unsafe { |
| 1818 | let max: u16x8 = _mm_max_epu16(a, b).as_u16x8(); |
| 1819 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_u16x8())) |
| 1820 | } |
| 1821 | } |
| 1822 | |
| 1823 | /// Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1824 | /// |
| 1825 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_max_epu16&expand=3602) |
| 1826 | #[inline ] |
| 1827 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1828 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1829 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
| 1830 | pub fn _mm_maskz_max_epu16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 1831 | unsafe { |
| 1832 | let max: u16x8 = _mm_max_epu16(a, b).as_u16x8(); |
| 1833 | transmute(src:simd_select_bitmask(m:k, yes:max, no:u16x8::ZERO)) |
| 1834 | } |
| 1835 | } |
| 1836 | |
| 1837 | /// Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst. |
| 1838 | /// |
| 1839 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_max_epu8&expand=3636) |
| 1840 | #[inline ] |
| 1841 | #[target_feature (enable = "avx512bw" )] |
| 1842 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1843 | #[cfg_attr (test, assert_instr(vpmaxub))] |
| 1844 | pub fn _mm512_max_epu8(a: __m512i, b: __m512i) -> __m512i { |
| 1845 | unsafe { |
| 1846 | let a: u8x64 = a.as_u8x64(); |
| 1847 | let b: u8x64 = b.as_u8x64(); |
| 1848 | transmute(src:simd_select::<i8x64, _>(mask:simd_gt(a, b), if_true:a, if_false:b)) |
| 1849 | } |
| 1850 | } |
| 1851 | |
| 1852 | /// Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1853 | /// |
| 1854 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_max_epu8&expand=3634) |
| 1855 | #[inline ] |
| 1856 | #[target_feature (enable = "avx512bw" )] |
| 1857 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1858 | #[cfg_attr (test, assert_instr(vpmaxub))] |
| 1859 | pub fn _mm512_mask_max_epu8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 1860 | unsafe { |
| 1861 | let max: u8x64 = _mm512_max_epu8(a, b).as_u8x64(); |
| 1862 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_u8x64())) |
| 1863 | } |
| 1864 | } |
| 1865 | |
| 1866 | /// Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1867 | /// |
| 1868 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_max_epu8&expand=3635) |
| 1869 | #[inline ] |
| 1870 | #[target_feature (enable = "avx512bw" )] |
| 1871 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1872 | #[cfg_attr (test, assert_instr(vpmaxub))] |
| 1873 | pub fn _mm512_maskz_max_epu8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 1874 | unsafe { |
| 1875 | let max: u8x64 = _mm512_max_epu8(a, b).as_u8x64(); |
| 1876 | transmute(src:simd_select_bitmask(m:k, yes:max, no:u8x64::ZERO)) |
| 1877 | } |
| 1878 | } |
| 1879 | |
| 1880 | /// Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1881 | /// |
| 1882 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_max_epu8&expand=3631) |
| 1883 | #[inline ] |
| 1884 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1885 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1886 | #[cfg_attr (test, assert_instr(vpmaxub))] |
| 1887 | pub fn _mm256_mask_max_epu8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 1888 | unsafe { |
| 1889 | let max: u8x32 = _mm256_max_epu8(a, b).as_u8x32(); |
| 1890 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_u8x32())) |
| 1891 | } |
| 1892 | } |
| 1893 | |
| 1894 | /// Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1895 | /// |
| 1896 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_max_epu8&expand=3632) |
| 1897 | #[inline ] |
| 1898 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1899 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1900 | #[cfg_attr (test, assert_instr(vpmaxub))] |
| 1901 | pub fn _mm256_maskz_max_epu8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 1902 | unsafe { |
| 1903 | let max: u8x32 = _mm256_max_epu8(a, b).as_u8x32(); |
| 1904 | transmute(src:simd_select_bitmask(m:k, yes:max, no:u8x32::ZERO)) |
| 1905 | } |
| 1906 | } |
| 1907 | |
| 1908 | /// Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1909 | /// |
| 1910 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_max_epu8&expand=3628) |
| 1911 | #[inline ] |
| 1912 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1913 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1914 | #[cfg_attr (test, assert_instr(vpmaxub))] |
| 1915 | pub fn _mm_mask_max_epu8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 1916 | unsafe { |
| 1917 | let max: u8x16 = _mm_max_epu8(a, b).as_u8x16(); |
| 1918 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_u8x16())) |
| 1919 | } |
| 1920 | } |
| 1921 | |
| 1922 | /// Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1923 | /// |
| 1924 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_max_epu8&expand=3629) |
| 1925 | #[inline ] |
| 1926 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1927 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1928 | #[cfg_attr (test, assert_instr(vpmaxub))] |
| 1929 | pub fn _mm_maskz_max_epu8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 1930 | unsafe { |
| 1931 | let max: u8x16 = _mm_max_epu8(a, b).as_u8x16(); |
| 1932 | transmute(src:simd_select_bitmask(m:k, yes:max, no:u8x16::ZERO)) |
| 1933 | } |
| 1934 | } |
| 1935 | |
| 1936 | /// Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst. |
| 1937 | /// |
| 1938 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_max_epi16&expand=3573) |
| 1939 | #[inline ] |
| 1940 | #[target_feature (enable = "avx512bw" )] |
| 1941 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1942 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
| 1943 | pub fn _mm512_max_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 1944 | unsafe { |
| 1945 | let a: i16x32 = a.as_i16x32(); |
| 1946 | let b: i16x32 = b.as_i16x32(); |
| 1947 | transmute(src:simd_select::<i16x32, _>(mask:simd_gt(a, b), if_true:a, if_false:b)) |
| 1948 | } |
| 1949 | } |
| 1950 | |
| 1951 | /// Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1952 | /// |
| 1953 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_max_epi16&expand=3571) |
| 1954 | #[inline ] |
| 1955 | #[target_feature (enable = "avx512bw" )] |
| 1956 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1957 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
| 1958 | pub fn _mm512_mask_max_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1959 | unsafe { |
| 1960 | let max: i16x32 = _mm512_max_epi16(a, b).as_i16x32(); |
| 1961 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_i16x32())) |
| 1962 | } |
| 1963 | } |
| 1964 | |
| 1965 | /// Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1966 | /// |
| 1967 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_max_epi16&expand=3572) |
| 1968 | #[inline ] |
| 1969 | #[target_feature (enable = "avx512bw" )] |
| 1970 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1971 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
| 1972 | pub fn _mm512_maskz_max_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 1973 | unsafe { |
| 1974 | let max: i16x32 = _mm512_max_epi16(a, b).as_i16x32(); |
| 1975 | transmute(src:simd_select_bitmask(m:k, yes:max, no:i16x32::ZERO)) |
| 1976 | } |
| 1977 | } |
| 1978 | |
| 1979 | /// Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 1980 | /// |
| 1981 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_max_epi16&expand=3568) |
| 1982 | #[inline ] |
| 1983 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1984 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1985 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
| 1986 | pub fn _mm256_mask_max_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 1987 | unsafe { |
| 1988 | let max: i16x16 = _mm256_max_epi16(a, b).as_i16x16(); |
| 1989 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_i16x16())) |
| 1990 | } |
| 1991 | } |
| 1992 | |
| 1993 | /// Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 1994 | /// |
| 1995 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_max_epi16&expand=3569) |
| 1996 | #[inline ] |
| 1997 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 1998 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 1999 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
| 2000 | pub fn _mm256_maskz_max_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 2001 | unsafe { |
| 2002 | let max: i16x16 = _mm256_max_epi16(a, b).as_i16x16(); |
| 2003 | transmute(src:simd_select_bitmask(m:k, yes:max, no:i16x16::ZERO)) |
| 2004 | } |
| 2005 | } |
| 2006 | |
| 2007 | /// Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2008 | /// |
| 2009 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_max_epi16&expand=3565) |
| 2010 | #[inline ] |
| 2011 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2012 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2013 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
| 2014 | pub fn _mm_mask_max_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 2015 | unsafe { |
| 2016 | let max: i16x8 = _mm_max_epi16(a, b).as_i16x8(); |
| 2017 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_i16x8())) |
| 2018 | } |
| 2019 | } |
| 2020 | |
| 2021 | /// Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2022 | /// |
| 2023 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_max_epi16&expand=3566) |
| 2024 | #[inline ] |
| 2025 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2026 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2027 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
| 2028 | pub fn _mm_maskz_max_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 2029 | unsafe { |
| 2030 | let max: i16x8 = _mm_max_epi16(a, b).as_i16x8(); |
| 2031 | transmute(src:simd_select_bitmask(m:k, yes:max, no:i16x8::ZERO)) |
| 2032 | } |
| 2033 | } |
| 2034 | |
| 2035 | /// Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst. |
| 2036 | /// |
| 2037 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_max_epi8&expand=3600) |
| 2038 | #[inline ] |
| 2039 | #[target_feature (enable = "avx512bw" )] |
| 2040 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2041 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
| 2042 | pub fn _mm512_max_epi8(a: __m512i, b: __m512i) -> __m512i { |
| 2043 | unsafe { |
| 2044 | let a: i8x64 = a.as_i8x64(); |
| 2045 | let b: i8x64 = b.as_i8x64(); |
| 2046 | transmute(src:simd_select::<i8x64, _>(mask:simd_gt(a, b), if_true:a, if_false:b)) |
| 2047 | } |
| 2048 | } |
| 2049 | |
| 2050 | /// Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2051 | /// |
| 2052 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_max_epi8&expand=3598) |
| 2053 | #[inline ] |
| 2054 | #[target_feature (enable = "avx512bw" )] |
| 2055 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2056 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
| 2057 | pub fn _mm512_mask_max_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 2058 | unsafe { |
| 2059 | let max: i8x64 = _mm512_max_epi8(a, b).as_i8x64(); |
| 2060 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_i8x64())) |
| 2061 | } |
| 2062 | } |
| 2063 | |
| 2064 | /// Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2065 | /// |
| 2066 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_max_epi8&expand=3599) |
| 2067 | #[inline ] |
| 2068 | #[target_feature (enable = "avx512bw" )] |
| 2069 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2070 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
| 2071 | pub fn _mm512_maskz_max_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 2072 | unsafe { |
| 2073 | let max: i8x64 = _mm512_max_epi8(a, b).as_i8x64(); |
| 2074 | transmute(src:simd_select_bitmask(m:k, yes:max, no:i8x64::ZERO)) |
| 2075 | } |
| 2076 | } |
| 2077 | |
| 2078 | /// Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2079 | /// |
| 2080 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_max_epi8&expand=3595) |
| 2081 | #[inline ] |
| 2082 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2083 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2084 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
| 2085 | pub fn _mm256_mask_max_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 2086 | unsafe { |
| 2087 | let max: i8x32 = _mm256_max_epi8(a, b).as_i8x32(); |
| 2088 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_i8x32())) |
| 2089 | } |
| 2090 | } |
| 2091 | |
| 2092 | /// Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2093 | /// |
| 2094 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_max_epi8&expand=3596) |
| 2095 | #[inline ] |
| 2096 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2097 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2098 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
| 2099 | pub fn _mm256_maskz_max_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 2100 | unsafe { |
| 2101 | let max: i8x32 = _mm256_max_epi8(a, b).as_i8x32(); |
| 2102 | transmute(src:simd_select_bitmask(m:k, yes:max, no:i8x32::ZERO)) |
| 2103 | } |
| 2104 | } |
| 2105 | |
| 2106 | /// Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2107 | /// |
| 2108 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_max_epi8&expand=3592) |
| 2109 | #[inline ] |
| 2110 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2111 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2112 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
| 2113 | pub fn _mm_mask_max_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 2114 | unsafe { |
| 2115 | let max: i8x16 = _mm_max_epi8(a, b).as_i8x16(); |
| 2116 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_i8x16())) |
| 2117 | } |
| 2118 | } |
| 2119 | |
| 2120 | /// Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2121 | /// |
| 2122 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_max_epi8&expand=3593) |
| 2123 | #[inline ] |
| 2124 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2125 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2126 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
| 2127 | pub fn _mm_maskz_max_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 2128 | unsafe { |
| 2129 | let max: i8x16 = _mm_max_epi8(a, b).as_i8x16(); |
| 2130 | transmute(src:simd_select_bitmask(m:k, yes:max, no:i8x16::ZERO)) |
| 2131 | } |
| 2132 | } |
| 2133 | |
| 2134 | /// Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst. |
| 2135 | /// |
| 2136 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_min_epu16&expand=3723) |
| 2137 | #[inline ] |
| 2138 | #[target_feature (enable = "avx512bw" )] |
| 2139 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2140 | #[cfg_attr (test, assert_instr(vpminuw))] |
| 2141 | pub fn _mm512_min_epu16(a: __m512i, b: __m512i) -> __m512i { |
| 2142 | unsafe { |
| 2143 | let a: u16x32 = a.as_u16x32(); |
| 2144 | let b: u16x32 = b.as_u16x32(); |
| 2145 | transmute(src:simd_select::<i16x32, _>(mask:simd_lt(a, b), if_true:a, if_false:b)) |
| 2146 | } |
| 2147 | } |
| 2148 | |
| 2149 | /// Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2150 | /// |
| 2151 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_min_epu16&expand=3721) |
| 2152 | #[inline ] |
| 2153 | #[target_feature (enable = "avx512bw" )] |
| 2154 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2155 | #[cfg_attr (test, assert_instr(vpminuw))] |
| 2156 | pub fn _mm512_mask_min_epu16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 2157 | unsafe { |
| 2158 | let min: u16x32 = _mm512_min_epu16(a, b).as_u16x32(); |
| 2159 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_u16x32())) |
| 2160 | } |
| 2161 | } |
| 2162 | |
| 2163 | /// Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2164 | /// |
| 2165 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_min_epu16&expand=3722) |
| 2166 | #[inline ] |
| 2167 | #[target_feature (enable = "avx512bw" )] |
| 2168 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2169 | #[cfg_attr (test, assert_instr(vpminuw))] |
| 2170 | pub fn _mm512_maskz_min_epu16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 2171 | unsafe { |
| 2172 | let min: u16x32 = _mm512_min_epu16(a, b).as_u16x32(); |
| 2173 | transmute(src:simd_select_bitmask(m:k, yes:min, no:u16x32::ZERO)) |
| 2174 | } |
| 2175 | } |
| 2176 | |
| 2177 | /// Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2178 | /// |
| 2179 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_min_epu16&expand=3718) |
| 2180 | #[inline ] |
| 2181 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2182 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2183 | #[cfg_attr (test, assert_instr(vpminuw))] |
| 2184 | pub fn _mm256_mask_min_epu16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 2185 | unsafe { |
| 2186 | let min: u16x16 = _mm256_min_epu16(a, b).as_u16x16(); |
| 2187 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_u16x16())) |
| 2188 | } |
| 2189 | } |
| 2190 | |
| 2191 | /// Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2192 | /// |
| 2193 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_min_epu16&expand=3719) |
| 2194 | #[inline ] |
| 2195 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2196 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2197 | #[cfg_attr (test, assert_instr(vpminuw))] |
| 2198 | pub fn _mm256_maskz_min_epu16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 2199 | unsafe { |
| 2200 | let min: u16x16 = _mm256_min_epu16(a, b).as_u16x16(); |
| 2201 | transmute(src:simd_select_bitmask(m:k, yes:min, no:u16x16::ZERO)) |
| 2202 | } |
| 2203 | } |
| 2204 | |
| 2205 | /// Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2206 | /// |
| 2207 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_min_epu16&expand=3715) |
| 2208 | #[inline ] |
| 2209 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2210 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2211 | #[cfg_attr (test, assert_instr(vpminuw))] |
| 2212 | pub fn _mm_mask_min_epu16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 2213 | unsafe { |
| 2214 | let min: u16x8 = _mm_min_epu16(a, b).as_u16x8(); |
| 2215 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_u16x8())) |
| 2216 | } |
| 2217 | } |
| 2218 | |
| 2219 | /// Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2220 | /// |
| 2221 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_min_epu16&expand=3716) |
| 2222 | #[inline ] |
| 2223 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2224 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2225 | #[cfg_attr (test, assert_instr(vpminuw))] |
| 2226 | pub fn _mm_maskz_min_epu16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 2227 | unsafe { |
| 2228 | let min: u16x8 = _mm_min_epu16(a, b).as_u16x8(); |
| 2229 | transmute(src:simd_select_bitmask(m:k, yes:min, no:u16x8::ZERO)) |
| 2230 | } |
| 2231 | } |
| 2232 | |
| 2233 | /// Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst. |
| 2234 | /// |
| 2235 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_min_epu8&expand=3750) |
| 2236 | #[inline ] |
| 2237 | #[target_feature (enable = "avx512bw" )] |
| 2238 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2239 | #[cfg_attr (test, assert_instr(vpminub))] |
| 2240 | pub fn _mm512_min_epu8(a: __m512i, b: __m512i) -> __m512i { |
| 2241 | unsafe { |
| 2242 | let a: u8x64 = a.as_u8x64(); |
| 2243 | let b: u8x64 = b.as_u8x64(); |
| 2244 | transmute(src:simd_select::<i8x64, _>(mask:simd_lt(a, b), if_true:a, if_false:b)) |
| 2245 | } |
| 2246 | } |
| 2247 | |
| 2248 | /// Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2249 | /// |
| 2250 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_min_epu8&expand=3748) |
| 2251 | #[inline ] |
| 2252 | #[target_feature (enable = "avx512bw" )] |
| 2253 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2254 | #[cfg_attr (test, assert_instr(vpminub))] |
| 2255 | pub fn _mm512_mask_min_epu8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 2256 | unsafe { |
| 2257 | let min: u8x64 = _mm512_min_epu8(a, b).as_u8x64(); |
| 2258 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_u8x64())) |
| 2259 | } |
| 2260 | } |
| 2261 | |
| 2262 | /// Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2263 | /// |
| 2264 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_min_epu8&expand=3749) |
| 2265 | #[inline ] |
| 2266 | #[target_feature (enable = "avx512bw" )] |
| 2267 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2268 | #[cfg_attr (test, assert_instr(vpminub))] |
| 2269 | pub fn _mm512_maskz_min_epu8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 2270 | unsafe { |
| 2271 | let min: u8x64 = _mm512_min_epu8(a, b).as_u8x64(); |
| 2272 | transmute(src:simd_select_bitmask(m:k, yes:min, no:u8x64::ZERO)) |
| 2273 | } |
| 2274 | } |
| 2275 | |
| 2276 | /// Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2277 | /// |
| 2278 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_min_epu8&expand=3745) |
| 2279 | #[inline ] |
| 2280 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2281 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2282 | #[cfg_attr (test, assert_instr(vpminub))] |
| 2283 | pub fn _mm256_mask_min_epu8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 2284 | unsafe { |
| 2285 | let min: u8x32 = _mm256_min_epu8(a, b).as_u8x32(); |
| 2286 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_u8x32())) |
| 2287 | } |
| 2288 | } |
| 2289 | |
| 2290 | /// Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2291 | /// |
| 2292 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_min_epu8&expand=3746) |
| 2293 | #[inline ] |
| 2294 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2295 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2296 | #[cfg_attr (test, assert_instr(vpminub))] |
| 2297 | pub fn _mm256_maskz_min_epu8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 2298 | unsafe { |
| 2299 | let min: u8x32 = _mm256_min_epu8(a, b).as_u8x32(); |
| 2300 | transmute(src:simd_select_bitmask(m:k, yes:min, no:u8x32::ZERO)) |
| 2301 | } |
| 2302 | } |
| 2303 | |
| 2304 | /// Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2305 | /// |
| 2306 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_min_epu8&expand=3742) |
| 2307 | #[inline ] |
| 2308 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2309 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2310 | #[cfg_attr (test, assert_instr(vpminub))] |
| 2311 | pub fn _mm_mask_min_epu8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 2312 | unsafe { |
| 2313 | let min: u8x16 = _mm_min_epu8(a, b).as_u8x16(); |
| 2314 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_u8x16())) |
| 2315 | } |
| 2316 | } |
| 2317 | |
| 2318 | /// Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2319 | /// |
| 2320 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_min_epu8&expand=3743) |
| 2321 | #[inline ] |
| 2322 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2323 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2324 | #[cfg_attr (test, assert_instr(vpminub))] |
| 2325 | pub fn _mm_maskz_min_epu8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 2326 | unsafe { |
| 2327 | let min: u8x16 = _mm_min_epu8(a, b).as_u8x16(); |
| 2328 | transmute(src:simd_select_bitmask(m:k, yes:min, no:u8x16::ZERO)) |
| 2329 | } |
| 2330 | } |
| 2331 | |
| 2332 | /// Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst. |
| 2333 | /// |
| 2334 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_min_epi16&expand=3687) |
| 2335 | #[inline ] |
| 2336 | #[target_feature (enable = "avx512bw" )] |
| 2337 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2338 | #[cfg_attr (test, assert_instr(vpminsw))] |
| 2339 | pub fn _mm512_min_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 2340 | unsafe { |
| 2341 | let a: i16x32 = a.as_i16x32(); |
| 2342 | let b: i16x32 = b.as_i16x32(); |
| 2343 | transmute(src:simd_select::<i16x32, _>(mask:simd_lt(a, b), if_true:a, if_false:b)) |
| 2344 | } |
| 2345 | } |
| 2346 | |
| 2347 | /// Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2348 | /// |
| 2349 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_min_epi16&expand=3685) |
| 2350 | #[inline ] |
| 2351 | #[target_feature (enable = "avx512bw" )] |
| 2352 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2353 | #[cfg_attr (test, assert_instr(vpminsw))] |
| 2354 | pub fn _mm512_mask_min_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 2355 | unsafe { |
| 2356 | let min: i16x32 = _mm512_min_epi16(a, b).as_i16x32(); |
| 2357 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_i16x32())) |
| 2358 | } |
| 2359 | } |
| 2360 | |
| 2361 | /// Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2362 | /// |
| 2363 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_min_epi16&expand=3686) |
| 2364 | #[inline ] |
| 2365 | #[target_feature (enable = "avx512bw" )] |
| 2366 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2367 | #[cfg_attr (test, assert_instr(vpminsw))] |
| 2368 | pub fn _mm512_maskz_min_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 2369 | unsafe { |
| 2370 | let min: i16x32 = _mm512_min_epi16(a, b).as_i16x32(); |
| 2371 | transmute(src:simd_select_bitmask(m:k, yes:min, no:i16x32::ZERO)) |
| 2372 | } |
| 2373 | } |
| 2374 | |
| 2375 | /// Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2376 | /// |
| 2377 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_min_epi16&expand=3682) |
| 2378 | #[inline ] |
| 2379 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2380 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2381 | #[cfg_attr (test, assert_instr(vpminsw))] |
| 2382 | pub fn _mm256_mask_min_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 2383 | unsafe { |
| 2384 | let min: i16x16 = _mm256_min_epi16(a, b).as_i16x16(); |
| 2385 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_i16x16())) |
| 2386 | } |
| 2387 | } |
| 2388 | |
| 2389 | /// Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2390 | /// |
| 2391 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_min_epi16&expand=3683) |
| 2392 | #[inline ] |
| 2393 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2394 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2395 | #[cfg_attr (test, assert_instr(vpminsw))] |
| 2396 | pub fn _mm256_maskz_min_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 2397 | unsafe { |
| 2398 | let min: i16x16 = _mm256_min_epi16(a, b).as_i16x16(); |
| 2399 | transmute(src:simd_select_bitmask(m:k, yes:min, no:i16x16::ZERO)) |
| 2400 | } |
| 2401 | } |
| 2402 | |
| 2403 | /// Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2404 | /// |
| 2405 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_min_epi16&expand=3679) |
| 2406 | #[inline ] |
| 2407 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2408 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2409 | #[cfg_attr (test, assert_instr(vpminsw))] |
| 2410 | pub fn _mm_mask_min_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 2411 | unsafe { |
| 2412 | let min: i16x8 = _mm_min_epi16(a, b).as_i16x8(); |
| 2413 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_i16x8())) |
| 2414 | } |
| 2415 | } |
| 2416 | |
| 2417 | /// Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2418 | /// |
| 2419 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_min_epi16&expand=3680) |
| 2420 | #[inline ] |
| 2421 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2422 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2423 | #[cfg_attr (test, assert_instr(vpminsw))] |
| 2424 | pub fn _mm_maskz_min_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 2425 | unsafe { |
| 2426 | let min: i16x8 = _mm_min_epi16(a, b).as_i16x8(); |
| 2427 | transmute(src:simd_select_bitmask(m:k, yes:min, no:i16x8::ZERO)) |
| 2428 | } |
| 2429 | } |
| 2430 | |
| 2431 | /// Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst. |
| 2432 | /// |
| 2433 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_min_epi8&expand=3714) |
| 2434 | #[inline ] |
| 2435 | #[target_feature (enable = "avx512bw" )] |
| 2436 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2437 | #[cfg_attr (test, assert_instr(vpminsb))] |
| 2438 | pub fn _mm512_min_epi8(a: __m512i, b: __m512i) -> __m512i { |
| 2439 | unsafe { |
| 2440 | let a: i8x64 = a.as_i8x64(); |
| 2441 | let b: i8x64 = b.as_i8x64(); |
| 2442 | transmute(src:simd_select::<i8x64, _>(mask:simd_lt(a, b), if_true:a, if_false:b)) |
| 2443 | } |
| 2444 | } |
| 2445 | |
| 2446 | /// Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2447 | /// |
| 2448 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_min_epi8&expand=3712) |
| 2449 | #[inline ] |
| 2450 | #[target_feature (enable = "avx512bw" )] |
| 2451 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2452 | #[cfg_attr (test, assert_instr(vpminsb))] |
| 2453 | pub fn _mm512_mask_min_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 2454 | unsafe { |
| 2455 | let min: i8x64 = _mm512_min_epi8(a, b).as_i8x64(); |
| 2456 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_i8x64())) |
| 2457 | } |
| 2458 | } |
| 2459 | |
| 2460 | /// Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2461 | /// |
| 2462 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_min_epi8&expand=3713) |
| 2463 | #[inline ] |
| 2464 | #[target_feature (enable = "avx512bw" )] |
| 2465 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2466 | #[cfg_attr (test, assert_instr(vpminsb))] |
| 2467 | pub fn _mm512_maskz_min_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 2468 | unsafe { |
| 2469 | let min: i8x64 = _mm512_min_epi8(a, b).as_i8x64(); |
| 2470 | transmute(src:simd_select_bitmask(m:k, yes:min, no:i8x64::ZERO)) |
| 2471 | } |
| 2472 | } |
| 2473 | |
| 2474 | /// Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2475 | /// |
| 2476 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_min_epi8&expand=3709) |
| 2477 | #[inline ] |
| 2478 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2479 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2480 | #[cfg_attr (test, assert_instr(vpminsb))] |
| 2481 | pub fn _mm256_mask_min_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 2482 | unsafe { |
| 2483 | let min: i8x32 = _mm256_min_epi8(a, b).as_i8x32(); |
| 2484 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_i8x32())) |
| 2485 | } |
| 2486 | } |
| 2487 | |
| 2488 | /// Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2489 | /// |
| 2490 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_min_epi8&expand=3710) |
| 2491 | #[inline ] |
| 2492 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2493 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2494 | #[cfg_attr (test, assert_instr(vpminsb))] |
| 2495 | pub fn _mm256_maskz_min_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 2496 | unsafe { |
| 2497 | let min: i8x32 = _mm256_min_epi8(a, b).as_i8x32(); |
| 2498 | transmute(src:simd_select_bitmask(m:k, yes:min, no:i8x32::ZERO)) |
| 2499 | } |
| 2500 | } |
| 2501 | |
| 2502 | /// Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 2503 | /// |
| 2504 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_min_epi8&expand=3706) |
| 2505 | #[inline ] |
| 2506 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2507 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2508 | #[cfg_attr (test, assert_instr(vpminsb))] |
| 2509 | pub fn _mm_mask_min_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 2510 | unsafe { |
| 2511 | let min: i8x16 = _mm_min_epi8(a, b).as_i8x16(); |
| 2512 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_i8x16())) |
| 2513 | } |
| 2514 | } |
| 2515 | |
| 2516 | /// Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 2517 | /// |
| 2518 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_min_epi8&expand=3707) |
| 2519 | #[inline ] |
| 2520 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2521 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2522 | #[cfg_attr (test, assert_instr(vpminsb))] |
| 2523 | pub fn _mm_maskz_min_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 2524 | unsafe { |
| 2525 | let min: i8x16 = _mm_min_epi8(a, b).as_i8x16(); |
| 2526 | transmute(src:simd_select_bitmask(m:k, yes:min, no:i8x16::ZERO)) |
| 2527 | } |
| 2528 | } |
| 2529 | |
| 2530 | /// Compare packed unsigned 16-bit integers in a and b for less-than, and store the results in mask vector k. |
| 2531 | /// |
| 2532 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmplt_epu16_mask&expand=1050) |
| 2533 | #[inline ] |
| 2534 | #[target_feature (enable = "avx512bw" )] |
| 2535 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2536 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2537 | pub fn _mm512_cmplt_epu16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 2538 | unsafe { simd_bitmask::<u16x32, _>(simd_lt(x:a.as_u16x32(), y:b.as_u16x32())) } |
| 2539 | } |
| 2540 | |
| 2541 | /// Compare packed unsigned 16-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2542 | /// |
| 2543 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmplt_epu16_mask&expand=1051) |
| 2544 | #[inline ] |
| 2545 | #[target_feature (enable = "avx512bw" )] |
| 2546 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2547 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2548 | pub fn _mm512_mask_cmplt_epu16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 2549 | _mm512_mask_cmp_epu16_mask::<_MM_CMPINT_LT>(k1, a, b) |
| 2550 | } |
| 2551 | |
| 2552 | /// Compare packed unsigned 16-bit integers in a and b for less-than, and store the results in mask vector k. |
| 2553 | /// |
| 2554 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmplt_epu16_mask&expand=1050) |
| 2555 | #[inline ] |
| 2556 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2557 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2558 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2559 | pub fn _mm256_cmplt_epu16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 2560 | unsafe { simd_bitmask::<u16x16, _>(simd_lt(x:a.as_u16x16(), y:b.as_u16x16())) } |
| 2561 | } |
| 2562 | |
| 2563 | /// Compare packed unsigned 16-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2564 | /// |
| 2565 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmplt_epu16_mask&expand=1049) |
| 2566 | #[inline ] |
| 2567 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2568 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2569 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2570 | pub fn _mm256_mask_cmplt_epu16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 2571 | _mm256_mask_cmp_epu16_mask::<_MM_CMPINT_LT>(k1, a, b) |
| 2572 | } |
| 2573 | |
| 2574 | /// Compare packed unsigned 16-bit integers in a and b for less-than, and store the results in mask vector k. |
| 2575 | /// |
| 2576 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmplt_epu16_mask&expand=1018) |
| 2577 | #[inline ] |
| 2578 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2579 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2580 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2581 | pub fn _mm_cmplt_epu16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 2582 | unsafe { simd_bitmask::<u16x8, _>(simd_lt(x:a.as_u16x8(), y:b.as_u16x8())) } |
| 2583 | } |
| 2584 | |
| 2585 | /// Compare packed unsigned 16-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2586 | /// |
| 2587 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmplt_epu16_mask&expand=1019) |
| 2588 | #[inline ] |
| 2589 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2590 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2591 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2592 | pub fn _mm_mask_cmplt_epu16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 2593 | _mm_mask_cmp_epu16_mask::<_MM_CMPINT_LT>(k1, a, b) |
| 2594 | } |
| 2595 | |
| 2596 | /// Compare packed unsigned 8-bit integers in a and b for less-than, and store the results in mask vector k. |
| 2597 | /// |
| 2598 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=mm512_cmplt_epu8_mask&expand=1068) |
| 2599 | #[inline ] |
| 2600 | #[target_feature (enable = "avx512bw" )] |
| 2601 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2602 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2603 | pub fn _mm512_cmplt_epu8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 2604 | unsafe { simd_bitmask::<u8x64, _>(simd_lt(x:a.as_u8x64(), y:b.as_u8x64())) } |
| 2605 | } |
| 2606 | |
| 2607 | /// Compare packed unsigned 8-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2608 | /// |
| 2609 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmplt_epu8_mask&expand=1069) |
| 2610 | #[inline ] |
| 2611 | #[target_feature (enable = "avx512bw" )] |
| 2612 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2613 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2614 | pub fn _mm512_mask_cmplt_epu8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 2615 | _mm512_mask_cmp_epu8_mask::<_MM_CMPINT_LT>(k1, a, b) |
| 2616 | } |
| 2617 | |
| 2618 | /// Compare packed unsigned 8-bit integers in a and b for less-than, and store the results in mask vector k. |
| 2619 | /// |
| 2620 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmplt_epu8_mask&expand=1066) |
| 2621 | #[inline ] |
| 2622 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2623 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2624 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2625 | pub fn _mm256_cmplt_epu8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 2626 | unsafe { simd_bitmask::<u8x32, _>(simd_lt(x:a.as_u8x32(), y:b.as_u8x32())) } |
| 2627 | } |
| 2628 | |
| 2629 | /// Compare packed unsigned 8-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2630 | /// |
| 2631 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmplt_epu8_mask&expand=1067) |
| 2632 | #[inline ] |
| 2633 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2634 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2635 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2636 | pub fn _mm256_mask_cmplt_epu8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 2637 | _mm256_mask_cmp_epu8_mask::<_MM_CMPINT_LT>(k1, a, b) |
| 2638 | } |
| 2639 | |
| 2640 | /// Compare packed unsigned 8-bit integers in a and b for less-than, and store the results in mask vector k. |
| 2641 | /// |
| 2642 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmplt_epu8_mask&expand=1064) |
| 2643 | #[inline ] |
| 2644 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2645 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2646 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2647 | pub fn _mm_cmplt_epu8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 2648 | unsafe { simd_bitmask::<u8x16, _>(simd_lt(x:a.as_u8x16(), y:b.as_u8x16())) } |
| 2649 | } |
| 2650 | |
| 2651 | /// Compare packed unsigned 8-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2652 | /// |
| 2653 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmplt_epu8_mask&expand=1065) |
| 2654 | #[inline ] |
| 2655 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2656 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2657 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2658 | pub fn _mm_mask_cmplt_epu8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 2659 | _mm_mask_cmp_epu8_mask::<_MM_CMPINT_LT>(k1, a, b) |
| 2660 | } |
| 2661 | |
| 2662 | /// Compare packed signed 16-bit integers in a and b for less-than, and store the results in mask vector k. |
| 2663 | /// |
| 2664 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmplt_epi16_mask&expand=1022) |
| 2665 | #[inline ] |
| 2666 | #[target_feature (enable = "avx512bw" )] |
| 2667 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2668 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2669 | pub fn _mm512_cmplt_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 2670 | unsafe { simd_bitmask::<i16x32, _>(simd_lt(x:a.as_i16x32(), y:b.as_i16x32())) } |
| 2671 | } |
| 2672 | |
| 2673 | /// Compare packed signed 16-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2674 | /// |
| 2675 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmplt_epi16_mask&expand=1023) |
| 2676 | #[inline ] |
| 2677 | #[target_feature (enable = "avx512bw" )] |
| 2678 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2679 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2680 | pub fn _mm512_mask_cmplt_epi16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 2681 | _mm512_mask_cmp_epi16_mask::<_MM_CMPINT_LT>(k1, a, b) |
| 2682 | } |
| 2683 | |
| 2684 | /// Compare packed signed 16-bit integers in a and b for less-than, and store the results in mask vector k. |
| 2685 | /// |
| 2686 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmplt_epi16_mask&expand=1020) |
| 2687 | #[inline ] |
| 2688 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2689 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2690 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2691 | pub fn _mm256_cmplt_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 2692 | unsafe { simd_bitmask::<i16x16, _>(simd_lt(x:a.as_i16x16(), y:b.as_i16x16())) } |
| 2693 | } |
| 2694 | |
| 2695 | /// Compare packed signed 16-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2696 | /// |
| 2697 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmplt_epi16_mask&expand=1021) |
| 2698 | #[inline ] |
| 2699 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2700 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2701 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2702 | pub fn _mm256_mask_cmplt_epi16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 2703 | _mm256_mask_cmp_epi16_mask::<_MM_CMPINT_LT>(k1, a, b) |
| 2704 | } |
| 2705 | |
| 2706 | /// Compare packed signed 16-bit integers in a and b for less-than, and store the results in mask vector k. |
| 2707 | /// |
| 2708 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmplt_epi16_mask&expand=1018) |
| 2709 | #[inline ] |
| 2710 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2711 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2712 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2713 | pub fn _mm_cmplt_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 2714 | unsafe { simd_bitmask::<i16x8, _>(simd_lt(x:a.as_i16x8(), y:b.as_i16x8())) } |
| 2715 | } |
| 2716 | |
| 2717 | /// Compare packed signed 16-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2718 | /// |
| 2719 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmplt_epi16_mask&expand=1019) |
| 2720 | #[inline ] |
| 2721 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2722 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2723 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2724 | pub fn _mm_mask_cmplt_epi16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 2725 | _mm_mask_cmp_epi16_mask::<_MM_CMPINT_LT>(k1, a, b) |
| 2726 | } |
| 2727 | |
| 2728 | /// Compare packed signed 8-bit integers in a and b for less-than, and store the results in mask vector k. |
| 2729 | /// |
| 2730 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmplt_epi8_mask&expand=1044) |
| 2731 | #[inline ] |
| 2732 | #[target_feature (enable = "avx512bw" )] |
| 2733 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2734 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2735 | pub fn _mm512_cmplt_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 2736 | unsafe { simd_bitmask::<i8x64, _>(simd_lt(x:a.as_i8x64(), y:b.as_i8x64())) } |
| 2737 | } |
| 2738 | |
| 2739 | /// Compare packed signed 8-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2740 | /// |
| 2741 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmplt_epi8_mask&expand=1045) |
| 2742 | #[inline ] |
| 2743 | #[target_feature (enable = "avx512bw" )] |
| 2744 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2745 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2746 | pub fn _mm512_mask_cmplt_epi8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 2747 | _mm512_mask_cmp_epi8_mask::<_MM_CMPINT_LT>(k1, a, b) |
| 2748 | } |
| 2749 | |
| 2750 | /// Compare packed signed 8-bit integers in a and b for less-than, and store the results in mask vector k. |
| 2751 | /// |
| 2752 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmplt_epi8_mask&expand=1042) |
| 2753 | #[inline ] |
| 2754 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2755 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2756 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2757 | pub fn _mm256_cmplt_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 2758 | unsafe { simd_bitmask::<i8x32, _>(simd_lt(x:a.as_i8x32(), y:b.as_i8x32())) } |
| 2759 | } |
| 2760 | |
| 2761 | /// Compare packed signed 8-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2762 | /// |
| 2763 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmplt_epi8_mask&expand=1043) |
| 2764 | #[inline ] |
| 2765 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2766 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2767 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2768 | pub fn _mm256_mask_cmplt_epi8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 2769 | _mm256_mask_cmp_epi8_mask::<_MM_CMPINT_LT>(k1, a, b) |
| 2770 | } |
| 2771 | |
| 2772 | /// Compare packed signed 8-bit integers in a and b for less-than, and store the results in mask vector k. |
| 2773 | /// |
| 2774 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmplt_epi8_mask&expand=1040) |
| 2775 | #[inline ] |
| 2776 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2777 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2778 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2779 | pub fn _mm_cmplt_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 2780 | unsafe { simd_bitmask::<i8x16, _>(simd_lt(x:a.as_i8x16(), y:b.as_i8x16())) } |
| 2781 | } |
| 2782 | |
| 2783 | /// Compare packed signed 8-bit integers in a and b for less-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2784 | /// |
| 2785 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmplt_epi8_mask&expand=1041) |
| 2786 | #[inline ] |
| 2787 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2788 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2789 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2790 | pub fn _mm_mask_cmplt_epi8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 2791 | _mm_mask_cmp_epi8_mask::<_MM_CMPINT_LT>(k1, a, b) |
| 2792 | } |
| 2793 | |
| 2794 | /// Compare packed unsigned 16-bit integers in a and b for greater-than, and store the results in mask vector k. |
| 2795 | /// |
| 2796 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpgt_epu16_mask&expand=927) |
| 2797 | #[inline ] |
| 2798 | #[target_feature (enable = "avx512bw" )] |
| 2799 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2800 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2801 | pub fn _mm512_cmpgt_epu16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 2802 | unsafe { simd_bitmask::<u16x32, _>(simd_gt(x:a.as_u16x32(), y:b.as_u16x32())) } |
| 2803 | } |
| 2804 | |
| 2805 | /// Compare packed unsigned 16-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2806 | /// |
| 2807 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpgt_epu16_mask&expand=928) |
| 2808 | #[inline ] |
| 2809 | #[target_feature (enable = "avx512bw" )] |
| 2810 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2811 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2812 | pub fn _mm512_mask_cmpgt_epu16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 2813 | _mm512_mask_cmp_epu16_mask::<_MM_CMPINT_NLE>(k1, a, b) |
| 2814 | } |
| 2815 | |
| 2816 | /// Compare packed unsigned 16-bit integers in a and b for greater-than, and store the results in mask vector k. |
| 2817 | /// |
| 2818 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpgt_epu16_mask&expand=925) |
| 2819 | #[inline ] |
| 2820 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2821 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2822 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2823 | pub fn _mm256_cmpgt_epu16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 2824 | unsafe { simd_bitmask::<u16x16, _>(simd_gt(x:a.as_u16x16(), y:b.as_u16x16())) } |
| 2825 | } |
| 2826 | |
| 2827 | /// Compare packed unsigned 16-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2828 | /// |
| 2829 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpgt_epu16_mask&expand=926) |
| 2830 | #[inline ] |
| 2831 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2832 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2833 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2834 | pub fn _mm256_mask_cmpgt_epu16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 2835 | _mm256_mask_cmp_epu16_mask::<_MM_CMPINT_NLE>(k1, a, b) |
| 2836 | } |
| 2837 | |
| 2838 | /// Compare packed unsigned 16-bit integers in a and b for greater-than, and store the results in mask vector k. |
| 2839 | /// |
| 2840 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpgt_epu16_mask&expand=923) |
| 2841 | #[inline ] |
| 2842 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2843 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2844 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2845 | pub fn _mm_cmpgt_epu16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 2846 | unsafe { simd_bitmask::<u16x8, _>(simd_gt(x:a.as_u16x8(), y:b.as_u16x8())) } |
| 2847 | } |
| 2848 | |
| 2849 | /// Compare packed unsigned 16-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2850 | /// |
| 2851 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpgt_epu16_mask&expand=924) |
| 2852 | #[inline ] |
| 2853 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2854 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2855 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2856 | pub fn _mm_mask_cmpgt_epu16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 2857 | _mm_mask_cmp_epu16_mask::<_MM_CMPINT_NLE>(k1, a, b) |
| 2858 | } |
| 2859 | |
| 2860 | /// Compare packed unsigned 8-bit integers in a and b for greater-than, and store the results in mask vector k. |
| 2861 | /// |
| 2862 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpgt_epu8_mask&expand=945) |
| 2863 | #[inline ] |
| 2864 | #[target_feature (enable = "avx512bw" )] |
| 2865 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2866 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2867 | pub fn _mm512_cmpgt_epu8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 2868 | unsafe { simd_bitmask::<u8x64, _>(simd_gt(x:a.as_u8x64(), y:b.as_u8x64())) } |
| 2869 | } |
| 2870 | |
| 2871 | /// Compare packed unsigned 8-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2872 | /// |
| 2873 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpgt_epu8_mask&expand=946) |
| 2874 | #[inline ] |
| 2875 | #[target_feature (enable = "avx512bw" )] |
| 2876 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2877 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2878 | pub fn _mm512_mask_cmpgt_epu8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 2879 | _mm512_mask_cmp_epu8_mask::<_MM_CMPINT_NLE>(k1, a, b) |
| 2880 | } |
| 2881 | |
| 2882 | /// Compare packed unsigned 8-bit integers in a and b for greater-than, and store the results in mask vector k. |
| 2883 | /// |
| 2884 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpgt_epu8_mask&expand=943) |
| 2885 | #[inline ] |
| 2886 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2887 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2888 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2889 | pub fn _mm256_cmpgt_epu8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 2890 | unsafe { simd_bitmask::<u8x32, _>(simd_gt(x:a.as_u8x32(), y:b.as_u8x32())) } |
| 2891 | } |
| 2892 | |
| 2893 | /// Compare packed unsigned 8-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2894 | /// |
| 2895 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpgt_epu8_mask&expand=944) |
| 2896 | #[inline ] |
| 2897 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2898 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2899 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2900 | pub fn _mm256_mask_cmpgt_epu8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 2901 | _mm256_mask_cmp_epu8_mask::<_MM_CMPINT_NLE>(k1, a, b) |
| 2902 | } |
| 2903 | |
| 2904 | /// Compare packed unsigned 8-bit integers in a and b for greater-than, and store the results in mask vector k. |
| 2905 | /// |
| 2906 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpgt_epu8_mask&expand=941) |
| 2907 | #[inline ] |
| 2908 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2909 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2910 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2911 | pub fn _mm_cmpgt_epu8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 2912 | unsafe { simd_bitmask::<u8x16, _>(simd_gt(x:a.as_u8x16(), y:b.as_u8x16())) } |
| 2913 | } |
| 2914 | |
| 2915 | /// Compare packed unsigned 8-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2916 | /// |
| 2917 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpgt_epu8_mask&expand=942) |
| 2918 | #[inline ] |
| 2919 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2920 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2921 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2922 | pub fn _mm_mask_cmpgt_epu8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 2923 | _mm_mask_cmp_epu8_mask::<_MM_CMPINT_NLE>(k1, a, b) |
| 2924 | } |
| 2925 | |
| 2926 | /// Compare packed signed 16-bit integers in a and b for greater-than, and store the results in mask vector k. |
| 2927 | /// |
| 2928 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpgt_epi16_mask&expand=897) |
| 2929 | #[inline ] |
| 2930 | #[target_feature (enable = "avx512bw" )] |
| 2931 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2932 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2933 | pub fn _mm512_cmpgt_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 2934 | unsafe { simd_bitmask::<i16x32, _>(simd_gt(x:a.as_i16x32(), y:b.as_i16x32())) } |
| 2935 | } |
| 2936 | |
| 2937 | /// Compare packed signed 16-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2938 | /// |
| 2939 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpgt_epi16_mask&expand=898) |
| 2940 | #[inline ] |
| 2941 | #[target_feature (enable = "avx512bw" )] |
| 2942 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2943 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2944 | pub fn _mm512_mask_cmpgt_epi16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 2945 | _mm512_mask_cmp_epi16_mask::<_MM_CMPINT_NLE>(k1, a, b) |
| 2946 | } |
| 2947 | |
| 2948 | /// Compare packed signed 16-bit integers in a and b for greater-than, and store the results in mask vector k. |
| 2949 | /// |
| 2950 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpgt_epi16_mask&expand=895) |
| 2951 | #[inline ] |
| 2952 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2953 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2954 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2955 | pub fn _mm256_cmpgt_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 2956 | unsafe { simd_bitmask::<i16x16, _>(simd_gt(x:a.as_i16x16(), y:b.as_i16x16())) } |
| 2957 | } |
| 2958 | |
| 2959 | /// Compare packed signed 16-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2960 | /// |
| 2961 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpgt_epi16_mask&expand=896) |
| 2962 | #[inline ] |
| 2963 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2964 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2965 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2966 | pub fn _mm256_mask_cmpgt_epi16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 2967 | _mm256_mask_cmp_epi16_mask::<_MM_CMPINT_NLE>(k1, a, b) |
| 2968 | } |
| 2969 | |
| 2970 | /// Compare packed signed 16-bit integers in a and b for greater-than, and store the results in mask vector k. |
| 2971 | /// |
| 2972 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpgt_epi16_mask&expand=893) |
| 2973 | #[inline ] |
| 2974 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2975 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2976 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2977 | pub fn _mm_cmpgt_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 2978 | unsafe { simd_bitmask::<i16x8, _>(simd_gt(x:a.as_i16x8(), y:b.as_i16x8())) } |
| 2979 | } |
| 2980 | |
| 2981 | /// Compare packed signed 16-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 2982 | /// |
| 2983 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpgt_epi16_mask&expand=894) |
| 2984 | #[inline ] |
| 2985 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 2986 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2987 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2988 | pub fn _mm_mask_cmpgt_epi16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 2989 | _mm_mask_cmp_epi16_mask::<_MM_CMPINT_NLE>(k1, a, b) |
| 2990 | } |
| 2991 | |
| 2992 | /// Compare packed signed 8-bit integers in a and b for greater-than, and store the results in mask vector k. |
| 2993 | /// |
| 2994 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpgt_epi8_mask&expand=921) |
| 2995 | #[inline ] |
| 2996 | #[target_feature (enable = "avx512bw" )] |
| 2997 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 2998 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 2999 | pub fn _mm512_cmpgt_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 3000 | unsafe { simd_bitmask::<i8x64, _>(simd_gt(x:a.as_i8x64(), y:b.as_i8x64())) } |
| 3001 | } |
| 3002 | |
| 3003 | /// Compare packed signed 8-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3004 | /// |
| 3005 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpgt_epi8_mask&expand=922) |
| 3006 | #[inline ] |
| 3007 | #[target_feature (enable = "avx512bw" )] |
| 3008 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3009 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3010 | pub fn _mm512_mask_cmpgt_epi8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 3011 | _mm512_mask_cmp_epi8_mask::<_MM_CMPINT_NLE>(k1, a, b) |
| 3012 | } |
| 3013 | |
| 3014 | /// Compare packed signed 8-bit integers in a and b for greater-than, and store the results in mask vector k. |
| 3015 | /// |
| 3016 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpgt_epi8_mask&expand=919) |
| 3017 | #[inline ] |
| 3018 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3019 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3020 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3021 | pub fn _mm256_cmpgt_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 3022 | unsafe { simd_bitmask::<i8x32, _>(simd_gt(x:a.as_i8x32(), y:b.as_i8x32())) } |
| 3023 | } |
| 3024 | |
| 3025 | /// Compare packed signed 8-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3026 | /// |
| 3027 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpgt_epi8_mask&expand=920) |
| 3028 | #[inline ] |
| 3029 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3030 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3031 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3032 | pub fn _mm256_mask_cmpgt_epi8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 3033 | _mm256_mask_cmp_epi8_mask::<_MM_CMPINT_NLE>(k1, a, b) |
| 3034 | } |
| 3035 | |
| 3036 | /// Compare packed signed 8-bit integers in a and b for greater-than, and store the results in mask vector k. |
| 3037 | /// |
| 3038 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpgt_epi8_mask&expand=917) |
| 3039 | #[inline ] |
| 3040 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3041 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3042 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3043 | pub fn _mm_cmpgt_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 3044 | unsafe { simd_bitmask::<i8x16, _>(simd_gt(x:a.as_i8x16(), y:b.as_i8x16())) } |
| 3045 | } |
| 3046 | |
| 3047 | /// Compare packed signed 8-bit integers in a and b for greater-than, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3048 | /// |
| 3049 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpgt_epi8_mask&expand=918) |
| 3050 | #[inline ] |
| 3051 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3052 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3053 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3054 | pub fn _mm_mask_cmpgt_epi8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 3055 | _mm_mask_cmp_epi8_mask::<_MM_CMPINT_NLE>(k1, a, b) |
| 3056 | } |
| 3057 | |
| 3058 | /// Compare packed unsigned 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
| 3059 | /// |
| 3060 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmple_epu16_mask&expand=989) |
| 3061 | #[inline ] |
| 3062 | #[target_feature (enable = "avx512bw" )] |
| 3063 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3064 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3065 | pub fn _mm512_cmple_epu16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 3066 | unsafe { simd_bitmask::<u16x32, _>(simd_le(x:a.as_u16x32(), y:b.as_u16x32())) } |
| 3067 | } |
| 3068 | |
| 3069 | /// Compare packed unsigned 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3070 | /// |
| 3071 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmple_epu16_mask&expand=990) |
| 3072 | #[inline ] |
| 3073 | #[target_feature (enable = "avx512bw" )] |
| 3074 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3075 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3076 | pub fn _mm512_mask_cmple_epu16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 3077 | _mm512_mask_cmp_epu16_mask::<_MM_CMPINT_LE>(k1, a, b) |
| 3078 | } |
| 3079 | |
| 3080 | /// Compare packed unsigned 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
| 3081 | /// |
| 3082 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmple_epu16_mask&expand=987) |
| 3083 | #[inline ] |
| 3084 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3085 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3086 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3087 | pub fn _mm256_cmple_epu16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 3088 | unsafe { simd_bitmask::<u16x16, _>(simd_le(x:a.as_u16x16(), y:b.as_u16x16())) } |
| 3089 | } |
| 3090 | |
| 3091 | /// Compare packed unsigned 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3092 | /// |
| 3093 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmple_epu16_mask&expand=988) |
| 3094 | #[inline ] |
| 3095 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3096 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3097 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3098 | pub fn _mm256_mask_cmple_epu16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 3099 | _mm256_mask_cmp_epu16_mask::<_MM_CMPINT_LE>(k1, a, b) |
| 3100 | } |
| 3101 | |
| 3102 | /// Compare packed unsigned 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
| 3103 | /// |
| 3104 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmple_epu16_mask&expand=985) |
| 3105 | #[inline ] |
| 3106 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3107 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3108 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3109 | pub fn _mm_cmple_epu16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 3110 | unsafe { simd_bitmask::<u16x8, _>(simd_le(x:a.as_u16x8(), y:b.as_u16x8())) } |
| 3111 | } |
| 3112 | |
| 3113 | /// Compare packed unsigned 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3114 | /// |
| 3115 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmple_epu16_mask&expand=986) |
| 3116 | #[inline ] |
| 3117 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3118 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3119 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3120 | pub fn _mm_mask_cmple_epu16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 3121 | _mm_mask_cmp_epu16_mask::<_MM_CMPINT_LE>(k1, a, b) |
| 3122 | } |
| 3123 | |
| 3124 | /// Compare packed unsigned 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
| 3125 | /// |
| 3126 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmple_epu8_mask&expand=1007) |
| 3127 | #[inline ] |
| 3128 | #[target_feature (enable = "avx512bw" )] |
| 3129 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3130 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3131 | pub fn _mm512_cmple_epu8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 3132 | unsafe { simd_bitmask::<u8x64, _>(simd_le(x:a.as_u8x64(), y:b.as_u8x64())) } |
| 3133 | } |
| 3134 | |
| 3135 | /// Compare packed unsigned 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3136 | /// |
| 3137 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmple_epu8_mask&expand=1008) |
| 3138 | #[inline ] |
| 3139 | #[target_feature (enable = "avx512bw" )] |
| 3140 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3141 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3142 | pub fn _mm512_mask_cmple_epu8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 3143 | _mm512_mask_cmp_epu8_mask::<_MM_CMPINT_LE>(k1, a, b) |
| 3144 | } |
| 3145 | |
| 3146 | /// Compare packed unsigned 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
| 3147 | /// |
| 3148 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmple_epu8_mask&expand=1005) |
| 3149 | #[inline ] |
| 3150 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3151 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3152 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3153 | pub fn _mm256_cmple_epu8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 3154 | unsafe { simd_bitmask::<u8x32, _>(simd_le(x:a.as_u8x32(), y:b.as_u8x32())) } |
| 3155 | } |
| 3156 | |
| 3157 | /// Compare packed unsigned 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3158 | /// |
| 3159 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmple_epu8_mask&expand=1006) |
| 3160 | #[inline ] |
| 3161 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3162 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3163 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3164 | pub fn _mm256_mask_cmple_epu8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 3165 | _mm256_mask_cmp_epu8_mask::<_MM_CMPINT_LE>(k1, a, b) |
| 3166 | } |
| 3167 | |
| 3168 | /// Compare packed unsigned 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
| 3169 | /// |
| 3170 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmple_epu8_mask&expand=1003) |
| 3171 | #[inline ] |
| 3172 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3173 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3174 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3175 | pub fn _mm_cmple_epu8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 3176 | unsafe { simd_bitmask::<u8x16, _>(simd_le(x:a.as_u8x16(), y:b.as_u8x16())) } |
| 3177 | } |
| 3178 | |
| 3179 | /// Compare packed unsigned 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3180 | /// |
| 3181 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmple_epu8_mask&expand=1004) |
| 3182 | #[inline ] |
| 3183 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3184 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3185 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3186 | pub fn _mm_mask_cmple_epu8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 3187 | _mm_mask_cmp_epu8_mask::<_MM_CMPINT_LE>(k1, a, b) |
| 3188 | } |
| 3189 | |
| 3190 | /// Compare packed signed 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
| 3191 | /// |
| 3192 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmple_epi16_mask&expand=965) |
| 3193 | #[inline ] |
| 3194 | #[target_feature (enable = "avx512bw" )] |
| 3195 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3196 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3197 | pub fn _mm512_cmple_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 3198 | unsafe { simd_bitmask::<i16x32, _>(simd_le(x:a.as_i16x32(), y:b.as_i16x32())) } |
| 3199 | } |
| 3200 | |
| 3201 | /// Compare packed signed 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3202 | /// |
| 3203 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmple_epi16_mask&expand=966) |
| 3204 | #[inline ] |
| 3205 | #[target_feature (enable = "avx512bw" )] |
| 3206 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3207 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3208 | pub fn _mm512_mask_cmple_epi16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 3209 | _mm512_mask_cmp_epi16_mask::<_MM_CMPINT_LE>(k1, a, b) |
| 3210 | } |
| 3211 | |
| 3212 | /// Compare packed signed 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
| 3213 | /// |
| 3214 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmple_epi16_mask&expand=963) |
| 3215 | #[inline ] |
| 3216 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3217 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3218 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3219 | pub fn _mm256_cmple_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 3220 | unsafe { simd_bitmask::<i16x16, _>(simd_le(x:a.as_i16x16(), y:b.as_i16x16())) } |
| 3221 | } |
| 3222 | |
| 3223 | /// Compare packed signed 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3224 | /// |
| 3225 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmple_epi16_mask&expand=964) |
| 3226 | #[inline ] |
| 3227 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3228 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3229 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3230 | pub fn _mm256_mask_cmple_epi16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 3231 | _mm256_mask_cmp_epi16_mask::<_MM_CMPINT_LE>(k1, a, b) |
| 3232 | } |
| 3233 | |
| 3234 | /// Compare packed signed 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
| 3235 | /// |
| 3236 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmple_epi16_mask&expand=961) |
| 3237 | #[inline ] |
| 3238 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3239 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3240 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3241 | pub fn _mm_cmple_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 3242 | unsafe { simd_bitmask::<i16x8, _>(simd_le(x:a.as_i16x8(), y:b.as_i16x8())) } |
| 3243 | } |
| 3244 | |
| 3245 | /// Compare packed signed 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3246 | /// |
| 3247 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmple_epi16_mask&expand=962) |
| 3248 | #[inline ] |
| 3249 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3250 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3251 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3252 | pub fn _mm_mask_cmple_epi16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 3253 | _mm_mask_cmp_epi16_mask::<_MM_CMPINT_LE>(k1, a, b) |
| 3254 | } |
| 3255 | |
| 3256 | /// Compare packed signed 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
| 3257 | /// |
| 3258 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmple_epi8_mask&expand=983) |
| 3259 | #[inline ] |
| 3260 | #[target_feature (enable = "avx512bw" )] |
| 3261 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3262 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3263 | pub fn _mm512_cmple_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 3264 | unsafe { simd_bitmask::<i8x64, _>(simd_le(x:a.as_i8x64(), y:b.as_i8x64())) } |
| 3265 | } |
| 3266 | |
| 3267 | /// Compare packed signed 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3268 | /// |
| 3269 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmple_epi8_mask&expand=984) |
| 3270 | #[inline ] |
| 3271 | #[target_feature (enable = "avx512bw" )] |
| 3272 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3273 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3274 | pub fn _mm512_mask_cmple_epi8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 3275 | _mm512_mask_cmp_epi8_mask::<_MM_CMPINT_LE>(k1, a, b) |
| 3276 | } |
| 3277 | |
| 3278 | /// Compare packed signed 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
| 3279 | /// |
| 3280 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmple_epi8_mask&expand=981) |
| 3281 | #[inline ] |
| 3282 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3283 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3284 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3285 | pub fn _mm256_cmple_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 3286 | unsafe { simd_bitmask::<i8x32, _>(simd_le(x:a.as_i8x32(), y:b.as_i8x32())) } |
| 3287 | } |
| 3288 | |
| 3289 | /// Compare packed signed 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3290 | /// |
| 3291 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmple_epi8_mask&expand=982) |
| 3292 | #[inline ] |
| 3293 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3294 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3295 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3296 | pub fn _mm256_mask_cmple_epi8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 3297 | _mm256_mask_cmp_epi8_mask::<_MM_CMPINT_LE>(k1, a, b) |
| 3298 | } |
| 3299 | |
| 3300 | /// Compare packed signed 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
| 3301 | /// |
| 3302 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmple_epi8_mask&expand=979) |
| 3303 | #[inline ] |
| 3304 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3305 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3306 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3307 | pub fn _mm_cmple_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 3308 | unsafe { simd_bitmask::<i8x16, _>(simd_le(x:a.as_i8x16(), y:b.as_i8x16())) } |
| 3309 | } |
| 3310 | |
| 3311 | /// Compare packed signed 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3312 | /// |
| 3313 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmple_epi8_mask&expand=980) |
| 3314 | #[inline ] |
| 3315 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3316 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3317 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3318 | pub fn _mm_mask_cmple_epi8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 3319 | _mm_mask_cmp_epi8_mask::<_MM_CMPINT_LE>(k1, a, b) |
| 3320 | } |
| 3321 | |
| 3322 | /// Compare packed unsigned 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
| 3323 | /// |
| 3324 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpge_epu16_mask&expand=867) |
| 3325 | #[inline ] |
| 3326 | #[target_feature (enable = "avx512bw" )] |
| 3327 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3328 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3329 | pub fn _mm512_cmpge_epu16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 3330 | unsafe { simd_bitmask::<u16x32, _>(simd_ge(x:a.as_u16x32(), y:b.as_u16x32())) } |
| 3331 | } |
| 3332 | |
| 3333 | /// Compare packed unsigned 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3334 | /// |
| 3335 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpge_epu16_mask&expand=868) |
| 3336 | #[inline ] |
| 3337 | #[target_feature (enable = "avx512bw" )] |
| 3338 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3339 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3340 | pub fn _mm512_mask_cmpge_epu16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 3341 | _mm512_mask_cmp_epu16_mask::<_MM_CMPINT_NLT>(k1, a, b) |
| 3342 | } |
| 3343 | |
| 3344 | /// Compare packed unsigned 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
| 3345 | /// |
| 3346 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpge_epu16_mask&expand=865) |
| 3347 | #[inline ] |
| 3348 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3349 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3350 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3351 | pub fn _mm256_cmpge_epu16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 3352 | unsafe { simd_bitmask::<u16x16, _>(simd_ge(x:a.as_u16x16(), y:b.as_u16x16())) } |
| 3353 | } |
| 3354 | |
| 3355 | /// Compare packed unsigned 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3356 | /// |
| 3357 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpge_epu16_mask&expand=866) |
| 3358 | #[inline ] |
| 3359 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3360 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3361 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3362 | pub fn _mm256_mask_cmpge_epu16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 3363 | _mm256_mask_cmp_epu16_mask::<_MM_CMPINT_NLT>(k1, a, b) |
| 3364 | } |
| 3365 | |
| 3366 | /// Compare packed unsigned 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
| 3367 | /// |
| 3368 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpge_epu16_mask&expand=863) |
| 3369 | #[inline ] |
| 3370 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3371 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3372 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3373 | pub fn _mm_cmpge_epu16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 3374 | unsafe { simd_bitmask::<u16x8, _>(simd_ge(x:a.as_u16x8(), y:b.as_u16x8())) } |
| 3375 | } |
| 3376 | |
| 3377 | /// Compare packed unsigned 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3378 | /// |
| 3379 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpge_epu16_mask&expand=864) |
| 3380 | #[inline ] |
| 3381 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3382 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3383 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3384 | pub fn _mm_mask_cmpge_epu16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 3385 | _mm_mask_cmp_epu16_mask::<_MM_CMPINT_NLT>(k1, a, b) |
| 3386 | } |
| 3387 | |
| 3388 | /// Compare packed unsigned 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
| 3389 | /// |
| 3390 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpge_epu8_mask&expand=885) |
| 3391 | #[inline ] |
| 3392 | #[target_feature (enable = "avx512bw" )] |
| 3393 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3394 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3395 | pub fn _mm512_cmpge_epu8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 3396 | unsafe { simd_bitmask::<u8x64, _>(simd_ge(x:a.as_u8x64(), y:b.as_u8x64())) } |
| 3397 | } |
| 3398 | |
| 3399 | /// Compare packed unsigned 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3400 | /// |
| 3401 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpge_epu8_mask&expand=886) |
| 3402 | #[inline ] |
| 3403 | #[target_feature (enable = "avx512bw" )] |
| 3404 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3405 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3406 | pub fn _mm512_mask_cmpge_epu8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 3407 | _mm512_mask_cmp_epu8_mask::<_MM_CMPINT_NLT>(k1, a, b) |
| 3408 | } |
| 3409 | |
| 3410 | /// Compare packed unsigned 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
| 3411 | /// |
| 3412 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpge_epu8_mask&expand=883) |
| 3413 | #[inline ] |
| 3414 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3415 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3416 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3417 | pub fn _mm256_cmpge_epu8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 3418 | unsafe { simd_bitmask::<u8x32, _>(simd_ge(x:a.as_u8x32(), y:b.as_u8x32())) } |
| 3419 | } |
| 3420 | |
| 3421 | /// Compare packed unsigned 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3422 | /// |
| 3423 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpge_epu8_mask&expand=884) |
| 3424 | #[inline ] |
| 3425 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3426 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3427 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3428 | pub fn _mm256_mask_cmpge_epu8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 3429 | _mm256_mask_cmp_epu8_mask::<_MM_CMPINT_NLT>(k1, a, b) |
| 3430 | } |
| 3431 | |
| 3432 | /// Compare packed unsigned 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
| 3433 | /// |
| 3434 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpge_epu8_mask&expand=881) |
| 3435 | #[inline ] |
| 3436 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3437 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3438 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3439 | pub fn _mm_cmpge_epu8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 3440 | unsafe { simd_bitmask::<u8x16, _>(simd_ge(x:a.as_u8x16(), y:b.as_u8x16())) } |
| 3441 | } |
| 3442 | |
| 3443 | /// Compare packed unsigned 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3444 | /// |
| 3445 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpge_epu8_mask&expand=882) |
| 3446 | #[inline ] |
| 3447 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3448 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3449 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3450 | pub fn _mm_mask_cmpge_epu8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 3451 | _mm_mask_cmp_epu8_mask::<_MM_CMPINT_NLT>(k1, a, b) |
| 3452 | } |
| 3453 | |
| 3454 | /// Compare packed signed 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
| 3455 | /// |
| 3456 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpge_epi16_mask&expand=843) |
| 3457 | #[inline ] |
| 3458 | #[target_feature (enable = "avx512bw" )] |
| 3459 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3460 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3461 | pub fn _mm512_cmpge_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 3462 | unsafe { simd_bitmask::<i16x32, _>(simd_ge(x:a.as_i16x32(), y:b.as_i16x32())) } |
| 3463 | } |
| 3464 | |
| 3465 | /// Compare packed signed 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3466 | /// |
| 3467 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpge_epi16_mask&expand=844) |
| 3468 | #[inline ] |
| 3469 | #[target_feature (enable = "avx512bw" )] |
| 3470 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3471 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3472 | pub fn _mm512_mask_cmpge_epi16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 3473 | _mm512_mask_cmp_epi16_mask::<_MM_CMPINT_NLT>(k1, a, b) |
| 3474 | } |
| 3475 | |
| 3476 | /// Compare packed signed 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
| 3477 | /// |
| 3478 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpge_epi16_mask&expand=841) |
| 3479 | #[inline ] |
| 3480 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3481 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3482 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3483 | pub fn _mm256_cmpge_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 3484 | unsafe { simd_bitmask::<i16x16, _>(simd_ge(x:a.as_i16x16(), y:b.as_i16x16())) } |
| 3485 | } |
| 3486 | |
| 3487 | /// Compare packed signed 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3488 | /// |
| 3489 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpge_epi16_mask&expand=842) |
| 3490 | #[inline ] |
| 3491 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3492 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3493 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3494 | pub fn _mm256_mask_cmpge_epi16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 3495 | _mm256_mask_cmp_epi16_mask::<_MM_CMPINT_NLT>(k1, a, b) |
| 3496 | } |
| 3497 | |
| 3498 | /// Compare packed signed 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
| 3499 | /// |
| 3500 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpge_epi16_mask&expand=839) |
| 3501 | #[inline ] |
| 3502 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3503 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3504 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3505 | pub fn _mm_cmpge_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 3506 | unsafe { simd_bitmask::<i16x8, _>(simd_ge(x:a.as_i16x8(), y:b.as_i16x8())) } |
| 3507 | } |
| 3508 | |
| 3509 | /// Compare packed signed 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3510 | /// |
| 3511 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpge_epi16_mask&expand=840) |
| 3512 | #[inline ] |
| 3513 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3514 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3515 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3516 | pub fn _mm_mask_cmpge_epi16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 3517 | _mm_mask_cmp_epi16_mask::<_MM_CMPINT_NLT>(k1, a, b) |
| 3518 | } |
| 3519 | |
| 3520 | /// Compare packed signed 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
| 3521 | /// |
| 3522 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpge_epi8_mask&expand=861) |
| 3523 | #[inline ] |
| 3524 | #[target_feature (enable = "avx512bw" )] |
| 3525 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3526 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3527 | pub fn _mm512_cmpge_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 3528 | unsafe { simd_bitmask::<i8x64, _>(simd_ge(x:a.as_i8x64(), y:b.as_i8x64())) } |
| 3529 | } |
| 3530 | |
| 3531 | /// Compare packed signed 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3532 | /// |
| 3533 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpge_epi8_mask&expand=862) |
| 3534 | #[inline ] |
| 3535 | #[target_feature (enable = "avx512bw" )] |
| 3536 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3537 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3538 | pub fn _mm512_mask_cmpge_epi8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 3539 | _mm512_mask_cmp_epi8_mask::<_MM_CMPINT_NLT>(k1, a, b) |
| 3540 | } |
| 3541 | |
| 3542 | /// Compare packed signed 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
| 3543 | /// |
| 3544 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpge_epi8_mask&expand=859) |
| 3545 | #[inline ] |
| 3546 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3547 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3548 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3549 | pub fn _mm256_cmpge_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 3550 | unsafe { simd_bitmask::<i8x32, _>(simd_ge(x:a.as_i8x32(), y:b.as_i8x32())) } |
| 3551 | } |
| 3552 | |
| 3553 | /// Compare packed signed 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3554 | /// |
| 3555 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpge_epi8_mask&expand=860) |
| 3556 | #[inline ] |
| 3557 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3558 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3559 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3560 | pub fn _mm256_mask_cmpge_epi8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 3561 | _mm256_mask_cmp_epi8_mask::<_MM_CMPINT_NLT>(k1, a, b) |
| 3562 | } |
| 3563 | |
| 3564 | /// Compare packed signed 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
| 3565 | /// |
| 3566 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpge_epi8_mask&expand=857) |
| 3567 | #[inline ] |
| 3568 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3569 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3570 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3571 | pub fn _mm_cmpge_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 3572 | unsafe { simd_bitmask::<i8x16, _>(simd_ge(x:a.as_i8x16(), y:b.as_i8x16())) } |
| 3573 | } |
| 3574 | |
| 3575 | /// Compare packed signed 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3576 | /// |
| 3577 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpge_epi8_mask&expand=858) |
| 3578 | #[inline ] |
| 3579 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3580 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3581 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3582 | pub fn _mm_mask_cmpge_epi8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 3583 | _mm_mask_cmp_epi8_mask::<_MM_CMPINT_NLT>(k1, a, b) |
| 3584 | } |
| 3585 | |
| 3586 | /// Compare packed unsigned 16-bit integers in a and b for equality, and store the results in mask vector k. |
| 3587 | /// |
| 3588 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpeq_epu16_mask&expand=801) |
| 3589 | #[inline ] |
| 3590 | #[target_feature (enable = "avx512bw" )] |
| 3591 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3592 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3593 | pub fn _mm512_cmpeq_epu16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 3594 | unsafe { simd_bitmask::<u16x32, _>(simd_eq(x:a.as_u16x32(), y:b.as_u16x32())) } |
| 3595 | } |
| 3596 | |
| 3597 | /// Compare packed unsigned 16-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3598 | /// |
| 3599 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpeq_epu16_mask&expand=802) |
| 3600 | #[inline ] |
| 3601 | #[target_feature (enable = "avx512bw" )] |
| 3602 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3603 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3604 | pub fn _mm512_mask_cmpeq_epu16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 3605 | _mm512_mask_cmp_epu16_mask::<_MM_CMPINT_EQ>(k1, a, b) |
| 3606 | } |
| 3607 | |
| 3608 | /// Compare packed unsigned 16-bit integers in a and b for equality, and store the results in mask vector k. |
| 3609 | /// |
| 3610 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpeq_epu16_mask&expand=799) |
| 3611 | #[inline ] |
| 3612 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3613 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3614 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3615 | pub fn _mm256_cmpeq_epu16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 3616 | unsafe { simd_bitmask::<u16x16, _>(simd_eq(x:a.as_u16x16(), y:b.as_u16x16())) } |
| 3617 | } |
| 3618 | |
| 3619 | /// Compare packed unsigned 16-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3620 | /// |
| 3621 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpeq_epu16_mask&expand=800) |
| 3622 | #[inline ] |
| 3623 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3624 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3625 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3626 | pub fn _mm256_mask_cmpeq_epu16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 3627 | _mm256_mask_cmp_epu16_mask::<_MM_CMPINT_EQ>(k1, a, b) |
| 3628 | } |
| 3629 | |
| 3630 | /// Compare packed unsigned 16-bit integers in a and b for equality, and store the results in mask vector k. |
| 3631 | /// |
| 3632 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpeq_epu16_mask&expand=797) |
| 3633 | #[inline ] |
| 3634 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3635 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3636 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3637 | pub fn _mm_cmpeq_epu16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 3638 | unsafe { simd_bitmask::<u16x8, _>(simd_eq(x:a.as_u16x8(), y:b.as_u16x8())) } |
| 3639 | } |
| 3640 | |
| 3641 | /// Compare packed unsigned 16-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3642 | /// |
| 3643 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpeq_epu16_mask&expand=798) |
| 3644 | #[inline ] |
| 3645 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3646 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3647 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3648 | pub fn _mm_mask_cmpeq_epu16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 3649 | _mm_mask_cmp_epu16_mask::<_MM_CMPINT_EQ>(k1, a, b) |
| 3650 | } |
| 3651 | |
| 3652 | /// Compare packed unsigned 8-bit integers in a and b for equality, and store the results in mask vector k. |
| 3653 | /// |
| 3654 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpeq_epu8_mask&expand=819) |
| 3655 | #[inline ] |
| 3656 | #[target_feature (enable = "avx512bw" )] |
| 3657 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3658 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3659 | pub fn _mm512_cmpeq_epu8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 3660 | unsafe { simd_bitmask::<u8x64, _>(simd_eq(x:a.as_u8x64(), y:b.as_u8x64())) } |
| 3661 | } |
| 3662 | |
| 3663 | /// Compare packed unsigned 8-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3664 | /// |
| 3665 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpeq_epu8_mask&expand=820) |
| 3666 | #[inline ] |
| 3667 | #[target_feature (enable = "avx512bw" )] |
| 3668 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3669 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3670 | pub fn _mm512_mask_cmpeq_epu8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 3671 | _mm512_mask_cmp_epu8_mask::<_MM_CMPINT_EQ>(k1, a, b) |
| 3672 | } |
| 3673 | |
| 3674 | /// Compare packed unsigned 8-bit integers in a and b for equality, and store the results in mask vector k. |
| 3675 | /// |
| 3676 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpeq_epu8_mask&expand=817) |
| 3677 | #[inline ] |
| 3678 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3679 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3680 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3681 | pub fn _mm256_cmpeq_epu8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 3682 | unsafe { simd_bitmask::<u8x32, _>(simd_eq(x:a.as_u8x32(), y:b.as_u8x32())) } |
| 3683 | } |
| 3684 | |
| 3685 | /// Compare packed unsigned 8-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3686 | /// |
| 3687 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpeq_epu8_mask&expand=818) |
| 3688 | #[inline ] |
| 3689 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3690 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3691 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3692 | pub fn _mm256_mask_cmpeq_epu8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 3693 | _mm256_mask_cmp_epu8_mask::<_MM_CMPINT_EQ>(k1, a, b) |
| 3694 | } |
| 3695 | |
| 3696 | /// Compare packed unsigned 8-bit integers in a and b for equality, and store the results in mask vector k. |
| 3697 | /// |
| 3698 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpeq_epu8_mask&expand=815) |
| 3699 | #[inline ] |
| 3700 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3701 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3702 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3703 | pub fn _mm_cmpeq_epu8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 3704 | unsafe { simd_bitmask::<u8x16, _>(simd_eq(x:a.as_u8x16(), y:b.as_u8x16())) } |
| 3705 | } |
| 3706 | |
| 3707 | /// Compare packed unsigned 8-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3708 | /// |
| 3709 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpeq_epu8_mask&expand=816) |
| 3710 | #[inline ] |
| 3711 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3712 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3713 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3714 | pub fn _mm_mask_cmpeq_epu8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 3715 | _mm_mask_cmp_epu8_mask::<_MM_CMPINT_EQ>(k1, a, b) |
| 3716 | } |
| 3717 | |
| 3718 | /// Compare packed signed 16-bit integers in a and b for equality, and store the results in mask vector k. |
| 3719 | /// |
| 3720 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpeq_epi16_mask&expand=771) |
| 3721 | #[inline ] |
| 3722 | #[target_feature (enable = "avx512bw" )] |
| 3723 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3724 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3725 | pub fn _mm512_cmpeq_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 3726 | unsafe { simd_bitmask::<i16x32, _>(simd_eq(x:a.as_i16x32(), y:b.as_i16x32())) } |
| 3727 | } |
| 3728 | |
| 3729 | /// Compare packed signed 16-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3730 | /// |
| 3731 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpeq_epi16_mask&expand=772) |
| 3732 | #[inline ] |
| 3733 | #[target_feature (enable = "avx512bw" )] |
| 3734 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3735 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3736 | pub fn _mm512_mask_cmpeq_epi16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 3737 | _mm512_mask_cmp_epi16_mask::<_MM_CMPINT_EQ>(k1, a, b) |
| 3738 | } |
| 3739 | |
| 3740 | /// Compare packed signed 16-bit integers in a and b for equality, and store the results in mask vector k. |
| 3741 | /// |
| 3742 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpeq_epi16_mask&expand=769) |
| 3743 | #[inline ] |
| 3744 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3745 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3746 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3747 | pub fn _mm256_cmpeq_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 3748 | unsafe { simd_bitmask::<i16x16, _>(simd_eq(x:a.as_i16x16(), y:b.as_i16x16())) } |
| 3749 | } |
| 3750 | |
| 3751 | /// Compare packed signed 16-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3752 | /// |
| 3753 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpeq_epi16_mask&expand=770) |
| 3754 | #[inline ] |
| 3755 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3756 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3757 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3758 | pub fn _mm256_mask_cmpeq_epi16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 3759 | _mm256_mask_cmp_epi16_mask::<_MM_CMPINT_EQ>(k1, a, b) |
| 3760 | } |
| 3761 | |
| 3762 | /// Compare packed signed 16-bit integers in a and b for equality, and store the results in mask vector k. |
| 3763 | /// |
| 3764 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpeq_epi16_mask&expand=767) |
| 3765 | #[inline ] |
| 3766 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3767 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3768 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3769 | pub fn _mm_cmpeq_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 3770 | unsafe { simd_bitmask::<i16x8, _>(simd_eq(x:a.as_i16x8(), y:b.as_i16x8())) } |
| 3771 | } |
| 3772 | |
| 3773 | /// Compare packed signed 16-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3774 | /// |
| 3775 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpeq_epi16_mask&expand=768) |
| 3776 | #[inline ] |
| 3777 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3778 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3779 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3780 | pub fn _mm_mask_cmpeq_epi16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 3781 | _mm_mask_cmp_epi16_mask::<_MM_CMPINT_EQ>(k1, a, b) |
| 3782 | } |
| 3783 | |
| 3784 | /// Compare packed signed 8-bit integers in a and b for equality, and store the results in mask vector k. |
| 3785 | /// |
| 3786 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpeq_epi8_mask&expand=795) |
| 3787 | #[inline ] |
| 3788 | #[target_feature (enable = "avx512bw" )] |
| 3789 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3790 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3791 | pub fn _mm512_cmpeq_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 3792 | unsafe { simd_bitmask::<i8x64, _>(simd_eq(x:a.as_i8x64(), y:b.as_i8x64())) } |
| 3793 | } |
| 3794 | |
| 3795 | /// Compare packed signed 8-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3796 | /// |
| 3797 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpeq_epi8_mask&expand=796) |
| 3798 | #[inline ] |
| 3799 | #[target_feature (enable = "avx512bw" )] |
| 3800 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3801 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3802 | pub fn _mm512_mask_cmpeq_epi8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 3803 | _mm512_mask_cmp_epi8_mask::<_MM_CMPINT_EQ>(k1, a, b) |
| 3804 | } |
| 3805 | |
| 3806 | /// Compare packed signed 8-bit integers in a and b for equality, and store the results in mask vector k. |
| 3807 | /// |
| 3808 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpeq_epi8_mask&expand=793) |
| 3809 | #[inline ] |
| 3810 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3811 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3812 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3813 | pub fn _mm256_cmpeq_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 3814 | unsafe { simd_bitmask::<i8x32, _>(simd_eq(x:a.as_i8x32(), y:b.as_i8x32())) } |
| 3815 | } |
| 3816 | |
| 3817 | /// Compare packed signed 8-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3818 | /// |
| 3819 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpeq_epi8_mask&expand=794) |
| 3820 | #[inline ] |
| 3821 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3822 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3823 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3824 | pub fn _mm256_mask_cmpeq_epi8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 3825 | _mm256_mask_cmp_epi8_mask::<_MM_CMPINT_EQ>(k1, a, b) |
| 3826 | } |
| 3827 | |
| 3828 | /// Compare packed signed 8-bit integers in a and b for equality, and store the results in mask vector k. |
| 3829 | /// |
| 3830 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpeq_epi8_mask&expand=791) |
| 3831 | #[inline ] |
| 3832 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3833 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3834 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3835 | pub fn _mm_cmpeq_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 3836 | unsafe { simd_bitmask::<i8x16, _>(simd_eq(x:a.as_i8x16(), y:b.as_i8x16())) } |
| 3837 | } |
| 3838 | |
| 3839 | /// Compare packed signed 8-bit integers in a and b for equality, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3840 | /// |
| 3841 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpeq_epi8_mask&expand=792) |
| 3842 | #[inline ] |
| 3843 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3844 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3845 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3846 | pub fn _mm_mask_cmpeq_epi8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 3847 | _mm_mask_cmp_epi8_mask::<_MM_CMPINT_EQ>(k1, a, b) |
| 3848 | } |
| 3849 | |
| 3850 | /// Compare packed unsigned 16-bit integers in a and b for not-equal, and store the results in mask vector k. |
| 3851 | /// |
| 3852 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpneq_epu16_mask&expand=1106) |
| 3853 | #[inline ] |
| 3854 | #[target_feature (enable = "avx512bw" )] |
| 3855 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3856 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3857 | pub fn _mm512_cmpneq_epu16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 3858 | unsafe { simd_bitmask::<u16x32, _>(simd_ne(x:a.as_u16x32(), y:b.as_u16x32())) } |
| 3859 | } |
| 3860 | |
| 3861 | /// Compare packed unsigned 16-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3862 | /// |
| 3863 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpneq_epu16_mask&expand=1107) |
| 3864 | #[inline ] |
| 3865 | #[target_feature (enable = "avx512bw" )] |
| 3866 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3867 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3868 | pub fn _mm512_mask_cmpneq_epu16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 3869 | _mm512_mask_cmp_epu16_mask::<_MM_CMPINT_NE>(k1, a, b) |
| 3870 | } |
| 3871 | |
| 3872 | /// Compare packed unsigned 16-bit integers in a and b for not-equal, and store the results in mask vector k. |
| 3873 | /// |
| 3874 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpneq_epu16_mask&expand=1104) |
| 3875 | #[inline ] |
| 3876 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3877 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3878 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3879 | pub fn _mm256_cmpneq_epu16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 3880 | unsafe { simd_bitmask::<u16x16, _>(simd_ne(x:a.as_u16x16(), y:b.as_u16x16())) } |
| 3881 | } |
| 3882 | |
| 3883 | /// Compare packed unsigned 16-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3884 | /// |
| 3885 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpneq_epu16_mask&expand=1105) |
| 3886 | #[inline ] |
| 3887 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3888 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3889 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3890 | pub fn _mm256_mask_cmpneq_epu16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 3891 | _mm256_mask_cmp_epu16_mask::<_MM_CMPINT_NE>(k1, a, b) |
| 3892 | } |
| 3893 | |
| 3894 | /// Compare packed unsigned 16-bit integers in a and b for not-equal, and store the results in mask vector k. |
| 3895 | /// |
| 3896 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpneq_epu16_mask&expand=1102) |
| 3897 | #[inline ] |
| 3898 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3899 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3900 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3901 | pub fn _mm_cmpneq_epu16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 3902 | unsafe { simd_bitmask::<u16x8, _>(simd_ne(x:a.as_u16x8(), y:b.as_u16x8())) } |
| 3903 | } |
| 3904 | |
| 3905 | /// Compare packed unsigned 16-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3906 | /// |
| 3907 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpneq_epu16_mask&expand=1103) |
| 3908 | #[inline ] |
| 3909 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3910 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3911 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3912 | pub fn _mm_mask_cmpneq_epu16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 3913 | _mm_mask_cmp_epu16_mask::<_MM_CMPINT_NE>(k1, a, b) |
| 3914 | } |
| 3915 | |
| 3916 | /// Compare packed unsigned 8-bit integers in a and b for not-equal, and store the results in mask vector k. |
| 3917 | /// |
| 3918 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpneq_epu8_mask&expand=1124) |
| 3919 | #[inline ] |
| 3920 | #[target_feature (enable = "avx512bw" )] |
| 3921 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3922 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3923 | pub fn _mm512_cmpneq_epu8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 3924 | unsafe { simd_bitmask::<u8x64, _>(simd_ne(x:a.as_u8x64(), y:b.as_u8x64())) } |
| 3925 | } |
| 3926 | |
| 3927 | /// Compare packed unsigned 8-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3928 | /// |
| 3929 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpneq_epu8_mask&expand=1125) |
| 3930 | #[inline ] |
| 3931 | #[target_feature (enable = "avx512bw" )] |
| 3932 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3933 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3934 | pub fn _mm512_mask_cmpneq_epu8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 3935 | _mm512_mask_cmp_epu8_mask::<_MM_CMPINT_NE>(k1, a, b) |
| 3936 | } |
| 3937 | |
| 3938 | /// Compare packed unsigned 8-bit integers in a and b for not-equal, and store the results in mask vector k. |
| 3939 | /// |
| 3940 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpneq_epu8_mask&expand=1122) |
| 3941 | #[inline ] |
| 3942 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3943 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3944 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3945 | pub fn _mm256_cmpneq_epu8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 3946 | unsafe { simd_bitmask::<u8x32, _>(simd_ne(x:a.as_u8x32(), y:b.as_u8x32())) } |
| 3947 | } |
| 3948 | |
| 3949 | /// Compare packed unsigned 8-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3950 | /// |
| 3951 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpneq_epu8_mask&expand=1123) |
| 3952 | #[inline ] |
| 3953 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3954 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3955 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3956 | pub fn _mm256_mask_cmpneq_epu8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 3957 | _mm256_mask_cmp_epu8_mask::<_MM_CMPINT_NE>(k1, a, b) |
| 3958 | } |
| 3959 | |
| 3960 | /// Compare packed unsigned 8-bit integers in a and b for not-equal, and store the results in mask vector k. |
| 3961 | /// |
| 3962 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpneq_epu8_mask&expand=1120) |
| 3963 | #[inline ] |
| 3964 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3965 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3966 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3967 | pub fn _mm_cmpneq_epu8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 3968 | unsafe { simd_bitmask::<u8x16, _>(simd_ne(x:a.as_u8x16(), y:b.as_u8x16())) } |
| 3969 | } |
| 3970 | |
| 3971 | /// Compare packed unsigned 8-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3972 | /// |
| 3973 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpneq_epu8_mask&expand=1121) |
| 3974 | #[inline ] |
| 3975 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 3976 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3977 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3978 | pub fn _mm_mask_cmpneq_epu8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 3979 | _mm_mask_cmp_epu8_mask::<_MM_CMPINT_NE>(k1, a, b) |
| 3980 | } |
| 3981 | |
| 3982 | /// Compare packed signed 16-bit integers in a and b for not-equal, and store the results in mask vector k. |
| 3983 | /// |
| 3984 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpneq_epi16_mask&expand=1082) |
| 3985 | #[inline ] |
| 3986 | #[target_feature (enable = "avx512bw" )] |
| 3987 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3988 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 3989 | pub fn _mm512_cmpneq_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 3990 | unsafe { simd_bitmask::<i16x32, _>(simd_ne(x:a.as_i16x32(), y:b.as_i16x32())) } |
| 3991 | } |
| 3992 | |
| 3993 | /// Compare packed signed 16-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 3994 | /// |
| 3995 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpneq_epi16_mask&expand=1083) |
| 3996 | #[inline ] |
| 3997 | #[target_feature (enable = "avx512bw" )] |
| 3998 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 3999 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 4000 | pub fn _mm512_mask_cmpneq_epi16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 4001 | _mm512_mask_cmp_epi16_mask::<_MM_CMPINT_NE>(k1, a, b) |
| 4002 | } |
| 4003 | |
| 4004 | /// Compare packed signed 16-bit integers in a and b for not-equal, and store the results in mask vector k. |
| 4005 | /// |
| 4006 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpneq_epi16_mask&expand=1080) |
| 4007 | #[inline ] |
| 4008 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4009 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4010 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 4011 | pub fn _mm256_cmpneq_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 4012 | unsafe { simd_bitmask::<i16x16, _>(simd_ne(x:a.as_i16x16(), y:b.as_i16x16())) } |
| 4013 | } |
| 4014 | |
| 4015 | /// Compare packed signed 16-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4016 | /// |
| 4017 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpneq_epi16_mask&expand=1081) |
| 4018 | #[inline ] |
| 4019 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4020 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4021 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 4022 | pub fn _mm256_mask_cmpneq_epi16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 4023 | _mm256_mask_cmp_epi16_mask::<_MM_CMPINT_NE>(k1, a, b) |
| 4024 | } |
| 4025 | |
| 4026 | /// Compare packed signed 16-bit integers in a and b for not-equal, and store the results in mask vector k. |
| 4027 | /// |
| 4028 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpneq_epi16_mask&expand=1078) |
| 4029 | #[inline ] |
| 4030 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4031 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4032 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 4033 | pub fn _mm_cmpneq_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 4034 | unsafe { simd_bitmask::<i16x8, _>(simd_ne(x:a.as_i16x8(), y:b.as_i16x8())) } |
| 4035 | } |
| 4036 | |
| 4037 | /// Compare packed signed 16-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4038 | /// |
| 4039 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpneq_epi16_mask&expand=1079) |
| 4040 | #[inline ] |
| 4041 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4042 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4043 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 4044 | pub fn _mm_mask_cmpneq_epi16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 4045 | _mm_mask_cmp_epi16_mask::<_MM_CMPINT_NE>(k1, a, b) |
| 4046 | } |
| 4047 | |
| 4048 | /// Compare packed signed 8-bit integers in a and b for not-equal, and store the results in mask vector k. |
| 4049 | /// |
| 4050 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpneq_epi8_mask&expand=1100) |
| 4051 | #[inline ] |
| 4052 | #[target_feature (enable = "avx512bw" )] |
| 4053 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4054 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 4055 | pub fn _mm512_cmpneq_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 4056 | unsafe { simd_bitmask::<i8x64, _>(simd_ne(x:a.as_i8x64(), y:b.as_i8x64())) } |
| 4057 | } |
| 4058 | |
| 4059 | /// Compare packed signed 8-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4060 | /// |
| 4061 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpneq_epi8_mask&expand=1101) |
| 4062 | #[inline ] |
| 4063 | #[target_feature (enable = "avx512bw" )] |
| 4064 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4065 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 4066 | pub fn _mm512_mask_cmpneq_epi8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 4067 | _mm512_mask_cmp_epi8_mask::<_MM_CMPINT_NE>(k1, a, b) |
| 4068 | } |
| 4069 | |
| 4070 | /// Compare packed signed 8-bit integers in a and b for not-equal, and store the results in mask vector k. |
| 4071 | /// |
| 4072 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpneq_epi8_mask&expand=1098) |
| 4073 | #[inline ] |
| 4074 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4075 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4076 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 4077 | pub fn _mm256_cmpneq_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 4078 | unsafe { simd_bitmask::<i8x32, _>(simd_ne(x:a.as_i8x32(), y:b.as_i8x32())) } |
| 4079 | } |
| 4080 | |
| 4081 | /// Compare packed signed 8-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4082 | /// |
| 4083 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpneq_epi8_mask&expand=1099) |
| 4084 | #[inline ] |
| 4085 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4086 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4087 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 4088 | pub fn _mm256_mask_cmpneq_epi8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 4089 | _mm256_mask_cmp_epi8_mask::<_MM_CMPINT_NE>(k1, a, b) |
| 4090 | } |
| 4091 | |
| 4092 | /// Compare packed signed 8-bit integers in a and b for not-equal, and store the results in mask vector k. |
| 4093 | /// |
| 4094 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpneq_epi8_mask&expand=1096) |
| 4095 | #[inline ] |
| 4096 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4097 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4098 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 4099 | pub fn _mm_cmpneq_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 4100 | unsafe { simd_bitmask::<i8x16, _>(simd_ne(x:a.as_i8x16(), y:b.as_i8x16())) } |
| 4101 | } |
| 4102 | |
| 4103 | /// Compare packed signed 8-bit integers in a and b for not-equal, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4104 | /// |
| 4105 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpneq_epi8_mask&expand=1097) |
| 4106 | #[inline ] |
| 4107 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4108 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4109 | #[cfg_attr (test, assert_instr(vpcmp))] |
| 4110 | pub fn _mm_mask_cmpneq_epi8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 4111 | _mm_mask_cmp_epi8_mask::<_MM_CMPINT_NE>(k1, a, b) |
| 4112 | } |
| 4113 | |
| 4114 | /// Compare packed unsigned 16-bit integers in a and b based on the comparison operand specified by `IMM8`, and store the results in mask vector k. |
| 4115 | /// |
| 4116 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmp_epu16_mask&expand=715) |
| 4117 | #[inline ] |
| 4118 | #[target_feature (enable = "avx512bw" )] |
| 4119 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4120 | #[rustc_legacy_const_generics (2)] |
| 4121 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4122 | pub fn _mm512_cmp_epu16_mask<const IMM8: i32>(a: __m512i, b: __m512i) -> __mmask32 { |
| 4123 | unsafe { |
| 4124 | static_assert_uimm_bits!(IMM8, 3); |
| 4125 | let a: u16x32 = a.as_u16x32(); |
| 4126 | let b: u16x32 = b.as_u16x32(); |
| 4127 | let r: i16x32 = match IMM8 { |
| 4128 | 0 => simd_eq(x:a, y:b), |
| 4129 | 1 => simd_lt(x:a, y:b), |
| 4130 | 2 => simd_le(x:a, y:b), |
| 4131 | 3 => i16x32::ZERO, |
| 4132 | 4 => simd_ne(x:a, y:b), |
| 4133 | 5 => simd_ge(x:a, y:b), |
| 4134 | 6 => simd_gt(x:a, y:b), |
| 4135 | _ => i16x32::splat(-1), |
| 4136 | }; |
| 4137 | simd_bitmask(r) |
| 4138 | } |
| 4139 | } |
| 4140 | |
| 4141 | /// Compare packed unsigned 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4142 | /// |
| 4143 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmp_epu16_mask&expand=716) |
| 4144 | #[inline ] |
| 4145 | #[target_feature (enable = "avx512bw" )] |
| 4146 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4147 | #[rustc_legacy_const_generics (3)] |
| 4148 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4149 | pub fn _mm512_mask_cmp_epu16_mask<const IMM8: i32>( |
| 4150 | k1: __mmask32, |
| 4151 | a: __m512i, |
| 4152 | b: __m512i, |
| 4153 | ) -> __mmask32 { |
| 4154 | unsafe { |
| 4155 | static_assert_uimm_bits!(IMM8, 3); |
| 4156 | let a: u16x32 = a.as_u16x32(); |
| 4157 | let b: u16x32 = b.as_u16x32(); |
| 4158 | let k1: i16x32 = simd_select_bitmask(m:k1, yes:i16x32::splat(-1), no:i16x32::ZERO); |
| 4159 | let r: i16x32 = match IMM8 { |
| 4160 | 0 => simd_and(x:k1, y:simd_eq(x:a, y:b)), |
| 4161 | 1 => simd_and(x:k1, y:simd_lt(x:a, y:b)), |
| 4162 | 2 => simd_and(x:k1, y:simd_le(x:a, y:b)), |
| 4163 | 3 => i16x32::ZERO, |
| 4164 | 4 => simd_and(x:k1, y:simd_ne(x:a, y:b)), |
| 4165 | 5 => simd_and(x:k1, y:simd_ge(x:a, y:b)), |
| 4166 | 6 => simd_and(x:k1, y:simd_gt(x:a, y:b)), |
| 4167 | _ => k1, |
| 4168 | }; |
| 4169 | simd_bitmask(r) |
| 4170 | } |
| 4171 | } |
| 4172 | |
| 4173 | /// Compare packed unsigned 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k. |
| 4174 | /// |
| 4175 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmp_epu16_mask&expand=713) |
| 4176 | #[inline ] |
| 4177 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4178 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4179 | #[rustc_legacy_const_generics (2)] |
| 4180 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4181 | pub fn _mm256_cmp_epu16_mask<const IMM8: i32>(a: __m256i, b: __m256i) -> __mmask16 { |
| 4182 | unsafe { |
| 4183 | static_assert_uimm_bits!(IMM8, 3); |
| 4184 | let a: u16x16 = a.as_u16x16(); |
| 4185 | let b: u16x16 = b.as_u16x16(); |
| 4186 | let r: i16x16 = match IMM8 { |
| 4187 | 0 => simd_eq(x:a, y:b), |
| 4188 | 1 => simd_lt(x:a, y:b), |
| 4189 | 2 => simd_le(x:a, y:b), |
| 4190 | 3 => i16x16::ZERO, |
| 4191 | 4 => simd_ne(x:a, y:b), |
| 4192 | 5 => simd_ge(x:a, y:b), |
| 4193 | 6 => simd_gt(x:a, y:b), |
| 4194 | _ => i16x16::splat(-1), |
| 4195 | }; |
| 4196 | simd_bitmask(r) |
| 4197 | } |
| 4198 | } |
| 4199 | |
| 4200 | /// Compare packed unsigned 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4201 | /// |
| 4202 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmp_epu16_mask&expand=714) |
| 4203 | #[inline ] |
| 4204 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4205 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4206 | #[rustc_legacy_const_generics (3)] |
| 4207 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4208 | pub fn _mm256_mask_cmp_epu16_mask<const IMM8: i32>( |
| 4209 | k1: __mmask16, |
| 4210 | a: __m256i, |
| 4211 | b: __m256i, |
| 4212 | ) -> __mmask16 { |
| 4213 | unsafe { |
| 4214 | static_assert_uimm_bits!(IMM8, 3); |
| 4215 | let a: u16x16 = a.as_u16x16(); |
| 4216 | let b: u16x16 = b.as_u16x16(); |
| 4217 | let k1: i16x16 = simd_select_bitmask(m:k1, yes:i16x16::splat(-1), no:i16x16::ZERO); |
| 4218 | let r: i16x16 = match IMM8 { |
| 4219 | 0 => simd_and(x:k1, y:simd_eq(x:a, y:b)), |
| 4220 | 1 => simd_and(x:k1, y:simd_lt(x:a, y:b)), |
| 4221 | 2 => simd_and(x:k1, y:simd_le(x:a, y:b)), |
| 4222 | 3 => i16x16::ZERO, |
| 4223 | 4 => simd_and(x:k1, y:simd_ne(x:a, y:b)), |
| 4224 | 5 => simd_and(x:k1, y:simd_ge(x:a, y:b)), |
| 4225 | 6 => simd_and(x:k1, y:simd_gt(x:a, y:b)), |
| 4226 | _ => k1, |
| 4227 | }; |
| 4228 | simd_bitmask(r) |
| 4229 | } |
| 4230 | } |
| 4231 | |
| 4232 | /// Compare packed unsigned 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k. |
| 4233 | /// |
| 4234 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmp_epu16_mask&expand=711) |
| 4235 | #[inline ] |
| 4236 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4237 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4238 | #[rustc_legacy_const_generics (2)] |
| 4239 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4240 | pub fn _mm_cmp_epu16_mask<const IMM8: i32>(a: __m128i, b: __m128i) -> __mmask8 { |
| 4241 | unsafe { |
| 4242 | static_assert_uimm_bits!(IMM8, 3); |
| 4243 | let a: u16x8 = a.as_u16x8(); |
| 4244 | let b: u16x8 = b.as_u16x8(); |
| 4245 | let r: i16x8 = match IMM8 { |
| 4246 | 0 => simd_eq(x:a, y:b), |
| 4247 | 1 => simd_lt(x:a, y:b), |
| 4248 | 2 => simd_le(x:a, y:b), |
| 4249 | 3 => i16x8::ZERO, |
| 4250 | 4 => simd_ne(x:a, y:b), |
| 4251 | 5 => simd_ge(x:a, y:b), |
| 4252 | 6 => simd_gt(x:a, y:b), |
| 4253 | _ => i16x8::splat(-1), |
| 4254 | }; |
| 4255 | simd_bitmask(r) |
| 4256 | } |
| 4257 | } |
| 4258 | |
| 4259 | /// Compare packed unsigned 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4260 | /// |
| 4261 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmp_epu16_mask&expand=712) |
| 4262 | #[inline ] |
| 4263 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4264 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4265 | #[rustc_legacy_const_generics (3)] |
| 4266 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4267 | pub fn _mm_mask_cmp_epu16_mask<const IMM8: i32>(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 4268 | unsafe { |
| 4269 | static_assert_uimm_bits!(IMM8, 3); |
| 4270 | let a: u16x8 = a.as_u16x8(); |
| 4271 | let b: u16x8 = b.as_u16x8(); |
| 4272 | let k1: i16x8 = simd_select_bitmask(m:k1, yes:i16x8::splat(-1), no:i16x8::ZERO); |
| 4273 | let r: i16x8 = match IMM8 { |
| 4274 | 0 => simd_and(x:k1, y:simd_eq(x:a, y:b)), |
| 4275 | 1 => simd_and(x:k1, y:simd_lt(x:a, y:b)), |
| 4276 | 2 => simd_and(x:k1, y:simd_le(x:a, y:b)), |
| 4277 | 3 => i16x8::ZERO, |
| 4278 | 4 => simd_and(x:k1, y:simd_ne(x:a, y:b)), |
| 4279 | 5 => simd_and(x:k1, y:simd_ge(x:a, y:b)), |
| 4280 | 6 => simd_and(x:k1, y:simd_gt(x:a, y:b)), |
| 4281 | _ => k1, |
| 4282 | }; |
| 4283 | simd_bitmask(r) |
| 4284 | } |
| 4285 | } |
| 4286 | |
| 4287 | /// Compare packed unsigned 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k. |
| 4288 | /// |
| 4289 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmp_epu8_mask&expand=733) |
| 4290 | #[inline ] |
| 4291 | #[target_feature (enable = "avx512bw" )] |
| 4292 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4293 | #[rustc_legacy_const_generics (2)] |
| 4294 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4295 | pub fn _mm512_cmp_epu8_mask<const IMM8: i32>(a: __m512i, b: __m512i) -> __mmask64 { |
| 4296 | unsafe { |
| 4297 | static_assert_uimm_bits!(IMM8, 3); |
| 4298 | let a: u8x64 = a.as_u8x64(); |
| 4299 | let b: u8x64 = b.as_u8x64(); |
| 4300 | let r: i8x64 = match IMM8 { |
| 4301 | 0 => simd_eq(x:a, y:b), |
| 4302 | 1 => simd_lt(x:a, y:b), |
| 4303 | 2 => simd_le(x:a, y:b), |
| 4304 | 3 => i8x64::ZERO, |
| 4305 | 4 => simd_ne(x:a, y:b), |
| 4306 | 5 => simd_ge(x:a, y:b), |
| 4307 | 6 => simd_gt(x:a, y:b), |
| 4308 | _ => i8x64::splat(-1), |
| 4309 | }; |
| 4310 | simd_bitmask(r) |
| 4311 | } |
| 4312 | } |
| 4313 | |
| 4314 | /// Compare packed unsigned 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4315 | /// |
| 4316 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmp_epu8_mask&expand=734) |
| 4317 | #[inline ] |
| 4318 | #[target_feature (enable = "avx512bw" )] |
| 4319 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4320 | #[rustc_legacy_const_generics (3)] |
| 4321 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4322 | pub fn _mm512_mask_cmp_epu8_mask<const IMM8: i32>( |
| 4323 | k1: __mmask64, |
| 4324 | a: __m512i, |
| 4325 | b: __m512i, |
| 4326 | ) -> __mmask64 { |
| 4327 | unsafe { |
| 4328 | static_assert_uimm_bits!(IMM8, 3); |
| 4329 | let a: u8x64 = a.as_u8x64(); |
| 4330 | let b: u8x64 = b.as_u8x64(); |
| 4331 | let k1: i8x64 = simd_select_bitmask(m:k1, yes:i8x64::splat(-1), no:i8x64::ZERO); |
| 4332 | let r: i8x64 = match IMM8 { |
| 4333 | 0 => simd_and(x:k1, y:simd_eq(x:a, y:b)), |
| 4334 | 1 => simd_and(x:k1, y:simd_lt(x:a, y:b)), |
| 4335 | 2 => simd_and(x:k1, y:simd_le(x:a, y:b)), |
| 4336 | 3 => i8x64::ZERO, |
| 4337 | 4 => simd_and(x:k1, y:simd_ne(x:a, y:b)), |
| 4338 | 5 => simd_and(x:k1, y:simd_ge(x:a, y:b)), |
| 4339 | 6 => simd_and(x:k1, y:simd_gt(x:a, y:b)), |
| 4340 | _ => k1, |
| 4341 | }; |
| 4342 | simd_bitmask(r) |
| 4343 | } |
| 4344 | } |
| 4345 | |
| 4346 | /// Compare packed unsigned 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k. |
| 4347 | /// |
| 4348 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmp_epu8_mask&expand=731) |
| 4349 | #[inline ] |
| 4350 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4351 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4352 | #[rustc_legacy_const_generics (2)] |
| 4353 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4354 | pub fn _mm256_cmp_epu8_mask<const IMM8: i32>(a: __m256i, b: __m256i) -> __mmask32 { |
| 4355 | unsafe { |
| 4356 | static_assert_uimm_bits!(IMM8, 3); |
| 4357 | let a: u8x32 = a.as_u8x32(); |
| 4358 | let b: u8x32 = b.as_u8x32(); |
| 4359 | let r: i8x32 = match IMM8 { |
| 4360 | 0 => simd_eq(x:a, y:b), |
| 4361 | 1 => simd_lt(x:a, y:b), |
| 4362 | 2 => simd_le(x:a, y:b), |
| 4363 | 3 => i8x32::ZERO, |
| 4364 | 4 => simd_ne(x:a, y:b), |
| 4365 | 5 => simd_ge(x:a, y:b), |
| 4366 | 6 => simd_gt(x:a, y:b), |
| 4367 | _ => i8x32::splat(-1), |
| 4368 | }; |
| 4369 | simd_bitmask(r) |
| 4370 | } |
| 4371 | } |
| 4372 | |
| 4373 | /// Compare packed unsigned 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4374 | /// |
| 4375 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmp_epu8_mask&expand=732) |
| 4376 | #[inline ] |
| 4377 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4378 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4379 | #[rustc_legacy_const_generics (3)] |
| 4380 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4381 | pub fn _mm256_mask_cmp_epu8_mask<const IMM8: i32>( |
| 4382 | k1: __mmask32, |
| 4383 | a: __m256i, |
| 4384 | b: __m256i, |
| 4385 | ) -> __mmask32 { |
| 4386 | unsafe { |
| 4387 | static_assert_uimm_bits!(IMM8, 3); |
| 4388 | let a: u8x32 = a.as_u8x32(); |
| 4389 | let b: u8x32 = b.as_u8x32(); |
| 4390 | let k1: i8x32 = simd_select_bitmask(m:k1, yes:i8x32::splat(-1), no:i8x32::ZERO); |
| 4391 | let r: i8x32 = match IMM8 { |
| 4392 | 0 => simd_and(x:k1, y:simd_eq(x:a, y:b)), |
| 4393 | 1 => simd_and(x:k1, y:simd_lt(x:a, y:b)), |
| 4394 | 2 => simd_and(x:k1, y:simd_le(x:a, y:b)), |
| 4395 | 3 => i8x32::ZERO, |
| 4396 | 4 => simd_and(x:k1, y:simd_ne(x:a, y:b)), |
| 4397 | 5 => simd_and(x:k1, y:simd_ge(x:a, y:b)), |
| 4398 | 6 => simd_and(x:k1, y:simd_gt(x:a, y:b)), |
| 4399 | _ => k1, |
| 4400 | }; |
| 4401 | simd_bitmask(r) |
| 4402 | } |
| 4403 | } |
| 4404 | |
| 4405 | /// Compare packed unsigned 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k. |
| 4406 | /// |
| 4407 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmp_epu8_mask&expand=729) |
| 4408 | #[inline ] |
| 4409 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4410 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4411 | #[rustc_legacy_const_generics (2)] |
| 4412 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4413 | pub fn _mm_cmp_epu8_mask<const IMM8: i32>(a: __m128i, b: __m128i) -> __mmask16 { |
| 4414 | unsafe { |
| 4415 | static_assert_uimm_bits!(IMM8, 3); |
| 4416 | let a: u8x16 = a.as_u8x16(); |
| 4417 | let b: u8x16 = b.as_u8x16(); |
| 4418 | let r: i8x16 = match IMM8 { |
| 4419 | 0 => simd_eq(x:a, y:b), |
| 4420 | 1 => simd_lt(x:a, y:b), |
| 4421 | 2 => simd_le(x:a, y:b), |
| 4422 | 3 => i8x16::ZERO, |
| 4423 | 4 => simd_ne(x:a, y:b), |
| 4424 | 5 => simd_ge(x:a, y:b), |
| 4425 | 6 => simd_gt(x:a, y:b), |
| 4426 | _ => i8x16::splat(-1), |
| 4427 | }; |
| 4428 | simd_bitmask(r) |
| 4429 | } |
| 4430 | } |
| 4431 | |
| 4432 | /// Compare packed unsigned 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4433 | /// |
| 4434 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmp_epu8_mask&expand=730) |
| 4435 | #[inline ] |
| 4436 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4437 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4438 | #[rustc_legacy_const_generics (3)] |
| 4439 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4440 | pub fn _mm_mask_cmp_epu8_mask<const IMM8: i32>(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 4441 | unsafe { |
| 4442 | static_assert_uimm_bits!(IMM8, 3); |
| 4443 | let a: u8x16 = a.as_u8x16(); |
| 4444 | let b: u8x16 = b.as_u8x16(); |
| 4445 | let k1: i8x16 = simd_select_bitmask(m:k1, yes:i8x16::splat(-1), no:i8x16::ZERO); |
| 4446 | let r: i8x16 = match IMM8 { |
| 4447 | 0 => simd_and(x:k1, y:simd_eq(x:a, y:b)), |
| 4448 | 1 => simd_and(x:k1, y:simd_lt(x:a, y:b)), |
| 4449 | 2 => simd_and(x:k1, y:simd_le(x:a, y:b)), |
| 4450 | 3 => i8x16::ZERO, |
| 4451 | 4 => simd_and(x:k1, y:simd_ne(x:a, y:b)), |
| 4452 | 5 => simd_and(x:k1, y:simd_ge(x:a, y:b)), |
| 4453 | 6 => simd_and(x:k1, y:simd_gt(x:a, y:b)), |
| 4454 | _ => k1, |
| 4455 | }; |
| 4456 | simd_bitmask(r) |
| 4457 | } |
| 4458 | } |
| 4459 | |
| 4460 | /// Compare packed signed 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k. |
| 4461 | /// |
| 4462 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmp_epi16_mask&expand=691) |
| 4463 | #[inline ] |
| 4464 | #[target_feature (enable = "avx512bw" )] |
| 4465 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4466 | #[rustc_legacy_const_generics (2)] |
| 4467 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4468 | pub fn _mm512_cmp_epi16_mask<const IMM8: i32>(a: __m512i, b: __m512i) -> __mmask32 { |
| 4469 | unsafe { |
| 4470 | static_assert_uimm_bits!(IMM8, 3); |
| 4471 | let a: i16x32 = a.as_i16x32(); |
| 4472 | let b: i16x32 = b.as_i16x32(); |
| 4473 | let r: i16x32 = match IMM8 { |
| 4474 | 0 => simd_eq(x:a, y:b), |
| 4475 | 1 => simd_lt(x:a, y:b), |
| 4476 | 2 => simd_le(x:a, y:b), |
| 4477 | 3 => i16x32::ZERO, |
| 4478 | 4 => simd_ne(x:a, y:b), |
| 4479 | 5 => simd_ge(x:a, y:b), |
| 4480 | 6 => simd_gt(x:a, y:b), |
| 4481 | _ => i16x32::splat(-1), |
| 4482 | }; |
| 4483 | simd_bitmask(r) |
| 4484 | } |
| 4485 | } |
| 4486 | |
| 4487 | /// Compare packed signed 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4488 | /// |
| 4489 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmp_epi16_mask&expand=692) |
| 4490 | #[inline ] |
| 4491 | #[target_feature (enable = "avx512bw" )] |
| 4492 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4493 | #[rustc_legacy_const_generics (3)] |
| 4494 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4495 | pub fn _mm512_mask_cmp_epi16_mask<const IMM8: i32>( |
| 4496 | k1: __mmask32, |
| 4497 | a: __m512i, |
| 4498 | b: __m512i, |
| 4499 | ) -> __mmask32 { |
| 4500 | unsafe { |
| 4501 | static_assert_uimm_bits!(IMM8, 3); |
| 4502 | let a: i16x32 = a.as_i16x32(); |
| 4503 | let b: i16x32 = b.as_i16x32(); |
| 4504 | let k1: i16x32 = simd_select_bitmask(m:k1, yes:i16x32::splat(-1), no:i16x32::ZERO); |
| 4505 | let r: i16x32 = match IMM8 { |
| 4506 | 0 => simd_and(x:k1, y:simd_eq(x:a, y:b)), |
| 4507 | 1 => simd_and(x:k1, y:simd_lt(x:a, y:b)), |
| 4508 | 2 => simd_and(x:k1, y:simd_le(x:a, y:b)), |
| 4509 | 3 => i16x32::ZERO, |
| 4510 | 4 => simd_and(x:k1, y:simd_ne(x:a, y:b)), |
| 4511 | 5 => simd_and(x:k1, y:simd_ge(x:a, y:b)), |
| 4512 | 6 => simd_and(x:k1, y:simd_gt(x:a, y:b)), |
| 4513 | _ => k1, |
| 4514 | }; |
| 4515 | simd_bitmask(r) |
| 4516 | } |
| 4517 | } |
| 4518 | |
| 4519 | /// Compare packed signed 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k. |
| 4520 | /// |
| 4521 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmp_epi16_mask&expand=689) |
| 4522 | #[inline ] |
| 4523 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4524 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4525 | #[rustc_legacy_const_generics (2)] |
| 4526 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4527 | pub fn _mm256_cmp_epi16_mask<const IMM8: i32>(a: __m256i, b: __m256i) -> __mmask16 { |
| 4528 | unsafe { |
| 4529 | static_assert_uimm_bits!(IMM8, 3); |
| 4530 | let a: i16x16 = a.as_i16x16(); |
| 4531 | let b: i16x16 = b.as_i16x16(); |
| 4532 | let r: i16x16 = match IMM8 { |
| 4533 | 0 => simd_eq(x:a, y:b), |
| 4534 | 1 => simd_lt(x:a, y:b), |
| 4535 | 2 => simd_le(x:a, y:b), |
| 4536 | 3 => i16x16::ZERO, |
| 4537 | 4 => simd_ne(x:a, y:b), |
| 4538 | 5 => simd_ge(x:a, y:b), |
| 4539 | 6 => simd_gt(x:a, y:b), |
| 4540 | _ => i16x16::splat(-1), |
| 4541 | }; |
| 4542 | simd_bitmask(r) |
| 4543 | } |
| 4544 | } |
| 4545 | |
| 4546 | /// Compare packed signed 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4547 | /// |
| 4548 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmp_epi16_mask&expand=690) |
| 4549 | #[inline ] |
| 4550 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4551 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4552 | #[rustc_legacy_const_generics (3)] |
| 4553 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4554 | pub fn _mm256_mask_cmp_epi16_mask<const IMM8: i32>( |
| 4555 | k1: __mmask16, |
| 4556 | a: __m256i, |
| 4557 | b: __m256i, |
| 4558 | ) -> __mmask16 { |
| 4559 | unsafe { |
| 4560 | static_assert_uimm_bits!(IMM8, 3); |
| 4561 | let a: i16x16 = a.as_i16x16(); |
| 4562 | let b: i16x16 = b.as_i16x16(); |
| 4563 | let k1: i16x16 = simd_select_bitmask(m:k1, yes:i16x16::splat(-1), no:i16x16::ZERO); |
| 4564 | let r: i16x16 = match IMM8 { |
| 4565 | 0 => simd_and(x:k1, y:simd_eq(x:a, y:b)), |
| 4566 | 1 => simd_and(x:k1, y:simd_lt(x:a, y:b)), |
| 4567 | 2 => simd_and(x:k1, y:simd_le(x:a, y:b)), |
| 4568 | 3 => i16x16::ZERO, |
| 4569 | 4 => simd_and(x:k1, y:simd_ne(x:a, y:b)), |
| 4570 | 5 => simd_and(x:k1, y:simd_ge(x:a, y:b)), |
| 4571 | 6 => simd_and(x:k1, y:simd_gt(x:a, y:b)), |
| 4572 | _ => k1, |
| 4573 | }; |
| 4574 | simd_bitmask(r) |
| 4575 | } |
| 4576 | } |
| 4577 | |
| 4578 | /// Compare packed signed 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k. |
| 4579 | /// |
| 4580 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmp_epi16_mask&expand=687) |
| 4581 | #[inline ] |
| 4582 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4583 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4584 | #[rustc_legacy_const_generics (2)] |
| 4585 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4586 | pub fn _mm_cmp_epi16_mask<const IMM8: i32>(a: __m128i, b: __m128i) -> __mmask8 { |
| 4587 | unsafe { |
| 4588 | static_assert_uimm_bits!(IMM8, 3); |
| 4589 | let a: i16x8 = a.as_i16x8(); |
| 4590 | let b: i16x8 = b.as_i16x8(); |
| 4591 | let r: i16x8 = match IMM8 { |
| 4592 | 0 => simd_eq(x:a, y:b), |
| 4593 | 1 => simd_lt(x:a, y:b), |
| 4594 | 2 => simd_le(x:a, y:b), |
| 4595 | 3 => i16x8::ZERO, |
| 4596 | 4 => simd_ne(x:a, y:b), |
| 4597 | 5 => simd_ge(x:a, y:b), |
| 4598 | 6 => simd_gt(x:a, y:b), |
| 4599 | _ => i16x8::splat(-1), |
| 4600 | }; |
| 4601 | simd_bitmask(r) |
| 4602 | } |
| 4603 | } |
| 4604 | |
| 4605 | /// Compare packed signed 16-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4606 | /// |
| 4607 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmp_epi16_mask&expand=688) |
| 4608 | #[inline ] |
| 4609 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4610 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4611 | #[rustc_legacy_const_generics (3)] |
| 4612 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4613 | pub fn _mm_mask_cmp_epi16_mask<const IMM8: i32>(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 4614 | unsafe { |
| 4615 | static_assert_uimm_bits!(IMM8, 3); |
| 4616 | let a: i16x8 = a.as_i16x8(); |
| 4617 | let b: i16x8 = b.as_i16x8(); |
| 4618 | let k1: i16x8 = simd_select_bitmask(m:k1, yes:i16x8::splat(-1), no:i16x8::ZERO); |
| 4619 | let r: i16x8 = match IMM8 { |
| 4620 | 0 => simd_and(x:k1, y:simd_eq(x:a, y:b)), |
| 4621 | 1 => simd_and(x:k1, y:simd_lt(x:a, y:b)), |
| 4622 | 2 => simd_and(x:k1, y:simd_le(x:a, y:b)), |
| 4623 | 3 => i16x8::ZERO, |
| 4624 | 4 => simd_and(x:k1, y:simd_ne(x:a, y:b)), |
| 4625 | 5 => simd_and(x:k1, y:simd_ge(x:a, y:b)), |
| 4626 | 6 => simd_and(x:k1, y:simd_gt(x:a, y:b)), |
| 4627 | _ => k1, |
| 4628 | }; |
| 4629 | simd_bitmask(r) |
| 4630 | } |
| 4631 | } |
| 4632 | |
| 4633 | /// Compare packed signed 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k. |
| 4634 | /// |
| 4635 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmp_epi8_mask&expand=709) |
| 4636 | #[inline ] |
| 4637 | #[target_feature (enable = "avx512bw" )] |
| 4638 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4639 | #[rustc_legacy_const_generics (2)] |
| 4640 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4641 | pub fn _mm512_cmp_epi8_mask<const IMM8: i32>(a: __m512i, b: __m512i) -> __mmask64 { |
| 4642 | unsafe { |
| 4643 | static_assert_uimm_bits!(IMM8, 3); |
| 4644 | let a: i8x64 = a.as_i8x64(); |
| 4645 | let b: i8x64 = b.as_i8x64(); |
| 4646 | let r: i8x64 = match IMM8 { |
| 4647 | 0 => simd_eq(x:a, y:b), |
| 4648 | 1 => simd_lt(x:a, y:b), |
| 4649 | 2 => simd_le(x:a, y:b), |
| 4650 | 3 => i8x64::ZERO, |
| 4651 | 4 => simd_ne(x:a, y:b), |
| 4652 | 5 => simd_ge(x:a, y:b), |
| 4653 | 6 => simd_gt(x:a, y:b), |
| 4654 | _ => i8x64::splat(-1), |
| 4655 | }; |
| 4656 | simd_bitmask(r) |
| 4657 | } |
| 4658 | } |
| 4659 | |
| 4660 | /// Compare packed signed 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4661 | /// |
| 4662 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmp_epi8_mask&expand=710) |
| 4663 | #[inline ] |
| 4664 | #[target_feature (enable = "avx512bw" )] |
| 4665 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4666 | #[rustc_legacy_const_generics (3)] |
| 4667 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4668 | pub fn _mm512_mask_cmp_epi8_mask<const IMM8: i32>( |
| 4669 | k1: __mmask64, |
| 4670 | a: __m512i, |
| 4671 | b: __m512i, |
| 4672 | ) -> __mmask64 { |
| 4673 | unsafe { |
| 4674 | static_assert_uimm_bits!(IMM8, 3); |
| 4675 | let a: i8x64 = a.as_i8x64(); |
| 4676 | let b: i8x64 = b.as_i8x64(); |
| 4677 | let k1: i8x64 = simd_select_bitmask(m:k1, yes:i8x64::splat(-1), no:i8x64::ZERO); |
| 4678 | let r: i8x64 = match IMM8 { |
| 4679 | 0 => simd_and(x:k1, y:simd_eq(x:a, y:b)), |
| 4680 | 1 => simd_and(x:k1, y:simd_lt(x:a, y:b)), |
| 4681 | 2 => simd_and(x:k1, y:simd_le(x:a, y:b)), |
| 4682 | 3 => i8x64::ZERO, |
| 4683 | 4 => simd_and(x:k1, y:simd_ne(x:a, y:b)), |
| 4684 | 5 => simd_and(x:k1, y:simd_ge(x:a, y:b)), |
| 4685 | 6 => simd_and(x:k1, y:simd_gt(x:a, y:b)), |
| 4686 | _ => k1, |
| 4687 | }; |
| 4688 | simd_bitmask(r) |
| 4689 | } |
| 4690 | } |
| 4691 | |
| 4692 | /// Compare packed signed 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k. |
| 4693 | /// |
| 4694 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmp_epi8_mask&expand=707) |
| 4695 | #[inline ] |
| 4696 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4697 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4698 | #[rustc_legacy_const_generics (2)] |
| 4699 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4700 | pub fn _mm256_cmp_epi8_mask<const IMM8: i32>(a: __m256i, b: __m256i) -> __mmask32 { |
| 4701 | unsafe { |
| 4702 | static_assert_uimm_bits!(IMM8, 3); |
| 4703 | let a: i8x32 = a.as_i8x32(); |
| 4704 | let b: i8x32 = b.as_i8x32(); |
| 4705 | let r: i8x32 = match IMM8 { |
| 4706 | 0 => simd_eq(x:a, y:b), |
| 4707 | 1 => simd_lt(x:a, y:b), |
| 4708 | 2 => simd_le(x:a, y:b), |
| 4709 | 3 => i8x32::ZERO, |
| 4710 | 4 => simd_ne(x:a, y:b), |
| 4711 | 5 => simd_ge(x:a, y:b), |
| 4712 | 6 => simd_gt(x:a, y:b), |
| 4713 | _ => i8x32::splat(-1), |
| 4714 | }; |
| 4715 | simd_bitmask(r) |
| 4716 | } |
| 4717 | } |
| 4718 | |
| 4719 | /// Compare packed signed 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4720 | /// |
| 4721 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmp_epi8_mask&expand=708) |
| 4722 | #[inline ] |
| 4723 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4724 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4725 | #[rustc_legacy_const_generics (3)] |
| 4726 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4727 | pub fn _mm256_mask_cmp_epi8_mask<const IMM8: i32>( |
| 4728 | k1: __mmask32, |
| 4729 | a: __m256i, |
| 4730 | b: __m256i, |
| 4731 | ) -> __mmask32 { |
| 4732 | unsafe { |
| 4733 | static_assert_uimm_bits!(IMM8, 3); |
| 4734 | let a: i8x32 = a.as_i8x32(); |
| 4735 | let b: i8x32 = b.as_i8x32(); |
| 4736 | let k1: i8x32 = simd_select_bitmask(m:k1, yes:i8x32::splat(-1), no:i8x32::ZERO); |
| 4737 | let r: i8x32 = match IMM8 { |
| 4738 | 0 => simd_and(x:k1, y:simd_eq(x:a, y:b)), |
| 4739 | 1 => simd_and(x:k1, y:simd_lt(x:a, y:b)), |
| 4740 | 2 => simd_and(x:k1, y:simd_le(x:a, y:b)), |
| 4741 | 3 => i8x32::ZERO, |
| 4742 | 4 => simd_and(x:k1, y:simd_ne(x:a, y:b)), |
| 4743 | 5 => simd_and(x:k1, y:simd_ge(x:a, y:b)), |
| 4744 | 6 => simd_and(x:k1, y:simd_gt(x:a, y:b)), |
| 4745 | _ => k1, |
| 4746 | }; |
| 4747 | simd_bitmask(r) |
| 4748 | } |
| 4749 | } |
| 4750 | |
| 4751 | /// Compare packed signed 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k. |
| 4752 | /// |
| 4753 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmp_epi8_mask&expand=705) |
| 4754 | #[inline ] |
| 4755 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4756 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4757 | #[rustc_legacy_const_generics (2)] |
| 4758 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4759 | pub fn _mm_cmp_epi8_mask<const IMM8: i32>(a: __m128i, b: __m128i) -> __mmask16 { |
| 4760 | unsafe { |
| 4761 | static_assert_uimm_bits!(IMM8, 3); |
| 4762 | let a: i8x16 = a.as_i8x16(); |
| 4763 | let b: i8x16 = b.as_i8x16(); |
| 4764 | let r: i8x16 = match IMM8 { |
| 4765 | 0 => simd_eq(x:a, y:b), |
| 4766 | 1 => simd_lt(x:a, y:b), |
| 4767 | 2 => simd_le(x:a, y:b), |
| 4768 | 3 => i8x16::ZERO, |
| 4769 | 4 => simd_ne(x:a, y:b), |
| 4770 | 5 => simd_ge(x:a, y:b), |
| 4771 | 6 => simd_gt(x:a, y:b), |
| 4772 | _ => i8x16::splat(-1), |
| 4773 | }; |
| 4774 | simd_bitmask(r) |
| 4775 | } |
| 4776 | } |
| 4777 | |
| 4778 | /// Compare packed signed 8-bit integers in a and b based on the comparison operand specified by imm8, and store the results in mask vector k using zeromask k1 (elements are zeroed out when the corresponding mask bit is not set). |
| 4779 | /// |
| 4780 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmp_epi8_mask&expand=706) |
| 4781 | #[inline ] |
| 4782 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4783 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4784 | #[rustc_legacy_const_generics (3)] |
| 4785 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
| 4786 | pub fn _mm_mask_cmp_epi8_mask<const IMM8: i32>(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 4787 | unsafe { |
| 4788 | static_assert_uimm_bits!(IMM8, 3); |
| 4789 | let a: i8x16 = a.as_i8x16(); |
| 4790 | let b: i8x16 = b.as_i8x16(); |
| 4791 | let k1: i8x16 = simd_select_bitmask(m:k1, yes:i8x16::splat(-1), no:i8x16::ZERO); |
| 4792 | let r: i8x16 = match IMM8 { |
| 4793 | 0 => simd_and(x:k1, y:simd_eq(x:a, y:b)), |
| 4794 | 1 => simd_and(x:k1, y:simd_lt(x:a, y:b)), |
| 4795 | 2 => simd_and(x:k1, y:simd_le(x:a, y:b)), |
| 4796 | 3 => i8x16::ZERO, |
| 4797 | 4 => simd_and(x:k1, y:simd_ne(x:a, y:b)), |
| 4798 | 5 => simd_and(x:k1, y:simd_ge(x:a, y:b)), |
| 4799 | 6 => simd_and(x:k1, y:simd_gt(x:a, y:b)), |
| 4800 | _ => k1, |
| 4801 | }; |
| 4802 | simd_bitmask(r) |
| 4803 | } |
| 4804 | } |
| 4805 | |
| 4806 | /// Reduce the packed 16-bit integers in a by addition. Returns the sum of all elements in a. |
| 4807 | /// |
| 4808 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_add_epi16) |
| 4809 | #[inline ] |
| 4810 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4811 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4812 | pub fn _mm256_reduce_add_epi16(a: __m256i) -> i16 { |
| 4813 | unsafe { simd_reduce_add_unordered(a.as_i16x16()) } |
| 4814 | } |
| 4815 | |
| 4816 | /// Reduce the packed 16-bit integers in a by addition using mask k. Returns the sum of all active elements in a. |
| 4817 | /// |
| 4818 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_add_epi16) |
| 4819 | #[inline ] |
| 4820 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4821 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4822 | pub fn _mm256_mask_reduce_add_epi16(k: __mmask16, a: __m256i) -> i16 { |
| 4823 | unsafe { simd_reduce_add_unordered(simd_select_bitmask(m:k, yes:a.as_i16x16(), no:i16x16::ZERO)) } |
| 4824 | } |
| 4825 | |
| 4826 | /// Reduce the packed 16-bit integers in a by addition. Returns the sum of all elements in a. |
| 4827 | /// |
| 4828 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_add_epi16) |
| 4829 | #[inline ] |
| 4830 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4831 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4832 | pub fn _mm_reduce_add_epi16(a: __m128i) -> i16 { |
| 4833 | unsafe { simd_reduce_add_unordered(a.as_i16x8()) } |
| 4834 | } |
| 4835 | |
| 4836 | /// Reduce the packed 16-bit integers in a by addition using mask k. Returns the sum of all active elements in a. |
| 4837 | /// |
| 4838 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_add_epi16) |
| 4839 | #[inline ] |
| 4840 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4841 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4842 | pub fn _mm_mask_reduce_add_epi16(k: __mmask8, a: __m128i) -> i16 { |
| 4843 | unsafe { simd_reduce_add_unordered(simd_select_bitmask(m:k, yes:a.as_i16x8(), no:i16x8::ZERO)) } |
| 4844 | } |
| 4845 | |
| 4846 | /// Reduce the packed 8-bit integers in a by addition. Returns the sum of all elements in a. |
| 4847 | /// |
| 4848 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_add_epi8) |
| 4849 | #[inline ] |
| 4850 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4851 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4852 | pub fn _mm256_reduce_add_epi8(a: __m256i) -> i8 { |
| 4853 | unsafe { simd_reduce_add_unordered(a.as_i8x32()) } |
| 4854 | } |
| 4855 | |
| 4856 | /// Reduce the packed 8-bit integers in a by addition using mask k. Returns the sum of all active elements in a. |
| 4857 | /// |
| 4858 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_add_epi8) |
| 4859 | #[inline ] |
| 4860 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4861 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4862 | pub fn _mm256_mask_reduce_add_epi8(k: __mmask32, a: __m256i) -> i8 { |
| 4863 | unsafe { simd_reduce_add_unordered(simd_select_bitmask(m:k, yes:a.as_i8x32(), no:i8x32::ZERO)) } |
| 4864 | } |
| 4865 | |
| 4866 | /// Reduce the packed 8-bit integers in a by addition. Returns the sum of all elements in a. |
| 4867 | /// |
| 4868 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_add_epi8) |
| 4869 | #[inline ] |
| 4870 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4871 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4872 | pub fn _mm_reduce_add_epi8(a: __m128i) -> i8 { |
| 4873 | unsafe { simd_reduce_add_unordered(a.as_i8x16()) } |
| 4874 | } |
| 4875 | |
| 4876 | /// Reduce the packed 8-bit integers in a by addition using mask k. Returns the sum of all active elements in a. |
| 4877 | /// |
| 4878 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_add_epi8) |
| 4879 | #[inline ] |
| 4880 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4881 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4882 | pub fn _mm_mask_reduce_add_epi8(k: __mmask16, a: __m128i) -> i8 { |
| 4883 | unsafe { simd_reduce_add_unordered(simd_select_bitmask(m:k, yes:a.as_i8x16(), no:i8x16::ZERO)) } |
| 4884 | } |
| 4885 | |
| 4886 | /// Reduce the packed 16-bit integers in a by bitwise AND. Returns the bitwise AND of all elements in a. |
| 4887 | /// |
| 4888 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_and_epi16) |
| 4889 | #[inline ] |
| 4890 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4891 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4892 | pub fn _mm256_reduce_and_epi16(a: __m256i) -> i16 { |
| 4893 | unsafe { simd_reduce_and(a.as_i16x16()) } |
| 4894 | } |
| 4895 | |
| 4896 | /// Reduce the packed 16-bit integers in a by bitwise AND using mask k. Returns the bitwise AND of all active elements in a. |
| 4897 | /// |
| 4898 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_and_epi16) |
| 4899 | #[inline ] |
| 4900 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4901 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4902 | pub fn _mm256_mask_reduce_and_epi16(k: __mmask16, a: __m256i) -> i16 { |
| 4903 | unsafe { |
| 4904 | simd_reduce_and(simd_select_bitmask( |
| 4905 | m:k, |
| 4906 | yes:a.as_i16x16(), |
| 4907 | no:_mm256_set1_epi64x(-1).as_i16x16(), |
| 4908 | )) |
| 4909 | } |
| 4910 | } |
| 4911 | |
| 4912 | /// Reduce the packed 16-bit integers in a by bitwise AND. Returns the bitwise AND of all elements in a. |
| 4913 | /// |
| 4914 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_and_epi16) |
| 4915 | #[inline ] |
| 4916 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4917 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4918 | pub fn _mm_reduce_and_epi16(a: __m128i) -> i16 { |
| 4919 | unsafe { simd_reduce_and(a.as_i16x8()) } |
| 4920 | } |
| 4921 | |
| 4922 | /// Reduce the packed 16-bit integers in a by bitwise AND using mask k. Returns the bitwise AND of all active elements in a. |
| 4923 | /// |
| 4924 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_and_epi16) |
| 4925 | #[inline ] |
| 4926 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4927 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4928 | pub fn _mm_mask_reduce_and_epi16(k: __mmask8, a: __m128i) -> i16 { |
| 4929 | unsafe { |
| 4930 | simd_reduce_and(simd_select_bitmask( |
| 4931 | m:k, |
| 4932 | yes:a.as_i16x8(), |
| 4933 | no:_mm_set1_epi64x(-1).as_i16x8(), |
| 4934 | )) |
| 4935 | } |
| 4936 | } |
| 4937 | |
| 4938 | /// Reduce the packed 8-bit integers in a by bitwise AND. Returns the bitwise AND of all elements in a. |
| 4939 | /// |
| 4940 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_and_epi8) |
| 4941 | #[inline ] |
| 4942 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4943 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4944 | pub fn _mm256_reduce_and_epi8(a: __m256i) -> i8 { |
| 4945 | unsafe { simd_reduce_and(a.as_i8x32()) } |
| 4946 | } |
| 4947 | |
| 4948 | /// Reduce the packed 8-bit integers in a by bitwise AND using mask k. Returns the bitwise AND of all active elements in a. |
| 4949 | /// |
| 4950 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_and_epi8) |
| 4951 | #[inline ] |
| 4952 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4953 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4954 | pub fn _mm256_mask_reduce_and_epi8(k: __mmask32, a: __m256i) -> i8 { |
| 4955 | unsafe { |
| 4956 | simd_reduce_and(simd_select_bitmask( |
| 4957 | m:k, |
| 4958 | yes:a.as_i8x32(), |
| 4959 | no:_mm256_set1_epi64x(-1).as_i8x32(), |
| 4960 | )) |
| 4961 | } |
| 4962 | } |
| 4963 | |
| 4964 | /// Reduce the packed 8-bit integers in a by bitwise AND. Returns the bitwise AND of all elements in a. |
| 4965 | /// |
| 4966 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_and_epi8) |
| 4967 | #[inline ] |
| 4968 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4969 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4970 | pub fn _mm_reduce_and_epi8(a: __m128i) -> i8 { |
| 4971 | unsafe { simd_reduce_and(a.as_i8x16()) } |
| 4972 | } |
| 4973 | |
| 4974 | /// Reduce the packed 8-bit integers in a by bitwise AND using mask k. Returns the bitwise AND of all active elements in a. |
| 4975 | /// |
| 4976 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_and_epi8) |
| 4977 | #[inline ] |
| 4978 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4979 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4980 | pub fn _mm_mask_reduce_and_epi8(k: __mmask16, a: __m128i) -> i8 { |
| 4981 | unsafe { |
| 4982 | simd_reduce_and(simd_select_bitmask( |
| 4983 | m:k, |
| 4984 | yes:a.as_i8x16(), |
| 4985 | no:_mm_set1_epi64x(-1).as_i8x16(), |
| 4986 | )) |
| 4987 | } |
| 4988 | } |
| 4989 | |
| 4990 | /// Reduce the packed 16-bit integers in a by maximum. Returns the maximum of all elements in a. |
| 4991 | /// |
| 4992 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_max_epi16) |
| 4993 | #[inline ] |
| 4994 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 4995 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 4996 | pub fn _mm256_reduce_max_epi16(a: __m256i) -> i16 { |
| 4997 | unsafe { simd_reduce_max(a.as_i16x16()) } |
| 4998 | } |
| 4999 | |
| 5000 | /// Reduce the packed 16-bit integers in a by maximum using mask k. Returns the maximum of all active elements in a. |
| 5001 | /// |
| 5002 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_max_epi16) |
| 5003 | #[inline ] |
| 5004 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5005 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5006 | pub fn _mm256_mask_reduce_max_epi16(k: __mmask16, a: __m256i) -> i16 { |
| 5007 | unsafe { simd_reduce_max(simd_select_bitmask(m:k, yes:a.as_i16x16(), no:i16x16::splat(-32768))) } |
| 5008 | } |
| 5009 | |
| 5010 | /// Reduce the packed 16-bit integers in a by maximum. Returns the maximum of all elements in a. |
| 5011 | /// |
| 5012 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_max_epi16) |
| 5013 | #[inline ] |
| 5014 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5015 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5016 | pub fn _mm_reduce_max_epi16(a: __m128i) -> i16 { |
| 5017 | unsafe { simd_reduce_max(a.as_i16x8()) } |
| 5018 | } |
| 5019 | |
| 5020 | /// Reduce the packed 16-bit integers in a by maximum using mask k. Returns the maximum of all active elements in a. |
| 5021 | /// |
| 5022 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_max_epi16) |
| 5023 | #[inline ] |
| 5024 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5025 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5026 | pub fn _mm_mask_reduce_max_epi16(k: __mmask8, a: __m128i) -> i16 { |
| 5027 | unsafe { simd_reduce_max(simd_select_bitmask(m:k, yes:a.as_i16x8(), no:i16x8::splat(-32768))) } |
| 5028 | } |
| 5029 | |
| 5030 | /// Reduce the packed 8-bit integers in a by maximum. Returns the maximum of all elements in a. |
| 5031 | /// |
| 5032 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_max_epi8) |
| 5033 | #[inline ] |
| 5034 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5035 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5036 | pub fn _mm256_reduce_max_epi8(a: __m256i) -> i8 { |
| 5037 | unsafe { simd_reduce_max(a.as_i8x32()) } |
| 5038 | } |
| 5039 | |
| 5040 | /// Reduce the packed 8-bit integers in a by maximum using mask k. Returns the maximum of all active elements in a. |
| 5041 | /// |
| 5042 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_max_epi8) |
| 5043 | #[inline ] |
| 5044 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5045 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5046 | pub fn _mm256_mask_reduce_max_epi8(k: __mmask32, a: __m256i) -> i8 { |
| 5047 | unsafe { simd_reduce_max(simd_select_bitmask(m:k, yes:a.as_i8x32(), no:i8x32::splat(-128))) } |
| 5048 | } |
| 5049 | |
| 5050 | /// Reduce the packed 8-bit integers in a by maximum. Returns the maximum of all elements in a. |
| 5051 | /// |
| 5052 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_max_epi8) |
| 5053 | #[inline ] |
| 5054 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5055 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5056 | pub fn _mm_reduce_max_epi8(a: __m128i) -> i8 { |
| 5057 | unsafe { simd_reduce_max(a.as_i8x16()) } |
| 5058 | } |
| 5059 | |
| 5060 | /// Reduce the packed 8-bit integers in a by maximum using mask k. Returns the maximum of all active elements in a. |
| 5061 | /// |
| 5062 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_max_epi8) |
| 5063 | #[inline ] |
| 5064 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5065 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5066 | pub fn _mm_mask_reduce_max_epi8(k: __mmask16, a: __m128i) -> i8 { |
| 5067 | unsafe { simd_reduce_max(simd_select_bitmask(m:k, yes:a.as_i8x16(), no:i8x16::splat(-128))) } |
| 5068 | } |
| 5069 | |
| 5070 | /// Reduce the packed unsigned 16-bit integers in a by maximum. Returns the maximum of all elements in a. |
| 5071 | /// |
| 5072 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_max_epu16) |
| 5073 | #[inline ] |
| 5074 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5075 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5076 | pub fn _mm256_reduce_max_epu16(a: __m256i) -> u16 { |
| 5077 | unsafe { simd_reduce_max(a.as_u16x16()) } |
| 5078 | } |
| 5079 | |
| 5080 | /// Reduce the packed unsigned 16-bit integers in a by maximum using mask k. Returns the maximum of all active elements in a. |
| 5081 | /// |
| 5082 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_max_epu16) |
| 5083 | #[inline ] |
| 5084 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5085 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5086 | pub fn _mm256_mask_reduce_max_epu16(k: __mmask16, a: __m256i) -> u16 { |
| 5087 | unsafe { simd_reduce_max(simd_select_bitmask(m:k, yes:a.as_u16x16(), no:u16x16::ZERO)) } |
| 5088 | } |
| 5089 | |
| 5090 | /// Reduce the packed unsigned 16-bit integers in a by maximum. Returns the maximum of all elements in a. |
| 5091 | /// |
| 5092 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_max_epu16) |
| 5093 | #[inline ] |
| 5094 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5095 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5096 | pub fn _mm_reduce_max_epu16(a: __m128i) -> u16 { |
| 5097 | unsafe { simd_reduce_max(a.as_u16x8()) } |
| 5098 | } |
| 5099 | |
| 5100 | /// Reduce the packed unsigned 16-bit integers in a by maximum using mask k. Returns the maximum of all active elements in a. |
| 5101 | /// |
| 5102 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_max_epu16) |
| 5103 | #[inline ] |
| 5104 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5105 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5106 | pub fn _mm_mask_reduce_max_epu16(k: __mmask8, a: __m128i) -> u16 { |
| 5107 | unsafe { simd_reduce_max(simd_select_bitmask(m:k, yes:a.as_u16x8(), no:u16x8::ZERO)) } |
| 5108 | } |
| 5109 | |
| 5110 | /// Reduce the packed unsigned 8-bit integers in a by maximum. Returns the maximum of all elements in a. |
| 5111 | /// |
| 5112 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_max_epu8) |
| 5113 | #[inline ] |
| 5114 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5115 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5116 | pub fn _mm256_reduce_max_epu8(a: __m256i) -> u8 { |
| 5117 | unsafe { simd_reduce_max(a.as_u8x32()) } |
| 5118 | } |
| 5119 | |
| 5120 | /// Reduce the packed unsigned 8-bit integers in a by maximum using mask k. Returns the maximum of all active elements in a. |
| 5121 | /// |
| 5122 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_max_epu8) |
| 5123 | #[inline ] |
| 5124 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5125 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5126 | pub fn _mm256_mask_reduce_max_epu8(k: __mmask32, a: __m256i) -> u8 { |
| 5127 | unsafe { simd_reduce_max(simd_select_bitmask(m:k, yes:a.as_u8x32(), no:u8x32::ZERO)) } |
| 5128 | } |
| 5129 | |
| 5130 | /// Reduce the packed unsigned 8-bit integers in a by maximum. Returns the maximum of all elements in a. |
| 5131 | /// |
| 5132 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_max_epu8) |
| 5133 | #[inline ] |
| 5134 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5135 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5136 | pub fn _mm_reduce_max_epu8(a: __m128i) -> u8 { |
| 5137 | unsafe { simd_reduce_max(a.as_u8x16()) } |
| 5138 | } |
| 5139 | |
| 5140 | /// Reduce the packed unsigned 8-bit integers in a by maximum using mask k. Returns the maximum of all active elements in a. |
| 5141 | /// |
| 5142 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_max_epu8) |
| 5143 | #[inline ] |
| 5144 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5145 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5146 | pub fn _mm_mask_reduce_max_epu8(k: __mmask16, a: __m128i) -> u8 { |
| 5147 | unsafe { simd_reduce_max(simd_select_bitmask(m:k, yes:a.as_u8x16(), no:u8x16::ZERO)) } |
| 5148 | } |
| 5149 | |
| 5150 | /// Reduce the packed 16-bit integers in a by minimum. Returns the minimum of all elements in a. |
| 5151 | /// |
| 5152 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_min_epi16) |
| 5153 | #[inline ] |
| 5154 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5155 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5156 | pub fn _mm256_reduce_min_epi16(a: __m256i) -> i16 { |
| 5157 | unsafe { simd_reduce_min(a.as_i16x16()) } |
| 5158 | } |
| 5159 | |
| 5160 | /// Reduce the packed 16-bit integers in a by minimum using mask k. Returns the minimum of all active elements in a. |
| 5161 | /// |
| 5162 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_min_epi16) |
| 5163 | #[inline ] |
| 5164 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5165 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5166 | pub fn _mm256_mask_reduce_min_epi16(k: __mmask16, a: __m256i) -> i16 { |
| 5167 | unsafe { simd_reduce_min(simd_select_bitmask(m:k, yes:a.as_i16x16(), no:i16x16::splat(0x7fff))) } |
| 5168 | } |
| 5169 | |
| 5170 | /// Reduce the packed 16-bit integers in a by minimum. Returns the minimum of all elements in a. |
| 5171 | /// |
| 5172 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_min_epi16) |
| 5173 | #[inline ] |
| 5174 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5175 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5176 | pub fn _mm_reduce_min_epi16(a: __m128i) -> i16 { |
| 5177 | unsafe { simd_reduce_min(a.as_i16x8()) } |
| 5178 | } |
| 5179 | |
| 5180 | /// Reduce the packed 16-bit integers in a by minimum using mask k. Returns the minimum of all active elements in a. |
| 5181 | /// |
| 5182 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_min_epi16) |
| 5183 | #[inline ] |
| 5184 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5185 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5186 | pub fn _mm_mask_reduce_min_epi16(k: __mmask8, a: __m128i) -> i16 { |
| 5187 | unsafe { simd_reduce_min(simd_select_bitmask(m:k, yes:a.as_i16x8(), no:i16x8::splat(0x7fff))) } |
| 5188 | } |
| 5189 | |
| 5190 | /// Reduce the packed 8-bit integers in a by minimum. Returns the minimum of all elements in a. |
| 5191 | /// |
| 5192 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_min_epi8) |
| 5193 | #[inline ] |
| 5194 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5195 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5196 | pub fn _mm256_reduce_min_epi8(a: __m256i) -> i8 { |
| 5197 | unsafe { simd_reduce_min(a.as_i8x32()) } |
| 5198 | } |
| 5199 | |
| 5200 | /// Reduce the packed 8-bit integers in a by minimum using mask k. Returns the minimum of all active elements in a. |
| 5201 | /// |
| 5202 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_min_epi8) |
| 5203 | #[inline ] |
| 5204 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5205 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5206 | pub fn _mm256_mask_reduce_min_epi8(k: __mmask32, a: __m256i) -> i8 { |
| 5207 | unsafe { simd_reduce_min(simd_select_bitmask(m:k, yes:a.as_i8x32(), no:i8x32::splat(0x7f))) } |
| 5208 | } |
| 5209 | |
| 5210 | /// Reduce the packed 8-bit integers in a by minimum. Returns the minimum of all elements in a. |
| 5211 | /// |
| 5212 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_min_epi8) |
| 5213 | #[inline ] |
| 5214 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5215 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5216 | pub fn _mm_reduce_min_epi8(a: __m128i) -> i8 { |
| 5217 | unsafe { simd_reduce_min(a.as_i8x16()) } |
| 5218 | } |
| 5219 | |
| 5220 | /// Reduce the packed 8-bit integers in a by minimum using mask k. Returns the minimum of all active elements in a. |
| 5221 | /// |
| 5222 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_min_epi8) |
| 5223 | #[inline ] |
| 5224 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5225 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5226 | pub fn _mm_mask_reduce_min_epi8(k: __mmask16, a: __m128i) -> i8 { |
| 5227 | unsafe { simd_reduce_min(simd_select_bitmask(m:k, yes:a.as_i8x16(), no:i8x16::splat(0x7f))) } |
| 5228 | } |
| 5229 | |
| 5230 | /// Reduce the packed unsigned 16-bit integers in a by minimum. Returns the minimum of all elements in a. |
| 5231 | /// |
| 5232 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_min_epu16) |
| 5233 | #[inline ] |
| 5234 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5235 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5236 | pub fn _mm256_reduce_min_epu16(a: __m256i) -> u16 { |
| 5237 | unsafe { simd_reduce_min(a.as_u16x16()) } |
| 5238 | } |
| 5239 | |
| 5240 | /// Reduce the packed unsigned 16-bit integers in a by minimum using mask k. Returns the minimum of all active elements in a. |
| 5241 | /// |
| 5242 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_min_epu16) |
| 5243 | #[inline ] |
| 5244 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5245 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5246 | pub fn _mm256_mask_reduce_min_epu16(k: __mmask16, a: __m256i) -> u16 { |
| 5247 | unsafe { simd_reduce_min(simd_select_bitmask(m:k, yes:a.as_u16x16(), no:u16x16::splat(0xffff))) } |
| 5248 | } |
| 5249 | |
| 5250 | /// Reduce the packed unsigned 16-bit integers in a by minimum. Returns the minimum of all elements in a. |
| 5251 | /// |
| 5252 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_min_epu16) |
| 5253 | #[inline ] |
| 5254 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5255 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5256 | pub fn _mm_reduce_min_epu16(a: __m128i) -> u16 { |
| 5257 | unsafe { simd_reduce_min(a.as_u16x8()) } |
| 5258 | } |
| 5259 | |
| 5260 | /// Reduce the packed unsigned 16-bit integers in a by minimum using mask k. Returns the minimum of all active elements in a. |
| 5261 | /// |
| 5262 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_min_epu16) |
| 5263 | #[inline ] |
| 5264 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5265 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5266 | pub fn _mm_mask_reduce_min_epu16(k: __mmask8, a: __m128i) -> u16 { |
| 5267 | unsafe { simd_reduce_min(simd_select_bitmask(m:k, yes:a.as_u16x8(), no:u16x8::splat(0xffff))) } |
| 5268 | } |
| 5269 | |
| 5270 | /// Reduce the packed unsigned 8-bit integers in a by minimum. Returns the minimum of all elements in a. |
| 5271 | /// |
| 5272 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_min_epu8) |
| 5273 | #[inline ] |
| 5274 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5275 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5276 | pub fn _mm256_reduce_min_epu8(a: __m256i) -> u8 { |
| 5277 | unsafe { simd_reduce_min(a.as_u8x32()) } |
| 5278 | } |
| 5279 | |
| 5280 | /// Reduce the packed unsigned 8-bit integers in a by minimum using mask k. Returns the minimum of all active elements in a. |
| 5281 | /// |
| 5282 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_min_epu8) |
| 5283 | #[inline ] |
| 5284 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5285 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5286 | pub fn _mm256_mask_reduce_min_epu8(k: __mmask32, a: __m256i) -> u8 { |
| 5287 | unsafe { simd_reduce_min(simd_select_bitmask(m:k, yes:a.as_u8x32(), no:u8x32::splat(0xff))) } |
| 5288 | } |
| 5289 | |
| 5290 | /// Reduce the packed unsigned 8-bit integers in a by minimum. Returns the minimum of all elements in a. |
| 5291 | /// |
| 5292 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_min_epu8) |
| 5293 | #[inline ] |
| 5294 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5295 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5296 | pub fn _mm_reduce_min_epu8(a: __m128i) -> u8 { |
| 5297 | unsafe { simd_reduce_min(a.as_u8x16()) } |
| 5298 | } |
| 5299 | |
| 5300 | /// Reduce the packed unsigned 8-bit integers in a by minimum using mask k. Returns the minimum of all active elements in a. |
| 5301 | /// |
| 5302 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_min_epu8) |
| 5303 | #[inline ] |
| 5304 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5305 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5306 | pub fn _mm_mask_reduce_min_epu8(k: __mmask16, a: __m128i) -> u8 { |
| 5307 | unsafe { simd_reduce_min(simd_select_bitmask(m:k, yes:a.as_u8x16(), no:u8x16::splat(0xff))) } |
| 5308 | } |
| 5309 | |
| 5310 | /// Reduce the packed 16-bit integers in a by multiplication. Returns the product of all elements in a. |
| 5311 | /// |
| 5312 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_mul_epi16) |
| 5313 | #[inline ] |
| 5314 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5315 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5316 | pub fn _mm256_reduce_mul_epi16(a: __m256i) -> i16 { |
| 5317 | unsafe { simd_reduce_mul_unordered(a.as_i16x16()) } |
| 5318 | } |
| 5319 | |
| 5320 | /// Reduce the packed 16-bit integers in a by multiplication using mask k. Returns the product of all active elements in a. |
| 5321 | /// |
| 5322 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_mul_epi16) |
| 5323 | #[inline ] |
| 5324 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5325 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5326 | pub fn _mm256_mask_reduce_mul_epi16(k: __mmask16, a: __m256i) -> i16 { |
| 5327 | unsafe { simd_reduce_mul_unordered(simd_select_bitmask(m:k, yes:a.as_i16x16(), no:i16x16::splat(1))) } |
| 5328 | } |
| 5329 | |
| 5330 | /// Reduce the packed 16-bit integers in a by multiplication. Returns the product of all elements in a. |
| 5331 | /// |
| 5332 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_mul_epi16) |
| 5333 | #[inline ] |
| 5334 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5335 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5336 | pub fn _mm_reduce_mul_epi16(a: __m128i) -> i16 { |
| 5337 | unsafe { simd_reduce_mul_unordered(a.as_i16x8()) } |
| 5338 | } |
| 5339 | |
| 5340 | /// Reduce the packed 16-bit integers in a by multiplication using mask k. Returns the product of all active elements in a. |
| 5341 | /// |
| 5342 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_mul_epi16) |
| 5343 | #[inline ] |
| 5344 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5345 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5346 | pub fn _mm_mask_reduce_mul_epi16(k: __mmask8, a: __m128i) -> i16 { |
| 5347 | unsafe { simd_reduce_mul_unordered(simd_select_bitmask(m:k, yes:a.as_i16x8(), no:i16x8::splat(1))) } |
| 5348 | } |
| 5349 | |
| 5350 | /// Reduce the packed 8-bit integers in a by multiplication. Returns the product of all elements in a. |
| 5351 | /// |
| 5352 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_mul_epi8) |
| 5353 | #[inline ] |
| 5354 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5355 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5356 | pub fn _mm256_reduce_mul_epi8(a: __m256i) -> i8 { |
| 5357 | unsafe { simd_reduce_mul_unordered(a.as_i8x32()) } |
| 5358 | } |
| 5359 | |
| 5360 | /// Reduce the packed 8-bit integers in a by multiplication using mask k. Returns the product of all active elements in a. |
| 5361 | /// |
| 5362 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_mul_epi8) |
| 5363 | #[inline ] |
| 5364 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5365 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5366 | pub fn _mm256_mask_reduce_mul_epi8(k: __mmask32, a: __m256i) -> i8 { |
| 5367 | unsafe { simd_reduce_mul_unordered(simd_select_bitmask(m:k, yes:a.as_i8x32(), no:i8x32::splat(1))) } |
| 5368 | } |
| 5369 | |
| 5370 | /// Reduce the packed 8-bit integers in a by multiplication. Returns the product of all elements in a. |
| 5371 | /// |
| 5372 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_mul_epi8) |
| 5373 | #[inline ] |
| 5374 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5375 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5376 | pub fn _mm_reduce_mul_epi8(a: __m128i) -> i8 { |
| 5377 | unsafe { simd_reduce_mul_unordered(a.as_i8x16()) } |
| 5378 | } |
| 5379 | |
| 5380 | /// Reduce the packed 8-bit integers in a by multiplication using mask k. Returns the product of all active elements in a. |
| 5381 | /// |
| 5382 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_mul_epi8) |
| 5383 | #[inline ] |
| 5384 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5385 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5386 | pub fn _mm_mask_reduce_mul_epi8(k: __mmask16, a: __m128i) -> i8 { |
| 5387 | unsafe { simd_reduce_mul_unordered(simd_select_bitmask(m:k, yes:a.as_i8x16(), no:i8x16::splat(1))) } |
| 5388 | } |
| 5389 | |
| 5390 | /// Reduce the packed 16-bit integers in a by bitwise OR. Returns the bitwise OR of all elements in a. |
| 5391 | /// |
| 5392 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_or_epi16) |
| 5393 | #[inline ] |
| 5394 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5395 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5396 | pub fn _mm256_reduce_or_epi16(a: __m256i) -> i16 { |
| 5397 | unsafe { simd_reduce_or(a.as_i16x16()) } |
| 5398 | } |
| 5399 | |
| 5400 | /// Reduce the packed 16-bit integers in a by bitwise OR using mask k. Returns the bitwise OR of all active elements in a. |
| 5401 | /// |
| 5402 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_or_epi16) |
| 5403 | #[inline ] |
| 5404 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5405 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5406 | pub fn _mm256_mask_reduce_or_epi16(k: __mmask16, a: __m256i) -> i16 { |
| 5407 | unsafe { simd_reduce_or(simd_select_bitmask(m:k, yes:a.as_i16x16(), no:i16x16::ZERO)) } |
| 5408 | } |
| 5409 | |
| 5410 | /// Reduce the packed 16-bit integers in a by bitwise OR. Returns the bitwise OR of all elements in a. |
| 5411 | /// |
| 5412 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_or_epi16) |
| 5413 | #[inline ] |
| 5414 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5415 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5416 | pub fn _mm_reduce_or_epi16(a: __m128i) -> i16 { |
| 5417 | unsafe { simd_reduce_or(a.as_i16x8()) } |
| 5418 | } |
| 5419 | |
| 5420 | /// Reduce the packed 16-bit integers in a by bitwise OR using mask k. Returns the bitwise OR of all active elements in a. |
| 5421 | /// |
| 5422 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_or_epi16) |
| 5423 | #[inline ] |
| 5424 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5425 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5426 | pub fn _mm_mask_reduce_or_epi16(k: __mmask8, a: __m128i) -> i16 { |
| 5427 | unsafe { simd_reduce_or(simd_select_bitmask(m:k, yes:a.as_i16x8(), no:i16x8::ZERO)) } |
| 5428 | } |
| 5429 | |
| 5430 | /// Reduce the packed 8-bit integers in a by bitwise OR. Returns the bitwise OR of all elements in a. |
| 5431 | /// |
| 5432 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_reduce_or_epi8) |
| 5433 | #[inline ] |
| 5434 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5435 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5436 | pub fn _mm256_reduce_or_epi8(a: __m256i) -> i8 { |
| 5437 | unsafe { simd_reduce_or(a.as_i8x32()) } |
| 5438 | } |
| 5439 | |
| 5440 | /// Reduce the packed 8-bit integers in a by bitwise OR using mask k. Returns the bitwise OR of all active elements in a. |
| 5441 | /// |
| 5442 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_reduce_or_epi8) |
| 5443 | #[inline ] |
| 5444 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5445 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5446 | pub fn _mm256_mask_reduce_or_epi8(k: __mmask32, a: __m256i) -> i8 { |
| 5447 | unsafe { simd_reduce_or(simd_select_bitmask(m:k, yes:a.as_i8x32(), no:i8x32::ZERO)) } |
| 5448 | } |
| 5449 | |
| 5450 | /// Reduce the packed 8-bit integers in a by bitwise OR. Returns the bitwise OR of all elements in a. |
| 5451 | /// |
| 5452 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_reduce_or_epi8) |
| 5453 | #[inline ] |
| 5454 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5455 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5456 | pub fn _mm_reduce_or_epi8(a: __m128i) -> i8 { |
| 5457 | unsafe { simd_reduce_or(a.as_i8x16()) } |
| 5458 | } |
| 5459 | |
| 5460 | /// Reduce the packed 8-bit integers in a by bitwise OR using mask k. Returns the bitwise OR of all active elements in a. |
| 5461 | /// |
| 5462 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_reduce_or_epi8) |
| 5463 | #[inline ] |
| 5464 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5465 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5466 | pub fn _mm_mask_reduce_or_epi8(k: __mmask16, a: __m128i) -> i8 { |
| 5467 | unsafe { simd_reduce_or(simd_select_bitmask(m:k, yes:a.as_i8x16(), no:i8x16::ZERO)) } |
| 5468 | } |
| 5469 | |
| 5470 | /// Load 512-bits (composed of 32 packed 16-bit integers) from memory into dst. mem_addr does not need to be aligned on any particular boundary. |
| 5471 | /// |
| 5472 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_loadu_epi16&expand=3368) |
| 5473 | #[inline ] |
| 5474 | #[target_feature (enable = "avx512bw" )] |
| 5475 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5476 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu16 |
| 5477 | pub unsafe fn _mm512_loadu_epi16(mem_addr: *const i16) -> __m512i { |
| 5478 | ptr::read_unaligned(src:mem_addr as *const __m512i) |
| 5479 | } |
| 5480 | |
| 5481 | /// Load 256-bits (composed of 16 packed 16-bit integers) from memory into dst. mem_addr does not need to be aligned on any particular boundary. |
| 5482 | /// |
| 5483 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_loadu_epi16&expand=3365) |
| 5484 | #[inline ] |
| 5485 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5486 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5487 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu16 |
| 5488 | pub unsafe fn _mm256_loadu_epi16(mem_addr: *const i16) -> __m256i { |
| 5489 | ptr::read_unaligned(src:mem_addr as *const __m256i) |
| 5490 | } |
| 5491 | |
| 5492 | /// Load 128-bits (composed of 8 packed 16-bit integers) from memory into dst. mem_addr does not need to be aligned on any particular boundary. |
| 5493 | /// |
| 5494 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_loadu_epi16&expand=3362) |
| 5495 | #[inline ] |
| 5496 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5497 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5498 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu16 |
| 5499 | pub unsafe fn _mm_loadu_epi16(mem_addr: *const i16) -> __m128i { |
| 5500 | ptr::read_unaligned(src:mem_addr as *const __m128i) |
| 5501 | } |
| 5502 | |
| 5503 | /// Load 512-bits (composed of 64 packed 8-bit integers) from memory into dst. mem_addr does not need to be aligned on any particular boundary. |
| 5504 | /// |
| 5505 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_loadu_epi8&expand=3395) |
| 5506 | #[inline ] |
| 5507 | #[target_feature (enable = "avx512bw" )] |
| 5508 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5509 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu8 |
| 5510 | pub unsafe fn _mm512_loadu_epi8(mem_addr: *const i8) -> __m512i { |
| 5511 | ptr::read_unaligned(src:mem_addr as *const __m512i) |
| 5512 | } |
| 5513 | |
| 5514 | /// Load 256-bits (composed of 32 packed 8-bit integers) from memory into dst. mem_addr does not need to be aligned on any particular boundary. |
| 5515 | /// |
| 5516 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_loadu_epi8&expand=3392) |
| 5517 | #[inline ] |
| 5518 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5519 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5520 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu8 |
| 5521 | pub unsafe fn _mm256_loadu_epi8(mem_addr: *const i8) -> __m256i { |
| 5522 | ptr::read_unaligned(src:mem_addr as *const __m256i) |
| 5523 | } |
| 5524 | |
| 5525 | /// Load 128-bits (composed of 16 packed 8-bit integers) from memory into dst. mem_addr does not need to be aligned on any particular boundary. |
| 5526 | /// |
| 5527 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_loadu_epi8&expand=3389) |
| 5528 | #[inline ] |
| 5529 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5530 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5531 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu8 |
| 5532 | pub unsafe fn _mm_loadu_epi8(mem_addr: *const i8) -> __m128i { |
| 5533 | ptr::read_unaligned(src:mem_addr as *const __m128i) |
| 5534 | } |
| 5535 | |
| 5536 | /// Store 512-bits (composed of 32 packed 16-bit integers) from a into memory. mem_addr does not need to be aligned on any particular boundary. |
| 5537 | /// |
| 5538 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_storeu_epi16&expand=5622) |
| 5539 | #[inline ] |
| 5540 | #[target_feature (enable = "avx512bw" )] |
| 5541 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5542 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu16 |
| 5543 | pub unsafe fn _mm512_storeu_epi16(mem_addr: *mut i16, a: __m512i) { |
| 5544 | ptr::write_unaligned(dst:mem_addr as *mut __m512i, src:a); |
| 5545 | } |
| 5546 | |
| 5547 | /// Store 256-bits (composed of 16 packed 16-bit integers) from a into memory. mem_addr does not need to be aligned on any particular boundary. |
| 5548 | /// |
| 5549 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_storeu_epi16&expand=5620) |
| 5550 | #[inline ] |
| 5551 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5552 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5553 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu16 |
| 5554 | pub unsafe fn _mm256_storeu_epi16(mem_addr: *mut i16, a: __m256i) { |
| 5555 | ptr::write_unaligned(dst:mem_addr as *mut __m256i, src:a); |
| 5556 | } |
| 5557 | |
| 5558 | /// Store 128-bits (composed of 8 packed 16-bit integers) from a into memory. mem_addr does not need to be aligned on any particular boundary. |
| 5559 | /// |
| 5560 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_storeu_epi16&expand=5618) |
| 5561 | #[inline ] |
| 5562 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5563 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5564 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu16 |
| 5565 | pub unsafe fn _mm_storeu_epi16(mem_addr: *mut i16, a: __m128i) { |
| 5566 | ptr::write_unaligned(dst:mem_addr as *mut __m128i, src:a); |
| 5567 | } |
| 5568 | |
| 5569 | /// Store 512-bits (composed of 64 packed 8-bit integers) from a into memory. mem_addr does not need to be aligned on any particular boundary. |
| 5570 | /// |
| 5571 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_storeu_epi8&expand=5640) |
| 5572 | #[inline ] |
| 5573 | #[target_feature (enable = "avx512bw" )] |
| 5574 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5575 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu8 |
| 5576 | pub unsafe fn _mm512_storeu_epi8(mem_addr: *mut i8, a: __m512i) { |
| 5577 | ptr::write_unaligned(dst:mem_addr as *mut __m512i, src:a); |
| 5578 | } |
| 5579 | |
| 5580 | /// Store 256-bits (composed of 32 packed 8-bit integers) from a into memory. mem_addr does not need to be aligned on any particular boundary. |
| 5581 | /// |
| 5582 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_storeu_epi8&expand=5638) |
| 5583 | #[inline ] |
| 5584 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5585 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5586 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu8 |
| 5587 | pub unsafe fn _mm256_storeu_epi8(mem_addr: *mut i8, a: __m256i) { |
| 5588 | ptr::write_unaligned(dst:mem_addr as *mut __m256i, src:a); |
| 5589 | } |
| 5590 | |
| 5591 | /// Store 128-bits (composed of 16 packed 8-bit integers) from a into memory. mem_addr does not need to be aligned on any particular boundary. |
| 5592 | /// |
| 5593 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_storeu_epi8&expand=5636) |
| 5594 | #[inline ] |
| 5595 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5596 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5597 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu8 |
| 5598 | pub unsafe fn _mm_storeu_epi8(mem_addr: *mut i8, a: __m128i) { |
| 5599 | ptr::write_unaligned(dst:mem_addr as *mut __m128i, src:a); |
| 5600 | } |
| 5601 | |
| 5602 | /// Load packed 16-bit integers from memory into dst using writemask k |
| 5603 | /// (elements are copied from src when the corresponding mask bit is not set). |
| 5604 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5605 | /// |
| 5606 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_loadu_epi16) |
| 5607 | #[inline ] |
| 5608 | #[target_feature (enable = "avx512bw" )] |
| 5609 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 5610 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5611 | pub unsafe fn _mm512_mask_loadu_epi16(src: __m512i, k: __mmask32, mem_addr: *const i16) -> __m512i { |
| 5612 | let mask: i16x32 = simd_select_bitmask(m:k, yes:i16x32::splat(!0), no:i16x32::ZERO); |
| 5613 | simd_masked_load!(SimdAlign::Unaligned, mask, mem_addr, src.as_i16x32()).as_m512i() |
| 5614 | } |
| 5615 | |
| 5616 | /// Load packed 16-bit integers from memory into dst using zeromask k |
| 5617 | /// (elements are zeroed out when the corresponding mask bit is not set). |
| 5618 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5619 | /// |
| 5620 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_loadu_epi16) |
| 5621 | #[inline ] |
| 5622 | #[target_feature (enable = "avx512bw" )] |
| 5623 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 5624 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5625 | pub unsafe fn _mm512_maskz_loadu_epi16(k: __mmask32, mem_addr: *const i16) -> __m512i { |
| 5626 | _mm512_mask_loadu_epi16(src:_mm512_setzero_si512(), k, mem_addr) |
| 5627 | } |
| 5628 | |
| 5629 | /// Load packed 8-bit integers from memory into dst using writemask k |
| 5630 | /// (elements are copied from src when the corresponding mask bit is not set). |
| 5631 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5632 | /// |
| 5633 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_loadu_epi8) |
| 5634 | #[inline ] |
| 5635 | #[target_feature (enable = "avx512bw" )] |
| 5636 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 5637 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5638 | pub unsafe fn _mm512_mask_loadu_epi8(src: __m512i, k: __mmask64, mem_addr: *const i8) -> __m512i { |
| 5639 | let mask: i8x64 = simd_select_bitmask(m:k, yes:i8x64::splat(!0), no:i8x64::ZERO); |
| 5640 | simd_masked_load!(SimdAlign::Unaligned, mask, mem_addr, src.as_i8x64()).as_m512i() |
| 5641 | } |
| 5642 | |
| 5643 | /// Load packed 8-bit integers from memory into dst using zeromask k |
| 5644 | /// (elements are zeroed out when the corresponding mask bit is not set). |
| 5645 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5646 | /// |
| 5647 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_loadu_epi8) |
| 5648 | #[inline ] |
| 5649 | #[target_feature (enable = "avx512bw" )] |
| 5650 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 5651 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5652 | pub unsafe fn _mm512_maskz_loadu_epi8(k: __mmask64, mem_addr: *const i8) -> __m512i { |
| 5653 | _mm512_mask_loadu_epi8(src:_mm512_setzero_si512(), k, mem_addr) |
| 5654 | } |
| 5655 | |
| 5656 | /// Load packed 16-bit integers from memory into dst using writemask k |
| 5657 | /// (elements are copied from src when the corresponding mask bit is not set). |
| 5658 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5659 | /// |
| 5660 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_loadu_epi16) |
| 5661 | #[inline ] |
| 5662 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5663 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 5664 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5665 | pub unsafe fn _mm256_mask_loadu_epi16(src: __m256i, k: __mmask16, mem_addr: *const i16) -> __m256i { |
| 5666 | let mask: i16x16 = simd_select_bitmask(m:k, yes:i16x16::splat(!0), no:i16x16::ZERO); |
| 5667 | simd_masked_load!(SimdAlign::Unaligned, mask, mem_addr, src.as_i16x16()).as_m256i() |
| 5668 | } |
| 5669 | |
| 5670 | /// Load packed 16-bit integers from memory into dst using zeromask k |
| 5671 | /// (elements are zeroed out when the corresponding mask bit is not set). |
| 5672 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5673 | /// |
| 5674 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_loadu_epi16) |
| 5675 | #[inline ] |
| 5676 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5677 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 5678 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5679 | pub unsafe fn _mm256_maskz_loadu_epi16(k: __mmask16, mem_addr: *const i16) -> __m256i { |
| 5680 | _mm256_mask_loadu_epi16(src:_mm256_setzero_si256(), k, mem_addr) |
| 5681 | } |
| 5682 | |
| 5683 | /// Load packed 8-bit integers from memory into dst using writemask k |
| 5684 | /// (elements are copied from src when the corresponding mask bit is not set). |
| 5685 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5686 | /// |
| 5687 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_loadu_epi8) |
| 5688 | #[inline ] |
| 5689 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5690 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 5691 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5692 | pub unsafe fn _mm256_mask_loadu_epi8(src: __m256i, k: __mmask32, mem_addr: *const i8) -> __m256i { |
| 5693 | let mask: i8x32 = simd_select_bitmask(m:k, yes:i8x32::splat(!0), no:i8x32::ZERO); |
| 5694 | simd_masked_load!(SimdAlign::Unaligned, mask, mem_addr, src.as_i8x32()).as_m256i() |
| 5695 | } |
| 5696 | |
| 5697 | /// Load packed 8-bit integers from memory into dst using zeromask k |
| 5698 | /// (elements are zeroed out when the corresponding mask bit is not set). |
| 5699 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5700 | /// |
| 5701 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_loadu_epi8) |
| 5702 | #[inline ] |
| 5703 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5704 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 5705 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5706 | pub unsafe fn _mm256_maskz_loadu_epi8(k: __mmask32, mem_addr: *const i8) -> __m256i { |
| 5707 | _mm256_mask_loadu_epi8(src:_mm256_setzero_si256(), k, mem_addr) |
| 5708 | } |
| 5709 | |
| 5710 | /// Load packed 16-bit integers from memory into dst using writemask k |
| 5711 | /// (elements are copied from src when the corresponding mask bit is not set). |
| 5712 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5713 | /// |
| 5714 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_loadu_epi16) |
| 5715 | #[inline ] |
| 5716 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5717 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 5718 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5719 | pub unsafe fn _mm_mask_loadu_epi16(src: __m128i, k: __mmask8, mem_addr: *const i16) -> __m128i { |
| 5720 | let mask: i16x8 = simd_select_bitmask(m:k, yes:i16x8::splat(!0), no:i16x8::ZERO); |
| 5721 | simd_masked_load!(SimdAlign::Unaligned, mask, mem_addr, src.as_i16x8()).as_m128i() |
| 5722 | } |
| 5723 | |
| 5724 | /// Load packed 16-bit integers from memory into dst using zeromask k |
| 5725 | /// (elements are zeroed out when the corresponding mask bit is not set). |
| 5726 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5727 | /// |
| 5728 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_loadu_epi16) |
| 5729 | #[inline ] |
| 5730 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5731 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 5732 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5733 | pub unsafe fn _mm_maskz_loadu_epi16(k: __mmask8, mem_addr: *const i16) -> __m128i { |
| 5734 | _mm_mask_loadu_epi16(src:_mm_setzero_si128(), k, mem_addr) |
| 5735 | } |
| 5736 | |
| 5737 | /// Load packed 8-bit integers from memory into dst using writemask k |
| 5738 | /// (elements are copied from src when the corresponding mask bit is not set). |
| 5739 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5740 | /// |
| 5741 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_loadu_epi8) |
| 5742 | #[inline ] |
| 5743 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5744 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 5745 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5746 | pub unsafe fn _mm_mask_loadu_epi8(src: __m128i, k: __mmask16, mem_addr: *const i8) -> __m128i { |
| 5747 | let mask: i8x16 = simd_select_bitmask(m:k, yes:i8x16::splat(!0), no:i8x16::ZERO); |
| 5748 | simd_masked_load!(SimdAlign::Unaligned, mask, mem_addr, src.as_i8x16()).as_m128i() |
| 5749 | } |
| 5750 | |
| 5751 | /// Load packed 8-bit integers from memory into dst using zeromask k |
| 5752 | /// (elements are zeroed out when the corresponding mask bit is not set). |
| 5753 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5754 | /// |
| 5755 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_loadu_epi8) |
| 5756 | #[inline ] |
| 5757 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5758 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 5759 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5760 | pub unsafe fn _mm_maskz_loadu_epi8(k: __mmask16, mem_addr: *const i8) -> __m128i { |
| 5761 | _mm_mask_loadu_epi8(src:_mm_setzero_si128(), k, mem_addr) |
| 5762 | } |
| 5763 | |
| 5764 | /// Store packed 16-bit integers from a into memory using writemask k. |
| 5765 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5766 | /// |
| 5767 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_storeu_epi16) |
| 5768 | #[inline ] |
| 5769 | #[target_feature (enable = "avx512bw" )] |
| 5770 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 5771 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5772 | pub unsafe fn _mm512_mask_storeu_epi16(mem_addr: *mut i16, mask: __mmask32, a: __m512i) { |
| 5773 | let mask: i16x32 = simd_select_bitmask(m:mask, yes:i16x32::splat(!0), no:i16x32::ZERO); |
| 5774 | simd_masked_store!(SimdAlign::Unaligned, mask, mem_addr, a.as_i16x32()); |
| 5775 | } |
| 5776 | |
| 5777 | /// Store packed 8-bit integers from a into memory using writemask k. |
| 5778 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5779 | /// |
| 5780 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_storeu_epi8) |
| 5781 | #[inline ] |
| 5782 | #[target_feature (enable = "avx512bw" )] |
| 5783 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 5784 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5785 | pub unsafe fn _mm512_mask_storeu_epi8(mem_addr: *mut i8, mask: __mmask64, a: __m512i) { |
| 5786 | let mask: i8x64 = simd_select_bitmask(m:mask, yes:i8x64::splat(!0), no:i8x64::ZERO); |
| 5787 | simd_masked_store!(SimdAlign::Unaligned, mask, mem_addr, a.as_i8x64()); |
| 5788 | } |
| 5789 | |
| 5790 | /// Store packed 16-bit integers from a into memory using writemask k. |
| 5791 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5792 | /// |
| 5793 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_storeu_epi16) |
| 5794 | #[inline ] |
| 5795 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5796 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 5797 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5798 | pub unsafe fn _mm256_mask_storeu_epi16(mem_addr: *mut i16, mask: __mmask16, a: __m256i) { |
| 5799 | let mask: i16x16 = simd_select_bitmask(m:mask, yes:i16x16::splat(!0), no:i16x16::ZERO); |
| 5800 | simd_masked_store!(SimdAlign::Unaligned, mask, mem_addr, a.as_i16x16()); |
| 5801 | } |
| 5802 | |
| 5803 | /// Store packed 8-bit integers from a into memory using writemask k. |
| 5804 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5805 | /// |
| 5806 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_storeu_epi8) |
| 5807 | #[inline ] |
| 5808 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5809 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 5810 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5811 | pub unsafe fn _mm256_mask_storeu_epi8(mem_addr: *mut i8, mask: __mmask32, a: __m256i) { |
| 5812 | let mask: i8x32 = simd_select_bitmask(m:mask, yes:i8x32::splat(!0), no:i8x32::ZERO); |
| 5813 | simd_masked_store!(SimdAlign::Unaligned, mask, mem_addr, a.as_i8x32()); |
| 5814 | } |
| 5815 | |
| 5816 | /// Store packed 16-bit integers from a into memory using writemask k. |
| 5817 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5818 | /// |
| 5819 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_storeu_epi16) |
| 5820 | #[inline ] |
| 5821 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5822 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 5823 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5824 | pub unsafe fn _mm_mask_storeu_epi16(mem_addr: *mut i16, mask: __mmask8, a: __m128i) { |
| 5825 | let mask: i16x8 = simd_select_bitmask(m:mask, yes:i16x8::splat(!0), no:i16x8::ZERO); |
| 5826 | simd_masked_store!(SimdAlign::Unaligned, mask, mem_addr, a.as_i16x8()); |
| 5827 | } |
| 5828 | |
| 5829 | /// Store packed 8-bit integers from a into memory using writemask k. |
| 5830 | /// mem_addr does not need to be aligned on any particular boundary. |
| 5831 | /// |
| 5832 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_storeu_epi8) |
| 5833 | #[inline ] |
| 5834 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5835 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 5836 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5837 | pub unsafe fn _mm_mask_storeu_epi8(mem_addr: *mut i8, mask: __mmask16, a: __m128i) { |
| 5838 | let mask: i8x16 = simd_select_bitmask(m:mask, yes:i8x16::splat(!0), no:i8x16::ZERO); |
| 5839 | simd_masked_store!(SimdAlign::Unaligned, mask, mem_addr, a.as_i8x16()); |
| 5840 | } |
| 5841 | |
| 5842 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst. |
| 5843 | /// |
| 5844 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_madd_epi16&expand=3511) |
| 5845 | #[inline ] |
| 5846 | #[target_feature (enable = "avx512bw" )] |
| 5847 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5848 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
| 5849 | pub fn _mm512_madd_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 5850 | // It's a trick used in the Adler-32 algorithm to perform a widening addition. |
| 5851 | // |
| 5852 | // ```rust |
| 5853 | // #[target_feature(enable = "avx512bw")] |
| 5854 | // unsafe fn widening_add(mad: __m512i) -> __m512i { |
| 5855 | // _mm512_madd_epi16(mad, _mm512_set1_epi16(1)) |
| 5856 | // } |
| 5857 | // ``` |
| 5858 | // |
| 5859 | // If we implement this using generic vector intrinsics, the optimizer |
| 5860 | // will eliminate this pattern, and `vpmaddwd` will no longer be emitted. |
| 5861 | // For this reason, we use x86 intrinsics. |
| 5862 | unsafe { transmute(src:vpmaddwd(a.as_i16x32(), b.as_i16x32())) } |
| 5863 | } |
| 5864 | |
| 5865 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 5866 | /// |
| 5867 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_madd_epi16&expand=3512) |
| 5868 | #[inline ] |
| 5869 | #[target_feature (enable = "avx512bw" )] |
| 5870 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5871 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
| 5872 | pub fn _mm512_mask_madd_epi16(src: __m512i, k: __mmask16, a: __m512i, b: __m512i) -> __m512i { |
| 5873 | unsafe { |
| 5874 | let madd: i32x16 = _mm512_madd_epi16(a, b).as_i32x16(); |
| 5875 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:src.as_i32x16())) |
| 5876 | } |
| 5877 | } |
| 5878 | |
| 5879 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 5880 | /// |
| 5881 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_madd_epi16&expand=3513) |
| 5882 | #[inline ] |
| 5883 | #[target_feature (enable = "avx512bw" )] |
| 5884 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5885 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
| 5886 | pub fn _mm512_maskz_madd_epi16(k: __mmask16, a: __m512i, b: __m512i) -> __m512i { |
| 5887 | unsafe { |
| 5888 | let madd: i32x16 = _mm512_madd_epi16(a, b).as_i32x16(); |
| 5889 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:i32x16::ZERO)) |
| 5890 | } |
| 5891 | } |
| 5892 | |
| 5893 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 5894 | /// |
| 5895 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_madd_epi16&expand=3509) |
| 5896 | #[inline ] |
| 5897 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5898 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5899 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
| 5900 | pub fn _mm256_mask_madd_epi16(src: __m256i, k: __mmask8, a: __m256i, b: __m256i) -> __m256i { |
| 5901 | unsafe { |
| 5902 | let madd: i32x8 = _mm256_madd_epi16(a, b).as_i32x8(); |
| 5903 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:src.as_i32x8())) |
| 5904 | } |
| 5905 | } |
| 5906 | |
| 5907 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 5908 | /// |
| 5909 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_madd_epi16&expand=3510) |
| 5910 | #[inline ] |
| 5911 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5912 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5913 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
| 5914 | pub fn _mm256_maskz_madd_epi16(k: __mmask8, a: __m256i, b: __m256i) -> __m256i { |
| 5915 | unsafe { |
| 5916 | let madd: i32x8 = _mm256_madd_epi16(a, b).as_i32x8(); |
| 5917 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:i32x8::ZERO)) |
| 5918 | } |
| 5919 | } |
| 5920 | |
| 5921 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 5922 | /// |
| 5923 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_madd_epi16&expand=3506) |
| 5924 | #[inline ] |
| 5925 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5926 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5927 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
| 5928 | pub fn _mm_mask_madd_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 5929 | unsafe { |
| 5930 | let madd: i32x4 = _mm_madd_epi16(a, b).as_i32x4(); |
| 5931 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:src.as_i32x4())) |
| 5932 | } |
| 5933 | } |
| 5934 | |
| 5935 | /// Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 5936 | /// |
| 5937 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_madd_epi16&expand=3507) |
| 5938 | #[inline ] |
| 5939 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5940 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5941 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
| 5942 | pub fn _mm_maskz_madd_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 5943 | unsafe { |
| 5944 | let madd: i32x4 = _mm_madd_epi16(a, b).as_i32x4(); |
| 5945 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:i32x4::ZERO)) |
| 5946 | } |
| 5947 | } |
| 5948 | |
| 5949 | /// Vertically multiply each unsigned 8-bit integer from a with the corresponding signed 8-bit integer from b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst. |
| 5950 | /// |
| 5951 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maddubs_epi16&expand=3539) |
| 5952 | #[inline ] |
| 5953 | #[target_feature (enable = "avx512bw" )] |
| 5954 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5955 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
| 5956 | pub fn _mm512_maddubs_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 5957 | unsafe { transmute(src:vpmaddubsw(a.as_u8x64(), b.as_i8x64())) } |
| 5958 | } |
| 5959 | |
| 5960 | /// Multiply packed unsigned 8-bit integers in a by packed signed 8-bit integers in b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 5961 | /// |
| 5962 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_maddubs_epi16&expand=3540) |
| 5963 | #[inline ] |
| 5964 | #[target_feature (enable = "avx512bw" )] |
| 5965 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5966 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
| 5967 | pub fn _mm512_mask_maddubs_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 5968 | unsafe { |
| 5969 | let madd: i16x32 = _mm512_maddubs_epi16(a, b).as_i16x32(); |
| 5970 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:src.as_i16x32())) |
| 5971 | } |
| 5972 | } |
| 5973 | |
| 5974 | /// Multiply packed unsigned 8-bit integers in a by packed signed 8-bit integers in b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 5975 | /// |
| 5976 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_maddubs_epi16&expand=3541) |
| 5977 | #[inline ] |
| 5978 | #[target_feature (enable = "avx512bw" )] |
| 5979 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5980 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
| 5981 | pub fn _mm512_maskz_maddubs_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 5982 | unsafe { |
| 5983 | let madd: i16x32 = _mm512_maddubs_epi16(a, b).as_i16x32(); |
| 5984 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:i16x32::ZERO)) |
| 5985 | } |
| 5986 | } |
| 5987 | |
| 5988 | /// Multiply packed unsigned 8-bit integers in a by packed signed 8-bit integers in b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 5989 | /// |
| 5990 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_maddubs_epi16&expand=3537) |
| 5991 | #[inline ] |
| 5992 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 5993 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 5994 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
| 5995 | pub fn _mm256_mask_maddubs_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 5996 | unsafe { |
| 5997 | let madd: i16x16 = _mm256_maddubs_epi16(a, b).as_i16x16(); |
| 5998 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:src.as_i16x16())) |
| 5999 | } |
| 6000 | } |
| 6001 | |
| 6002 | /// Multiply packed unsigned 8-bit integers in a by packed signed 8-bit integers in b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6003 | /// |
| 6004 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_maddubs_epi16&expand=3538) |
| 6005 | #[inline ] |
| 6006 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6007 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6008 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
| 6009 | pub fn _mm256_maskz_maddubs_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 6010 | unsafe { |
| 6011 | let madd: i16x16 = _mm256_maddubs_epi16(a, b).as_i16x16(); |
| 6012 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:i16x16::ZERO)) |
| 6013 | } |
| 6014 | } |
| 6015 | |
| 6016 | /// Multiply packed unsigned 8-bit integers in a by packed signed 8-bit integers in b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6017 | /// |
| 6018 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_maddubs_epi16&expand=3534) |
| 6019 | #[inline ] |
| 6020 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6021 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6022 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
| 6023 | pub fn _mm_mask_maddubs_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 6024 | unsafe { |
| 6025 | let madd: i16x8 = _mm_maddubs_epi16(a, b).as_i16x8(); |
| 6026 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:src.as_i16x8())) |
| 6027 | } |
| 6028 | } |
| 6029 | |
| 6030 | /// Multiply packed unsigned 8-bit integers in a by packed signed 8-bit integers in b, producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6031 | /// |
| 6032 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_maddubs_epi16&expand=3535) |
| 6033 | #[inline ] |
| 6034 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6035 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6036 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
| 6037 | pub fn _mm_maskz_maddubs_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 6038 | unsafe { |
| 6039 | let madd: i16x8 = _mm_maddubs_epi16(a, b).as_i16x8(); |
| 6040 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:i16x8::ZERO)) |
| 6041 | } |
| 6042 | } |
| 6043 | |
| 6044 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst. |
| 6045 | /// |
| 6046 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_packs_epi32&expand=4091) |
| 6047 | #[inline ] |
| 6048 | #[target_feature (enable = "avx512bw" )] |
| 6049 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6050 | #[cfg_attr (test, assert_instr(vpackssdw))] |
| 6051 | pub fn _mm512_packs_epi32(a: __m512i, b: __m512i) -> __m512i { |
| 6052 | unsafe { transmute(src:vpackssdw(a.as_i32x16(), b.as_i32x16())) } |
| 6053 | } |
| 6054 | |
| 6055 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6056 | /// |
| 6057 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_packs_epi32&expand=4089) |
| 6058 | #[inline ] |
| 6059 | #[target_feature (enable = "avx512bw" )] |
| 6060 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6061 | #[cfg_attr (test, assert_instr(vpackssdw))] |
| 6062 | pub fn _mm512_mask_packs_epi32(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 6063 | unsafe { |
| 6064 | let pack: i16x32 = _mm512_packs_epi32(a, b).as_i16x32(); |
| 6065 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i16x32())) |
| 6066 | } |
| 6067 | } |
| 6068 | |
| 6069 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6070 | /// |
| 6071 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_packs_epi32&expand=4090) |
| 6072 | #[inline ] |
| 6073 | #[target_feature (enable = "avx512bw" )] |
| 6074 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6075 | #[cfg_attr (test, assert_instr(vpackssdw))] |
| 6076 | pub fn _mm512_maskz_packs_epi32(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 6077 | unsafe { |
| 6078 | let pack: i16x32 = _mm512_packs_epi32(a, b).as_i16x32(); |
| 6079 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:i16x32::ZERO)) |
| 6080 | } |
| 6081 | } |
| 6082 | |
| 6083 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6084 | /// |
| 6085 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_packs_epi32&expand=4086) |
| 6086 | #[inline ] |
| 6087 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6088 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6089 | #[cfg_attr (test, assert_instr(vpackssdw))] |
| 6090 | pub fn _mm256_mask_packs_epi32(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 6091 | unsafe { |
| 6092 | let pack: i16x16 = _mm256_packs_epi32(a, b).as_i16x16(); |
| 6093 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i16x16())) |
| 6094 | } |
| 6095 | } |
| 6096 | |
| 6097 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6098 | /// |
| 6099 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_packs_epi32&expand=4087) |
| 6100 | #[inline ] |
| 6101 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6102 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6103 | #[cfg_attr (test, assert_instr(vpackssdw))] |
| 6104 | pub fn _mm256_maskz_packs_epi32(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 6105 | unsafe { |
| 6106 | let pack: i16x16 = _mm256_packs_epi32(a, b).as_i16x16(); |
| 6107 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:i16x16::ZERO)) |
| 6108 | } |
| 6109 | } |
| 6110 | |
| 6111 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6112 | /// |
| 6113 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_packs_epi32&expand=4083) |
| 6114 | #[inline ] |
| 6115 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6116 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6117 | #[cfg_attr (test, assert_instr(vpackssdw))] |
| 6118 | pub fn _mm_mask_packs_epi32(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 6119 | unsafe { |
| 6120 | let pack: i16x8 = _mm_packs_epi32(a, b).as_i16x8(); |
| 6121 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i16x8())) |
| 6122 | } |
| 6123 | } |
| 6124 | |
| 6125 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6126 | /// |
| 6127 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_packs_epi32&expand=4084) |
| 6128 | #[inline ] |
| 6129 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6130 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6131 | #[cfg_attr (test, assert_instr(vpackssdw))] |
| 6132 | pub fn _mm_maskz_packs_epi32(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 6133 | unsafe { |
| 6134 | let pack: i16x8 = _mm_packs_epi32(a, b).as_i16x8(); |
| 6135 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:i16x8::ZERO)) |
| 6136 | } |
| 6137 | } |
| 6138 | |
| 6139 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst. |
| 6140 | /// |
| 6141 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_packs_epi16&expand=4082) |
| 6142 | #[inline ] |
| 6143 | #[target_feature (enable = "avx512bw" )] |
| 6144 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6145 | #[cfg_attr (test, assert_instr(vpacksswb))] |
| 6146 | pub fn _mm512_packs_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 6147 | unsafe { transmute(src:vpacksswb(a.as_i16x32(), b.as_i16x32())) } |
| 6148 | } |
| 6149 | |
| 6150 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6151 | /// |
| 6152 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_packs_epi16&expand=4080) |
| 6153 | #[inline ] |
| 6154 | #[target_feature (enable = "avx512bw" )] |
| 6155 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6156 | #[cfg_attr (test, assert_instr(vpacksswb))] |
| 6157 | pub fn _mm512_mask_packs_epi16(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 6158 | unsafe { |
| 6159 | let pack: i8x64 = _mm512_packs_epi16(a, b).as_i8x64(); |
| 6160 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i8x64())) |
| 6161 | } |
| 6162 | } |
| 6163 | |
| 6164 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6165 | /// |
| 6166 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_packs_epi16&expand=4081) |
| 6167 | #[inline ] |
| 6168 | #[target_feature (enable = "avx512bw" )] |
| 6169 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6170 | #[cfg_attr (test, assert_instr(vpacksswb))] |
| 6171 | pub fn _mm512_maskz_packs_epi16(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 6172 | unsafe { |
| 6173 | let pack: i8x64 = _mm512_packs_epi16(a, b).as_i8x64(); |
| 6174 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:i8x64::ZERO)) |
| 6175 | } |
| 6176 | } |
| 6177 | |
| 6178 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6179 | /// |
| 6180 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_packs_epi16&expand=4077) |
| 6181 | #[inline ] |
| 6182 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6183 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6184 | #[cfg_attr (test, assert_instr(vpacksswb))] |
| 6185 | pub fn _mm256_mask_packs_epi16(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 6186 | unsafe { |
| 6187 | let pack: i8x32 = _mm256_packs_epi16(a, b).as_i8x32(); |
| 6188 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i8x32())) |
| 6189 | } |
| 6190 | } |
| 6191 | |
| 6192 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6193 | /// |
| 6194 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=#text=_mm256_maskz_packs_epi16&expand=4078) |
| 6195 | #[inline ] |
| 6196 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6197 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6198 | #[cfg_attr (test, assert_instr(vpacksswb))] |
| 6199 | pub fn _mm256_maskz_packs_epi16(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 6200 | unsafe { |
| 6201 | let pack: i8x32 = _mm256_packs_epi16(a, b).as_i8x32(); |
| 6202 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:i8x32::ZERO)) |
| 6203 | } |
| 6204 | } |
| 6205 | |
| 6206 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6207 | /// |
| 6208 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_packs_epi16&expand=4074) |
| 6209 | #[inline ] |
| 6210 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6211 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6212 | #[cfg_attr (test, assert_instr(vpacksswb))] |
| 6213 | pub fn _mm_mask_packs_epi16(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 6214 | unsafe { |
| 6215 | let pack: i8x16 = _mm_packs_epi16(a, b).as_i8x16(); |
| 6216 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i8x16())) |
| 6217 | } |
| 6218 | } |
| 6219 | |
| 6220 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6221 | /// |
| 6222 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_packs_epi16&expand=4075) |
| 6223 | #[inline ] |
| 6224 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6225 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6226 | #[cfg_attr (test, assert_instr(vpacksswb))] |
| 6227 | pub fn _mm_maskz_packs_epi16(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 6228 | unsafe { |
| 6229 | let pack: i8x16 = _mm_packs_epi16(a, b).as_i8x16(); |
| 6230 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:i8x16::ZERO)) |
| 6231 | } |
| 6232 | } |
| 6233 | |
| 6234 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst. |
| 6235 | /// |
| 6236 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_packus_epi32&expand=4130) |
| 6237 | #[inline ] |
| 6238 | #[target_feature (enable = "avx512bw" )] |
| 6239 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6240 | #[cfg_attr (test, assert_instr(vpackusdw))] |
| 6241 | pub fn _mm512_packus_epi32(a: __m512i, b: __m512i) -> __m512i { |
| 6242 | unsafe { transmute(src:vpackusdw(a.as_i32x16(), b.as_i32x16())) } |
| 6243 | } |
| 6244 | |
| 6245 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6246 | /// |
| 6247 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_packus_epi32&expand=4128) |
| 6248 | #[inline ] |
| 6249 | #[target_feature (enable = "avx512bw" )] |
| 6250 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6251 | #[cfg_attr (test, assert_instr(vpackusdw))] |
| 6252 | pub fn _mm512_mask_packus_epi32(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 6253 | unsafe { |
| 6254 | let pack: i16x32 = _mm512_packus_epi32(a, b).as_i16x32(); |
| 6255 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i16x32())) |
| 6256 | } |
| 6257 | } |
| 6258 | |
| 6259 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6260 | /// |
| 6261 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_packus_epi32&expand=4129) |
| 6262 | #[inline ] |
| 6263 | #[target_feature (enable = "avx512bw" )] |
| 6264 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6265 | #[cfg_attr (test, assert_instr(vpackusdw))] |
| 6266 | pub fn _mm512_maskz_packus_epi32(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 6267 | unsafe { |
| 6268 | let pack: i16x32 = _mm512_packus_epi32(a, b).as_i16x32(); |
| 6269 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:i16x32::ZERO)) |
| 6270 | } |
| 6271 | } |
| 6272 | |
| 6273 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6274 | /// |
| 6275 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_packus_epi32&expand=4125) |
| 6276 | #[inline ] |
| 6277 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6278 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6279 | #[cfg_attr (test, assert_instr(vpackusdw))] |
| 6280 | pub fn _mm256_mask_packus_epi32(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 6281 | unsafe { |
| 6282 | let pack: i16x16 = _mm256_packus_epi32(a, b).as_i16x16(); |
| 6283 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i16x16())) |
| 6284 | } |
| 6285 | } |
| 6286 | |
| 6287 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6288 | /// |
| 6289 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_packus_epi32&expand=4126) |
| 6290 | #[inline ] |
| 6291 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6292 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6293 | #[cfg_attr (test, assert_instr(vpackusdw))] |
| 6294 | pub fn _mm256_maskz_packus_epi32(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 6295 | unsafe { |
| 6296 | let pack: i16x16 = _mm256_packus_epi32(a, b).as_i16x16(); |
| 6297 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:i16x16::ZERO)) |
| 6298 | } |
| 6299 | } |
| 6300 | |
| 6301 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6302 | /// |
| 6303 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_packus_epi32&expand=4122) |
| 6304 | #[inline ] |
| 6305 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6306 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6307 | #[cfg_attr (test, assert_instr(vpackusdw))] |
| 6308 | pub fn _mm_mask_packus_epi32(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 6309 | unsafe { |
| 6310 | let pack: i16x8 = _mm_packus_epi32(a, b).as_i16x8(); |
| 6311 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i16x8())) |
| 6312 | } |
| 6313 | } |
| 6314 | |
| 6315 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6316 | /// |
| 6317 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_packus_epi32&expand=4123) |
| 6318 | #[inline ] |
| 6319 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6320 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6321 | #[cfg_attr (test, assert_instr(vpackusdw))] |
| 6322 | pub fn _mm_maskz_packus_epi32(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 6323 | unsafe { |
| 6324 | let pack: i16x8 = _mm_packus_epi32(a, b).as_i16x8(); |
| 6325 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:i16x8::ZERO)) |
| 6326 | } |
| 6327 | } |
| 6328 | |
| 6329 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst. |
| 6330 | /// |
| 6331 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_packus_epi16&expand=4121) |
| 6332 | #[inline ] |
| 6333 | #[target_feature (enable = "avx512bw" )] |
| 6334 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6335 | #[cfg_attr (test, assert_instr(vpackuswb))] |
| 6336 | pub fn _mm512_packus_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 6337 | unsafe { transmute(src:vpackuswb(a.as_i16x32(), b.as_i16x32())) } |
| 6338 | } |
| 6339 | |
| 6340 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6341 | /// |
| 6342 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_packus_epi16&expand=4119) |
| 6343 | #[inline ] |
| 6344 | #[target_feature (enable = "avx512bw" )] |
| 6345 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6346 | #[cfg_attr (test, assert_instr(vpackuswb))] |
| 6347 | pub fn _mm512_mask_packus_epi16(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 6348 | unsafe { |
| 6349 | let pack: i8x64 = _mm512_packus_epi16(a, b).as_i8x64(); |
| 6350 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i8x64())) |
| 6351 | } |
| 6352 | } |
| 6353 | |
| 6354 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6355 | /// |
| 6356 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_packus_epi16&expand=4120) |
| 6357 | #[inline ] |
| 6358 | #[target_feature (enable = "avx512bw" )] |
| 6359 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6360 | #[cfg_attr (test, assert_instr(vpackuswb))] |
| 6361 | pub fn _mm512_maskz_packus_epi16(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 6362 | unsafe { |
| 6363 | let pack: i8x64 = _mm512_packus_epi16(a, b).as_i8x64(); |
| 6364 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:i8x64::ZERO)) |
| 6365 | } |
| 6366 | } |
| 6367 | |
| 6368 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6369 | /// |
| 6370 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_packus_epi16&expand=4116) |
| 6371 | #[inline ] |
| 6372 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6373 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6374 | #[cfg_attr (test, assert_instr(vpackuswb))] |
| 6375 | pub fn _mm256_mask_packus_epi16(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 6376 | unsafe { |
| 6377 | let pack: i8x32 = _mm256_packus_epi16(a, b).as_i8x32(); |
| 6378 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i8x32())) |
| 6379 | } |
| 6380 | } |
| 6381 | |
| 6382 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6383 | /// |
| 6384 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_packus_epi16&expand=4117) |
| 6385 | #[inline ] |
| 6386 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6387 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6388 | #[cfg_attr (test, assert_instr(vpackuswb))] |
| 6389 | pub fn _mm256_maskz_packus_epi16(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 6390 | unsafe { |
| 6391 | let pack: i8x32 = _mm256_packus_epi16(a, b).as_i8x32(); |
| 6392 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:i8x32::ZERO)) |
| 6393 | } |
| 6394 | } |
| 6395 | |
| 6396 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6397 | /// |
| 6398 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_packus_epi16&expand=4113) |
| 6399 | #[inline ] |
| 6400 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6401 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6402 | #[cfg_attr (test, assert_instr(vpackuswb))] |
| 6403 | pub fn _mm_mask_packus_epi16(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 6404 | unsafe { |
| 6405 | let pack: i8x16 = _mm_packus_epi16(a, b).as_i8x16(); |
| 6406 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i8x16())) |
| 6407 | } |
| 6408 | } |
| 6409 | |
| 6410 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6411 | /// |
| 6412 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_packus_epi16&expand=4114) |
| 6413 | #[inline ] |
| 6414 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6415 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6416 | #[cfg_attr (test, assert_instr(vpackuswb))] |
| 6417 | pub fn _mm_maskz_packus_epi16(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 6418 | unsafe { |
| 6419 | let pack: i8x16 = _mm_packus_epi16(a, b).as_i8x16(); |
| 6420 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:i8x16::ZERO)) |
| 6421 | } |
| 6422 | } |
| 6423 | |
| 6424 | /// Average packed unsigned 16-bit integers in a and b, and store the results in dst. |
| 6425 | /// |
| 6426 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_avg_epu16&expand=388) |
| 6427 | #[inline ] |
| 6428 | #[target_feature (enable = "avx512bw" )] |
| 6429 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6430 | #[cfg_attr (test, assert_instr(vpavgw))] |
| 6431 | pub fn _mm512_avg_epu16(a: __m512i, b: __m512i) -> __m512i { |
| 6432 | unsafe { |
| 6433 | let a: u32x32 = simd_cast::<_, u32x32>(a.as_u16x32()); |
| 6434 | let b: u32x32 = simd_cast::<_, u32x32>(b.as_u16x32()); |
| 6435 | let r: u32x32 = simd_shr(lhs:simd_add(simd_add(a, b), u32x32::splat(1)), rhs:u32x32::splat(1)); |
| 6436 | transmute(src:simd_cast::<_, u16x32>(r)) |
| 6437 | } |
| 6438 | } |
| 6439 | |
| 6440 | /// Average packed unsigned 16-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6441 | /// |
| 6442 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_avg_epu16&expand=389) |
| 6443 | #[inline ] |
| 6444 | #[target_feature (enable = "avx512bw" )] |
| 6445 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6446 | #[cfg_attr (test, assert_instr(vpavgw))] |
| 6447 | pub fn _mm512_mask_avg_epu16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 6448 | unsafe { |
| 6449 | let avg: u16x32 = _mm512_avg_epu16(a, b).as_u16x32(); |
| 6450 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:src.as_u16x32())) |
| 6451 | } |
| 6452 | } |
| 6453 | |
| 6454 | /// Average packed unsigned 16-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6455 | /// |
| 6456 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_avg_epu16&expand=390) |
| 6457 | #[inline ] |
| 6458 | #[target_feature (enable = "avx512bw" )] |
| 6459 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6460 | #[cfg_attr (test, assert_instr(vpavgw))] |
| 6461 | pub fn _mm512_maskz_avg_epu16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 6462 | unsafe { |
| 6463 | let avg: u16x32 = _mm512_avg_epu16(a, b).as_u16x32(); |
| 6464 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:u16x32::ZERO)) |
| 6465 | } |
| 6466 | } |
| 6467 | |
| 6468 | /// Average packed unsigned 16-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6469 | /// |
| 6470 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_avg_epu16&expand=386) |
| 6471 | #[inline ] |
| 6472 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6473 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6474 | #[cfg_attr (test, assert_instr(vpavgw))] |
| 6475 | pub fn _mm256_mask_avg_epu16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 6476 | unsafe { |
| 6477 | let avg: u16x16 = _mm256_avg_epu16(a, b).as_u16x16(); |
| 6478 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:src.as_u16x16())) |
| 6479 | } |
| 6480 | } |
| 6481 | |
| 6482 | /// Average packed unsigned 16-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6483 | /// |
| 6484 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_avg_epu16&expand=387) |
| 6485 | #[inline ] |
| 6486 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6487 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6488 | #[cfg_attr (test, assert_instr(vpavgw))] |
| 6489 | pub fn _mm256_maskz_avg_epu16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 6490 | unsafe { |
| 6491 | let avg: u16x16 = _mm256_avg_epu16(a, b).as_u16x16(); |
| 6492 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:u16x16::ZERO)) |
| 6493 | } |
| 6494 | } |
| 6495 | |
| 6496 | /// Average packed unsigned 16-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6497 | /// |
| 6498 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_avg_epu16&expand=383) |
| 6499 | #[inline ] |
| 6500 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6501 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6502 | #[cfg_attr (test, assert_instr(vpavgw))] |
| 6503 | pub fn _mm_mask_avg_epu16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 6504 | unsafe { |
| 6505 | let avg: u16x8 = _mm_avg_epu16(a, b).as_u16x8(); |
| 6506 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:src.as_u16x8())) |
| 6507 | } |
| 6508 | } |
| 6509 | |
| 6510 | /// Average packed unsigned 16-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6511 | /// |
| 6512 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_avg_epu16&expand=384) |
| 6513 | #[inline ] |
| 6514 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6515 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6516 | #[cfg_attr (test, assert_instr(vpavgw))] |
| 6517 | pub fn _mm_maskz_avg_epu16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 6518 | unsafe { |
| 6519 | let avg: u16x8 = _mm_avg_epu16(a, b).as_u16x8(); |
| 6520 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:u16x8::ZERO)) |
| 6521 | } |
| 6522 | } |
| 6523 | |
| 6524 | /// Average packed unsigned 8-bit integers in a and b, and store the results in dst. |
| 6525 | /// |
| 6526 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_avg_epu8&expand=397) |
| 6527 | #[inline ] |
| 6528 | #[target_feature (enable = "avx512bw" )] |
| 6529 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6530 | #[cfg_attr (test, assert_instr(vpavgb))] |
| 6531 | pub fn _mm512_avg_epu8(a: __m512i, b: __m512i) -> __m512i { |
| 6532 | unsafe { |
| 6533 | let a: u16x64 = simd_cast::<_, u16x64>(a.as_u8x64()); |
| 6534 | let b: u16x64 = simd_cast::<_, u16x64>(b.as_u8x64()); |
| 6535 | let r: u16x64 = simd_shr(lhs:simd_add(simd_add(a, b), u16x64::splat(1)), rhs:u16x64::splat(1)); |
| 6536 | transmute(src:simd_cast::<_, u8x64>(r)) |
| 6537 | } |
| 6538 | } |
| 6539 | |
| 6540 | /// Average packed unsigned 8-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6541 | /// |
| 6542 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_avg_epu8&expand=398) |
| 6543 | #[inline ] |
| 6544 | #[target_feature (enable = "avx512bw" )] |
| 6545 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6546 | #[cfg_attr (test, assert_instr(vpavgb))] |
| 6547 | pub fn _mm512_mask_avg_epu8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 6548 | unsafe { |
| 6549 | let avg: u8x64 = _mm512_avg_epu8(a, b).as_u8x64(); |
| 6550 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:src.as_u8x64())) |
| 6551 | } |
| 6552 | } |
| 6553 | |
| 6554 | /// Average packed unsigned 8-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6555 | /// |
| 6556 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_avg_epu8&expand=399) |
| 6557 | #[inline ] |
| 6558 | #[target_feature (enable = "avx512bw" )] |
| 6559 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6560 | #[cfg_attr (test, assert_instr(vpavgb))] |
| 6561 | pub fn _mm512_maskz_avg_epu8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 6562 | unsafe { |
| 6563 | let avg: u8x64 = _mm512_avg_epu8(a, b).as_u8x64(); |
| 6564 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:u8x64::ZERO)) |
| 6565 | } |
| 6566 | } |
| 6567 | |
| 6568 | /// Average packed unsigned 8-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6569 | /// |
| 6570 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_avg_epu8&expand=395) |
| 6571 | #[inline ] |
| 6572 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6573 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6574 | #[cfg_attr (test, assert_instr(vpavgb))] |
| 6575 | pub fn _mm256_mask_avg_epu8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 6576 | unsafe { |
| 6577 | let avg: u8x32 = _mm256_avg_epu8(a, b).as_u8x32(); |
| 6578 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:src.as_u8x32())) |
| 6579 | } |
| 6580 | } |
| 6581 | |
| 6582 | /// Average packed unsigned 8-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6583 | /// |
| 6584 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_avg_epu8&expand=396) |
| 6585 | #[inline ] |
| 6586 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6587 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6588 | #[cfg_attr (test, assert_instr(vpavgb))] |
| 6589 | pub fn _mm256_maskz_avg_epu8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 6590 | unsafe { |
| 6591 | let avg: u8x32 = _mm256_avg_epu8(a, b).as_u8x32(); |
| 6592 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:u8x32::ZERO)) |
| 6593 | } |
| 6594 | } |
| 6595 | |
| 6596 | /// Average packed unsigned 8-bit integers in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6597 | /// |
| 6598 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_avg_epu8&expand=392) |
| 6599 | #[inline ] |
| 6600 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6601 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6602 | #[cfg_attr (test, assert_instr(vpavgb))] |
| 6603 | pub fn _mm_mask_avg_epu8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 6604 | unsafe { |
| 6605 | let avg: u8x16 = _mm_avg_epu8(a, b).as_u8x16(); |
| 6606 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:src.as_u8x16())) |
| 6607 | } |
| 6608 | } |
| 6609 | |
| 6610 | /// Average packed unsigned 8-bit integers in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6611 | /// |
| 6612 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_avg_epu8&expand=393) |
| 6613 | #[inline ] |
| 6614 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6615 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6616 | #[cfg_attr (test, assert_instr(vpavgb))] |
| 6617 | pub fn _mm_maskz_avg_epu8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 6618 | unsafe { |
| 6619 | let avg: u8x16 = _mm_avg_epu8(a, b).as_u8x16(); |
| 6620 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:u8x16::ZERO)) |
| 6621 | } |
| 6622 | } |
| 6623 | |
| 6624 | /// Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst. |
| 6625 | /// |
| 6626 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_sll_epi16&expand=5271) |
| 6627 | #[inline ] |
| 6628 | #[target_feature (enable = "avx512bw" )] |
| 6629 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6630 | #[cfg_attr (test, assert_instr(vpsllw))] |
| 6631 | pub fn _mm512_sll_epi16(a: __m512i, count: __m128i) -> __m512i { |
| 6632 | unsafe { transmute(src:vpsllw(a.as_i16x32(), count.as_i16x8())) } |
| 6633 | } |
| 6634 | |
| 6635 | /// Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6636 | /// |
| 6637 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_sll_epi16&expand=5269) |
| 6638 | #[inline ] |
| 6639 | #[target_feature (enable = "avx512bw" )] |
| 6640 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6641 | #[cfg_attr (test, assert_instr(vpsllw))] |
| 6642 | pub fn _mm512_mask_sll_epi16(src: __m512i, k: __mmask32, a: __m512i, count: __m128i) -> __m512i { |
| 6643 | unsafe { |
| 6644 | let shf: i16x32 = _mm512_sll_epi16(a, count).as_i16x32(); |
| 6645 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
| 6646 | } |
| 6647 | } |
| 6648 | |
| 6649 | /// Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6650 | /// |
| 6651 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_sll_epi16&expand=5270) |
| 6652 | #[inline ] |
| 6653 | #[target_feature (enable = "avx512bw" )] |
| 6654 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6655 | #[cfg_attr (test, assert_instr(vpsllw))] |
| 6656 | pub fn _mm512_maskz_sll_epi16(k: __mmask32, a: __m512i, count: __m128i) -> __m512i { |
| 6657 | unsafe { |
| 6658 | let shf: i16x32 = _mm512_sll_epi16(a, count).as_i16x32(); |
| 6659 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x32::ZERO)) |
| 6660 | } |
| 6661 | } |
| 6662 | |
| 6663 | /// Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6664 | /// |
| 6665 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_sll_epi16&expand=5266) |
| 6666 | #[inline ] |
| 6667 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6668 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6669 | #[cfg_attr (test, assert_instr(vpsllw))] |
| 6670 | pub fn _mm256_mask_sll_epi16(src: __m256i, k: __mmask16, a: __m256i, count: __m128i) -> __m256i { |
| 6671 | unsafe { |
| 6672 | let shf: i16x16 = _mm256_sll_epi16(a, count).as_i16x16(); |
| 6673 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x16())) |
| 6674 | } |
| 6675 | } |
| 6676 | |
| 6677 | /// Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6678 | /// |
| 6679 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_sll_epi16&expand=5267) |
| 6680 | #[inline ] |
| 6681 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6682 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6683 | #[cfg_attr (test, assert_instr(vpsllw))] |
| 6684 | pub fn _mm256_maskz_sll_epi16(k: __mmask16, a: __m256i, count: __m128i) -> __m256i { |
| 6685 | unsafe { |
| 6686 | let shf: i16x16 = _mm256_sll_epi16(a, count).as_i16x16(); |
| 6687 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x16::ZERO)) |
| 6688 | } |
| 6689 | } |
| 6690 | |
| 6691 | /// Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6692 | /// |
| 6693 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_sll_epi16&expand=5263) |
| 6694 | #[inline ] |
| 6695 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6696 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6697 | #[cfg_attr (test, assert_instr(vpsllw))] |
| 6698 | pub fn _mm_mask_sll_epi16(src: __m128i, k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
| 6699 | unsafe { |
| 6700 | let shf: i16x8 = _mm_sll_epi16(a, count).as_i16x8(); |
| 6701 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x8())) |
| 6702 | } |
| 6703 | } |
| 6704 | |
| 6705 | /// Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6706 | /// |
| 6707 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_sll_epi16&expand=5264) |
| 6708 | #[inline ] |
| 6709 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6710 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6711 | #[cfg_attr (test, assert_instr(vpsllw))] |
| 6712 | pub fn _mm_maskz_sll_epi16(k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
| 6713 | unsafe { |
| 6714 | let shf: i16x8 = _mm_sll_epi16(a, count).as_i16x8(); |
| 6715 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x8::ZERO)) |
| 6716 | } |
| 6717 | } |
| 6718 | |
| 6719 | /// Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst. |
| 6720 | /// |
| 6721 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_slli_epi16&expand=5301) |
| 6722 | #[inline ] |
| 6723 | #[target_feature (enable = "avx512bw" )] |
| 6724 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6725 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
| 6726 | #[rustc_legacy_const_generics (1)] |
| 6727 | pub fn _mm512_slli_epi16<const IMM8: u32>(a: __m512i) -> __m512i { |
| 6728 | unsafe { |
| 6729 | static_assert_uimm_bits!(IMM8, 8); |
| 6730 | if IMM8 >= 16 { |
| 6731 | _mm512_setzero_si512() |
| 6732 | } else { |
| 6733 | transmute(src:simd_shl(lhs:a.as_u16x32(), rhs:u16x32::splat(IMM8 as u16))) |
| 6734 | } |
| 6735 | } |
| 6736 | } |
| 6737 | |
| 6738 | /// Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6739 | /// |
| 6740 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_slli_epi16&expand=5299) |
| 6741 | #[inline ] |
| 6742 | #[target_feature (enable = "avx512bw" )] |
| 6743 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6744 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
| 6745 | #[rustc_legacy_const_generics (3)] |
| 6746 | pub fn _mm512_mask_slli_epi16<const IMM8: u32>(src: __m512i, k: __mmask32, a: __m512i) -> __m512i { |
| 6747 | unsafe { |
| 6748 | static_assert_uimm_bits!(IMM8, 8); |
| 6749 | let shf: u16x32 = if IMM8 >= 16 { |
| 6750 | u16x32::ZERO |
| 6751 | } else { |
| 6752 | simd_shl(lhs:a.as_u16x32(), rhs:u16x32::splat(IMM8 as u16)) |
| 6753 | }; |
| 6754 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_u16x32())) |
| 6755 | } |
| 6756 | } |
| 6757 | |
| 6758 | /// Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6759 | /// |
| 6760 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_slli_epi16&expand=5300) |
| 6761 | #[inline ] |
| 6762 | #[target_feature (enable = "avx512bw" )] |
| 6763 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6764 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
| 6765 | #[rustc_legacy_const_generics (2)] |
| 6766 | pub fn _mm512_maskz_slli_epi16<const IMM8: u32>(k: __mmask32, a: __m512i) -> __m512i { |
| 6767 | unsafe { |
| 6768 | static_assert_uimm_bits!(IMM8, 8); |
| 6769 | if IMM8 >= 16 { |
| 6770 | _mm512_setzero_si512() |
| 6771 | } else { |
| 6772 | let shf: u16x32 = simd_shl(lhs:a.as_u16x32(), rhs:u16x32::splat(IMM8 as u16)); |
| 6773 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:u16x32::ZERO)) |
| 6774 | } |
| 6775 | } |
| 6776 | } |
| 6777 | |
| 6778 | /// Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6779 | /// |
| 6780 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_slli_epi16&expand=5296) |
| 6781 | #[inline ] |
| 6782 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6783 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6784 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
| 6785 | #[rustc_legacy_const_generics (3)] |
| 6786 | pub fn _mm256_mask_slli_epi16<const IMM8: u32>(src: __m256i, k: __mmask16, a: __m256i) -> __m256i { |
| 6787 | unsafe { |
| 6788 | static_assert_uimm_bits!(IMM8, 8); |
| 6789 | let shf: u16x16 = if IMM8 >= 16 { |
| 6790 | u16x16::ZERO |
| 6791 | } else { |
| 6792 | simd_shl(lhs:a.as_u16x16(), rhs:u16x16::splat(IMM8 as u16)) |
| 6793 | }; |
| 6794 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_u16x16())) |
| 6795 | } |
| 6796 | } |
| 6797 | |
| 6798 | /// Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6799 | /// |
| 6800 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_slli_epi16&expand=5297) |
| 6801 | #[inline ] |
| 6802 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6803 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6804 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
| 6805 | #[rustc_legacy_const_generics (2)] |
| 6806 | pub fn _mm256_maskz_slli_epi16<const IMM8: u32>(k: __mmask16, a: __m256i) -> __m256i { |
| 6807 | unsafe { |
| 6808 | static_assert_uimm_bits!(IMM8, 8); |
| 6809 | if IMM8 >= 16 { |
| 6810 | _mm256_setzero_si256() |
| 6811 | } else { |
| 6812 | let shf: u16x16 = simd_shl(lhs:a.as_u16x16(), rhs:u16x16::splat(IMM8 as u16)); |
| 6813 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:u16x16::ZERO)) |
| 6814 | } |
| 6815 | } |
| 6816 | } |
| 6817 | |
| 6818 | /// Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6819 | /// |
| 6820 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_slli_epi16&expand=5293) |
| 6821 | #[inline ] |
| 6822 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6823 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6824 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
| 6825 | #[rustc_legacy_const_generics (3)] |
| 6826 | pub fn _mm_mask_slli_epi16<const IMM8: u32>(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 6827 | unsafe { |
| 6828 | static_assert_uimm_bits!(IMM8, 8); |
| 6829 | let shf: u16x8 = if IMM8 >= 16 { |
| 6830 | u16x8::ZERO |
| 6831 | } else { |
| 6832 | simd_shl(lhs:a.as_u16x8(), rhs:u16x8::splat(IMM8 as u16)) |
| 6833 | }; |
| 6834 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_u16x8())) |
| 6835 | } |
| 6836 | } |
| 6837 | |
| 6838 | /// Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6839 | /// |
| 6840 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_slli_epi16&expand=5294) |
| 6841 | #[inline ] |
| 6842 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6843 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6844 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
| 6845 | #[rustc_legacy_const_generics (2)] |
| 6846 | pub fn _mm_maskz_slli_epi16<const IMM8: u32>(k: __mmask8, a: __m128i) -> __m128i { |
| 6847 | unsafe { |
| 6848 | static_assert_uimm_bits!(IMM8, 8); |
| 6849 | if IMM8 >= 16 { |
| 6850 | _mm_setzero_si128() |
| 6851 | } else { |
| 6852 | let shf: u16x8 = simd_shl(lhs:a.as_u16x8(), rhs:u16x8::splat(IMM8 as u16)); |
| 6853 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:u16x8::ZERO)) |
| 6854 | } |
| 6855 | } |
| 6856 | } |
| 6857 | |
| 6858 | /// Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst. |
| 6859 | /// |
| 6860 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_sllv_epi16&expand=5333) |
| 6861 | #[inline ] |
| 6862 | #[target_feature (enable = "avx512bw" )] |
| 6863 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6864 | #[cfg_attr (test, assert_instr(vpsllvw))] |
| 6865 | pub fn _mm512_sllv_epi16(a: __m512i, count: __m512i) -> __m512i { |
| 6866 | unsafe { |
| 6867 | let count: u16x32 = count.as_u16x32(); |
| 6868 | let no_overflow: u16x32 = simd_lt(x:count, y:u16x32::splat(u16::BITS as u16)); |
| 6869 | let count: u16x32 = simd_select(mask:no_overflow, if_true:count, if_false:u16x32::ZERO); |
| 6870 | simd_select(mask:no_overflow, if_true:simd_shl(a.as_u16x32(), count), if_false:u16x32::ZERO).as_m512i() |
| 6871 | } |
| 6872 | } |
| 6873 | |
| 6874 | /// Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6875 | /// |
| 6876 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_sllv_epi16&expand=5331) |
| 6877 | #[inline ] |
| 6878 | #[target_feature (enable = "avx512bw" )] |
| 6879 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6880 | #[cfg_attr (test, assert_instr(vpsllvw))] |
| 6881 | pub fn _mm512_mask_sllv_epi16(src: __m512i, k: __mmask32, a: __m512i, count: __m512i) -> __m512i { |
| 6882 | unsafe { |
| 6883 | let shf: i16x32 = _mm512_sllv_epi16(a, count).as_i16x32(); |
| 6884 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
| 6885 | } |
| 6886 | } |
| 6887 | |
| 6888 | /// Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6889 | /// |
| 6890 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_sllv_epi16&expand=5332) |
| 6891 | #[inline ] |
| 6892 | #[target_feature (enable = "avx512bw" )] |
| 6893 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6894 | #[cfg_attr (test, assert_instr(vpsllvw))] |
| 6895 | pub fn _mm512_maskz_sllv_epi16(k: __mmask32, a: __m512i, count: __m512i) -> __m512i { |
| 6896 | unsafe { |
| 6897 | let shf: i16x32 = _mm512_sllv_epi16(a, count).as_i16x32(); |
| 6898 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x32::ZERO)) |
| 6899 | } |
| 6900 | } |
| 6901 | |
| 6902 | /// Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst. |
| 6903 | /// |
| 6904 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_sllv_epi16&expand=5330) |
| 6905 | #[inline ] |
| 6906 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6907 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6908 | #[cfg_attr (test, assert_instr(vpsllvw))] |
| 6909 | pub fn _mm256_sllv_epi16(a: __m256i, count: __m256i) -> __m256i { |
| 6910 | unsafe { |
| 6911 | let count: u16x16 = count.as_u16x16(); |
| 6912 | let no_overflow: u16x16 = simd_lt(x:count, y:u16x16::splat(u16::BITS as u16)); |
| 6913 | let count: u16x16 = simd_select(mask:no_overflow, if_true:count, if_false:u16x16::ZERO); |
| 6914 | simd_select(mask:no_overflow, if_true:simd_shl(a.as_u16x16(), count), if_false:u16x16::ZERO).as_m256i() |
| 6915 | } |
| 6916 | } |
| 6917 | |
| 6918 | /// Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6919 | /// |
| 6920 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_sllv_epi16&expand=5328) |
| 6921 | #[inline ] |
| 6922 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6923 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6924 | #[cfg_attr (test, assert_instr(vpsllvw))] |
| 6925 | pub fn _mm256_mask_sllv_epi16(src: __m256i, k: __mmask16, a: __m256i, count: __m256i) -> __m256i { |
| 6926 | unsafe { |
| 6927 | let shf: i16x16 = _mm256_sllv_epi16(a, count).as_i16x16(); |
| 6928 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x16())) |
| 6929 | } |
| 6930 | } |
| 6931 | |
| 6932 | /// Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6933 | /// |
| 6934 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_sllv_epi16&expand=5329) |
| 6935 | #[inline ] |
| 6936 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6937 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6938 | #[cfg_attr (test, assert_instr(vpsllvw))] |
| 6939 | pub fn _mm256_maskz_sllv_epi16(k: __mmask16, a: __m256i, count: __m256i) -> __m256i { |
| 6940 | unsafe { |
| 6941 | let shf: i16x16 = _mm256_sllv_epi16(a, count).as_i16x16(); |
| 6942 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x16::ZERO)) |
| 6943 | } |
| 6944 | } |
| 6945 | |
| 6946 | /// Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst. |
| 6947 | /// |
| 6948 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_sllv_epi16&expand=5327) |
| 6949 | #[inline ] |
| 6950 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6951 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6952 | #[cfg_attr (test, assert_instr(vpsllvw))] |
| 6953 | pub fn _mm_sllv_epi16(a: __m128i, count: __m128i) -> __m128i { |
| 6954 | unsafe { |
| 6955 | let count: u16x8 = count.as_u16x8(); |
| 6956 | let no_overflow: u16x8 = simd_lt(x:count, y:u16x8::splat(u16::BITS as u16)); |
| 6957 | let count: u16x8 = simd_select(mask:no_overflow, if_true:count, if_false:u16x8::ZERO); |
| 6958 | simd_select(mask:no_overflow, if_true:simd_shl(a.as_u16x8(), count), if_false:u16x8::ZERO).as_m128i() |
| 6959 | } |
| 6960 | } |
| 6961 | |
| 6962 | /// Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 6963 | /// |
| 6964 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_sllv_epi16&expand=5325) |
| 6965 | #[inline ] |
| 6966 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6967 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6968 | #[cfg_attr (test, assert_instr(vpsllvw))] |
| 6969 | pub fn _mm_mask_sllv_epi16(src: __m128i, k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
| 6970 | unsafe { |
| 6971 | let shf: i16x8 = _mm_sllv_epi16(a, count).as_i16x8(); |
| 6972 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x8())) |
| 6973 | } |
| 6974 | } |
| 6975 | |
| 6976 | /// Shift packed 16-bit integers in a left by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 6977 | /// |
| 6978 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_sllv_epi16&expand=5326) |
| 6979 | #[inline ] |
| 6980 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 6981 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6982 | #[cfg_attr (test, assert_instr(vpsllvw))] |
| 6983 | pub fn _mm_maskz_sllv_epi16(k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
| 6984 | unsafe { |
| 6985 | let shf: i16x8 = _mm_sllv_epi16(a, count).as_i16x8(); |
| 6986 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x8::ZERO)) |
| 6987 | } |
| 6988 | } |
| 6989 | |
| 6990 | /// Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst. |
| 6991 | /// |
| 6992 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_srl_epi16&expand=5483) |
| 6993 | #[inline ] |
| 6994 | #[target_feature (enable = "avx512bw" )] |
| 6995 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 6996 | #[cfg_attr (test, assert_instr(vpsrlw))] |
| 6997 | pub fn _mm512_srl_epi16(a: __m512i, count: __m128i) -> __m512i { |
| 6998 | unsafe { transmute(src:vpsrlw(a.as_i16x32(), count.as_i16x8())) } |
| 6999 | } |
| 7000 | |
| 7001 | /// Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7002 | /// |
| 7003 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_srl_epi16&expand=5481) |
| 7004 | #[inline ] |
| 7005 | #[target_feature (enable = "avx512bw" )] |
| 7006 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7007 | #[cfg_attr (test, assert_instr(vpsrlw))] |
| 7008 | pub fn _mm512_mask_srl_epi16(src: __m512i, k: __mmask32, a: __m512i, count: __m128i) -> __m512i { |
| 7009 | unsafe { |
| 7010 | let shf: i16x32 = _mm512_srl_epi16(a, count).as_i16x32(); |
| 7011 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
| 7012 | } |
| 7013 | } |
| 7014 | |
| 7015 | /// Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7016 | /// |
| 7017 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_srl_epi16&expand=5482) |
| 7018 | #[inline ] |
| 7019 | #[target_feature (enable = "avx512bw" )] |
| 7020 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7021 | #[cfg_attr (test, assert_instr(vpsrlw))] |
| 7022 | pub fn _mm512_maskz_srl_epi16(k: __mmask32, a: __m512i, count: __m128i) -> __m512i { |
| 7023 | unsafe { |
| 7024 | let shf: i16x32 = _mm512_srl_epi16(a, count).as_i16x32(); |
| 7025 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x32::ZERO)) |
| 7026 | } |
| 7027 | } |
| 7028 | |
| 7029 | /// Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7030 | /// |
| 7031 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_srl_epi16&expand=5478) |
| 7032 | #[inline ] |
| 7033 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7034 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7035 | #[cfg_attr (test, assert_instr(vpsrlw))] |
| 7036 | pub fn _mm256_mask_srl_epi16(src: __m256i, k: __mmask16, a: __m256i, count: __m128i) -> __m256i { |
| 7037 | unsafe { |
| 7038 | let shf: i16x16 = _mm256_srl_epi16(a, count).as_i16x16(); |
| 7039 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x16())) |
| 7040 | } |
| 7041 | } |
| 7042 | |
| 7043 | /// Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7044 | /// |
| 7045 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_srl_epi16&expand=5479) |
| 7046 | #[inline ] |
| 7047 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7048 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7049 | #[cfg_attr (test, assert_instr(vpsrlw))] |
| 7050 | pub fn _mm256_maskz_srl_epi16(k: __mmask16, a: __m256i, count: __m128i) -> __m256i { |
| 7051 | unsafe { |
| 7052 | let shf: i16x16 = _mm256_srl_epi16(a, count).as_i16x16(); |
| 7053 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x16::ZERO)) |
| 7054 | } |
| 7055 | } |
| 7056 | |
| 7057 | /// Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7058 | /// |
| 7059 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_srl_epi16&expand=5475) |
| 7060 | #[inline ] |
| 7061 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7062 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7063 | #[cfg_attr (test, assert_instr(vpsrlw))] |
| 7064 | pub fn _mm_mask_srl_epi16(src: __m128i, k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
| 7065 | unsafe { |
| 7066 | let shf: i16x8 = _mm_srl_epi16(a, count).as_i16x8(); |
| 7067 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x8())) |
| 7068 | } |
| 7069 | } |
| 7070 | |
| 7071 | /// Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7072 | /// |
| 7073 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_srl_epi16&expand=5476) |
| 7074 | #[inline ] |
| 7075 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7076 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7077 | #[cfg_attr (test, assert_instr(vpsrlw))] |
| 7078 | pub fn _mm_maskz_srl_epi16(k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
| 7079 | unsafe { |
| 7080 | let shf: i16x8 = _mm_srl_epi16(a, count).as_i16x8(); |
| 7081 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x8::ZERO)) |
| 7082 | } |
| 7083 | } |
| 7084 | |
| 7085 | /// Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst. |
| 7086 | /// |
| 7087 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_srli_epi16&expand=5513) |
| 7088 | #[inline ] |
| 7089 | #[target_feature (enable = "avx512bw" )] |
| 7090 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7091 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
| 7092 | #[rustc_legacy_const_generics (1)] |
| 7093 | pub fn _mm512_srli_epi16<const IMM8: u32>(a: __m512i) -> __m512i { |
| 7094 | unsafe { |
| 7095 | static_assert_uimm_bits!(IMM8, 8); |
| 7096 | if IMM8 >= 16 { |
| 7097 | _mm512_setzero_si512() |
| 7098 | } else { |
| 7099 | transmute(src:simd_shr(lhs:a.as_u16x32(), rhs:u16x32::splat(IMM8 as u16))) |
| 7100 | } |
| 7101 | } |
| 7102 | } |
| 7103 | |
| 7104 | /// Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7105 | /// |
| 7106 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_srli_epi16&expand=5511) |
| 7107 | #[inline ] |
| 7108 | #[target_feature (enable = "avx512bw" )] |
| 7109 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7110 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
| 7111 | #[rustc_legacy_const_generics (3)] |
| 7112 | pub fn _mm512_mask_srli_epi16<const IMM8: u32>(src: __m512i, k: __mmask32, a: __m512i) -> __m512i { |
| 7113 | unsafe { |
| 7114 | static_assert_uimm_bits!(IMM8, 8); |
| 7115 | let shf: u16x32 = if IMM8 >= 16 { |
| 7116 | u16x32::ZERO |
| 7117 | } else { |
| 7118 | simd_shr(lhs:a.as_u16x32(), rhs:u16x32::splat(IMM8 as u16)) |
| 7119 | }; |
| 7120 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_u16x32())) |
| 7121 | } |
| 7122 | } |
| 7123 | |
| 7124 | /// Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7125 | /// |
| 7126 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_srli_epi16&expand=5512) |
| 7127 | #[inline ] |
| 7128 | #[target_feature (enable = "avx512bw" )] |
| 7129 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7130 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
| 7131 | #[rustc_legacy_const_generics (2)] |
| 7132 | pub fn _mm512_maskz_srli_epi16<const IMM8: i32>(k: __mmask32, a: __m512i) -> __m512i { |
| 7133 | unsafe { |
| 7134 | static_assert_uimm_bits!(IMM8, 8); |
| 7135 | //imm8 should be u32, it seems the document to verify is incorrect |
| 7136 | if IMM8 >= 16 { |
| 7137 | _mm512_setzero_si512() |
| 7138 | } else { |
| 7139 | let shf: u16x32 = simd_shr(lhs:a.as_u16x32(), rhs:u16x32::splat(IMM8 as u16)); |
| 7140 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:u16x32::ZERO)) |
| 7141 | } |
| 7142 | } |
| 7143 | } |
| 7144 | |
| 7145 | /// Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7146 | /// |
| 7147 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_srli_epi16&expand=5508) |
| 7148 | #[inline ] |
| 7149 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7150 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7151 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
| 7152 | #[rustc_legacy_const_generics (3)] |
| 7153 | pub fn _mm256_mask_srli_epi16<const IMM8: i32>(src: __m256i, k: __mmask16, a: __m256i) -> __m256i { |
| 7154 | unsafe { |
| 7155 | static_assert_uimm_bits!(IMM8, 8); |
| 7156 | let shf: __m256i = _mm256_srli_epi16::<IMM8>(a); |
| 7157 | transmute(src:simd_select_bitmask(m:k, yes:shf.as_i16x16(), no:src.as_i16x16())) |
| 7158 | } |
| 7159 | } |
| 7160 | |
| 7161 | /// Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7162 | /// |
| 7163 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_srli_epi16&expand=5509) |
| 7164 | #[inline ] |
| 7165 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7166 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7167 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
| 7168 | #[rustc_legacy_const_generics (2)] |
| 7169 | pub fn _mm256_maskz_srli_epi16<const IMM8: i32>(k: __mmask16, a: __m256i) -> __m256i { |
| 7170 | unsafe { |
| 7171 | static_assert_uimm_bits!(IMM8, 8); |
| 7172 | let shf: __m256i = _mm256_srli_epi16::<IMM8>(a); |
| 7173 | transmute(src:simd_select_bitmask(m:k, yes:shf.as_i16x16(), no:i16x16::ZERO)) |
| 7174 | } |
| 7175 | } |
| 7176 | |
| 7177 | /// Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7178 | /// |
| 7179 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_srli_epi16&expand=5505) |
| 7180 | #[inline ] |
| 7181 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7182 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7183 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
| 7184 | #[rustc_legacy_const_generics (3)] |
| 7185 | pub fn _mm_mask_srli_epi16<const IMM8: i32>(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 7186 | unsafe { |
| 7187 | static_assert_uimm_bits!(IMM8, 8); |
| 7188 | let shf: __m128i = _mm_srli_epi16::<IMM8>(a); |
| 7189 | transmute(src:simd_select_bitmask(m:k, yes:shf.as_i16x8(), no:src.as_i16x8())) |
| 7190 | } |
| 7191 | } |
| 7192 | |
| 7193 | /// Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7194 | /// |
| 7195 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_srli_epi16&expand=5506) |
| 7196 | #[inline ] |
| 7197 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7198 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7199 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
| 7200 | #[rustc_legacy_const_generics (2)] |
| 7201 | pub fn _mm_maskz_srli_epi16<const IMM8: i32>(k: __mmask8, a: __m128i) -> __m128i { |
| 7202 | unsafe { |
| 7203 | static_assert_uimm_bits!(IMM8, 8); |
| 7204 | let shf: __m128i = _mm_srli_epi16::<IMM8>(a); |
| 7205 | transmute(src:simd_select_bitmask(m:k, yes:shf.as_i16x8(), no:i16x8::ZERO)) |
| 7206 | } |
| 7207 | } |
| 7208 | |
| 7209 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst. |
| 7210 | /// |
| 7211 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_srlv_epi16&expand=5545) |
| 7212 | #[inline ] |
| 7213 | #[target_feature (enable = "avx512bw" )] |
| 7214 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7215 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
| 7216 | pub fn _mm512_srlv_epi16(a: __m512i, count: __m512i) -> __m512i { |
| 7217 | unsafe { |
| 7218 | let count: u16x32 = count.as_u16x32(); |
| 7219 | let no_overflow: u16x32 = simd_lt(x:count, y:u16x32::splat(u16::BITS as u16)); |
| 7220 | let count: u16x32 = simd_select(mask:no_overflow, if_true:count, if_false:u16x32::ZERO); |
| 7221 | simd_select(mask:no_overflow, if_true:simd_shr(a.as_u16x32(), count), if_false:u16x32::ZERO).as_m512i() |
| 7222 | } |
| 7223 | } |
| 7224 | |
| 7225 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7226 | /// |
| 7227 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_srlv_epi16&expand=5543) |
| 7228 | #[inline ] |
| 7229 | #[target_feature (enable = "avx512bw" )] |
| 7230 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7231 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
| 7232 | pub fn _mm512_mask_srlv_epi16(src: __m512i, k: __mmask32, a: __m512i, count: __m512i) -> __m512i { |
| 7233 | unsafe { |
| 7234 | let shf: i16x32 = _mm512_srlv_epi16(a, count).as_i16x32(); |
| 7235 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
| 7236 | } |
| 7237 | } |
| 7238 | |
| 7239 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7240 | /// |
| 7241 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_srlv_epi16&expand=5544) |
| 7242 | #[inline ] |
| 7243 | #[target_feature (enable = "avx512bw" )] |
| 7244 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7245 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
| 7246 | pub fn _mm512_maskz_srlv_epi16(k: __mmask32, a: __m512i, count: __m512i) -> __m512i { |
| 7247 | unsafe { |
| 7248 | let shf: i16x32 = _mm512_srlv_epi16(a, count).as_i16x32(); |
| 7249 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x32::ZERO)) |
| 7250 | } |
| 7251 | } |
| 7252 | |
| 7253 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst. |
| 7254 | /// |
| 7255 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_srlv_epi16&expand=5542) |
| 7256 | #[inline ] |
| 7257 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7258 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7259 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
| 7260 | pub fn _mm256_srlv_epi16(a: __m256i, count: __m256i) -> __m256i { |
| 7261 | unsafe { |
| 7262 | let count: u16x16 = count.as_u16x16(); |
| 7263 | let no_overflow: u16x16 = simd_lt(x:count, y:u16x16::splat(u16::BITS as u16)); |
| 7264 | let count: u16x16 = simd_select(mask:no_overflow, if_true:count, if_false:u16x16::ZERO); |
| 7265 | simd_select(mask:no_overflow, if_true:simd_shr(a.as_u16x16(), count), if_false:u16x16::ZERO).as_m256i() |
| 7266 | } |
| 7267 | } |
| 7268 | |
| 7269 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7270 | /// |
| 7271 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_srlv_epi16&expand=5540) |
| 7272 | #[inline ] |
| 7273 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7274 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7275 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
| 7276 | pub fn _mm256_mask_srlv_epi16(src: __m256i, k: __mmask16, a: __m256i, count: __m256i) -> __m256i { |
| 7277 | unsafe { |
| 7278 | let shf: i16x16 = _mm256_srlv_epi16(a, count).as_i16x16(); |
| 7279 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x16())) |
| 7280 | } |
| 7281 | } |
| 7282 | |
| 7283 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7284 | /// |
| 7285 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_srlv_epi16&expand=5541) |
| 7286 | #[inline ] |
| 7287 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7288 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7289 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
| 7290 | pub fn _mm256_maskz_srlv_epi16(k: __mmask16, a: __m256i, count: __m256i) -> __m256i { |
| 7291 | unsafe { |
| 7292 | let shf: i16x16 = _mm256_srlv_epi16(a, count).as_i16x16(); |
| 7293 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x16::ZERO)) |
| 7294 | } |
| 7295 | } |
| 7296 | |
| 7297 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst. |
| 7298 | /// |
| 7299 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_srlv_epi16&expand=5539) |
| 7300 | #[inline ] |
| 7301 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7302 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7303 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
| 7304 | pub fn _mm_srlv_epi16(a: __m128i, count: __m128i) -> __m128i { |
| 7305 | unsafe { |
| 7306 | let count: u16x8 = count.as_u16x8(); |
| 7307 | let no_overflow: u16x8 = simd_lt(x:count, y:u16x8::splat(u16::BITS as u16)); |
| 7308 | let count: u16x8 = simd_select(mask:no_overflow, if_true:count, if_false:u16x8::ZERO); |
| 7309 | simd_select(mask:no_overflow, if_true:simd_shr(a.as_u16x8(), count), if_false:u16x8::ZERO).as_m128i() |
| 7310 | } |
| 7311 | } |
| 7312 | |
| 7313 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7314 | /// |
| 7315 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_srlv_epi16&expand=5537) |
| 7316 | #[inline ] |
| 7317 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7318 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7319 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
| 7320 | pub fn _mm_mask_srlv_epi16(src: __m128i, k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
| 7321 | unsafe { |
| 7322 | let shf: i16x8 = _mm_srlv_epi16(a, count).as_i16x8(); |
| 7323 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x8())) |
| 7324 | } |
| 7325 | } |
| 7326 | |
| 7327 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in zeros, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7328 | /// |
| 7329 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_srlv_epi16&expand=5538) |
| 7330 | #[inline ] |
| 7331 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7332 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7333 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
| 7334 | pub fn _mm_maskz_srlv_epi16(k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
| 7335 | unsafe { |
| 7336 | let shf: i16x8 = _mm_srlv_epi16(a, count).as_i16x8(); |
| 7337 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x8::ZERO)) |
| 7338 | } |
| 7339 | } |
| 7340 | |
| 7341 | /// Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst. |
| 7342 | /// |
| 7343 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_sra_epi16&expand=5398) |
| 7344 | #[inline ] |
| 7345 | #[target_feature (enable = "avx512bw" )] |
| 7346 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7347 | #[cfg_attr (test, assert_instr(vpsraw))] |
| 7348 | pub fn _mm512_sra_epi16(a: __m512i, count: __m128i) -> __m512i { |
| 7349 | unsafe { transmute(src:vpsraw(a.as_i16x32(), count.as_i16x8())) } |
| 7350 | } |
| 7351 | |
| 7352 | /// Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7353 | /// |
| 7354 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_sra_epi16&expand=5396) |
| 7355 | #[inline ] |
| 7356 | #[target_feature (enable = "avx512bw" )] |
| 7357 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7358 | #[cfg_attr (test, assert_instr(vpsraw))] |
| 7359 | pub fn _mm512_mask_sra_epi16(src: __m512i, k: __mmask32, a: __m512i, count: __m128i) -> __m512i { |
| 7360 | unsafe { |
| 7361 | let shf: i16x32 = _mm512_sra_epi16(a, count).as_i16x32(); |
| 7362 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
| 7363 | } |
| 7364 | } |
| 7365 | |
| 7366 | /// Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7367 | /// |
| 7368 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_sra_epi16&expand=5397) |
| 7369 | #[inline ] |
| 7370 | #[target_feature (enable = "avx512bw" )] |
| 7371 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7372 | #[cfg_attr (test, assert_instr(vpsraw))] |
| 7373 | pub fn _mm512_maskz_sra_epi16(k: __mmask32, a: __m512i, count: __m128i) -> __m512i { |
| 7374 | unsafe { |
| 7375 | let shf: i16x32 = _mm512_sra_epi16(a, count).as_i16x32(); |
| 7376 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x32::ZERO)) |
| 7377 | } |
| 7378 | } |
| 7379 | |
| 7380 | /// Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7381 | /// |
| 7382 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_sra_epi16&expand=5393) |
| 7383 | #[inline ] |
| 7384 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7385 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7386 | #[cfg_attr (test, assert_instr(vpsraw))] |
| 7387 | pub fn _mm256_mask_sra_epi16(src: __m256i, k: __mmask16, a: __m256i, count: __m128i) -> __m256i { |
| 7388 | unsafe { |
| 7389 | let shf: i16x16 = _mm256_sra_epi16(a, count).as_i16x16(); |
| 7390 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x16())) |
| 7391 | } |
| 7392 | } |
| 7393 | |
| 7394 | /// Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7395 | /// |
| 7396 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_sra_epi16&expand=5394) |
| 7397 | #[inline ] |
| 7398 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7399 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7400 | #[cfg_attr (test, assert_instr(vpsraw))] |
| 7401 | pub fn _mm256_maskz_sra_epi16(k: __mmask16, a: __m256i, count: __m128i) -> __m256i { |
| 7402 | unsafe { |
| 7403 | let shf: i16x16 = _mm256_sra_epi16(a, count).as_i16x16(); |
| 7404 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x16::ZERO)) |
| 7405 | } |
| 7406 | } |
| 7407 | |
| 7408 | /// Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7409 | /// |
| 7410 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_sra_epi16&expand=5390) |
| 7411 | #[inline ] |
| 7412 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7413 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7414 | #[cfg_attr (test, assert_instr(vpsraw))] |
| 7415 | pub fn _mm_mask_sra_epi16(src: __m128i, k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
| 7416 | unsafe { |
| 7417 | let shf: i16x8 = _mm_sra_epi16(a, count).as_i16x8(); |
| 7418 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x8())) |
| 7419 | } |
| 7420 | } |
| 7421 | |
| 7422 | /// Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7423 | /// |
| 7424 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_sra_epi16&expand=5391) |
| 7425 | #[inline ] |
| 7426 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7427 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7428 | #[cfg_attr (test, assert_instr(vpsraw))] |
| 7429 | pub fn _mm_maskz_sra_epi16(k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
| 7430 | unsafe { |
| 7431 | let shf: i16x8 = _mm_sra_epi16(a, count).as_i16x8(); |
| 7432 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x8::ZERO)) |
| 7433 | } |
| 7434 | } |
| 7435 | |
| 7436 | /// Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst. |
| 7437 | /// |
| 7438 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_srai_epi16&expand=5427) |
| 7439 | #[inline ] |
| 7440 | #[target_feature (enable = "avx512bw" )] |
| 7441 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7442 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
| 7443 | #[rustc_legacy_const_generics (1)] |
| 7444 | pub fn _mm512_srai_epi16<const IMM8: u32>(a: __m512i) -> __m512i { |
| 7445 | unsafe { |
| 7446 | static_assert_uimm_bits!(IMM8, 8); |
| 7447 | transmute(src:simd_shr(lhs:a.as_i16x32(), rhs:i16x32::splat(IMM8.min(15) as i16))) |
| 7448 | } |
| 7449 | } |
| 7450 | |
| 7451 | /// Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7452 | /// |
| 7453 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_srai_epi16&expand=5425) |
| 7454 | #[inline ] |
| 7455 | #[target_feature (enable = "avx512bw" )] |
| 7456 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7457 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
| 7458 | #[rustc_legacy_const_generics (3)] |
| 7459 | pub fn _mm512_mask_srai_epi16<const IMM8: u32>(src: __m512i, k: __mmask32, a: __m512i) -> __m512i { |
| 7460 | unsafe { |
| 7461 | static_assert_uimm_bits!(IMM8, 8); |
| 7462 | let shf: i16x32 = simd_shr(lhs:a.as_i16x32(), rhs:i16x32::splat(IMM8.min(15) as i16)); |
| 7463 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
| 7464 | } |
| 7465 | } |
| 7466 | |
| 7467 | /// Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7468 | /// |
| 7469 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_srai_epi16&expand=5426) |
| 7470 | #[inline ] |
| 7471 | #[target_feature (enable = "avx512bw" )] |
| 7472 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7473 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
| 7474 | #[rustc_legacy_const_generics (2)] |
| 7475 | pub fn _mm512_maskz_srai_epi16<const IMM8: u32>(k: __mmask32, a: __m512i) -> __m512i { |
| 7476 | unsafe { |
| 7477 | static_assert_uimm_bits!(IMM8, 8); |
| 7478 | let shf: i16x32 = simd_shr(lhs:a.as_i16x32(), rhs:i16x32::splat(IMM8.min(15) as i16)); |
| 7479 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x32::ZERO)) |
| 7480 | } |
| 7481 | } |
| 7482 | |
| 7483 | /// Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7484 | /// |
| 7485 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_srai_epi16&expand=5422) |
| 7486 | #[inline ] |
| 7487 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7488 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7489 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
| 7490 | #[rustc_legacy_const_generics (3)] |
| 7491 | pub fn _mm256_mask_srai_epi16<const IMM8: u32>(src: __m256i, k: __mmask16, a: __m256i) -> __m256i { |
| 7492 | unsafe { |
| 7493 | static_assert_uimm_bits!(IMM8, 8); |
| 7494 | let r: i16x16 = simd_shr(lhs:a.as_i16x16(), rhs:i16x16::splat(IMM8.min(15) as i16)); |
| 7495 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i16x16())) |
| 7496 | } |
| 7497 | } |
| 7498 | |
| 7499 | /// Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7500 | /// |
| 7501 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_srai_epi16&expand=5423) |
| 7502 | #[inline ] |
| 7503 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7504 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7505 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
| 7506 | #[rustc_legacy_const_generics (2)] |
| 7507 | pub fn _mm256_maskz_srai_epi16<const IMM8: u32>(k: __mmask16, a: __m256i) -> __m256i { |
| 7508 | unsafe { |
| 7509 | static_assert_uimm_bits!(IMM8, 8); |
| 7510 | let r: i16x16 = simd_shr(lhs:a.as_i16x16(), rhs:i16x16::splat(IMM8.min(15) as i16)); |
| 7511 | transmute(src:simd_select_bitmask(m:k, yes:r, no:i16x16::ZERO)) |
| 7512 | } |
| 7513 | } |
| 7514 | |
| 7515 | /// Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7516 | /// |
| 7517 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_srai_epi16&expand=5419) |
| 7518 | #[inline ] |
| 7519 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7520 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7521 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
| 7522 | #[rustc_legacy_const_generics (3)] |
| 7523 | pub fn _mm_mask_srai_epi16<const IMM8: u32>(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 7524 | unsafe { |
| 7525 | static_assert_uimm_bits!(IMM8, 8); |
| 7526 | let r: i16x8 = simd_shr(lhs:a.as_i16x8(), rhs:i16x8::splat(IMM8.min(15) as i16)); |
| 7527 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i16x8())) |
| 7528 | } |
| 7529 | } |
| 7530 | |
| 7531 | /// Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7532 | /// |
| 7533 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_srai_epi16&expand=5420) |
| 7534 | #[inline ] |
| 7535 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7536 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7537 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
| 7538 | #[rustc_legacy_const_generics (2)] |
| 7539 | pub fn _mm_maskz_srai_epi16<const IMM8: u32>(k: __mmask8, a: __m128i) -> __m128i { |
| 7540 | unsafe { |
| 7541 | static_assert_uimm_bits!(IMM8, 8); |
| 7542 | let r: i16x8 = simd_shr(lhs:a.as_i16x8(), rhs:i16x8::splat(IMM8.min(15) as i16)); |
| 7543 | transmute(src:simd_select_bitmask(m:k, yes:r, no:i16x8::ZERO)) |
| 7544 | } |
| 7545 | } |
| 7546 | |
| 7547 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst. |
| 7548 | /// |
| 7549 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_srav_epi16&expand=5456) |
| 7550 | #[inline ] |
| 7551 | #[target_feature (enable = "avx512bw" )] |
| 7552 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7553 | #[cfg_attr (test, assert_instr(vpsravw))] |
| 7554 | pub fn _mm512_srav_epi16(a: __m512i, count: __m512i) -> __m512i { |
| 7555 | unsafe { |
| 7556 | let count: u16x32 = count.as_u16x32(); |
| 7557 | let no_overflow: u16x32 = simd_lt(x:count, y:u16x32::splat(u16::BITS as u16)); |
| 7558 | let count: i16x32 = simd_select(mask:no_overflow, if_true:transmute(count), if_false:i16x32::splat(15)); |
| 7559 | simd_shr(lhs:a.as_i16x32(), rhs:count).as_m512i() |
| 7560 | } |
| 7561 | } |
| 7562 | |
| 7563 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7564 | /// |
| 7565 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_srav_epi16&expand=5454) |
| 7566 | #[inline ] |
| 7567 | #[target_feature (enable = "avx512bw" )] |
| 7568 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7569 | #[cfg_attr (test, assert_instr(vpsravw))] |
| 7570 | pub fn _mm512_mask_srav_epi16(src: __m512i, k: __mmask32, a: __m512i, count: __m512i) -> __m512i { |
| 7571 | unsafe { |
| 7572 | let shf: i16x32 = _mm512_srav_epi16(a, count).as_i16x32(); |
| 7573 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
| 7574 | } |
| 7575 | } |
| 7576 | |
| 7577 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7578 | /// |
| 7579 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_srav_epi16&expand=5455) |
| 7580 | #[inline ] |
| 7581 | #[target_feature (enable = "avx512bw" )] |
| 7582 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7583 | #[cfg_attr (test, assert_instr(vpsravw))] |
| 7584 | pub fn _mm512_maskz_srav_epi16(k: __mmask32, a: __m512i, count: __m512i) -> __m512i { |
| 7585 | unsafe { |
| 7586 | let shf: i16x32 = _mm512_srav_epi16(a, count).as_i16x32(); |
| 7587 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x32::ZERO)) |
| 7588 | } |
| 7589 | } |
| 7590 | |
| 7591 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst. |
| 7592 | /// |
| 7593 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_srav_epi16&expand=5453) |
| 7594 | #[inline ] |
| 7595 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7596 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7597 | #[cfg_attr (test, assert_instr(vpsravw))] |
| 7598 | pub fn _mm256_srav_epi16(a: __m256i, count: __m256i) -> __m256i { |
| 7599 | unsafe { |
| 7600 | let count: u16x16 = count.as_u16x16(); |
| 7601 | let no_overflow: u16x16 = simd_lt(x:count, y:u16x16::splat(u16::BITS as u16)); |
| 7602 | let count: i16x16 = simd_select(mask:no_overflow, if_true:transmute(count), if_false:i16x16::splat(15)); |
| 7603 | simd_shr(lhs:a.as_i16x16(), rhs:count).as_m256i() |
| 7604 | } |
| 7605 | } |
| 7606 | |
| 7607 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7608 | /// |
| 7609 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_srav_epi16&expand=5451) |
| 7610 | #[inline ] |
| 7611 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7612 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7613 | #[cfg_attr (test, assert_instr(vpsravw))] |
| 7614 | pub fn _mm256_mask_srav_epi16(src: __m256i, k: __mmask16, a: __m256i, count: __m256i) -> __m256i { |
| 7615 | unsafe { |
| 7616 | let shf: i16x16 = _mm256_srav_epi16(a, count).as_i16x16(); |
| 7617 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x16())) |
| 7618 | } |
| 7619 | } |
| 7620 | |
| 7621 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7622 | /// |
| 7623 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_srav_epi16&expand=5452) |
| 7624 | #[inline ] |
| 7625 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7626 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7627 | #[cfg_attr (test, assert_instr(vpsravw))] |
| 7628 | pub fn _mm256_maskz_srav_epi16(k: __mmask16, a: __m256i, count: __m256i) -> __m256i { |
| 7629 | unsafe { |
| 7630 | let shf: i16x16 = _mm256_srav_epi16(a, count).as_i16x16(); |
| 7631 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x16::ZERO)) |
| 7632 | } |
| 7633 | } |
| 7634 | |
| 7635 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst. |
| 7636 | /// |
| 7637 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_srav_epi16&expand=5450) |
| 7638 | #[inline ] |
| 7639 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7640 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7641 | #[cfg_attr (test, assert_instr(vpsravw))] |
| 7642 | pub fn _mm_srav_epi16(a: __m128i, count: __m128i) -> __m128i { |
| 7643 | unsafe { |
| 7644 | let count: u16x8 = count.as_u16x8(); |
| 7645 | let no_overflow: u16x8 = simd_lt(x:count, y:u16x8::splat(u16::BITS as u16)); |
| 7646 | let count: i16x8 = simd_select(mask:no_overflow, if_true:transmute(count), if_false:i16x8::splat(15)); |
| 7647 | simd_shr(lhs:a.as_i16x8(), rhs:count).as_m128i() |
| 7648 | } |
| 7649 | } |
| 7650 | |
| 7651 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7652 | /// |
| 7653 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_srav_epi16&expand=5448) |
| 7654 | #[inline ] |
| 7655 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7656 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7657 | #[cfg_attr (test, assert_instr(vpsravw))] |
| 7658 | pub fn _mm_mask_srav_epi16(src: __m128i, k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
| 7659 | unsafe { |
| 7660 | let shf: i16x8 = _mm_srav_epi16(a, count).as_i16x8(); |
| 7661 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x8())) |
| 7662 | } |
| 7663 | } |
| 7664 | |
| 7665 | /// Shift packed 16-bit integers in a right by the amount specified by the corresponding element in count while shifting in sign bits, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7666 | /// |
| 7667 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_srav_epi16&expand=5449) |
| 7668 | #[inline ] |
| 7669 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7670 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7671 | #[cfg_attr (test, assert_instr(vpsravw))] |
| 7672 | pub fn _mm_maskz_srav_epi16(k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
| 7673 | unsafe { |
| 7674 | let shf: i16x8 = _mm_srav_epi16(a, count).as_i16x8(); |
| 7675 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:i16x8::ZERO)) |
| 7676 | } |
| 7677 | } |
| 7678 | |
| 7679 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst. |
| 7680 | /// |
| 7681 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_permutex2var_epi16&expand=4226) |
| 7682 | #[inline ] |
| 7683 | #[target_feature (enable = "avx512bw" )] |
| 7684 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7685 | #[cfg_attr (test, assert_instr(vperm))] //vpermi2w or vpermt2w |
| 7686 | pub fn _mm512_permutex2var_epi16(a: __m512i, idx: __m512i, b: __m512i) -> __m512i { |
| 7687 | unsafe { transmute(src:vpermi2w(a.as_i16x32(), idx.as_i16x32(), b.as_i16x32())) } |
| 7688 | } |
| 7689 | |
| 7690 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set). |
| 7691 | /// |
| 7692 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_permutex2var_epi16&expand=4223) |
| 7693 | #[inline ] |
| 7694 | #[target_feature (enable = "avx512bw" )] |
| 7695 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7696 | #[cfg_attr (test, assert_instr(vpermt2w))] |
| 7697 | pub fn _mm512_mask_permutex2var_epi16( |
| 7698 | a: __m512i, |
| 7699 | k: __mmask32, |
| 7700 | idx: __m512i, |
| 7701 | b: __m512i, |
| 7702 | ) -> __m512i { |
| 7703 | unsafe { |
| 7704 | let permute: i16x32 = _mm512_permutex2var_epi16(a, idx, b).as_i16x32(); |
| 7705 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:a.as_i16x32())) |
| 7706 | } |
| 7707 | } |
| 7708 | |
| 7709 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7710 | /// |
| 7711 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_permutex2var_epi16&expand=4225) |
| 7712 | #[inline ] |
| 7713 | #[target_feature (enable = "avx512bw" )] |
| 7714 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7715 | #[cfg_attr (test, assert_instr(vperm))] //vpermi2w or vpermt2w |
| 7716 | pub fn _mm512_maskz_permutex2var_epi16( |
| 7717 | k: __mmask32, |
| 7718 | a: __m512i, |
| 7719 | idx: __m512i, |
| 7720 | b: __m512i, |
| 7721 | ) -> __m512i { |
| 7722 | unsafe { |
| 7723 | let permute: i16x32 = _mm512_permutex2var_epi16(a, idx, b).as_i16x32(); |
| 7724 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:i16x32::ZERO)) |
| 7725 | } |
| 7726 | } |
| 7727 | |
| 7728 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set). |
| 7729 | /// |
| 7730 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask2_permutex2var_epi16&expand=4224) |
| 7731 | #[inline ] |
| 7732 | #[target_feature (enable = "avx512bw" )] |
| 7733 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7734 | #[cfg_attr (test, assert_instr(vpermi2w))] |
| 7735 | pub fn _mm512_mask2_permutex2var_epi16( |
| 7736 | a: __m512i, |
| 7737 | idx: __m512i, |
| 7738 | k: __mmask32, |
| 7739 | b: __m512i, |
| 7740 | ) -> __m512i { |
| 7741 | unsafe { |
| 7742 | let permute: i16x32 = _mm512_permutex2var_epi16(a, idx, b).as_i16x32(); |
| 7743 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:idx.as_i16x32())) |
| 7744 | } |
| 7745 | } |
| 7746 | |
| 7747 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst. |
| 7748 | /// |
| 7749 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_permutex2var_epi16&expand=4222) |
| 7750 | #[inline ] |
| 7751 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7752 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7753 | #[cfg_attr (test, assert_instr(vperm))] //vpermi2w or vpermt2w |
| 7754 | pub fn _mm256_permutex2var_epi16(a: __m256i, idx: __m256i, b: __m256i) -> __m256i { |
| 7755 | unsafe { transmute(src:vpermi2w256(a.as_i16x16(), idx.as_i16x16(), b.as_i16x16())) } |
| 7756 | } |
| 7757 | |
| 7758 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set). |
| 7759 | /// |
| 7760 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_permutex2var_epi16&expand=4219) |
| 7761 | #[inline ] |
| 7762 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7763 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7764 | #[cfg_attr (test, assert_instr(vpermt2w))] |
| 7765 | pub fn _mm256_mask_permutex2var_epi16( |
| 7766 | a: __m256i, |
| 7767 | k: __mmask16, |
| 7768 | idx: __m256i, |
| 7769 | b: __m256i, |
| 7770 | ) -> __m256i { |
| 7771 | unsafe { |
| 7772 | let permute: i16x16 = _mm256_permutex2var_epi16(a, idx, b).as_i16x16(); |
| 7773 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:a.as_i16x16())) |
| 7774 | } |
| 7775 | } |
| 7776 | |
| 7777 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7778 | /// |
| 7779 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_permutex2var_epi16&expand=4221) |
| 7780 | #[inline ] |
| 7781 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7782 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7783 | #[cfg_attr (test, assert_instr(vperm))] //vpermi2w or vpermt2w |
| 7784 | pub fn _mm256_maskz_permutex2var_epi16( |
| 7785 | k: __mmask16, |
| 7786 | a: __m256i, |
| 7787 | idx: __m256i, |
| 7788 | b: __m256i, |
| 7789 | ) -> __m256i { |
| 7790 | unsafe { |
| 7791 | let permute: i16x16 = _mm256_permutex2var_epi16(a, idx, b).as_i16x16(); |
| 7792 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:i16x16::ZERO)) |
| 7793 | } |
| 7794 | } |
| 7795 | |
| 7796 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set). |
| 7797 | /// |
| 7798 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask2_permutex2var_epi16&expand=4220) |
| 7799 | #[inline ] |
| 7800 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7801 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7802 | #[cfg_attr (test, assert_instr(vpermi2w))] |
| 7803 | pub fn _mm256_mask2_permutex2var_epi16( |
| 7804 | a: __m256i, |
| 7805 | idx: __m256i, |
| 7806 | k: __mmask16, |
| 7807 | b: __m256i, |
| 7808 | ) -> __m256i { |
| 7809 | unsafe { |
| 7810 | let permute: i16x16 = _mm256_permutex2var_epi16(a, idx, b).as_i16x16(); |
| 7811 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:idx.as_i16x16())) |
| 7812 | } |
| 7813 | } |
| 7814 | |
| 7815 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst. |
| 7816 | /// |
| 7817 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_permutex2var_epi16&expand=4218) |
| 7818 | #[inline ] |
| 7819 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7820 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7821 | #[cfg_attr (test, assert_instr(vperm))] //vpermi2w or vpermt2w |
| 7822 | pub fn _mm_permutex2var_epi16(a: __m128i, idx: __m128i, b: __m128i) -> __m128i { |
| 7823 | unsafe { transmute(src:vpermi2w128(a.as_i16x8(), idx.as_i16x8(), b.as_i16x8())) } |
| 7824 | } |
| 7825 | |
| 7826 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from a when the corresponding mask bit is not set). |
| 7827 | /// |
| 7828 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_permutex2var_epi16&expand=4215) |
| 7829 | #[inline ] |
| 7830 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7831 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7832 | #[cfg_attr (test, assert_instr(vpermt2w))] |
| 7833 | pub fn _mm_mask_permutex2var_epi16(a: __m128i, k: __mmask8, idx: __m128i, b: __m128i) -> __m128i { |
| 7834 | unsafe { |
| 7835 | let permute: i16x8 = _mm_permutex2var_epi16(a, idx, b).as_i16x8(); |
| 7836 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:a.as_i16x8())) |
| 7837 | } |
| 7838 | } |
| 7839 | |
| 7840 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7841 | /// |
| 7842 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_permutex2var_epi16&expand=4217) |
| 7843 | #[inline ] |
| 7844 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7845 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7846 | #[cfg_attr (test, assert_instr(vperm))] //vpermi2w or vpermt2w |
| 7847 | pub fn _mm_maskz_permutex2var_epi16(k: __mmask8, a: __m128i, idx: __m128i, b: __m128i) -> __m128i { |
| 7848 | unsafe { |
| 7849 | let permute: i16x8 = _mm_permutex2var_epi16(a, idx, b).as_i16x8(); |
| 7850 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:i16x8::ZERO)) |
| 7851 | } |
| 7852 | } |
| 7853 | |
| 7854 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst using writemask k (elements are copied from idx when the corresponding mask bit is not set). |
| 7855 | /// |
| 7856 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask2_permutex2var_epi16&expand=4216) |
| 7857 | #[inline ] |
| 7858 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7859 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7860 | #[cfg_attr (test, assert_instr(vpermi2w))] |
| 7861 | pub fn _mm_mask2_permutex2var_epi16(a: __m128i, idx: __m128i, k: __mmask8, b: __m128i) -> __m128i { |
| 7862 | unsafe { |
| 7863 | let permute: i16x8 = _mm_permutex2var_epi16(a, idx, b).as_i16x8(); |
| 7864 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:idx.as_i16x8())) |
| 7865 | } |
| 7866 | } |
| 7867 | |
| 7868 | /// Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst. |
| 7869 | /// |
| 7870 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_permutexvar_epi16&expand=4295) |
| 7871 | #[inline ] |
| 7872 | #[target_feature (enable = "avx512bw" )] |
| 7873 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7874 | #[cfg_attr (test, assert_instr(vpermw))] |
| 7875 | pub fn _mm512_permutexvar_epi16(idx: __m512i, a: __m512i) -> __m512i { |
| 7876 | unsafe { transmute(src:vpermw(a.as_i16x32(), idx.as_i16x32())) } |
| 7877 | } |
| 7878 | |
| 7879 | /// Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7880 | /// |
| 7881 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_permutexvar_epi16&expand=4293) |
| 7882 | #[inline ] |
| 7883 | #[target_feature (enable = "avx512bw" )] |
| 7884 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7885 | #[cfg_attr (test, assert_instr(vpermw))] |
| 7886 | pub fn _mm512_mask_permutexvar_epi16( |
| 7887 | src: __m512i, |
| 7888 | k: __mmask32, |
| 7889 | idx: __m512i, |
| 7890 | a: __m512i, |
| 7891 | ) -> __m512i { |
| 7892 | unsafe { |
| 7893 | let permute: i16x32 = _mm512_permutexvar_epi16(idx, a).as_i16x32(); |
| 7894 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:src.as_i16x32())) |
| 7895 | } |
| 7896 | } |
| 7897 | |
| 7898 | /// Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7899 | /// |
| 7900 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_permutexvar_epi16&expand=4294) |
| 7901 | #[inline ] |
| 7902 | #[target_feature (enable = "avx512bw" )] |
| 7903 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7904 | #[cfg_attr (test, assert_instr(vpermw))] |
| 7905 | pub fn _mm512_maskz_permutexvar_epi16(k: __mmask32, idx: __m512i, a: __m512i) -> __m512i { |
| 7906 | unsafe { |
| 7907 | let permute: i16x32 = _mm512_permutexvar_epi16(idx, a).as_i16x32(); |
| 7908 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:i16x32::ZERO)) |
| 7909 | } |
| 7910 | } |
| 7911 | |
| 7912 | /// Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst. |
| 7913 | /// |
| 7914 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_permutexvar_epi16&expand=4292) |
| 7915 | #[inline ] |
| 7916 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7917 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7918 | #[cfg_attr (test, assert_instr(vpermw))] |
| 7919 | pub fn _mm256_permutexvar_epi16(idx: __m256i, a: __m256i) -> __m256i { |
| 7920 | unsafe { transmute(src:vpermw256(a.as_i16x16(), idx.as_i16x16())) } |
| 7921 | } |
| 7922 | |
| 7923 | /// Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7924 | /// |
| 7925 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_permutexvar_epi16&expand=4290) |
| 7926 | #[inline ] |
| 7927 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7928 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7929 | #[cfg_attr (test, assert_instr(vpermw))] |
| 7930 | pub fn _mm256_mask_permutexvar_epi16( |
| 7931 | src: __m256i, |
| 7932 | k: __mmask16, |
| 7933 | idx: __m256i, |
| 7934 | a: __m256i, |
| 7935 | ) -> __m256i { |
| 7936 | unsafe { |
| 7937 | let permute: i16x16 = _mm256_permutexvar_epi16(idx, a).as_i16x16(); |
| 7938 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:src.as_i16x16())) |
| 7939 | } |
| 7940 | } |
| 7941 | |
| 7942 | /// Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7943 | /// |
| 7944 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_permutexvar_epi16&expand=4291) |
| 7945 | #[inline ] |
| 7946 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7947 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7948 | #[cfg_attr (test, assert_instr(vpermw))] |
| 7949 | pub fn _mm256_maskz_permutexvar_epi16(k: __mmask16, idx: __m256i, a: __m256i) -> __m256i { |
| 7950 | unsafe { |
| 7951 | let permute: i16x16 = _mm256_permutexvar_epi16(idx, a).as_i16x16(); |
| 7952 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:i16x16::ZERO)) |
| 7953 | } |
| 7954 | } |
| 7955 | |
| 7956 | /// Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst. |
| 7957 | /// |
| 7958 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_permutexvar_epi16&expand=4289) |
| 7959 | #[inline ] |
| 7960 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7961 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7962 | #[cfg_attr (test, assert_instr(vpermw))] |
| 7963 | pub fn _mm_permutexvar_epi16(idx: __m128i, a: __m128i) -> __m128i { |
| 7964 | unsafe { transmute(src:vpermw128(a.as_i16x8(), idx.as_i16x8())) } |
| 7965 | } |
| 7966 | |
| 7967 | /// Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 7968 | /// |
| 7969 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_permutexvar_epi16&expand=4287) |
| 7970 | #[inline ] |
| 7971 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7972 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7973 | #[cfg_attr (test, assert_instr(vpermw))] |
| 7974 | pub fn _mm_mask_permutexvar_epi16(src: __m128i, k: __mmask8, idx: __m128i, a: __m128i) -> __m128i { |
| 7975 | unsafe { |
| 7976 | let permute: i16x8 = _mm_permutexvar_epi16(idx, a).as_i16x8(); |
| 7977 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:src.as_i16x8())) |
| 7978 | } |
| 7979 | } |
| 7980 | |
| 7981 | /// Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 7982 | /// |
| 7983 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_permutexvar_epi16&expand=4288) |
| 7984 | #[inline ] |
| 7985 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 7986 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 7987 | #[cfg_attr (test, assert_instr(vpermw))] |
| 7988 | pub fn _mm_maskz_permutexvar_epi16(k: __mmask8, idx: __m128i, a: __m128i) -> __m128i { |
| 7989 | unsafe { |
| 7990 | let permute: i16x8 = _mm_permutexvar_epi16(idx, a).as_i16x8(); |
| 7991 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:i16x8::ZERO)) |
| 7992 | } |
| 7993 | } |
| 7994 | |
| 7995 | /// Blend packed 16-bit integers from a and b using control mask k, and store the results in dst. |
| 7996 | /// |
| 7997 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_blend_epi16&expand=430) |
| 7998 | #[inline ] |
| 7999 | #[target_feature (enable = "avx512bw" )] |
| 8000 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8001 | #[cfg_attr (test, assert_instr(vmovdqu16))] //should be vpblendmw |
| 8002 | pub fn _mm512_mask_blend_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 8003 | unsafe { transmute(src:simd_select_bitmask(m:k, yes:b.as_i16x32(), no:a.as_i16x32())) } |
| 8004 | } |
| 8005 | |
| 8006 | /// Blend packed 16-bit integers from a and b using control mask k, and store the results in dst. |
| 8007 | /// |
| 8008 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_blend_epi16&expand=429) |
| 8009 | #[inline ] |
| 8010 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8011 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8012 | #[cfg_attr (test, assert_instr(vmovdqu16))] //should be vpblendmw |
| 8013 | pub fn _mm256_mask_blend_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 8014 | unsafe { transmute(src:simd_select_bitmask(m:k, yes:b.as_i16x16(), no:a.as_i16x16())) } |
| 8015 | } |
| 8016 | |
| 8017 | /// Blend packed 16-bit integers from a and b using control mask k, and store the results in dst. |
| 8018 | /// |
| 8019 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_blend_epi16&expand=427) |
| 8020 | #[inline ] |
| 8021 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8022 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8023 | #[cfg_attr (test, assert_instr(vmovdqu16))] //should be vpblendmw |
| 8024 | pub fn _mm_mask_blend_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 8025 | unsafe { transmute(src:simd_select_bitmask(m:k, yes:b.as_i16x8(), no:a.as_i16x8())) } |
| 8026 | } |
| 8027 | |
| 8028 | /// Blend packed 8-bit integers from a and b using control mask k, and store the results in dst. |
| 8029 | /// |
| 8030 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_blend_epi8&expand=441) |
| 8031 | #[inline ] |
| 8032 | #[target_feature (enable = "avx512bw" )] |
| 8033 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8034 | #[cfg_attr (test, assert_instr(vmovdqu8))] //should be vpblendmb |
| 8035 | pub fn _mm512_mask_blend_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 8036 | unsafe { transmute(src:simd_select_bitmask(m:k, yes:b.as_i8x64(), no:a.as_i8x64())) } |
| 8037 | } |
| 8038 | |
| 8039 | /// Blend packed 8-bit integers from a and b using control mask k, and store the results in dst. |
| 8040 | /// |
| 8041 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_blend_epi8&expand=440) |
| 8042 | #[inline ] |
| 8043 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8044 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8045 | #[cfg_attr (test, assert_instr(vmovdqu8))] //should be vpblendmb |
| 8046 | pub fn _mm256_mask_blend_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 8047 | unsafe { transmute(src:simd_select_bitmask(m:k, yes:b.as_i8x32(), no:a.as_i8x32())) } |
| 8048 | } |
| 8049 | |
| 8050 | /// Blend packed 8-bit integers from a and b using control mask k, and store the results in dst. |
| 8051 | /// |
| 8052 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_blend_epi8&expand=439) |
| 8053 | #[inline ] |
| 8054 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8055 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8056 | #[cfg_attr (test, assert_instr(vmovdqu8))] //should be vpblendmb |
| 8057 | pub fn _mm_mask_blend_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 8058 | unsafe { transmute(src:simd_select_bitmask(m:k, yes:b.as_i8x16(), no:a.as_i8x16())) } |
| 8059 | } |
| 8060 | |
| 8061 | /// Broadcast the low packed 16-bit integer from a to all elements of dst. |
| 8062 | /// |
| 8063 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_broadcastw_epi16&expand=587) |
| 8064 | #[inline ] |
| 8065 | #[target_feature (enable = "avx512bw" )] |
| 8066 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8067 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8068 | pub fn _mm512_broadcastw_epi16(a: __m128i) -> __m512i { |
| 8069 | unsafe { |
| 8070 | let a: i16x32 = _mm512_castsi128_si512(a).as_i16x32(); |
| 8071 | let ret: i16x32 = simd_shuffle!( |
| 8072 | a, |
| 8073 | a, |
| 8074 | [ |
| 8075 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 8076 | 0, 0, 0, 0, |
| 8077 | ], |
| 8078 | ); |
| 8079 | transmute(src:ret) |
| 8080 | } |
| 8081 | } |
| 8082 | |
| 8083 | /// Broadcast the low packed 16-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8084 | /// |
| 8085 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_broadcastw_epi16&expand=588) |
| 8086 | #[inline ] |
| 8087 | #[target_feature (enable = "avx512bw" )] |
| 8088 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8089 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8090 | pub fn _mm512_mask_broadcastw_epi16(src: __m512i, k: __mmask32, a: __m128i) -> __m512i { |
| 8091 | unsafe { |
| 8092 | let broadcast: i16x32 = _mm512_broadcastw_epi16(a).as_i16x32(); |
| 8093 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:src.as_i16x32())) |
| 8094 | } |
| 8095 | } |
| 8096 | |
| 8097 | /// Broadcast the low packed 16-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8098 | /// |
| 8099 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_broadcastw_epi16&expand=589) |
| 8100 | #[inline ] |
| 8101 | #[target_feature (enable = "avx512bw" )] |
| 8102 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8103 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8104 | pub fn _mm512_maskz_broadcastw_epi16(k: __mmask32, a: __m128i) -> __m512i { |
| 8105 | unsafe { |
| 8106 | let broadcast: i16x32 = _mm512_broadcastw_epi16(a).as_i16x32(); |
| 8107 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:i16x32::ZERO)) |
| 8108 | } |
| 8109 | } |
| 8110 | |
| 8111 | /// Broadcast the low packed 16-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8112 | /// |
| 8113 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_broadcastw_epi16&expand=585) |
| 8114 | #[inline ] |
| 8115 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8116 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8117 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8118 | pub fn _mm256_mask_broadcastw_epi16(src: __m256i, k: __mmask16, a: __m128i) -> __m256i { |
| 8119 | unsafe { |
| 8120 | let broadcast: i16x16 = _mm256_broadcastw_epi16(a).as_i16x16(); |
| 8121 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:src.as_i16x16())) |
| 8122 | } |
| 8123 | } |
| 8124 | |
| 8125 | /// Broadcast the low packed 16-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8126 | /// |
| 8127 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_broadcastw_epi16&expand=586) |
| 8128 | #[inline ] |
| 8129 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8130 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8131 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8132 | pub fn _mm256_maskz_broadcastw_epi16(k: __mmask16, a: __m128i) -> __m256i { |
| 8133 | unsafe { |
| 8134 | let broadcast: i16x16 = _mm256_broadcastw_epi16(a).as_i16x16(); |
| 8135 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:i16x16::ZERO)) |
| 8136 | } |
| 8137 | } |
| 8138 | |
| 8139 | /// Broadcast the low packed 16-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8140 | /// |
| 8141 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_broadcastw_epi16&expand=582) |
| 8142 | #[inline ] |
| 8143 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8144 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8145 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8146 | pub fn _mm_mask_broadcastw_epi16(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 8147 | unsafe { |
| 8148 | let broadcast: i16x8 = _mm_broadcastw_epi16(a).as_i16x8(); |
| 8149 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:src.as_i16x8())) |
| 8150 | } |
| 8151 | } |
| 8152 | |
| 8153 | /// Broadcast the low packed 16-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8154 | /// |
| 8155 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_broadcastw_epi16&expand=583) |
| 8156 | #[inline ] |
| 8157 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8158 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8159 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8160 | pub fn _mm_maskz_broadcastw_epi16(k: __mmask8, a: __m128i) -> __m128i { |
| 8161 | unsafe { |
| 8162 | let broadcast: i16x8 = _mm_broadcastw_epi16(a).as_i16x8(); |
| 8163 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:i16x8::ZERO)) |
| 8164 | } |
| 8165 | } |
| 8166 | |
| 8167 | /// Broadcast the low packed 8-bit integer from a to all elements of dst. |
| 8168 | /// |
| 8169 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_broadcastb_epi8&expand=536) |
| 8170 | #[inline ] |
| 8171 | #[target_feature (enable = "avx512bw" )] |
| 8172 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8173 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
| 8174 | pub fn _mm512_broadcastb_epi8(a: __m128i) -> __m512i { |
| 8175 | unsafe { |
| 8176 | let a: i8x64 = _mm512_castsi128_si512(a).as_i8x64(); |
| 8177 | let ret: i8x64 = simd_shuffle!( |
| 8178 | a, |
| 8179 | a, |
| 8180 | [ |
| 8181 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 8182 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 8183 | 0, 0, 0, 0, 0, 0, 0, 0, |
| 8184 | ], |
| 8185 | ); |
| 8186 | transmute(src:ret) |
| 8187 | } |
| 8188 | } |
| 8189 | |
| 8190 | /// Broadcast the low packed 8-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8191 | /// |
| 8192 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_broadcastb_epi8&expand=537) |
| 8193 | #[inline ] |
| 8194 | #[target_feature (enable = "avx512bw" )] |
| 8195 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8196 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
| 8197 | pub fn _mm512_mask_broadcastb_epi8(src: __m512i, k: __mmask64, a: __m128i) -> __m512i { |
| 8198 | unsafe { |
| 8199 | let broadcast: i8x64 = _mm512_broadcastb_epi8(a).as_i8x64(); |
| 8200 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:src.as_i8x64())) |
| 8201 | } |
| 8202 | } |
| 8203 | |
| 8204 | /// Broadcast the low packed 8-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8205 | /// |
| 8206 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_broadcastb_epi8&expand=538) |
| 8207 | #[inline ] |
| 8208 | #[target_feature (enable = "avx512bw" )] |
| 8209 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8210 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
| 8211 | pub fn _mm512_maskz_broadcastb_epi8(k: __mmask64, a: __m128i) -> __m512i { |
| 8212 | unsafe { |
| 8213 | let broadcast: i8x64 = _mm512_broadcastb_epi8(a).as_i8x64(); |
| 8214 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:i8x64::ZERO)) |
| 8215 | } |
| 8216 | } |
| 8217 | |
| 8218 | /// Broadcast the low packed 8-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8219 | /// |
| 8220 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_broadcastb_epi8&expand=534) |
| 8221 | #[inline ] |
| 8222 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8223 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8224 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
| 8225 | pub fn _mm256_mask_broadcastb_epi8(src: __m256i, k: __mmask32, a: __m128i) -> __m256i { |
| 8226 | unsafe { |
| 8227 | let broadcast: i8x32 = _mm256_broadcastb_epi8(a).as_i8x32(); |
| 8228 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:src.as_i8x32())) |
| 8229 | } |
| 8230 | } |
| 8231 | |
| 8232 | /// Broadcast the low packed 8-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8233 | /// |
| 8234 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_broadcastb_epi8&expand=535) |
| 8235 | #[inline ] |
| 8236 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8237 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8238 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
| 8239 | pub fn _mm256_maskz_broadcastb_epi8(k: __mmask32, a: __m128i) -> __m256i { |
| 8240 | unsafe { |
| 8241 | let broadcast: i8x32 = _mm256_broadcastb_epi8(a).as_i8x32(); |
| 8242 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:i8x32::ZERO)) |
| 8243 | } |
| 8244 | } |
| 8245 | |
| 8246 | /// Broadcast the low packed 8-bit integer from a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8247 | /// |
| 8248 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_broadcastb_epi8&expand=531) |
| 8249 | #[inline ] |
| 8250 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8251 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8252 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
| 8253 | pub fn _mm_mask_broadcastb_epi8(src: __m128i, k: __mmask16, a: __m128i) -> __m128i { |
| 8254 | unsafe { |
| 8255 | let broadcast: i8x16 = _mm_broadcastb_epi8(a).as_i8x16(); |
| 8256 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:src.as_i8x16())) |
| 8257 | } |
| 8258 | } |
| 8259 | |
| 8260 | /// Broadcast the low packed 8-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8261 | /// |
| 8262 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_broadcastb_epi8&expand=532) |
| 8263 | #[inline ] |
| 8264 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8265 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8266 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
| 8267 | pub fn _mm_maskz_broadcastb_epi8(k: __mmask16, a: __m128i) -> __m128i { |
| 8268 | unsafe { |
| 8269 | let broadcast: i8x16 = _mm_broadcastb_epi8(a).as_i8x16(); |
| 8270 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:i8x16::ZERO)) |
| 8271 | } |
| 8272 | } |
| 8273 | |
| 8274 | /// Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst. |
| 8275 | /// |
| 8276 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_unpackhi_epi16&expand=6012) |
| 8277 | #[inline ] |
| 8278 | #[target_feature (enable = "avx512bw" )] |
| 8279 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8280 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
| 8281 | pub fn _mm512_unpackhi_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 8282 | unsafe { |
| 8283 | let a: i16x32 = a.as_i16x32(); |
| 8284 | let b: i16x32 = b.as_i16x32(); |
| 8285 | #[rustfmt::skip] |
| 8286 | let r: i16x32 = simd_shuffle!( |
| 8287 | a, |
| 8288 | b, |
| 8289 | [ |
| 8290 | 4, 32 + 4, 5, 32 + 5, |
| 8291 | 6, 32 + 6, 7, 32 + 7, |
| 8292 | 12, 32 + 12, 13, 32 + 13, |
| 8293 | 14, 32 + 14, 15, 32 + 15, |
| 8294 | 20, 32 + 20, 21, 32 + 21, |
| 8295 | 22, 32 + 22, 23, 32 + 23, |
| 8296 | 28, 32 + 28, 29, 32 + 29, |
| 8297 | 30, 32 + 30, 31, 32 + 31, |
| 8298 | ], |
| 8299 | ); |
| 8300 | transmute(src:r) |
| 8301 | } |
| 8302 | } |
| 8303 | |
| 8304 | /// Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8305 | /// |
| 8306 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_unpackhi_epi16&expand=6010) |
| 8307 | #[inline ] |
| 8308 | #[target_feature (enable = "avx512bw" )] |
| 8309 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8310 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
| 8311 | pub fn _mm512_mask_unpackhi_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 8312 | unsafe { |
| 8313 | let unpackhi: i16x32 = _mm512_unpackhi_epi16(a, b).as_i16x32(); |
| 8314 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:src.as_i16x32())) |
| 8315 | } |
| 8316 | } |
| 8317 | |
| 8318 | /// Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8319 | /// |
| 8320 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_unpackhi_epi16&expand=6011) |
| 8321 | #[inline ] |
| 8322 | #[target_feature (enable = "avx512bw" )] |
| 8323 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8324 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
| 8325 | pub fn _mm512_maskz_unpackhi_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 8326 | unsafe { |
| 8327 | let unpackhi: i16x32 = _mm512_unpackhi_epi16(a, b).as_i16x32(); |
| 8328 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:i16x32::ZERO)) |
| 8329 | } |
| 8330 | } |
| 8331 | |
| 8332 | /// Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8333 | /// |
| 8334 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_unpackhi_epi16&expand=6007) |
| 8335 | #[inline ] |
| 8336 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8337 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8338 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
| 8339 | pub fn _mm256_mask_unpackhi_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 8340 | unsafe { |
| 8341 | let unpackhi: i16x16 = _mm256_unpackhi_epi16(a, b).as_i16x16(); |
| 8342 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:src.as_i16x16())) |
| 8343 | } |
| 8344 | } |
| 8345 | |
| 8346 | /// Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8347 | /// |
| 8348 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_unpackhi_epi16&expand=6008) |
| 8349 | #[inline ] |
| 8350 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8351 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8352 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
| 8353 | pub fn _mm256_maskz_unpackhi_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 8354 | unsafe { |
| 8355 | let unpackhi: i16x16 = _mm256_unpackhi_epi16(a, b).as_i16x16(); |
| 8356 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:i16x16::ZERO)) |
| 8357 | } |
| 8358 | } |
| 8359 | |
| 8360 | /// Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8361 | /// |
| 8362 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_unpackhi_epi16&expand=6004) |
| 8363 | #[inline ] |
| 8364 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8365 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8366 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
| 8367 | pub fn _mm_mask_unpackhi_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 8368 | unsafe { |
| 8369 | let unpackhi: i16x8 = _mm_unpackhi_epi16(a, b).as_i16x8(); |
| 8370 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:src.as_i16x8())) |
| 8371 | } |
| 8372 | } |
| 8373 | |
| 8374 | /// Unpack and interleave 16-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8375 | /// |
| 8376 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_unpackhi_epi16&expand=6005) |
| 8377 | #[inline ] |
| 8378 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8379 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8380 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
| 8381 | pub fn _mm_maskz_unpackhi_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 8382 | unsafe { |
| 8383 | let unpackhi: i16x8 = _mm_unpackhi_epi16(a, b).as_i16x8(); |
| 8384 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:i16x8::ZERO)) |
| 8385 | } |
| 8386 | } |
| 8387 | |
| 8388 | /// Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst. |
| 8389 | /// |
| 8390 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_unpackhi_epi8&expand=6039) |
| 8391 | #[inline ] |
| 8392 | #[target_feature (enable = "avx512bw" )] |
| 8393 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8394 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
| 8395 | pub fn _mm512_unpackhi_epi8(a: __m512i, b: __m512i) -> __m512i { |
| 8396 | unsafe { |
| 8397 | let a = a.as_i8x64(); |
| 8398 | let b = b.as_i8x64(); |
| 8399 | #[rustfmt::skip] |
| 8400 | let r: i8x64 = simd_shuffle!( |
| 8401 | a, |
| 8402 | b, |
| 8403 | [ |
| 8404 | 8, 64 + 8, 9, 64 + 9, |
| 8405 | 10, 64 + 10, 11, 64 + 11, |
| 8406 | 12, 64 + 12, 13, 64 + 13, |
| 8407 | 14, 64 + 14, 15, 64 + 15, |
| 8408 | 24, 64 + 24, 25, 64 + 25, |
| 8409 | 26, 64 + 26, 27, 64 + 27, |
| 8410 | 28, 64 + 28, 29, 64 + 29, |
| 8411 | 30, 64 + 30, 31, 64 + 31, |
| 8412 | 40, 64 + 40, 41, 64 + 41, |
| 8413 | 42, 64 + 42, 43, 64 + 43, |
| 8414 | 44, 64 + 44, 45, 64 + 45, |
| 8415 | 46, 64 + 46, 47, 64 + 47, |
| 8416 | 56, 64 + 56, 57, 64 + 57, |
| 8417 | 58, 64 + 58, 59, 64 + 59, |
| 8418 | 60, 64 + 60, 61, 64 + 61, |
| 8419 | 62, 64 + 62, 63, 64 + 63, |
| 8420 | ], |
| 8421 | ); |
| 8422 | transmute(r) |
| 8423 | } |
| 8424 | } |
| 8425 | |
| 8426 | /// Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8427 | /// |
| 8428 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_unpackhi_epi8&expand=6037) |
| 8429 | #[inline ] |
| 8430 | #[target_feature (enable = "avx512bw" )] |
| 8431 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8432 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
| 8433 | pub fn _mm512_mask_unpackhi_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 8434 | unsafe { |
| 8435 | let unpackhi: i8x64 = _mm512_unpackhi_epi8(a, b).as_i8x64(); |
| 8436 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:src.as_i8x64())) |
| 8437 | } |
| 8438 | } |
| 8439 | |
| 8440 | /// Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8441 | /// |
| 8442 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_unpackhi_epi8&expand=6038) |
| 8443 | #[inline ] |
| 8444 | #[target_feature (enable = "avx512bw" )] |
| 8445 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8446 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
| 8447 | pub fn _mm512_maskz_unpackhi_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 8448 | unsafe { |
| 8449 | let unpackhi: i8x64 = _mm512_unpackhi_epi8(a, b).as_i8x64(); |
| 8450 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:i8x64::ZERO)) |
| 8451 | } |
| 8452 | } |
| 8453 | |
| 8454 | /// Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8455 | /// |
| 8456 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_unpackhi_epi8&expand=6034) |
| 8457 | #[inline ] |
| 8458 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8459 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8460 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
| 8461 | pub fn _mm256_mask_unpackhi_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 8462 | unsafe { |
| 8463 | let unpackhi: i8x32 = _mm256_unpackhi_epi8(a, b).as_i8x32(); |
| 8464 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:src.as_i8x32())) |
| 8465 | } |
| 8466 | } |
| 8467 | |
| 8468 | /// Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8469 | /// |
| 8470 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_unpackhi_epi8&expand=6035) |
| 8471 | #[inline ] |
| 8472 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8473 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8474 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
| 8475 | pub fn _mm256_maskz_unpackhi_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 8476 | unsafe { |
| 8477 | let unpackhi: i8x32 = _mm256_unpackhi_epi8(a, b).as_i8x32(); |
| 8478 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:i8x32::ZERO)) |
| 8479 | } |
| 8480 | } |
| 8481 | |
| 8482 | /// Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8483 | /// |
| 8484 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_unpackhi_epi8&expand=6031) |
| 8485 | #[inline ] |
| 8486 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8487 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8488 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
| 8489 | pub fn _mm_mask_unpackhi_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 8490 | unsafe { |
| 8491 | let unpackhi: i8x16 = _mm_unpackhi_epi8(a, b).as_i8x16(); |
| 8492 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:src.as_i8x16())) |
| 8493 | } |
| 8494 | } |
| 8495 | |
| 8496 | /// Unpack and interleave 8-bit integers from the high half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8497 | /// |
| 8498 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_unpackhi_epi8&expand=6032) |
| 8499 | #[inline ] |
| 8500 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8501 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8502 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
| 8503 | pub fn _mm_maskz_unpackhi_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 8504 | unsafe { |
| 8505 | let unpackhi: i8x16 = _mm_unpackhi_epi8(a, b).as_i8x16(); |
| 8506 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:i8x16::ZERO)) |
| 8507 | } |
| 8508 | } |
| 8509 | |
| 8510 | /// Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst. |
| 8511 | /// |
| 8512 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_unpacklo_epi16&expand=6069) |
| 8513 | #[inline ] |
| 8514 | #[target_feature (enable = "avx512bw" )] |
| 8515 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8516 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
| 8517 | pub fn _mm512_unpacklo_epi16(a: __m512i, b: __m512i) -> __m512i { |
| 8518 | unsafe { |
| 8519 | let a: i16x32 = a.as_i16x32(); |
| 8520 | let b: i16x32 = b.as_i16x32(); |
| 8521 | #[rustfmt::skip] |
| 8522 | let r: i16x32 = simd_shuffle!( |
| 8523 | a, |
| 8524 | b, |
| 8525 | [ |
| 8526 | 0, 32+0, 1, 32+1, |
| 8527 | 2, 32+2, 3, 32+3, |
| 8528 | 8, 32+8, 9, 32+9, |
| 8529 | 10, 32+10, 11, 32+11, |
| 8530 | 16, 32+16, 17, 32+17, |
| 8531 | 18, 32+18, 19, 32+19, |
| 8532 | 24, 32+24, 25, 32+25, |
| 8533 | 26, 32+26, 27, 32+27 |
| 8534 | ], |
| 8535 | ); |
| 8536 | transmute(src:r) |
| 8537 | } |
| 8538 | } |
| 8539 | |
| 8540 | /// Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8541 | /// |
| 8542 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_unpacklo_epi16&expand=6067) |
| 8543 | #[inline ] |
| 8544 | #[target_feature (enable = "avx512bw" )] |
| 8545 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8546 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
| 8547 | pub fn _mm512_mask_unpacklo_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 8548 | unsafe { |
| 8549 | let unpacklo: i16x32 = _mm512_unpacklo_epi16(a, b).as_i16x32(); |
| 8550 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:src.as_i16x32())) |
| 8551 | } |
| 8552 | } |
| 8553 | |
| 8554 | /// Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8555 | /// |
| 8556 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_unpacklo_epi16&expand=6068) |
| 8557 | #[inline ] |
| 8558 | #[target_feature (enable = "avx512bw" )] |
| 8559 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8560 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
| 8561 | pub fn _mm512_maskz_unpacklo_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 8562 | unsafe { |
| 8563 | let unpacklo: i16x32 = _mm512_unpacklo_epi16(a, b).as_i16x32(); |
| 8564 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:i16x32::ZERO)) |
| 8565 | } |
| 8566 | } |
| 8567 | |
| 8568 | /// Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8569 | /// |
| 8570 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_unpacklo_epi16&expand=6064) |
| 8571 | #[inline ] |
| 8572 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8573 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8574 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
| 8575 | pub fn _mm256_mask_unpacklo_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 8576 | unsafe { |
| 8577 | let unpacklo: i16x16 = _mm256_unpacklo_epi16(a, b).as_i16x16(); |
| 8578 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:src.as_i16x16())) |
| 8579 | } |
| 8580 | } |
| 8581 | |
| 8582 | /// Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8583 | /// |
| 8584 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_unpacklo_epi16&expand=6065) |
| 8585 | #[inline ] |
| 8586 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8587 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8588 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
| 8589 | pub fn _mm256_maskz_unpacklo_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 8590 | unsafe { |
| 8591 | let unpacklo: i16x16 = _mm256_unpacklo_epi16(a, b).as_i16x16(); |
| 8592 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:i16x16::ZERO)) |
| 8593 | } |
| 8594 | } |
| 8595 | |
| 8596 | /// Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8597 | /// |
| 8598 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_unpacklo_epi16&expand=6061) |
| 8599 | #[inline ] |
| 8600 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8601 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8602 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
| 8603 | pub fn _mm_mask_unpacklo_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 8604 | unsafe { |
| 8605 | let unpacklo: i16x8 = _mm_unpacklo_epi16(a, b).as_i16x8(); |
| 8606 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:src.as_i16x8())) |
| 8607 | } |
| 8608 | } |
| 8609 | |
| 8610 | /// Unpack and interleave 16-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8611 | /// |
| 8612 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_unpacklo_epi16&expand=6062) |
| 8613 | #[inline ] |
| 8614 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8615 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8616 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
| 8617 | pub fn _mm_maskz_unpacklo_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 8618 | unsafe { |
| 8619 | let unpacklo: i16x8 = _mm_unpacklo_epi16(a, b).as_i16x8(); |
| 8620 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:i16x8::ZERO)) |
| 8621 | } |
| 8622 | } |
| 8623 | |
| 8624 | /// Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst. |
| 8625 | /// |
| 8626 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_unpacklo_epi8&expand=6096) |
| 8627 | #[inline ] |
| 8628 | #[target_feature (enable = "avx512bw" )] |
| 8629 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8630 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
| 8631 | pub fn _mm512_unpacklo_epi8(a: __m512i, b: __m512i) -> __m512i { |
| 8632 | unsafe { |
| 8633 | let a = a.as_i8x64(); |
| 8634 | let b = b.as_i8x64(); |
| 8635 | #[rustfmt::skip] |
| 8636 | let r: i8x64 = simd_shuffle!( |
| 8637 | a, |
| 8638 | b, |
| 8639 | [ |
| 8640 | 0, 64+0, 1, 64+1, |
| 8641 | 2, 64+2, 3, 64+3, |
| 8642 | 4, 64+4, 5, 64+5, |
| 8643 | 6, 64+6, 7, 64+7, |
| 8644 | 16, 64+16, 17, 64+17, |
| 8645 | 18, 64+18, 19, 64+19, |
| 8646 | 20, 64+20, 21, 64+21, |
| 8647 | 22, 64+22, 23, 64+23, |
| 8648 | 32, 64+32, 33, 64+33, |
| 8649 | 34, 64+34, 35, 64+35, |
| 8650 | 36, 64+36, 37, 64+37, |
| 8651 | 38, 64+38, 39, 64+39, |
| 8652 | 48, 64+48, 49, 64+49, |
| 8653 | 50, 64+50, 51, 64+51, |
| 8654 | 52, 64+52, 53, 64+53, |
| 8655 | 54, 64+54, 55, 64+55, |
| 8656 | ], |
| 8657 | ); |
| 8658 | transmute(r) |
| 8659 | } |
| 8660 | } |
| 8661 | |
| 8662 | /// Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8663 | /// |
| 8664 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_unpacklo_epi8&expand=6094) |
| 8665 | #[inline ] |
| 8666 | #[target_feature (enable = "avx512bw" )] |
| 8667 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8668 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
| 8669 | pub fn _mm512_mask_unpacklo_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 8670 | unsafe { |
| 8671 | let unpacklo: i8x64 = _mm512_unpacklo_epi8(a, b).as_i8x64(); |
| 8672 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:src.as_i8x64())) |
| 8673 | } |
| 8674 | } |
| 8675 | |
| 8676 | /// Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8677 | /// |
| 8678 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_unpacklo_epi8&expand=6095) |
| 8679 | #[inline ] |
| 8680 | #[target_feature (enable = "avx512bw" )] |
| 8681 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8682 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
| 8683 | pub fn _mm512_maskz_unpacklo_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 8684 | unsafe { |
| 8685 | let unpacklo: i8x64 = _mm512_unpacklo_epi8(a, b).as_i8x64(); |
| 8686 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:i8x64::ZERO)) |
| 8687 | } |
| 8688 | } |
| 8689 | |
| 8690 | /// Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8691 | /// |
| 8692 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_unpacklo_epi8&expand=6091) |
| 8693 | #[inline ] |
| 8694 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8695 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8696 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
| 8697 | pub fn _mm256_mask_unpacklo_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 8698 | unsafe { |
| 8699 | let unpacklo: i8x32 = _mm256_unpacklo_epi8(a, b).as_i8x32(); |
| 8700 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:src.as_i8x32())) |
| 8701 | } |
| 8702 | } |
| 8703 | |
| 8704 | /// Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8705 | /// |
| 8706 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_unpacklo_epi8&expand=6092) |
| 8707 | #[inline ] |
| 8708 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8709 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8710 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
| 8711 | pub fn _mm256_maskz_unpacklo_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 8712 | unsafe { |
| 8713 | let unpacklo: i8x32 = _mm256_unpacklo_epi8(a, b).as_i8x32(); |
| 8714 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:i8x32::ZERO)) |
| 8715 | } |
| 8716 | } |
| 8717 | |
| 8718 | /// Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8719 | /// |
| 8720 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_unpacklo_epi8&expand=6088) |
| 8721 | #[inline ] |
| 8722 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8723 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8724 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
| 8725 | pub fn _mm_mask_unpacklo_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 8726 | unsafe { |
| 8727 | let unpacklo: i8x16 = _mm_unpacklo_epi8(a, b).as_i8x16(); |
| 8728 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:src.as_i8x16())) |
| 8729 | } |
| 8730 | } |
| 8731 | |
| 8732 | /// Unpack and interleave 8-bit integers from the low half of each 128-bit lane in a and b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8733 | /// |
| 8734 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_unpacklo_epi8&expand=6089) |
| 8735 | #[inline ] |
| 8736 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8737 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8738 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
| 8739 | pub fn _mm_maskz_unpacklo_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 8740 | unsafe { |
| 8741 | let unpacklo: i8x16 = _mm_unpacklo_epi8(a, b).as_i8x16(); |
| 8742 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:i8x16::ZERO)) |
| 8743 | } |
| 8744 | } |
| 8745 | |
| 8746 | /// Move packed 16-bit integers from a into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8747 | /// |
| 8748 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_mov_epi16&expand=3795) |
| 8749 | #[inline ] |
| 8750 | #[target_feature (enable = "avx512bw" )] |
| 8751 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8752 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 8753 | pub fn _mm512_mask_mov_epi16(src: __m512i, k: __mmask32, a: __m512i) -> __m512i { |
| 8754 | unsafe { |
| 8755 | let mov: i16x32 = a.as_i16x32(); |
| 8756 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:src.as_i16x32())) |
| 8757 | } |
| 8758 | } |
| 8759 | |
| 8760 | /// Move packed 16-bit integers from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8761 | /// |
| 8762 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_mov_epi16&expand=3796) |
| 8763 | #[inline ] |
| 8764 | #[target_feature (enable = "avx512bw" )] |
| 8765 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8766 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 8767 | pub fn _mm512_maskz_mov_epi16(k: __mmask32, a: __m512i) -> __m512i { |
| 8768 | unsafe { |
| 8769 | let mov: i16x32 = a.as_i16x32(); |
| 8770 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:i16x32::ZERO)) |
| 8771 | } |
| 8772 | } |
| 8773 | |
| 8774 | /// Move packed 16-bit integers from a into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8775 | /// |
| 8776 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_mov_epi16&expand=3793) |
| 8777 | #[inline ] |
| 8778 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8779 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8780 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 8781 | pub fn _mm256_mask_mov_epi16(src: __m256i, k: __mmask16, a: __m256i) -> __m256i { |
| 8782 | unsafe { |
| 8783 | let mov: i16x16 = a.as_i16x16(); |
| 8784 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:src.as_i16x16())) |
| 8785 | } |
| 8786 | } |
| 8787 | |
| 8788 | /// Move packed 16-bit integers from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8789 | /// |
| 8790 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_mov_epi16&expand=3794) |
| 8791 | #[inline ] |
| 8792 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8793 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8794 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 8795 | pub fn _mm256_maskz_mov_epi16(k: __mmask16, a: __m256i) -> __m256i { |
| 8796 | unsafe { |
| 8797 | let mov: i16x16 = a.as_i16x16(); |
| 8798 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:i16x16::ZERO)) |
| 8799 | } |
| 8800 | } |
| 8801 | |
| 8802 | /// Move packed 16-bit integers from a into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8803 | /// |
| 8804 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_mov_epi16&expand=3791) |
| 8805 | #[inline ] |
| 8806 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8807 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8808 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 8809 | pub fn _mm_mask_mov_epi16(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 8810 | unsafe { |
| 8811 | let mov: i16x8 = a.as_i16x8(); |
| 8812 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:src.as_i16x8())) |
| 8813 | } |
| 8814 | } |
| 8815 | |
| 8816 | /// Move packed 16-bit integers from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8817 | /// |
| 8818 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_mov_epi16&expand=3792) |
| 8819 | #[inline ] |
| 8820 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8821 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8822 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
| 8823 | pub fn _mm_maskz_mov_epi16(k: __mmask8, a: __m128i) -> __m128i { |
| 8824 | unsafe { |
| 8825 | let mov: i16x8 = a.as_i16x8(); |
| 8826 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:i16x8::ZERO)) |
| 8827 | } |
| 8828 | } |
| 8829 | |
| 8830 | /// Move packed 8-bit integers from a into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8831 | /// |
| 8832 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_mov_epi8&expand=3813) |
| 8833 | #[inline ] |
| 8834 | #[target_feature (enable = "avx512bw" )] |
| 8835 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8836 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 8837 | pub fn _mm512_mask_mov_epi8(src: __m512i, k: __mmask64, a: __m512i) -> __m512i { |
| 8838 | unsafe { |
| 8839 | let mov: i8x64 = a.as_i8x64(); |
| 8840 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:src.as_i8x64())) |
| 8841 | } |
| 8842 | } |
| 8843 | |
| 8844 | /// Move packed 8-bit integers from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8845 | /// |
| 8846 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_mov_epi8&expand=3814) |
| 8847 | #[inline ] |
| 8848 | #[target_feature (enable = "avx512bw" )] |
| 8849 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8850 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 8851 | pub fn _mm512_maskz_mov_epi8(k: __mmask64, a: __m512i) -> __m512i { |
| 8852 | unsafe { |
| 8853 | let mov: i8x64 = a.as_i8x64(); |
| 8854 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:i8x64::ZERO)) |
| 8855 | } |
| 8856 | } |
| 8857 | |
| 8858 | /// Move packed 8-bit integers from a into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8859 | /// |
| 8860 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_mov_epi8&expand=3811) |
| 8861 | #[inline ] |
| 8862 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8863 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8864 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 8865 | pub fn _mm256_mask_mov_epi8(src: __m256i, k: __mmask32, a: __m256i) -> __m256i { |
| 8866 | unsafe { |
| 8867 | let mov: i8x32 = a.as_i8x32(); |
| 8868 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:src.as_i8x32())) |
| 8869 | } |
| 8870 | } |
| 8871 | |
| 8872 | /// Move packed 8-bit integers from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8873 | /// |
| 8874 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_mov_epi8&expand=3812) |
| 8875 | #[inline ] |
| 8876 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8877 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8878 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 8879 | pub fn _mm256_maskz_mov_epi8(k: __mmask32, a: __m256i) -> __m256i { |
| 8880 | unsafe { |
| 8881 | let mov: i8x32 = a.as_i8x32(); |
| 8882 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:i8x32::ZERO)) |
| 8883 | } |
| 8884 | } |
| 8885 | |
| 8886 | /// Move packed 8-bit integers from a into dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8887 | /// |
| 8888 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_mov_epi8&expand=3809) |
| 8889 | #[inline ] |
| 8890 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8891 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8892 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 8893 | pub fn _mm_mask_mov_epi8(src: __m128i, k: __mmask16, a: __m128i) -> __m128i { |
| 8894 | unsafe { |
| 8895 | let mov: i8x16 = a.as_i8x16(); |
| 8896 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:src.as_i8x16())) |
| 8897 | } |
| 8898 | } |
| 8899 | |
| 8900 | /// Move packed 8-bit integers from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8901 | /// |
| 8902 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_mov_epi8&expand=3810) |
| 8903 | #[inline ] |
| 8904 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8905 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8906 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
| 8907 | pub fn _mm_maskz_mov_epi8(k: __mmask16, a: __m128i) -> __m128i { |
| 8908 | unsafe { |
| 8909 | let mov: i8x16 = a.as_i8x16(); |
| 8910 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:i8x16::ZERO)) |
| 8911 | } |
| 8912 | } |
| 8913 | |
| 8914 | /// Broadcast 16-bit integer a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8915 | /// |
| 8916 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_set1_epi16&expand=4942) |
| 8917 | #[inline ] |
| 8918 | #[target_feature (enable = "avx512bw" )] |
| 8919 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8920 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8921 | pub fn _mm512_mask_set1_epi16(src: __m512i, k: __mmask32, a: i16) -> __m512i { |
| 8922 | unsafe { |
| 8923 | let r: i16x32 = _mm512_set1_epi16(a).as_i16x32(); |
| 8924 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i16x32())) |
| 8925 | } |
| 8926 | } |
| 8927 | |
| 8928 | /// Broadcast the low packed 16-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8929 | /// |
| 8930 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_set1_epi16&expand=4943) |
| 8931 | #[inline ] |
| 8932 | #[target_feature (enable = "avx512bw" )] |
| 8933 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8934 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8935 | pub fn _mm512_maskz_set1_epi16(k: __mmask32, a: i16) -> __m512i { |
| 8936 | unsafe { |
| 8937 | let r: i16x32 = _mm512_set1_epi16(a).as_i16x32(); |
| 8938 | transmute(src:simd_select_bitmask(m:k, yes:r, no:i16x32::ZERO)) |
| 8939 | } |
| 8940 | } |
| 8941 | |
| 8942 | /// Broadcast 16-bit integer a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8943 | /// |
| 8944 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_set1_epi16&expand=4939) |
| 8945 | #[inline ] |
| 8946 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8947 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8948 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8949 | pub fn _mm256_mask_set1_epi16(src: __m256i, k: __mmask16, a: i16) -> __m256i { |
| 8950 | unsafe { |
| 8951 | let r: i16x16 = _mm256_set1_epi16(a).as_i16x16(); |
| 8952 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i16x16())) |
| 8953 | } |
| 8954 | } |
| 8955 | |
| 8956 | /// Broadcast the low packed 16-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8957 | /// |
| 8958 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_set1_epi16&expand=4940) |
| 8959 | #[inline ] |
| 8960 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8961 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8962 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8963 | pub fn _mm256_maskz_set1_epi16(k: __mmask16, a: i16) -> __m256i { |
| 8964 | unsafe { |
| 8965 | let r: i16x16 = _mm256_set1_epi16(a).as_i16x16(); |
| 8966 | transmute(src:simd_select_bitmask(m:k, yes:r, no:i16x16::ZERO)) |
| 8967 | } |
| 8968 | } |
| 8969 | |
| 8970 | /// Broadcast 16-bit integer a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8971 | /// |
| 8972 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_set1_epi16&expand=4936) |
| 8973 | #[inline ] |
| 8974 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8975 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8976 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8977 | pub fn _mm_mask_set1_epi16(src: __m128i, k: __mmask8, a: i16) -> __m128i { |
| 8978 | unsafe { |
| 8979 | let r: i16x8 = _mm_set1_epi16(a).as_i16x8(); |
| 8980 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i16x8())) |
| 8981 | } |
| 8982 | } |
| 8983 | |
| 8984 | /// Broadcast the low packed 16-bit integer from a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 8985 | /// |
| 8986 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_set1_epi16&expand=4937) |
| 8987 | #[inline ] |
| 8988 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 8989 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 8990 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
| 8991 | pub fn _mm_maskz_set1_epi16(k: __mmask8, a: i16) -> __m128i { |
| 8992 | unsafe { |
| 8993 | let r: i16x8 = _mm_set1_epi16(a).as_i16x8(); |
| 8994 | transmute(src:simd_select_bitmask(m:k, yes:r, no:i16x8::ZERO)) |
| 8995 | } |
| 8996 | } |
| 8997 | |
| 8998 | /// Broadcast 8-bit integer a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 8999 | /// |
| 9000 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_set1_epi8&expand=4970) |
| 9001 | #[inline ] |
| 9002 | #[target_feature (enable = "avx512bw" )] |
| 9003 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9004 | #[cfg_attr (test, assert_instr(vpbroadcast))] |
| 9005 | pub fn _mm512_mask_set1_epi8(src: __m512i, k: __mmask64, a: i8) -> __m512i { |
| 9006 | unsafe { |
| 9007 | let r: i8x64 = _mm512_set1_epi8(a).as_i8x64(); |
| 9008 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i8x64())) |
| 9009 | } |
| 9010 | } |
| 9011 | |
| 9012 | /// Broadcast 8-bit integer a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 9013 | /// |
| 9014 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_set1_epi8&expand=4971) |
| 9015 | #[inline ] |
| 9016 | #[target_feature (enable = "avx512bw" )] |
| 9017 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9018 | #[cfg_attr (test, assert_instr(vpbroadcast))] |
| 9019 | pub fn _mm512_maskz_set1_epi8(k: __mmask64, a: i8) -> __m512i { |
| 9020 | unsafe { |
| 9021 | let r: i8x64 = _mm512_set1_epi8(a).as_i8x64(); |
| 9022 | transmute(src:simd_select_bitmask(m:k, yes:r, no:i8x64::ZERO)) |
| 9023 | } |
| 9024 | } |
| 9025 | |
| 9026 | /// Broadcast 8-bit integer a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9027 | /// |
| 9028 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_set1_epi8&expand=4967) |
| 9029 | #[inline ] |
| 9030 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9031 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9032 | #[cfg_attr (test, assert_instr(vpbroadcast))] |
| 9033 | pub fn _mm256_mask_set1_epi8(src: __m256i, k: __mmask32, a: i8) -> __m256i { |
| 9034 | unsafe { |
| 9035 | let r: i8x32 = _mm256_set1_epi8(a).as_i8x32(); |
| 9036 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i8x32())) |
| 9037 | } |
| 9038 | } |
| 9039 | |
| 9040 | /// Broadcast 8-bit integer a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 9041 | /// |
| 9042 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_set1_epi8&expand=4968) |
| 9043 | #[inline ] |
| 9044 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9045 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9046 | #[cfg_attr (test, assert_instr(vpbroadcast))] |
| 9047 | pub fn _mm256_maskz_set1_epi8(k: __mmask32, a: i8) -> __m256i { |
| 9048 | unsafe { |
| 9049 | let r: i8x32 = _mm256_set1_epi8(a).as_i8x32(); |
| 9050 | transmute(src:simd_select_bitmask(m:k, yes:r, no:i8x32::ZERO)) |
| 9051 | } |
| 9052 | } |
| 9053 | |
| 9054 | /// Broadcast 8-bit integer a to all elements of dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9055 | /// |
| 9056 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_set1_epi8&expand=4964) |
| 9057 | #[inline ] |
| 9058 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9059 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9060 | #[cfg_attr (test, assert_instr(vpbroadcast))] |
| 9061 | pub fn _mm_mask_set1_epi8(src: __m128i, k: __mmask16, a: i8) -> __m128i { |
| 9062 | unsafe { |
| 9063 | let r: i8x16 = _mm_set1_epi8(a).as_i8x16(); |
| 9064 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i8x16())) |
| 9065 | } |
| 9066 | } |
| 9067 | |
| 9068 | /// Broadcast 8-bit integer a to all elements of dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 9069 | /// |
| 9070 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_set1_epi8&expand=4965) |
| 9071 | #[inline ] |
| 9072 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9073 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9074 | #[cfg_attr (test, assert_instr(vpbroadcast))] |
| 9075 | pub fn _mm_maskz_set1_epi8(k: __mmask16, a: i8) -> __m128i { |
| 9076 | unsafe { |
| 9077 | let r: i8x16 = _mm_set1_epi8(a).as_i8x16(); |
| 9078 | transmute(src:simd_select_bitmask(m:k, yes:r, no:i8x16::ZERO)) |
| 9079 | } |
| 9080 | } |
| 9081 | |
| 9082 | /// Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst. |
| 9083 | /// |
| 9084 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_shufflelo_epi16&expand=5221) |
| 9085 | #[inline ] |
| 9086 | #[target_feature (enable = "avx512bw" )] |
| 9087 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9088 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 0))] |
| 9089 | #[rustc_legacy_const_generics (1)] |
| 9090 | pub fn _mm512_shufflelo_epi16<const IMM8: i32>(a: __m512i) -> __m512i { |
| 9091 | unsafe { |
| 9092 | static_assert_uimm_bits!(IMM8, 8); |
| 9093 | let a = a.as_i16x32(); |
| 9094 | let r: i16x32 = simd_shuffle!( |
| 9095 | a, |
| 9096 | a, |
| 9097 | [ |
| 9098 | IMM8 as u32 & 0b11, |
| 9099 | (IMM8 as u32 >> 2) & 0b11, |
| 9100 | (IMM8 as u32 >> 4) & 0b11, |
| 9101 | (IMM8 as u32 >> 6) & 0b11, |
| 9102 | 4, |
| 9103 | 5, |
| 9104 | 6, |
| 9105 | 7, |
| 9106 | (IMM8 as u32 & 0b11) + 8, |
| 9107 | ((IMM8 as u32 >> 2) & 0b11) + 8, |
| 9108 | ((IMM8 as u32 >> 4) & 0b11) + 8, |
| 9109 | ((IMM8 as u32 >> 6) & 0b11) + 8, |
| 9110 | 12, |
| 9111 | 13, |
| 9112 | 14, |
| 9113 | 15, |
| 9114 | (IMM8 as u32 & 0b11) + 16, |
| 9115 | ((IMM8 as u32 >> 2) & 0b11) + 16, |
| 9116 | ((IMM8 as u32 >> 4) & 0b11) + 16, |
| 9117 | ((IMM8 as u32 >> 6) & 0b11) + 16, |
| 9118 | 20, |
| 9119 | 21, |
| 9120 | 22, |
| 9121 | 23, |
| 9122 | (IMM8 as u32 & 0b11) + 24, |
| 9123 | ((IMM8 as u32 >> 2) & 0b11) + 24, |
| 9124 | ((IMM8 as u32 >> 4) & 0b11) + 24, |
| 9125 | ((IMM8 as u32 >> 6) & 0b11) + 24, |
| 9126 | 28, |
| 9127 | 29, |
| 9128 | 30, |
| 9129 | 31, |
| 9130 | ], |
| 9131 | ); |
| 9132 | transmute(r) |
| 9133 | } |
| 9134 | } |
| 9135 | |
| 9136 | /// Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9137 | /// |
| 9138 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_shufflelo_epi16&expand=5219) |
| 9139 | #[inline ] |
| 9140 | #[target_feature (enable = "avx512bw" )] |
| 9141 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9142 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 0))] |
| 9143 | #[rustc_legacy_const_generics (3)] |
| 9144 | pub fn _mm512_mask_shufflelo_epi16<const IMM8: i32>( |
| 9145 | src: __m512i, |
| 9146 | k: __mmask32, |
| 9147 | a: __m512i, |
| 9148 | ) -> __m512i { |
| 9149 | unsafe { |
| 9150 | static_assert_uimm_bits!(IMM8, 8); |
| 9151 | let r: __m512i = _mm512_shufflelo_epi16::<IMM8>(a); |
| 9152 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i16x32(), no:src.as_i16x32())) |
| 9153 | } |
| 9154 | } |
| 9155 | |
| 9156 | /// Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst, using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 9157 | /// |
| 9158 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_shufflelo_epi16&expand=5220) |
| 9159 | #[inline ] |
| 9160 | #[target_feature (enable = "avx512bw" )] |
| 9161 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9162 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 0))] |
| 9163 | #[rustc_legacy_const_generics (2)] |
| 9164 | pub fn _mm512_maskz_shufflelo_epi16<const IMM8: i32>(k: __mmask32, a: __m512i) -> __m512i { |
| 9165 | unsafe { |
| 9166 | static_assert_uimm_bits!(IMM8, 8); |
| 9167 | let r: __m512i = _mm512_shufflelo_epi16::<IMM8>(a); |
| 9168 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i16x32(), no:i16x32::ZERO)) |
| 9169 | } |
| 9170 | } |
| 9171 | |
| 9172 | /// Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9173 | /// |
| 9174 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_shufflelo_epi16&expand=5216) |
| 9175 | #[inline ] |
| 9176 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9177 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9178 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 5))] |
| 9179 | #[rustc_legacy_const_generics (3)] |
| 9180 | pub fn _mm256_mask_shufflelo_epi16<const IMM8: i32>( |
| 9181 | src: __m256i, |
| 9182 | k: __mmask16, |
| 9183 | a: __m256i, |
| 9184 | ) -> __m256i { |
| 9185 | unsafe { |
| 9186 | static_assert_uimm_bits!(IMM8, 8); |
| 9187 | let shuffle: __m256i = _mm256_shufflelo_epi16::<IMM8>(a); |
| 9188 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x16(), no:src.as_i16x16())) |
| 9189 | } |
| 9190 | } |
| 9191 | |
| 9192 | /// Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9193 | /// |
| 9194 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_shufflelo_epi16&expand=5217) |
| 9195 | #[inline ] |
| 9196 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9197 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9198 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 5))] |
| 9199 | #[rustc_legacy_const_generics (2)] |
| 9200 | pub fn _mm256_maskz_shufflelo_epi16<const IMM8: i32>(k: __mmask16, a: __m256i) -> __m256i { |
| 9201 | unsafe { |
| 9202 | static_assert_uimm_bits!(IMM8, 8); |
| 9203 | let shuffle: __m256i = _mm256_shufflelo_epi16::<IMM8>(a); |
| 9204 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x16(), no:i16x16::ZERO)) |
| 9205 | } |
| 9206 | } |
| 9207 | |
| 9208 | /// Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9209 | /// |
| 9210 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_shufflelo_epi16&expand=5213) |
| 9211 | #[inline ] |
| 9212 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9213 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9214 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 5))] |
| 9215 | #[rustc_legacy_const_generics (3)] |
| 9216 | pub fn _mm_mask_shufflelo_epi16<const IMM8: i32>(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 9217 | unsafe { |
| 9218 | static_assert_uimm_bits!(IMM8, 8); |
| 9219 | let shuffle: __m128i = _mm_shufflelo_epi16::<IMM8>(a); |
| 9220 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x8(), no:src.as_i16x8())) |
| 9221 | } |
| 9222 | } |
| 9223 | |
| 9224 | /// Shuffle 16-bit integers in the low 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the low 64 bits of 128-bit lanes of dst, with the high 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9225 | /// |
| 9226 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_shufflelo_epi16&expand=5214) |
| 9227 | #[inline ] |
| 9228 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9229 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9230 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 5))] |
| 9231 | #[rustc_legacy_const_generics (2)] |
| 9232 | pub fn _mm_maskz_shufflelo_epi16<const IMM8: i32>(k: __mmask8, a: __m128i) -> __m128i { |
| 9233 | unsafe { |
| 9234 | static_assert_uimm_bits!(IMM8, 8); |
| 9235 | let shuffle: __m128i = _mm_shufflelo_epi16::<IMM8>(a); |
| 9236 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x8(), no:i16x8::ZERO)) |
| 9237 | } |
| 9238 | } |
| 9239 | |
| 9240 | /// Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst. |
| 9241 | /// |
| 9242 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_shufflehi_epi16&expand=5212) |
| 9243 | #[inline ] |
| 9244 | #[target_feature (enable = "avx512bw" )] |
| 9245 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9246 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 0))] |
| 9247 | #[rustc_legacy_const_generics (1)] |
| 9248 | pub fn _mm512_shufflehi_epi16<const IMM8: i32>(a: __m512i) -> __m512i { |
| 9249 | unsafe { |
| 9250 | static_assert_uimm_bits!(IMM8, 8); |
| 9251 | let a = a.as_i16x32(); |
| 9252 | let r: i16x32 = simd_shuffle!( |
| 9253 | a, |
| 9254 | a, |
| 9255 | [ |
| 9256 | 0, |
| 9257 | 1, |
| 9258 | 2, |
| 9259 | 3, |
| 9260 | (IMM8 as u32 & 0b11) + 4, |
| 9261 | ((IMM8 as u32 >> 2) & 0b11) + 4, |
| 9262 | ((IMM8 as u32 >> 4) & 0b11) + 4, |
| 9263 | ((IMM8 as u32 >> 6) & 0b11) + 4, |
| 9264 | 8, |
| 9265 | 9, |
| 9266 | 10, |
| 9267 | 11, |
| 9268 | (IMM8 as u32 & 0b11) + 12, |
| 9269 | ((IMM8 as u32 >> 2) & 0b11) + 12, |
| 9270 | ((IMM8 as u32 >> 4) & 0b11) + 12, |
| 9271 | ((IMM8 as u32 >> 6) & 0b11) + 12, |
| 9272 | 16, |
| 9273 | 17, |
| 9274 | 18, |
| 9275 | 19, |
| 9276 | (IMM8 as u32 & 0b11) + 20, |
| 9277 | ((IMM8 as u32 >> 2) & 0b11) + 20, |
| 9278 | ((IMM8 as u32 >> 4) & 0b11) + 20, |
| 9279 | ((IMM8 as u32 >> 6) & 0b11) + 20, |
| 9280 | 24, |
| 9281 | 25, |
| 9282 | 26, |
| 9283 | 27, |
| 9284 | (IMM8 as u32 & 0b11) + 28, |
| 9285 | ((IMM8 as u32 >> 2) & 0b11) + 28, |
| 9286 | ((IMM8 as u32 >> 4) & 0b11) + 28, |
| 9287 | ((IMM8 as u32 >> 6) & 0b11) + 28, |
| 9288 | ], |
| 9289 | ); |
| 9290 | transmute(r) |
| 9291 | } |
| 9292 | } |
| 9293 | |
| 9294 | /// Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9295 | /// |
| 9296 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_shufflehi_epi16&expand=5210) |
| 9297 | #[inline ] |
| 9298 | #[target_feature (enable = "avx512bw" )] |
| 9299 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9300 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 0))] |
| 9301 | #[rustc_legacy_const_generics (3)] |
| 9302 | pub fn _mm512_mask_shufflehi_epi16<const IMM8: i32>( |
| 9303 | src: __m512i, |
| 9304 | k: __mmask32, |
| 9305 | a: __m512i, |
| 9306 | ) -> __m512i { |
| 9307 | unsafe { |
| 9308 | static_assert_uimm_bits!(IMM8, 8); |
| 9309 | let r: __m512i = _mm512_shufflehi_epi16::<IMM8>(a); |
| 9310 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i16x32(), no:src.as_i16x32())) |
| 9311 | } |
| 9312 | } |
| 9313 | |
| 9314 | /// Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst, using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 9315 | /// |
| 9316 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_shufflehi_epi16&expand=5211) |
| 9317 | #[inline ] |
| 9318 | #[target_feature (enable = "avx512bw" )] |
| 9319 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9320 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 0))] |
| 9321 | #[rustc_legacy_const_generics (2)] |
| 9322 | pub fn _mm512_maskz_shufflehi_epi16<const IMM8: i32>(k: __mmask32, a: __m512i) -> __m512i { |
| 9323 | unsafe { |
| 9324 | static_assert_uimm_bits!(IMM8, 8); |
| 9325 | let r: __m512i = _mm512_shufflehi_epi16::<IMM8>(a); |
| 9326 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i16x32(), no:i16x32::ZERO)) |
| 9327 | } |
| 9328 | } |
| 9329 | |
| 9330 | /// Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9331 | /// |
| 9332 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_shufflehi_epi16&expand=5207) |
| 9333 | #[inline ] |
| 9334 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9335 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9336 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 5))] |
| 9337 | #[rustc_legacy_const_generics (3)] |
| 9338 | pub fn _mm256_mask_shufflehi_epi16<const IMM8: i32>( |
| 9339 | src: __m256i, |
| 9340 | k: __mmask16, |
| 9341 | a: __m256i, |
| 9342 | ) -> __m256i { |
| 9343 | unsafe { |
| 9344 | static_assert_uimm_bits!(IMM8, 8); |
| 9345 | let shuffle: __m256i = _mm256_shufflehi_epi16::<IMM8>(a); |
| 9346 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x16(), no:src.as_i16x16())) |
| 9347 | } |
| 9348 | } |
| 9349 | |
| 9350 | /// Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst, using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 9351 | /// |
| 9352 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_shufflehi_epi16&expand=5208) |
| 9353 | #[inline ] |
| 9354 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9355 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9356 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 5))] |
| 9357 | #[rustc_legacy_const_generics (2)] |
| 9358 | pub fn _mm256_maskz_shufflehi_epi16<const IMM8: i32>(k: __mmask16, a: __m256i) -> __m256i { |
| 9359 | unsafe { |
| 9360 | static_assert_uimm_bits!(IMM8, 8); |
| 9361 | let shuffle: __m256i = _mm256_shufflehi_epi16::<IMM8>(a); |
| 9362 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x16(), no:i16x16::ZERO)) |
| 9363 | } |
| 9364 | } |
| 9365 | |
| 9366 | /// Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst, using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9367 | /// |
| 9368 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_shufflehi_epi16&expand=5204) |
| 9369 | #[inline ] |
| 9370 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9371 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9372 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 5))] |
| 9373 | #[rustc_legacy_const_generics (3)] |
| 9374 | pub fn _mm_mask_shufflehi_epi16<const IMM8: i32>(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 9375 | unsafe { |
| 9376 | static_assert_uimm_bits!(IMM8, 8); |
| 9377 | let shuffle: __m128i = _mm_shufflehi_epi16::<IMM8>(a); |
| 9378 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x8(), no:src.as_i16x8())) |
| 9379 | } |
| 9380 | } |
| 9381 | |
| 9382 | /// Shuffle 16-bit integers in the high 64 bits of 128-bit lanes of a using the control in imm8. Store the results in the high 64 bits of 128-bit lanes of dst, with the low 64 bits of 128-bit lanes being copied from a to dst, using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 9383 | /// |
| 9384 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_shufflehi_epi16&expand=5205) |
| 9385 | #[inline ] |
| 9386 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9387 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9388 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 5))] |
| 9389 | #[rustc_legacy_const_generics (2)] |
| 9390 | pub fn _mm_maskz_shufflehi_epi16<const IMM8: i32>(k: __mmask8, a: __m128i) -> __m128i { |
| 9391 | unsafe { |
| 9392 | static_assert_uimm_bits!(IMM8, 8); |
| 9393 | let shuffle: __m128i = _mm_shufflehi_epi16::<IMM8>(a); |
| 9394 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x8(), no:i16x8::ZERO)) |
| 9395 | } |
| 9396 | } |
| 9397 | |
| 9398 | /// Shuffle packed 8-bit integers in a according to shuffle control mask in the corresponding 8-bit element of b, and store the results in dst. |
| 9399 | /// |
| 9400 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_shuffle_epi8&expand=5159) |
| 9401 | #[inline ] |
| 9402 | #[target_feature (enable = "avx512bw" )] |
| 9403 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9404 | #[cfg_attr (test, assert_instr(vpshufb))] |
| 9405 | pub fn _mm512_shuffle_epi8(a: __m512i, b: __m512i) -> __m512i { |
| 9406 | unsafe { transmute(src:vpshufb(a.as_i8x64(), b.as_i8x64())) } |
| 9407 | } |
| 9408 | |
| 9409 | /// Shuffle 8-bit integers in a within 128-bit lanes using the control in the corresponding 8-bit element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9410 | /// |
| 9411 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_shuffle_epi8&expand=5157) |
| 9412 | #[inline ] |
| 9413 | #[target_feature (enable = "avx512bw" )] |
| 9414 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9415 | #[cfg_attr (test, assert_instr(vpshufb))] |
| 9416 | pub fn _mm512_mask_shuffle_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 9417 | unsafe { |
| 9418 | let shuffle: i8x64 = _mm512_shuffle_epi8(a, b).as_i8x64(); |
| 9419 | transmute(src:simd_select_bitmask(m:k, yes:shuffle, no:src.as_i8x64())) |
| 9420 | } |
| 9421 | } |
| 9422 | |
| 9423 | /// Shuffle packed 8-bit integers in a according to shuffle control mask in the corresponding 8-bit element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 9424 | /// |
| 9425 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_shuffle_epi8&expand=5158) |
| 9426 | #[inline ] |
| 9427 | #[target_feature (enable = "avx512bw" )] |
| 9428 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9429 | #[cfg_attr (test, assert_instr(vpshufb))] |
| 9430 | pub fn _mm512_maskz_shuffle_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 9431 | unsafe { |
| 9432 | let shuffle: i8x64 = _mm512_shuffle_epi8(a, b).as_i8x64(); |
| 9433 | transmute(src:simd_select_bitmask(m:k, yes:shuffle, no:i8x64::ZERO)) |
| 9434 | } |
| 9435 | } |
| 9436 | |
| 9437 | /// Shuffle 8-bit integers in a within 128-bit lanes using the control in the corresponding 8-bit element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9438 | /// |
| 9439 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_shuffle_epi8&expand=5154) |
| 9440 | #[inline ] |
| 9441 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9442 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9443 | #[cfg_attr (test, assert_instr(vpshufb))] |
| 9444 | pub fn _mm256_mask_shuffle_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 9445 | unsafe { |
| 9446 | let shuffle: i8x32 = _mm256_shuffle_epi8(a, b).as_i8x32(); |
| 9447 | transmute(src:simd_select_bitmask(m:k, yes:shuffle, no:src.as_i8x32())) |
| 9448 | } |
| 9449 | } |
| 9450 | |
| 9451 | /// Shuffle packed 8-bit integers in a according to shuffle control mask in the corresponding 8-bit element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 9452 | /// |
| 9453 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_shuffle_epi8&expand=5155) |
| 9454 | #[inline ] |
| 9455 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9456 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9457 | #[cfg_attr (test, assert_instr(vpshufb))] |
| 9458 | pub fn _mm256_maskz_shuffle_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 9459 | unsafe { |
| 9460 | let shuffle: i8x32 = _mm256_shuffle_epi8(a, b).as_i8x32(); |
| 9461 | transmute(src:simd_select_bitmask(m:k, yes:shuffle, no:i8x32::ZERO)) |
| 9462 | } |
| 9463 | } |
| 9464 | |
| 9465 | /// Shuffle 8-bit integers in a within 128-bit lanes using the control in the corresponding 8-bit element of b, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 9466 | /// |
| 9467 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_shuffle_epi8&expand=5151) |
| 9468 | #[inline ] |
| 9469 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9470 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9471 | #[cfg_attr (test, assert_instr(vpshufb))] |
| 9472 | pub fn _mm_mask_shuffle_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 9473 | unsafe { |
| 9474 | let shuffle: i8x16 = _mm_shuffle_epi8(a, b).as_i8x16(); |
| 9475 | transmute(src:simd_select_bitmask(m:k, yes:shuffle, no:src.as_i8x16())) |
| 9476 | } |
| 9477 | } |
| 9478 | |
| 9479 | /// Shuffle packed 8-bit integers in a according to shuffle control mask in the corresponding 8-bit element of b, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 9480 | /// |
| 9481 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_shuffle_epi8&expand=5152) |
| 9482 | #[inline ] |
| 9483 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9484 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9485 | #[cfg_attr (test, assert_instr(vpshufb))] |
| 9486 | pub fn _mm_maskz_shuffle_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 9487 | unsafe { |
| 9488 | let shuffle: i8x16 = _mm_shuffle_epi8(a, b).as_i8x16(); |
| 9489 | transmute(src:simd_select_bitmask(m:k, yes:shuffle, no:i8x16::ZERO)) |
| 9490 | } |
| 9491 | } |
| 9492 | |
| 9493 | /// Compute the bitwise AND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero. |
| 9494 | /// |
| 9495 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_test_epi16_mask&expand=5884) |
| 9496 | #[inline ] |
| 9497 | #[target_feature (enable = "avx512bw" )] |
| 9498 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9499 | #[cfg_attr (test, assert_instr(vptestmw))] |
| 9500 | pub fn _mm512_test_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 9501 | let and: __m512i = _mm512_and_si512(a, b); |
| 9502 | let zero: __m512i = _mm512_setzero_si512(); |
| 9503 | _mm512_cmpneq_epi16_mask(a:and, b:zero) |
| 9504 | } |
| 9505 | |
| 9506 | /// Compute the bitwise AND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero. |
| 9507 | /// |
| 9508 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_test_epi16_mask&expand=5883) |
| 9509 | #[inline ] |
| 9510 | #[target_feature (enable = "avx512bw" )] |
| 9511 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9512 | #[cfg_attr (test, assert_instr(vptestmw))] |
| 9513 | pub fn _mm512_mask_test_epi16_mask(k: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 9514 | let and: __m512i = _mm512_and_si512(a, b); |
| 9515 | let zero: __m512i = _mm512_setzero_si512(); |
| 9516 | _mm512_mask_cmpneq_epi16_mask(k1:k, a:and, b:zero) |
| 9517 | } |
| 9518 | |
| 9519 | /// Compute the bitwise AND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero. |
| 9520 | /// |
| 9521 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_test_epi16_mask&expand=5882) |
| 9522 | #[inline ] |
| 9523 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9524 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9525 | #[cfg_attr (test, assert_instr(vptestmw))] |
| 9526 | pub fn _mm256_test_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 9527 | let and: __m256i = _mm256_and_si256(a, b); |
| 9528 | let zero: __m256i = _mm256_setzero_si256(); |
| 9529 | _mm256_cmpneq_epi16_mask(a:and, b:zero) |
| 9530 | } |
| 9531 | |
| 9532 | /// Compute the bitwise AND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero. |
| 9533 | /// |
| 9534 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_test_epi16_mask&expand=5881) |
| 9535 | #[inline ] |
| 9536 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9537 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9538 | #[cfg_attr (test, assert_instr(vptestmw))] |
| 9539 | pub fn _mm256_mask_test_epi16_mask(k: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 9540 | let and: __m256i = _mm256_and_si256(a, b); |
| 9541 | let zero: __m256i = _mm256_setzero_si256(); |
| 9542 | _mm256_mask_cmpneq_epi16_mask(k1:k, a:and, b:zero) |
| 9543 | } |
| 9544 | |
| 9545 | /// Compute the bitwise AND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero. |
| 9546 | /// |
| 9547 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_test_epi16_mask&expand=5880) |
| 9548 | #[inline ] |
| 9549 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9550 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9551 | #[cfg_attr (test, assert_instr(vptestmw))] |
| 9552 | pub fn _mm_test_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 9553 | let and: __m128i = _mm_and_si128(a, b); |
| 9554 | let zero: __m128i = _mm_setzero_si128(); |
| 9555 | _mm_cmpneq_epi16_mask(a:and, b:zero) |
| 9556 | } |
| 9557 | |
| 9558 | /// Compute the bitwise AND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero. |
| 9559 | /// |
| 9560 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_test_epi16_mask&expand=5879) |
| 9561 | #[inline ] |
| 9562 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9563 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9564 | #[cfg_attr (test, assert_instr(vptestmw))] |
| 9565 | pub fn _mm_mask_test_epi16_mask(k: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 9566 | let and: __m128i = _mm_and_si128(a, b); |
| 9567 | let zero: __m128i = _mm_setzero_si128(); |
| 9568 | _mm_mask_cmpneq_epi16_mask(k1:k, a:and, b:zero) |
| 9569 | } |
| 9570 | |
| 9571 | /// Compute the bitwise AND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero. |
| 9572 | /// |
| 9573 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_test_epi8_mask&expand=5902) |
| 9574 | #[inline ] |
| 9575 | #[target_feature (enable = "avx512bw" )] |
| 9576 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9577 | #[cfg_attr (test, assert_instr(vptestmb))] |
| 9578 | pub fn _mm512_test_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 9579 | let and: __m512i = _mm512_and_si512(a, b); |
| 9580 | let zero: __m512i = _mm512_setzero_si512(); |
| 9581 | _mm512_cmpneq_epi8_mask(a:and, b:zero) |
| 9582 | } |
| 9583 | |
| 9584 | /// Compute the bitwise AND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero. |
| 9585 | /// |
| 9586 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_test_epi8_mask&expand=5901) |
| 9587 | #[inline ] |
| 9588 | #[target_feature (enable = "avx512bw" )] |
| 9589 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9590 | #[cfg_attr (test, assert_instr(vptestmb))] |
| 9591 | pub fn _mm512_mask_test_epi8_mask(k: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 9592 | let and: __m512i = _mm512_and_si512(a, b); |
| 9593 | let zero: __m512i = _mm512_setzero_si512(); |
| 9594 | _mm512_mask_cmpneq_epi8_mask(k1:k, a:and, b:zero) |
| 9595 | } |
| 9596 | |
| 9597 | /// Compute the bitwise AND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero. |
| 9598 | /// |
| 9599 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_test_epi8_mask&expand=5900) |
| 9600 | #[inline ] |
| 9601 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9602 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9603 | #[cfg_attr (test, assert_instr(vptestmb))] |
| 9604 | pub fn _mm256_test_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 9605 | let and: __m256i = _mm256_and_si256(a, b); |
| 9606 | let zero: __m256i = _mm256_setzero_si256(); |
| 9607 | _mm256_cmpneq_epi8_mask(a:and, b:zero) |
| 9608 | } |
| 9609 | |
| 9610 | /// Compute the bitwise AND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero. |
| 9611 | /// |
| 9612 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_test_epi8_mask&expand=5899) |
| 9613 | #[inline ] |
| 9614 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9615 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9616 | #[cfg_attr (test, assert_instr(vptestmb))] |
| 9617 | pub fn _mm256_mask_test_epi8_mask(k: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 9618 | let and: __m256i = _mm256_and_si256(a, b); |
| 9619 | let zero: __m256i = _mm256_setzero_si256(); |
| 9620 | _mm256_mask_cmpneq_epi8_mask(k1:k, a:and, b:zero) |
| 9621 | } |
| 9622 | |
| 9623 | /// Compute the bitwise AND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k if the intermediate value is non-zero. |
| 9624 | /// |
| 9625 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_test_epi8_mask&expand=5898) |
| 9626 | #[inline ] |
| 9627 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9628 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9629 | #[cfg_attr (test, assert_instr(vptestmb))] |
| 9630 | pub fn _mm_test_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 9631 | let and: __m128i = _mm_and_si128(a, b); |
| 9632 | let zero: __m128i = _mm_setzero_si128(); |
| 9633 | _mm_cmpneq_epi8_mask(a:and, b:zero) |
| 9634 | } |
| 9635 | |
| 9636 | /// Compute the bitwise AND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is non-zero. |
| 9637 | /// |
| 9638 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_test_epi8_mask&expand=5897) |
| 9639 | #[inline ] |
| 9640 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9641 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9642 | #[cfg_attr (test, assert_instr(vptestmb))] |
| 9643 | pub fn _mm_mask_test_epi8_mask(k: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 9644 | let and: __m128i = _mm_and_si128(a, b); |
| 9645 | let zero: __m128i = _mm_setzero_si128(); |
| 9646 | _mm_mask_cmpneq_epi8_mask(k1:k, a:and, b:zero) |
| 9647 | } |
| 9648 | |
| 9649 | /// Compute the bitwise NAND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k if the intermediate value is zero. |
| 9650 | /// |
| 9651 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_testn_epi16_mask&expand=5915) |
| 9652 | #[inline ] |
| 9653 | #[target_feature (enable = "avx512bw" )] |
| 9654 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9655 | #[cfg_attr (test, assert_instr(vptestnmw))] |
| 9656 | pub fn _mm512_testn_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
| 9657 | let and: __m512i = _mm512_and_si512(a, b); |
| 9658 | let zero: __m512i = _mm512_setzero_si512(); |
| 9659 | _mm512_cmpeq_epi16_mask(a:and, b:zero) |
| 9660 | } |
| 9661 | |
| 9662 | /// Compute the bitwise NAND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero. |
| 9663 | /// |
| 9664 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_testn_epi16_mask&expand=5914) |
| 9665 | #[inline ] |
| 9666 | #[target_feature (enable = "avx512bw" )] |
| 9667 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9668 | #[cfg_attr (test, assert_instr(vptestnmw))] |
| 9669 | pub fn _mm512_mask_testn_epi16_mask(k: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
| 9670 | let and: __m512i = _mm512_and_si512(a, b); |
| 9671 | let zero: __m512i = _mm512_setzero_si512(); |
| 9672 | _mm512_mask_cmpeq_epi16_mask(k1:k, a:and, b:zero) |
| 9673 | } |
| 9674 | |
| 9675 | /// Compute the bitwise NAND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k if the intermediate value is zero. |
| 9676 | /// |
| 9677 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_testn_epi16_mask&expand=5913) |
| 9678 | #[inline ] |
| 9679 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9680 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9681 | #[cfg_attr (test, assert_instr(vptestnmw))] |
| 9682 | pub fn _mm256_testn_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
| 9683 | let and: __m256i = _mm256_and_si256(a, b); |
| 9684 | let zero: __m256i = _mm256_setzero_si256(); |
| 9685 | _mm256_cmpeq_epi16_mask(a:and, b:zero) |
| 9686 | } |
| 9687 | |
| 9688 | /// Compute the bitwise NAND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero. |
| 9689 | /// |
| 9690 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_testn_epi16_mask&expand=5912) |
| 9691 | #[inline ] |
| 9692 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9693 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9694 | #[cfg_attr (test, assert_instr(vptestnmw))] |
| 9695 | pub fn _mm256_mask_testn_epi16_mask(k: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
| 9696 | let and: __m256i = _mm256_and_si256(a, b); |
| 9697 | let zero: __m256i = _mm256_setzero_si256(); |
| 9698 | _mm256_mask_cmpeq_epi16_mask(k1:k, a:and, b:zero) |
| 9699 | } |
| 9700 | |
| 9701 | /// Compute the bitwise NAND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k if the intermediate value is zero. |
| 9702 | /// |
| 9703 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_testn_epi16_mask&expand=5911) |
| 9704 | #[inline ] |
| 9705 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9706 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9707 | #[cfg_attr (test, assert_instr(vptestnmw))] |
| 9708 | pub fn _mm_testn_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
| 9709 | let and: __m128i = _mm_and_si128(a, b); |
| 9710 | let zero: __m128i = _mm_setzero_si128(); |
| 9711 | _mm_cmpeq_epi16_mask(a:and, b:zero) |
| 9712 | } |
| 9713 | |
| 9714 | /// Compute the bitwise NAND of packed 16-bit integers in a and b, producing intermediate 16-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero. |
| 9715 | /// |
| 9716 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_testn_epi16_mask&expand=5910) |
| 9717 | #[inline ] |
| 9718 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9719 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9720 | #[cfg_attr (test, assert_instr(vptestnmw))] |
| 9721 | pub fn _mm_mask_testn_epi16_mask(k: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
| 9722 | let and: __m128i = _mm_and_si128(a, b); |
| 9723 | let zero: __m128i = _mm_setzero_si128(); |
| 9724 | _mm_mask_cmpeq_epi16_mask(k1:k, a:and, b:zero) |
| 9725 | } |
| 9726 | |
| 9727 | /// Compute the bitwise NAND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k if the intermediate value is zero. |
| 9728 | /// |
| 9729 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_testn_epi8_mask&expand=5933) |
| 9730 | #[inline ] |
| 9731 | #[target_feature (enable = "avx512bw" )] |
| 9732 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9733 | #[cfg_attr (test, assert_instr(vptestnmb))] |
| 9734 | pub fn _mm512_testn_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
| 9735 | let and: __m512i = _mm512_and_si512(a, b); |
| 9736 | let zero: __m512i = _mm512_setzero_si512(); |
| 9737 | _mm512_cmpeq_epi8_mask(a:and, b:zero) |
| 9738 | } |
| 9739 | |
| 9740 | /// Compute the bitwise NAND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero. |
| 9741 | /// |
| 9742 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_testn_epi8_mask&expand=5932) |
| 9743 | #[inline ] |
| 9744 | #[target_feature (enable = "avx512bw" )] |
| 9745 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9746 | #[cfg_attr (test, assert_instr(vptestnmb))] |
| 9747 | pub fn _mm512_mask_testn_epi8_mask(k: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
| 9748 | let and: __m512i = _mm512_and_si512(a, b); |
| 9749 | let zero: __m512i = _mm512_setzero_si512(); |
| 9750 | _mm512_mask_cmpeq_epi8_mask(k1:k, a:and, b:zero) |
| 9751 | } |
| 9752 | |
| 9753 | /// Compute the bitwise NAND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k if the intermediate value is zero. |
| 9754 | /// |
| 9755 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_testn_epi8_mask&expand=5931) |
| 9756 | #[inline ] |
| 9757 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9758 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9759 | #[cfg_attr (test, assert_instr(vptestnmb))] |
| 9760 | pub fn _mm256_testn_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
| 9761 | let and: __m256i = _mm256_and_si256(a, b); |
| 9762 | let zero: __m256i = _mm256_setzero_si256(); |
| 9763 | _mm256_cmpeq_epi8_mask(a:and, b:zero) |
| 9764 | } |
| 9765 | |
| 9766 | /// Compute the bitwise NAND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero. |
| 9767 | /// |
| 9768 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_testn_epi8_mask&expand=5930) |
| 9769 | #[inline ] |
| 9770 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9771 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9772 | #[cfg_attr (test, assert_instr(vptestnmb))] |
| 9773 | pub fn _mm256_mask_testn_epi8_mask(k: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
| 9774 | let and: __m256i = _mm256_and_si256(a, b); |
| 9775 | let zero: __m256i = _mm256_setzero_si256(); |
| 9776 | _mm256_mask_cmpeq_epi8_mask(k1:k, a:and, b:zero) |
| 9777 | } |
| 9778 | |
| 9779 | /// Compute the bitwise NAND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k if the intermediate value is zero. |
| 9780 | /// |
| 9781 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_testn_epi8_mask&expand=5929) |
| 9782 | #[inline ] |
| 9783 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9784 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9785 | #[cfg_attr (test, assert_instr(vptestnmb))] |
| 9786 | pub fn _mm_testn_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
| 9787 | let and: __m128i = _mm_and_si128(a, b); |
| 9788 | let zero: __m128i = _mm_setzero_si128(); |
| 9789 | _mm_cmpeq_epi8_mask(a:and, b:zero) |
| 9790 | } |
| 9791 | |
| 9792 | /// Compute the bitwise NAND of packed 8-bit integers in a and b, producing intermediate 8-bit values, and set the corresponding bit in result mask k (subject to writemask k) if the intermediate value is zero. |
| 9793 | /// |
| 9794 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_testn_epi8_mask&expand=5928) |
| 9795 | #[inline ] |
| 9796 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9797 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9798 | #[cfg_attr (test, assert_instr(vptestnmb))] |
| 9799 | pub fn _mm_mask_testn_epi8_mask(k: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
| 9800 | let and: __m128i = _mm_and_si128(a, b); |
| 9801 | let zero: __m128i = _mm_setzero_si128(); |
| 9802 | _mm_mask_cmpeq_epi8_mask(k1:k, a:and, b:zero) |
| 9803 | } |
| 9804 | |
| 9805 | /// Store 64-bit mask from a into memory. |
| 9806 | /// |
| 9807 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_store_mask64&expand=5578) |
| 9808 | #[inline ] |
| 9809 | #[target_feature (enable = "avx512bw" )] |
| 9810 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9811 | #[cfg_attr (test, assert_instr(mov))] //should be kmovq |
| 9812 | pub unsafe fn _store_mask64(mem_addr: *mut __mmask64, a: __mmask64) { |
| 9813 | ptr::write(dst:mem_addr as *mut __mmask64, src:a); |
| 9814 | } |
| 9815 | |
| 9816 | /// Store 32-bit mask from a into memory. |
| 9817 | /// |
| 9818 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_store_mask32&expand=5577) |
| 9819 | #[inline ] |
| 9820 | #[target_feature (enable = "avx512bw" )] |
| 9821 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9822 | #[cfg_attr (test, assert_instr(mov))] //should be kmovd |
| 9823 | pub unsafe fn _store_mask32(mem_addr: *mut __mmask32, a: __mmask32) { |
| 9824 | ptr::write(dst:mem_addr as *mut __mmask32, src:a); |
| 9825 | } |
| 9826 | |
| 9827 | /// Load 64-bit mask from memory into k. |
| 9828 | /// |
| 9829 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_load_mask64&expand=3318) |
| 9830 | #[inline ] |
| 9831 | #[target_feature (enable = "avx512bw" )] |
| 9832 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9833 | #[cfg_attr (test, assert_instr(mov))] //should be kmovq |
| 9834 | pub unsafe fn _load_mask64(mem_addr: *const __mmask64) -> __mmask64 { |
| 9835 | ptr::read(src:mem_addr as *const __mmask64) |
| 9836 | } |
| 9837 | |
| 9838 | /// Load 32-bit mask from memory into k. |
| 9839 | /// |
| 9840 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_load_mask32&expand=3317) |
| 9841 | #[inline ] |
| 9842 | #[target_feature (enable = "avx512bw" )] |
| 9843 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9844 | #[cfg_attr (test, assert_instr(mov))] //should be kmovd |
| 9845 | pub unsafe fn _load_mask32(mem_addr: *const __mmask32) -> __mmask32 { |
| 9846 | ptr::read(src:mem_addr as *const __mmask32) |
| 9847 | } |
| 9848 | |
| 9849 | /// Compute the absolute differences of packed unsigned 8-bit integers in a and b, then horizontally sum each consecutive 8 differences to produce eight unsigned 16-bit integers, and pack these unsigned 16-bit integers in the low 16 bits of 64-bit elements in dst. |
| 9850 | /// |
| 9851 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_sad_epu8&expand=4855) |
| 9852 | #[inline ] |
| 9853 | #[target_feature (enable = "avx512bw" )] |
| 9854 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9855 | #[cfg_attr (test, assert_instr(vpsadbw))] |
| 9856 | pub fn _mm512_sad_epu8(a: __m512i, b: __m512i) -> __m512i { |
| 9857 | unsafe { transmute(src:vpsadbw(a.as_u8x64(), b.as_u8x64())) } |
| 9858 | } |
| 9859 | |
| 9860 | /// Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst. Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets. |
| 9861 | /// |
| 9862 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_dbsad_epu8&expand=2114) |
| 9863 | #[inline ] |
| 9864 | #[target_feature (enable = "avx512bw" )] |
| 9865 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9866 | #[rustc_legacy_const_generics (2)] |
| 9867 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
| 9868 | pub fn _mm512_dbsad_epu8<const IMM8: i32>(a: __m512i, b: __m512i) -> __m512i { |
| 9869 | unsafe { |
| 9870 | static_assert_uimm_bits!(IMM8, 8); |
| 9871 | let a: u8x64 = a.as_u8x64(); |
| 9872 | let b: u8x64 = b.as_u8x64(); |
| 9873 | let r: u16x32 = vdbpsadbw(a, b, IMM8); |
| 9874 | transmute(src:r) |
| 9875 | } |
| 9876 | } |
| 9877 | |
| 9878 | /// Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets. |
| 9879 | /// |
| 9880 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_dbsad_epu8&expand=2115) |
| 9881 | #[inline ] |
| 9882 | #[target_feature (enable = "avx512bw" )] |
| 9883 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9884 | #[rustc_legacy_const_generics (4)] |
| 9885 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
| 9886 | pub fn _mm512_mask_dbsad_epu8<const IMM8: i32>( |
| 9887 | src: __m512i, |
| 9888 | k: __mmask32, |
| 9889 | a: __m512i, |
| 9890 | b: __m512i, |
| 9891 | ) -> __m512i { |
| 9892 | unsafe { |
| 9893 | static_assert_uimm_bits!(IMM8, 8); |
| 9894 | let a: u8x64 = a.as_u8x64(); |
| 9895 | let b: u8x64 = b.as_u8x64(); |
| 9896 | let r: u16x32 = vdbpsadbw(a, b, IMM8); |
| 9897 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_u16x32())) |
| 9898 | } |
| 9899 | } |
| 9900 | |
| 9901 | /// Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets. |
| 9902 | /// |
| 9903 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_dbsad_epu8&expand=2116) |
| 9904 | #[inline ] |
| 9905 | #[target_feature (enable = "avx512bw" )] |
| 9906 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9907 | #[rustc_legacy_const_generics (3)] |
| 9908 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
| 9909 | pub fn _mm512_maskz_dbsad_epu8<const IMM8: i32>(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
| 9910 | unsafe { |
| 9911 | static_assert_uimm_bits!(IMM8, 8); |
| 9912 | let a: u8x64 = a.as_u8x64(); |
| 9913 | let b: u8x64 = b.as_u8x64(); |
| 9914 | let r: u16x32 = vdbpsadbw(a, b, IMM8); |
| 9915 | transmute(src:simd_select_bitmask(m:k, yes:r, no:u16x32::ZERO)) |
| 9916 | } |
| 9917 | } |
| 9918 | |
| 9919 | /// Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst. Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets. |
| 9920 | /// |
| 9921 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_dbsad_epu8&expand=2111) |
| 9922 | #[inline ] |
| 9923 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9924 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9925 | #[rustc_legacy_const_generics (2)] |
| 9926 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
| 9927 | pub fn _mm256_dbsad_epu8<const IMM8: i32>(a: __m256i, b: __m256i) -> __m256i { |
| 9928 | unsafe { |
| 9929 | static_assert_uimm_bits!(IMM8, 8); |
| 9930 | let a: u8x32 = a.as_u8x32(); |
| 9931 | let b: u8x32 = b.as_u8x32(); |
| 9932 | let r: u16x16 = vdbpsadbw256(a, b, IMM8); |
| 9933 | transmute(src:r) |
| 9934 | } |
| 9935 | } |
| 9936 | |
| 9937 | /// Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets. |
| 9938 | /// |
| 9939 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_dbsad_epu8&expand=2112) |
| 9940 | #[inline ] |
| 9941 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9942 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9943 | #[rustc_legacy_const_generics (4)] |
| 9944 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
| 9945 | pub fn _mm256_mask_dbsad_epu8<const IMM8: i32>( |
| 9946 | src: __m256i, |
| 9947 | k: __mmask16, |
| 9948 | a: __m256i, |
| 9949 | b: __m256i, |
| 9950 | ) -> __m256i { |
| 9951 | unsafe { |
| 9952 | static_assert_uimm_bits!(IMM8, 8); |
| 9953 | let a: u8x32 = a.as_u8x32(); |
| 9954 | let b: u8x32 = b.as_u8x32(); |
| 9955 | let r: u16x16 = vdbpsadbw256(a, b, IMM8); |
| 9956 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_u16x16())) |
| 9957 | } |
| 9958 | } |
| 9959 | |
| 9960 | /// Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets. |
| 9961 | /// |
| 9962 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_dbsad_epu8&expand=2113) |
| 9963 | #[inline ] |
| 9964 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9965 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9966 | #[rustc_legacy_const_generics (3)] |
| 9967 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
| 9968 | pub fn _mm256_maskz_dbsad_epu8<const IMM8: i32>(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
| 9969 | unsafe { |
| 9970 | static_assert_uimm_bits!(IMM8, 8); |
| 9971 | let a: u8x32 = a.as_u8x32(); |
| 9972 | let b: u8x32 = b.as_u8x32(); |
| 9973 | let r: u16x16 = vdbpsadbw256(a, b, IMM8); |
| 9974 | transmute(src:simd_select_bitmask(m:k, yes:r, no:u16x16::ZERO)) |
| 9975 | } |
| 9976 | } |
| 9977 | |
| 9978 | /// Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst. Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets. |
| 9979 | /// |
| 9980 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_dbsad_epu8&expand=2108) |
| 9981 | #[inline ] |
| 9982 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 9983 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 9984 | #[rustc_legacy_const_generics (2)] |
| 9985 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
| 9986 | pub fn _mm_dbsad_epu8<const IMM8: i32>(a: __m128i, b: __m128i) -> __m128i { |
| 9987 | unsafe { |
| 9988 | static_assert_uimm_bits!(IMM8, 8); |
| 9989 | let a: u8x16 = a.as_u8x16(); |
| 9990 | let b: u8x16 = b.as_u8x16(); |
| 9991 | let r: u16x8 = vdbpsadbw128(a, b, IMM8); |
| 9992 | transmute(src:r) |
| 9993 | } |
| 9994 | } |
| 9995 | |
| 9996 | /// Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets. |
| 9997 | /// |
| 9998 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_dbsad_epu8&expand=2109) |
| 9999 | #[inline ] |
| 10000 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10001 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10002 | #[rustc_legacy_const_generics (4)] |
| 10003 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
| 10004 | pub fn _mm_mask_dbsad_epu8<const IMM8: i32>( |
| 10005 | src: __m128i, |
| 10006 | k: __mmask8, |
| 10007 | a: __m128i, |
| 10008 | b: __m128i, |
| 10009 | ) -> __m128i { |
| 10010 | unsafe { |
| 10011 | static_assert_uimm_bits!(IMM8, 8); |
| 10012 | let a: u8x16 = a.as_u8x16(); |
| 10013 | let b: u8x16 = b.as_u8x16(); |
| 10014 | let r: u16x8 = vdbpsadbw128(a, b, IMM8); |
| 10015 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_u16x8())) |
| 10016 | } |
| 10017 | } |
| 10018 | |
| 10019 | /// Compute the sum of absolute differences (SADs) of quadruplets of unsigned 8-bit integers in a compared to those in b, and store the 16-bit results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). Four SADs are performed on four 8-bit quadruplets for each 64-bit lane. The first two SADs use the lower 8-bit quadruplet of the lane from a, and the last two SADs use the uppper 8-bit quadruplet of the lane from a. Quadruplets from b are selected from within 128-bit lanes according to the control in imm8, and each SAD in each 64-bit lane uses the selected quadruplet at 8-bit offsets. |
| 10020 | /// |
| 10021 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_dbsad_epu8&expand=2110) |
| 10022 | #[inline ] |
| 10023 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10024 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10025 | #[rustc_legacy_const_generics (3)] |
| 10026 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
| 10027 | pub fn _mm_maskz_dbsad_epu8<const IMM8: i32>(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
| 10028 | unsafe { |
| 10029 | static_assert_uimm_bits!(IMM8, 8); |
| 10030 | let a: u8x16 = a.as_u8x16(); |
| 10031 | let b: u8x16 = b.as_u8x16(); |
| 10032 | let r: u16x8 = vdbpsadbw128(a, b, IMM8); |
| 10033 | transmute(src:simd_select_bitmask(m:k, yes:r, no:u16x8::ZERO)) |
| 10034 | } |
| 10035 | } |
| 10036 | |
| 10037 | /// Set each bit of mask register k based on the most significant bit of the corresponding packed 16-bit integer in a. |
| 10038 | /// |
| 10039 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_movepi16_mask&expand=3873) |
| 10040 | #[inline ] |
| 10041 | #[target_feature (enable = "avx512bw" )] |
| 10042 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10043 | #[cfg_attr (test, assert_instr(vpmovw2m))] |
| 10044 | pub fn _mm512_movepi16_mask(a: __m512i) -> __mmask32 { |
| 10045 | let filter: __m512i = _mm512_set1_epi16(1 << 15); |
| 10046 | let a: __m512i = _mm512_and_si512(a, b:filter); |
| 10047 | _mm512_cmpeq_epi16_mask(a, b:filter) |
| 10048 | } |
| 10049 | |
| 10050 | /// Set each bit of mask register k based on the most significant bit of the corresponding packed 16-bit integer in a. |
| 10051 | /// |
| 10052 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_movepi16_mask&expand=3872) |
| 10053 | #[inline ] |
| 10054 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10055 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10056 | #[cfg_attr (test, assert_instr(vpmovw2m))] |
| 10057 | pub fn _mm256_movepi16_mask(a: __m256i) -> __mmask16 { |
| 10058 | let filter: __m256i = _mm256_set1_epi16(1 << 15); |
| 10059 | let a: __m256i = _mm256_and_si256(a, b:filter); |
| 10060 | _mm256_cmpeq_epi16_mask(a, b:filter) |
| 10061 | } |
| 10062 | |
| 10063 | /// Set each bit of mask register k based on the most significant bit of the corresponding packed 16-bit integer in a. |
| 10064 | /// |
| 10065 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_movepi16_mask&expand=3871) |
| 10066 | #[inline ] |
| 10067 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10068 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10069 | #[cfg_attr (test, assert_instr(vpmovw2m))] |
| 10070 | pub fn _mm_movepi16_mask(a: __m128i) -> __mmask8 { |
| 10071 | let filter: __m128i = _mm_set1_epi16(1 << 15); |
| 10072 | let a: __m128i = _mm_and_si128(a, b:filter); |
| 10073 | _mm_cmpeq_epi16_mask(a, b:filter) |
| 10074 | } |
| 10075 | |
| 10076 | /// Set each bit of mask register k based on the most significant bit of the corresponding packed 8-bit integer in a. |
| 10077 | /// |
| 10078 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_movepi8_mask&expand=3883) |
| 10079 | #[inline ] |
| 10080 | #[target_feature (enable = "avx512bw" )] |
| 10081 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10082 | #[cfg_attr (test, assert_instr(vpmovb2m))] |
| 10083 | pub fn _mm512_movepi8_mask(a: __m512i) -> __mmask64 { |
| 10084 | let filter: __m512i = _mm512_set1_epi8(1 << 7); |
| 10085 | let a: __m512i = _mm512_and_si512(a, b:filter); |
| 10086 | _mm512_cmpeq_epi8_mask(a, b:filter) |
| 10087 | } |
| 10088 | |
| 10089 | /// Set each bit of mask register k based on the most significant bit of the corresponding packed 8-bit integer in a. |
| 10090 | /// |
| 10091 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_movepi8_mask&expand=3882) |
| 10092 | #[inline ] |
| 10093 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10094 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10095 | #[cfg_attr (test, assert_instr(vpmovmskb))] // should be vpmovb2m but compiled to vpmovmskb in the test shim because that takes less cycles than |
| 10096 | // using vpmovb2m plus converting the mask register to a standard register. |
| 10097 | pub fn _mm256_movepi8_mask(a: __m256i) -> __mmask32 { |
| 10098 | let filter: __m256i = _mm256_set1_epi8(1 << 7); |
| 10099 | let a: __m256i = _mm256_and_si256(a, b:filter); |
| 10100 | _mm256_cmpeq_epi8_mask(a, b:filter) |
| 10101 | } |
| 10102 | |
| 10103 | /// Set each bit of mask register k based on the most significant bit of the corresponding packed 8-bit integer in a. |
| 10104 | /// |
| 10105 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_movepi8_mask&expand=3881) |
| 10106 | #[inline ] |
| 10107 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10108 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10109 | #[cfg_attr (test, assert_instr(vpmovmskb))] // should be vpmovb2m but compiled to vpmovmskb in the test shim because that takes less cycles than |
| 10110 | // using vpmovb2m plus converting the mask register to a standard register. |
| 10111 | pub fn _mm_movepi8_mask(a: __m128i) -> __mmask16 { |
| 10112 | let filter: __m128i = _mm_set1_epi8(1 << 7); |
| 10113 | let a: __m128i = _mm_and_si128(a, b:filter); |
| 10114 | _mm_cmpeq_epi8_mask(a, b:filter) |
| 10115 | } |
| 10116 | |
| 10117 | /// Set each packed 16-bit integer in dst to all ones or all zeros based on the value of the corresponding bit in k. |
| 10118 | /// |
| 10119 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_movm_epi16&expand=3886) |
| 10120 | #[inline ] |
| 10121 | #[target_feature (enable = "avx512bw" )] |
| 10122 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10123 | #[cfg_attr (test, assert_instr(vpmovm2w))] |
| 10124 | pub fn _mm512_movm_epi16(k: __mmask32) -> __m512i { |
| 10125 | unsafe { |
| 10126 | let one: i16x32 = _mm512_set1_epi16__m512i( |
| 10127 | 1 << 15 |
| 10128 | | 1 << 14 |
| 10129 | | 1 << 13 |
| 10130 | | 1 << 12 |
| 10131 | | 1 << 11 |
| 10132 | | 1 << 10 |
| 10133 | | 1 << 9 |
| 10134 | | 1 << 8 |
| 10135 | | 1 << 7 |
| 10136 | | 1 << 6 |
| 10137 | | 1 << 5 |
| 10138 | | 1 << 4 |
| 10139 | | 1 << 3 |
| 10140 | | 1 << 2 |
| 10141 | | 1 << 1 |
| 10142 | | 1 << 0, |
| 10143 | ) |
| 10144 | .as_i16x32(); |
| 10145 | transmute(src:simd_select_bitmask(m:k, yes:one, no:i16x32::ZERO)) |
| 10146 | } |
| 10147 | } |
| 10148 | |
| 10149 | /// Set each packed 16-bit integer in dst to all ones or all zeros based on the value of the corresponding bit in k. |
| 10150 | /// |
| 10151 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_movm_epi16&expand=3885) |
| 10152 | #[inline ] |
| 10153 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10154 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10155 | #[cfg_attr (test, assert_instr(vpmovm2w))] |
| 10156 | pub fn _mm256_movm_epi16(k: __mmask16) -> __m256i { |
| 10157 | unsafe { |
| 10158 | let one: i16x16 = _mm256_set1_epi16__m256i( |
| 10159 | 1 << 15 |
| 10160 | | 1 << 14 |
| 10161 | | 1 << 13 |
| 10162 | | 1 << 12 |
| 10163 | | 1 << 11 |
| 10164 | | 1 << 10 |
| 10165 | | 1 << 9 |
| 10166 | | 1 << 8 |
| 10167 | | 1 << 7 |
| 10168 | | 1 << 6 |
| 10169 | | 1 << 5 |
| 10170 | | 1 << 4 |
| 10171 | | 1 << 3 |
| 10172 | | 1 << 2 |
| 10173 | | 1 << 1 |
| 10174 | | 1 << 0, |
| 10175 | ) |
| 10176 | .as_i16x16(); |
| 10177 | transmute(src:simd_select_bitmask(m:k, yes:one, no:i16x16::ZERO)) |
| 10178 | } |
| 10179 | } |
| 10180 | |
| 10181 | /// Set each packed 16-bit integer in dst to all ones or all zeros based on the value of the corresponding bit in k. |
| 10182 | /// |
| 10183 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_movm_epi16&expand=3884) |
| 10184 | #[inline ] |
| 10185 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10186 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10187 | #[cfg_attr (test, assert_instr(vpmovm2w))] |
| 10188 | pub fn _mm_movm_epi16(k: __mmask8) -> __m128i { |
| 10189 | unsafe { |
| 10190 | let one: i16x8 = _mm_set1_epi16__m128i( |
| 10191 | 1 << 15 |
| 10192 | | 1 << 14 |
| 10193 | | 1 << 13 |
| 10194 | | 1 << 12 |
| 10195 | | 1 << 11 |
| 10196 | | 1 << 10 |
| 10197 | | 1 << 9 |
| 10198 | | 1 << 8 |
| 10199 | | 1 << 7 |
| 10200 | | 1 << 6 |
| 10201 | | 1 << 5 |
| 10202 | | 1 << 4 |
| 10203 | | 1 << 3 |
| 10204 | | 1 << 2 |
| 10205 | | 1 << 1 |
| 10206 | | 1 << 0, |
| 10207 | ) |
| 10208 | .as_i16x8(); |
| 10209 | transmute(src:simd_select_bitmask(m:k, yes:one, no:i16x8::ZERO)) |
| 10210 | } |
| 10211 | } |
| 10212 | |
| 10213 | /// Set each packed 8-bit integer in dst to all ones or all zeros based on the value of the corresponding bit in k. |
| 10214 | /// |
| 10215 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_movm_epi8&expand=3895) |
| 10216 | #[inline ] |
| 10217 | #[target_feature (enable = "avx512bw" )] |
| 10218 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10219 | #[cfg_attr (test, assert_instr(vpmovm2b))] |
| 10220 | pub fn _mm512_movm_epi8(k: __mmask64) -> __m512i { |
| 10221 | unsafe { |
| 10222 | let one: i8x64 = |
| 10223 | _mm512_set1_epi8__m512i(1 << 7 | 1 << 6 | 1 << 5 | 1 << 4 | 1 << 3 | 1 << 2 | 1 << 1 | 1 << 0) |
| 10224 | .as_i8x64(); |
| 10225 | transmute(src:simd_select_bitmask(m:k, yes:one, no:i8x64::ZERO)) |
| 10226 | } |
| 10227 | } |
| 10228 | |
| 10229 | /// Set each packed 8-bit integer in dst to all ones or all zeros based on the value of the corresponding bit in k. |
| 10230 | /// |
| 10231 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_movm_epi8&expand=3894) |
| 10232 | #[inline ] |
| 10233 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10234 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10235 | #[cfg_attr (test, assert_instr(vpmovm2b))] |
| 10236 | pub fn _mm256_movm_epi8(k: __mmask32) -> __m256i { |
| 10237 | unsafe { |
| 10238 | let one: i8x32 = |
| 10239 | _mm256_set1_epi8__m256i(1 << 7 | 1 << 6 | 1 << 5 | 1 << 4 | 1 << 3 | 1 << 2 | 1 << 1 | 1 << 0) |
| 10240 | .as_i8x32(); |
| 10241 | transmute(src:simd_select_bitmask(m:k, yes:one, no:i8x32::ZERO)) |
| 10242 | } |
| 10243 | } |
| 10244 | |
| 10245 | /// Set each packed 8-bit integer in dst to all ones or all zeros based on the value of the corresponding bit in k. |
| 10246 | /// |
| 10247 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_movm_epi8&expand=3893) |
| 10248 | #[inline ] |
| 10249 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10250 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10251 | #[cfg_attr (test, assert_instr(vpmovm2b))] |
| 10252 | pub fn _mm_movm_epi8(k: __mmask16) -> __m128i { |
| 10253 | unsafe { |
| 10254 | let one: i8x16 = |
| 10255 | _mm_set1_epi8__m128i(1 << 7 | 1 << 6 | 1 << 5 | 1 << 4 | 1 << 3 | 1 << 2 | 1 << 1 | 1 << 0) |
| 10256 | .as_i8x16(); |
| 10257 | transmute(src:simd_select_bitmask(m:k, yes:one, no:i8x16::ZERO)) |
| 10258 | } |
| 10259 | } |
| 10260 | |
| 10261 | /// Convert 32-bit mask a into an integer value, and store the result in dst. |
| 10262 | /// |
| 10263 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#_cvtmask32_u32) |
| 10264 | #[inline ] |
| 10265 | #[target_feature (enable = "avx512bw" )] |
| 10266 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10267 | pub fn _cvtmask32_u32(a: __mmask32) -> u32 { |
| 10268 | a |
| 10269 | } |
| 10270 | |
| 10271 | /// Convert integer value a into an 32-bit mask, and store the result in k. |
| 10272 | /// |
| 10273 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_cvtu32_mask32) |
| 10274 | #[inline ] |
| 10275 | #[target_feature (enable = "avx512bw" )] |
| 10276 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10277 | pub fn _cvtu32_mask32(a: u32) -> __mmask32 { |
| 10278 | a |
| 10279 | } |
| 10280 | |
| 10281 | /// Add 32-bit masks in a and b, and store the result in k. |
| 10282 | /// |
| 10283 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kadd_mask32&expand=3207) |
| 10284 | #[inline ] |
| 10285 | #[target_feature (enable = "avx512bw" )] |
| 10286 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10287 | pub fn _kadd_mask32(a: __mmask32, b: __mmask32) -> __mmask32 { |
| 10288 | a + b |
| 10289 | } |
| 10290 | |
| 10291 | /// Add 64-bit masks in a and b, and store the result in k. |
| 10292 | /// |
| 10293 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kadd_mask64&expand=3208) |
| 10294 | #[inline ] |
| 10295 | #[target_feature (enable = "avx512bw" )] |
| 10296 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10297 | pub fn _kadd_mask64(a: __mmask64, b: __mmask64) -> __mmask64 { |
| 10298 | a + b |
| 10299 | } |
| 10300 | |
| 10301 | /// Compute the bitwise AND of 32-bit masks a and b, and store the result in k. |
| 10302 | /// |
| 10303 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kand_mask32&expand=3213) |
| 10304 | #[inline ] |
| 10305 | #[target_feature (enable = "avx512bw" )] |
| 10306 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10307 | pub fn _kand_mask32(a: __mmask32, b: __mmask32) -> __mmask32 { |
| 10308 | a & b |
| 10309 | } |
| 10310 | |
| 10311 | /// Compute the bitwise AND of 64-bit masks a and b, and store the result in k. |
| 10312 | /// |
| 10313 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kand_mask64&expand=3214) |
| 10314 | #[inline ] |
| 10315 | #[target_feature (enable = "avx512bw" )] |
| 10316 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10317 | pub fn _kand_mask64(a: __mmask64, b: __mmask64) -> __mmask64 { |
| 10318 | a & b |
| 10319 | } |
| 10320 | |
| 10321 | /// Compute the bitwise NOT of 32-bit mask a, and store the result in k. |
| 10322 | /// |
| 10323 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_knot_mask32&expand=3234) |
| 10324 | #[inline ] |
| 10325 | #[target_feature (enable = "avx512bw" )] |
| 10326 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10327 | pub fn _knot_mask32(a: __mmask32) -> __mmask32 { |
| 10328 | !a |
| 10329 | } |
| 10330 | |
| 10331 | /// Compute the bitwise NOT of 64-bit mask a, and store the result in k. |
| 10332 | /// |
| 10333 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_knot_mask64&expand=3235) |
| 10334 | #[inline ] |
| 10335 | #[target_feature (enable = "avx512bw" )] |
| 10336 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10337 | pub fn _knot_mask64(a: __mmask64) -> __mmask64 { |
| 10338 | !a |
| 10339 | } |
| 10340 | |
| 10341 | /// Compute the bitwise NOT of 32-bit masks a and then AND with b, and store the result in k. |
| 10342 | /// |
| 10343 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kandn_mask32&expand=3219) |
| 10344 | #[inline ] |
| 10345 | #[target_feature (enable = "avx512bw" )] |
| 10346 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10347 | pub fn _kandn_mask32(a: __mmask32, b: __mmask32) -> __mmask32 { |
| 10348 | _knot_mask32(a) & b |
| 10349 | } |
| 10350 | |
| 10351 | /// Compute the bitwise NOT of 64-bit masks a and then AND with b, and store the result in k. |
| 10352 | /// |
| 10353 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kandn_mask64&expand=3220) |
| 10354 | #[inline ] |
| 10355 | #[target_feature (enable = "avx512bw" )] |
| 10356 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10357 | pub fn _kandn_mask64(a: __mmask64, b: __mmask64) -> __mmask64 { |
| 10358 | _knot_mask64(a) & b |
| 10359 | } |
| 10360 | |
| 10361 | /// Compute the bitwise OR of 32-bit masks a and b, and store the result in k. |
| 10362 | /// |
| 10363 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kor_mask32&expand=3240) |
| 10364 | #[inline ] |
| 10365 | #[target_feature (enable = "avx512bw" )] |
| 10366 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10367 | pub fn _kor_mask32(a: __mmask32, b: __mmask32) -> __mmask32 { |
| 10368 | a | b |
| 10369 | } |
| 10370 | |
| 10371 | /// Compute the bitwise OR of 64-bit masks a and b, and store the result in k. |
| 10372 | /// |
| 10373 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kor_mask64&expand=3241) |
| 10374 | #[inline ] |
| 10375 | #[target_feature (enable = "avx512bw" )] |
| 10376 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10377 | pub fn _kor_mask64(a: __mmask64, b: __mmask64) -> __mmask64 { |
| 10378 | a | b |
| 10379 | } |
| 10380 | |
| 10381 | /// Compute the bitwise XOR of 32-bit masks a and b, and store the result in k. |
| 10382 | /// |
| 10383 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kxor_mask32&expand=3292) |
| 10384 | #[inline ] |
| 10385 | #[target_feature (enable = "avx512bw" )] |
| 10386 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10387 | pub fn _kxor_mask32(a: __mmask32, b: __mmask32) -> __mmask32 { |
| 10388 | a ^ b |
| 10389 | } |
| 10390 | |
| 10391 | /// Compute the bitwise XOR of 64-bit masks a and b, and store the result in k. |
| 10392 | /// |
| 10393 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kxor_mask64&expand=3293) |
| 10394 | #[inline ] |
| 10395 | #[target_feature (enable = "avx512bw" )] |
| 10396 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10397 | pub fn _kxor_mask64(a: __mmask64, b: __mmask64) -> __mmask64 { |
| 10398 | a ^ b |
| 10399 | } |
| 10400 | |
| 10401 | /// Compute the bitwise XNOR of 32-bit masks a and b, and store the result in k. |
| 10402 | /// |
| 10403 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kxnor_mask32&expand=3286) |
| 10404 | #[inline ] |
| 10405 | #[target_feature (enable = "avx512bw" )] |
| 10406 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10407 | pub fn _kxnor_mask32(a: __mmask32, b: __mmask32) -> __mmask32 { |
| 10408 | _knot_mask32(a ^ b) |
| 10409 | } |
| 10410 | |
| 10411 | /// Compute the bitwise XNOR of 64-bit masks a and b, and store the result in k. |
| 10412 | /// |
| 10413 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kxnor_mask64&expand=3287) |
| 10414 | #[inline ] |
| 10415 | #[target_feature (enable = "avx512bw" )] |
| 10416 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10417 | pub fn _kxnor_mask64(a: __mmask64, b: __mmask64) -> __mmask64 { |
| 10418 | _knot_mask64(a ^ b) |
| 10419 | } |
| 10420 | |
| 10421 | /// Compute the bitwise OR of 32-bit masks a and b. If the result is all zeros, store 1 in dst, otherwise |
| 10422 | /// store 0 in dst. If the result is all ones, store 1 in all_ones, otherwise store 0 in all_ones. |
| 10423 | /// |
| 10424 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kortest_mask32_u8) |
| 10425 | #[inline ] |
| 10426 | #[target_feature (enable = "avx512bw" )] |
| 10427 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10428 | pub unsafe fn _kortest_mask32_u8(a: __mmask32, b: __mmask32, all_ones: *mut u8) -> u8 { |
| 10429 | let tmp: u32 = _kor_mask32(a, b); |
| 10430 | *all_ones = (tmp == 0xffffffff) as u8; |
| 10431 | (tmp == 0) as u8 |
| 10432 | } |
| 10433 | |
| 10434 | /// Compute the bitwise OR of 64-bit masks a and b. If the result is all zeros, store 1 in dst, otherwise |
| 10435 | /// store 0 in dst. If the result is all ones, store 1 in all_ones, otherwise store 0 in all_ones. |
| 10436 | /// |
| 10437 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kortest_mask64_u8) |
| 10438 | #[inline ] |
| 10439 | #[target_feature (enable = "avx512bw" )] |
| 10440 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10441 | pub unsafe fn _kortest_mask64_u8(a: __mmask64, b: __mmask64, all_ones: *mut u8) -> u8 { |
| 10442 | let tmp: u64 = _kor_mask64(a, b); |
| 10443 | *all_ones = (tmp == 0xffffffff_ffffffff) as u8; |
| 10444 | (tmp == 0) as u8 |
| 10445 | } |
| 10446 | |
| 10447 | /// Compute the bitwise OR of 32-bit masks a and b. If the result is all ones, store 1 in dst, otherwise |
| 10448 | /// store 0 in dst. |
| 10449 | /// |
| 10450 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kortestc_mask32_u8) |
| 10451 | #[inline ] |
| 10452 | #[target_feature (enable = "avx512bw" )] |
| 10453 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10454 | pub fn _kortestc_mask32_u8(a: __mmask32, b: __mmask32) -> u8 { |
| 10455 | (_kor_mask32(a, b) == 0xffffffff) as u8 |
| 10456 | } |
| 10457 | |
| 10458 | /// Compute the bitwise OR of 64-bit masks a and b. If the result is all ones, store 1 in dst, otherwise |
| 10459 | /// store 0 in dst. |
| 10460 | /// |
| 10461 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kortestc_mask64_u8) |
| 10462 | #[inline ] |
| 10463 | #[target_feature (enable = "avx512bw" )] |
| 10464 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10465 | pub fn _kortestc_mask64_u8(a: __mmask64, b: __mmask64) -> u8 { |
| 10466 | (_kor_mask64(a, b) == 0xffffffff_ffffffff) as u8 |
| 10467 | } |
| 10468 | |
| 10469 | /// Compute the bitwise OR of 32-bit masks a and b. If the result is all zeros, store 1 in dst, otherwise |
| 10470 | /// store 0 in dst. |
| 10471 | /// |
| 10472 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kortestz_mask32_u8) |
| 10473 | #[inline ] |
| 10474 | #[target_feature (enable = "avx512bw" )] |
| 10475 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10476 | pub fn _kortestz_mask32_u8(a: __mmask32, b: __mmask32) -> u8 { |
| 10477 | (_kor_mask32(a, b) == 0) as u8 |
| 10478 | } |
| 10479 | |
| 10480 | /// Compute the bitwise OR of 64-bit masks a and b. If the result is all zeros, store 1 in dst, otherwise |
| 10481 | /// store 0 in dst. |
| 10482 | /// |
| 10483 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kortestz_mask64_u8) |
| 10484 | #[inline ] |
| 10485 | #[target_feature (enable = "avx512bw" )] |
| 10486 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10487 | pub fn _kortestz_mask64_u8(a: __mmask64, b: __mmask64) -> u8 { |
| 10488 | (_kor_mask64(a, b) == 0) as u8 |
| 10489 | } |
| 10490 | |
| 10491 | /// Shift the bits of 32-bit mask a left by count while shifting in zeros, and store the least significant 32 bits of the result in k. |
| 10492 | /// |
| 10493 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kshiftli_mask32) |
| 10494 | #[inline ] |
| 10495 | #[target_feature (enable = "avx512bw" )] |
| 10496 | #[rustc_legacy_const_generics (1)] |
| 10497 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10498 | pub fn _kshiftli_mask32<const COUNT: u32>(a: __mmask32) -> __mmask32 { |
| 10499 | a.unbounded_shl(COUNT) |
| 10500 | } |
| 10501 | |
| 10502 | /// Shift the bits of 64-bit mask a left by count while shifting in zeros, and store the least significant 32 bits of the result in k. |
| 10503 | /// |
| 10504 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kshiftli_mask64) |
| 10505 | #[inline ] |
| 10506 | #[target_feature (enable = "avx512bw" )] |
| 10507 | #[rustc_legacy_const_generics (1)] |
| 10508 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10509 | pub fn _kshiftli_mask64<const COUNT: u32>(a: __mmask64) -> __mmask64 { |
| 10510 | a.unbounded_shl(COUNT) |
| 10511 | } |
| 10512 | |
| 10513 | /// Shift the bits of 32-bit mask a right by count while shifting in zeros, and store the least significant 32 bits of the result in k. |
| 10514 | /// |
| 10515 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kshiftri_mask32) |
| 10516 | #[inline ] |
| 10517 | #[target_feature (enable = "avx512bw" )] |
| 10518 | #[rustc_legacy_const_generics (1)] |
| 10519 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10520 | pub fn _kshiftri_mask32<const COUNT: u32>(a: __mmask32) -> __mmask32 { |
| 10521 | a.unbounded_shr(COUNT) |
| 10522 | } |
| 10523 | |
| 10524 | /// Shift the bits of 64-bit mask a right by count while shifting in zeros, and store the least significant 32 bits of the result in k. |
| 10525 | /// |
| 10526 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kshiftri_mask64) |
| 10527 | #[inline ] |
| 10528 | #[target_feature (enable = "avx512bw" )] |
| 10529 | #[rustc_legacy_const_generics (1)] |
| 10530 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10531 | pub fn _kshiftri_mask64<const COUNT: u32>(a: __mmask64) -> __mmask64 { |
| 10532 | a.unbounded_shr(COUNT) |
| 10533 | } |
| 10534 | |
| 10535 | /// Compute the bitwise AND of 32-bit masks a and b, and if the result is all zeros, store 1 in dst, |
| 10536 | /// otherwise store 0 in dst. Compute the bitwise NOT of a and then AND with b, if the result is all |
| 10537 | /// zeros, store 1 in and_not, otherwise store 0 in and_not. |
| 10538 | /// |
| 10539 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_ktest_mask32_u8) |
| 10540 | #[inline ] |
| 10541 | #[target_feature (enable = "avx512bw" )] |
| 10542 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10543 | pub unsafe fn _ktest_mask32_u8(a: __mmask32, b: __mmask32, and_not: *mut u8) -> u8 { |
| 10544 | *and_not = (_kandn_mask32(a, b) == 0) as u8; |
| 10545 | (_kand_mask32(a, b) == 0) as u8 |
| 10546 | } |
| 10547 | |
| 10548 | /// Compute the bitwise AND of 64-bit masks a and b, and if the result is all zeros, store 1 in dst, |
| 10549 | /// otherwise store 0 in dst. Compute the bitwise NOT of a and then AND with b, if the result is all |
| 10550 | /// zeros, store 1 in and_not, otherwise store 0 in and_not. |
| 10551 | /// |
| 10552 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_ktest_mask64_u8) |
| 10553 | #[inline ] |
| 10554 | #[target_feature (enable = "avx512bw" )] |
| 10555 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10556 | pub unsafe fn _ktest_mask64_u8(a: __mmask64, b: __mmask64, and_not: *mut u8) -> u8 { |
| 10557 | *and_not = (_kandn_mask64(a, b) == 0) as u8; |
| 10558 | (_kand_mask64(a, b) == 0) as u8 |
| 10559 | } |
| 10560 | |
| 10561 | /// Compute the bitwise NOT of 32-bit mask a and then AND with 16-bit mask b, if the result is all |
| 10562 | /// zeros, store 1 in dst, otherwise store 0 in dst. |
| 10563 | /// |
| 10564 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_ktestc_mask32_u8) |
| 10565 | #[inline ] |
| 10566 | #[target_feature (enable = "avx512bw" )] |
| 10567 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10568 | pub fn _ktestc_mask32_u8(a: __mmask32, b: __mmask32) -> u8 { |
| 10569 | (_kandn_mask32(a, b) == 0) as u8 |
| 10570 | } |
| 10571 | |
| 10572 | /// Compute the bitwise NOT of 64-bit mask a and then AND with 8-bit mask b, if the result is all |
| 10573 | /// zeros, store 1 in dst, otherwise store 0 in dst. |
| 10574 | /// |
| 10575 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_ktestc_mask64_u8) |
| 10576 | #[inline ] |
| 10577 | #[target_feature (enable = "avx512bw" )] |
| 10578 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10579 | pub fn _ktestc_mask64_u8(a: __mmask64, b: __mmask64) -> u8 { |
| 10580 | (_kandn_mask64(a, b) == 0) as u8 |
| 10581 | } |
| 10582 | |
| 10583 | /// Compute the bitwise AND of 32-bit masks a and b, if the result is all zeros, store 1 in dst, otherwise |
| 10584 | /// store 0 in dst. |
| 10585 | /// |
| 10586 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_ktestz_mask32_u8) |
| 10587 | #[inline ] |
| 10588 | #[target_feature (enable = "avx512bw" )] |
| 10589 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10590 | pub fn _ktestz_mask32_u8(a: __mmask32, b: __mmask32) -> u8 { |
| 10591 | (_kand_mask32(a, b) == 0) as u8 |
| 10592 | } |
| 10593 | |
| 10594 | /// Compute the bitwise AND of 64-bit masks a and b, if the result is all zeros, store 1 in dst, otherwise |
| 10595 | /// store 0 in dst. |
| 10596 | /// |
| 10597 | /// [Intel's Documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_ktestz_mask64_u8) |
| 10598 | #[inline ] |
| 10599 | #[target_feature (enable = "avx512bw" )] |
| 10600 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10601 | pub fn _ktestz_mask64_u8(a: __mmask64, b: __mmask64) -> u8 { |
| 10602 | (_kand_mask64(a, b) == 0) as u8 |
| 10603 | } |
| 10604 | |
| 10605 | /// Unpack and interleave 16 bits from masks a and b, and store the 32-bit result in k. |
| 10606 | /// |
| 10607 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_kunpackw) |
| 10608 | #[inline ] |
| 10609 | #[target_feature (enable = "avx512bw" )] |
| 10610 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10611 | #[cfg_attr (test, assert_instr(mov))] // generate normal and code instead of kunpckwd |
| 10612 | pub fn _mm512_kunpackw(a: __mmask32, b: __mmask32) -> __mmask32 { |
| 10613 | ((a & 0xffff) << 16) | (b & 0xffff) |
| 10614 | } |
| 10615 | |
| 10616 | /// Unpack and interleave 32 bits from masks a and b, and store the 64-bit result in k. |
| 10617 | /// |
| 10618 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_kunpackd) |
| 10619 | #[inline ] |
| 10620 | #[target_feature (enable = "avx512bw" )] |
| 10621 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10622 | #[cfg_attr (test, assert_instr(mov))] // generate normal and code instead of kunpckdq |
| 10623 | pub fn _mm512_kunpackd(a: __mmask64, b: __mmask64) -> __mmask64 { |
| 10624 | ((a & 0xffffffff) << 32) | (b & 0xffffffff) |
| 10625 | } |
| 10626 | |
| 10627 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst. |
| 10628 | /// |
| 10629 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cvtepi16_epi8&expand=1407) |
| 10630 | #[inline ] |
| 10631 | #[target_feature (enable = "avx512bw" )] |
| 10632 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10633 | #[cfg_attr (test, assert_instr(vpmovwb))] |
| 10634 | pub fn _mm512_cvtepi16_epi8(a: __m512i) -> __m256i { |
| 10635 | unsafe { |
| 10636 | let a: i16x32 = a.as_i16x32(); |
| 10637 | transmute::<i8x32, _>(src:simd_cast(a)) |
| 10638 | } |
| 10639 | } |
| 10640 | |
| 10641 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 10642 | /// |
| 10643 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtepi16_epi8&expand=1408) |
| 10644 | #[inline ] |
| 10645 | #[target_feature (enable = "avx512bw" )] |
| 10646 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10647 | #[cfg_attr (test, assert_instr(vpmovwb))] |
| 10648 | pub fn _mm512_mask_cvtepi16_epi8(src: __m256i, k: __mmask32, a: __m512i) -> __m256i { |
| 10649 | unsafe { |
| 10650 | let convert: i8x32 = _mm512_cvtepi16_epi8(a).as_i8x32(); |
| 10651 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i8x32())) |
| 10652 | } |
| 10653 | } |
| 10654 | |
| 10655 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 10656 | /// |
| 10657 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_cvtepi16_epi8&expand=1409) |
| 10658 | #[inline ] |
| 10659 | #[target_feature (enable = "avx512bw" )] |
| 10660 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10661 | #[cfg_attr (test, assert_instr(vpmovwb))] |
| 10662 | pub fn _mm512_maskz_cvtepi16_epi8(k: __mmask32, a: __m512i) -> __m256i { |
| 10663 | unsafe { |
| 10664 | let convert: i8x32 = _mm512_cvtepi16_epi8(a).as_i8x32(); |
| 10665 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:i8x32::ZERO)) |
| 10666 | } |
| 10667 | } |
| 10668 | |
| 10669 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst. |
| 10670 | /// |
| 10671 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cvtepi16_epi8&expand=1404) |
| 10672 | #[inline ] |
| 10673 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10674 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10675 | #[cfg_attr (test, assert_instr(vpmovwb))] |
| 10676 | pub fn _mm256_cvtepi16_epi8(a: __m256i) -> __m128i { |
| 10677 | unsafe { |
| 10678 | let a: i16x16 = a.as_i16x16(); |
| 10679 | transmute::<i8x16, _>(src:simd_cast(a)) |
| 10680 | } |
| 10681 | } |
| 10682 | |
| 10683 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 10684 | /// |
| 10685 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtepi16_epi8&expand=1405) |
| 10686 | #[inline ] |
| 10687 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10688 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10689 | #[cfg_attr (test, assert_instr(vpmovwb))] |
| 10690 | pub fn _mm256_mask_cvtepi16_epi8(src: __m128i, k: __mmask16, a: __m256i) -> __m128i { |
| 10691 | unsafe { |
| 10692 | let convert: i8x16 = _mm256_cvtepi16_epi8(a).as_i8x16(); |
| 10693 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i8x16())) |
| 10694 | } |
| 10695 | } |
| 10696 | |
| 10697 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 10698 | /// |
| 10699 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_cvtepi16_epi8&expand=1406) |
| 10700 | #[inline ] |
| 10701 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10702 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10703 | #[cfg_attr (test, assert_instr(vpmovwb))] |
| 10704 | pub fn _mm256_maskz_cvtepi16_epi8(k: __mmask16, a: __m256i) -> __m128i { |
| 10705 | unsafe { |
| 10706 | let convert: i8x16 = _mm256_cvtepi16_epi8(a).as_i8x16(); |
| 10707 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:i8x16::ZERO)) |
| 10708 | } |
| 10709 | } |
| 10710 | |
| 10711 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst. |
| 10712 | /// |
| 10713 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtepi16_epi8&expand=1401) |
| 10714 | #[inline ] |
| 10715 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10716 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10717 | #[cfg_attr (test, assert_instr(vpmovwb))] |
| 10718 | pub fn _mm_cvtepi16_epi8(a: __m128i) -> __m128i { |
| 10719 | unsafe { |
| 10720 | let a: i16x8 = a.as_i16x8(); |
| 10721 | let v256: i16x16 = simd_shuffle!( |
| 10722 | a, |
| 10723 | i16x8::ZERO, |
| 10724 | [0, 1, 2, 3, 4, 5, 6, 7, 8, 8, 8, 8, 8, 8, 8, 8] |
| 10725 | ); |
| 10726 | transmute::<i8x16, _>(src:simd_cast(v256)) |
| 10727 | } |
| 10728 | } |
| 10729 | |
| 10730 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 10731 | /// |
| 10732 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtepi16_epi8&expand=1402) |
| 10733 | #[inline ] |
| 10734 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10735 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10736 | #[cfg_attr (test, assert_instr(vpmovwb))] |
| 10737 | pub fn _mm_mask_cvtepi16_epi8(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 10738 | unsafe { |
| 10739 | let convert: i8x16 = _mm_cvtepi16_epi8(a).as_i8x16(); |
| 10740 | let k: __mmask16 = 0b11111111_11111111 & k as __mmask16; |
| 10741 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i8x16())) |
| 10742 | } |
| 10743 | } |
| 10744 | |
| 10745 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 10746 | /// |
| 10747 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_cvtepi16_epi8&expand=1403) |
| 10748 | #[inline ] |
| 10749 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10750 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10751 | #[cfg_attr (test, assert_instr(vpmovwb))] |
| 10752 | pub fn _mm_maskz_cvtepi16_epi8(k: __mmask8, a: __m128i) -> __m128i { |
| 10753 | unsafe { |
| 10754 | let convert: i8x16 = _mm_cvtepi16_epi8(a).as_i8x16(); |
| 10755 | let k: __mmask16 = 0b11111111_11111111 & k as __mmask16; |
| 10756 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:i8x16::ZERO)) |
| 10757 | } |
| 10758 | } |
| 10759 | |
| 10760 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst. |
| 10761 | /// |
| 10762 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cvtsepi16_epi8&expand=1807) |
| 10763 | #[inline ] |
| 10764 | #[target_feature (enable = "avx512bw" )] |
| 10765 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10766 | #[cfg_attr (test, assert_instr(vpmovswb))] |
| 10767 | pub fn _mm512_cvtsepi16_epi8(a: __m512i) -> __m256i { |
| 10768 | unsafe { |
| 10769 | transmute(src:vpmovswb( |
| 10770 | a.as_i16x32(), |
| 10771 | src:i8x32::ZERO, |
| 10772 | mask:0b11111111_11111111_11111111_11111111, |
| 10773 | )) |
| 10774 | } |
| 10775 | } |
| 10776 | |
| 10777 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 10778 | /// |
| 10779 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtsepi16_epi8&expand=1808) |
| 10780 | #[inline ] |
| 10781 | #[target_feature (enable = "avx512bw" )] |
| 10782 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10783 | #[cfg_attr (test, assert_instr(vpmovswb))] |
| 10784 | pub fn _mm512_mask_cvtsepi16_epi8(src: __m256i, k: __mmask32, a: __m512i) -> __m256i { |
| 10785 | unsafe { transmute(src:vpmovswb(a.as_i16x32(), src.as_i8x32(), mask:k)) } |
| 10786 | } |
| 10787 | |
| 10788 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 10789 | /// |
| 10790 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_cvtsepi16_epi8&expand=1809) |
| 10791 | #[inline ] |
| 10792 | #[target_feature (enable = "avx512bw" )] |
| 10793 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10794 | #[cfg_attr (test, assert_instr(vpmovswb))] |
| 10795 | pub fn _mm512_maskz_cvtsepi16_epi8(k: __mmask32, a: __m512i) -> __m256i { |
| 10796 | unsafe { transmute(src:vpmovswb(a.as_i16x32(), src:i8x32::ZERO, mask:k)) } |
| 10797 | } |
| 10798 | |
| 10799 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst. |
| 10800 | /// |
| 10801 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cvtsepi16_epi8&expand=1804) |
| 10802 | #[inline ] |
| 10803 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10804 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10805 | #[cfg_attr (test, assert_instr(vpmovswb))] |
| 10806 | pub fn _mm256_cvtsepi16_epi8(a: __m256i) -> __m128i { |
| 10807 | unsafe { transmute(src:vpmovswb256(a.as_i16x16(), src:i8x16::ZERO, mask:0b11111111_11111111)) } |
| 10808 | } |
| 10809 | |
| 10810 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 10811 | /// |
| 10812 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtsepi16_epi8&expand=1805) |
| 10813 | #[inline ] |
| 10814 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10815 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10816 | #[cfg_attr (test, assert_instr(vpmovswb))] |
| 10817 | pub fn _mm256_mask_cvtsepi16_epi8(src: __m128i, k: __mmask16, a: __m256i) -> __m128i { |
| 10818 | unsafe { transmute(src:vpmovswb256(a.as_i16x16(), src.as_i8x16(), mask:k)) } |
| 10819 | } |
| 10820 | |
| 10821 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 10822 | /// |
| 10823 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_cvtsepi16_epi8&expand=1806) |
| 10824 | #[inline ] |
| 10825 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10826 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10827 | #[cfg_attr (test, assert_instr(vpmovswb))] |
| 10828 | pub fn _mm256_maskz_cvtsepi16_epi8(k: __mmask16, a: __m256i) -> __m128i { |
| 10829 | unsafe { transmute(src:vpmovswb256(a.as_i16x16(), src:i8x16::ZERO, mask:k)) } |
| 10830 | } |
| 10831 | |
| 10832 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst. |
| 10833 | /// |
| 10834 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtsepi16_epi8&expand=1801) |
| 10835 | #[inline ] |
| 10836 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10837 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10838 | #[cfg_attr (test, assert_instr(vpmovswb))] |
| 10839 | pub fn _mm_cvtsepi16_epi8(a: __m128i) -> __m128i { |
| 10840 | unsafe { transmute(src:vpmovswb128(a.as_i16x8(), src:i8x16::ZERO, mask:0b11111111)) } |
| 10841 | } |
| 10842 | |
| 10843 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 10844 | /// |
| 10845 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtsepi16_epi8&expand=1802) |
| 10846 | #[inline ] |
| 10847 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10848 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10849 | #[cfg_attr (test, assert_instr(vpmovswb))] |
| 10850 | pub fn _mm_mask_cvtsepi16_epi8(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 10851 | unsafe { transmute(src:vpmovswb128(a.as_i16x8(), src.as_i8x16(), mask:k)) } |
| 10852 | } |
| 10853 | |
| 10854 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 10855 | /// |
| 10856 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_cvtsepi16_epi8&expand=1803) |
| 10857 | #[inline ] |
| 10858 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10859 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10860 | #[cfg_attr (test, assert_instr(vpmovswb))] |
| 10861 | pub fn _mm_maskz_cvtsepi16_epi8(k: __mmask8, a: __m128i) -> __m128i { |
| 10862 | unsafe { transmute(src:vpmovswb128(a.as_i16x8(), src:i8x16::ZERO, mask:k)) } |
| 10863 | } |
| 10864 | |
| 10865 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst. |
| 10866 | /// |
| 10867 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cvtusepi16_epi8&expand=2042) |
| 10868 | #[inline ] |
| 10869 | #[target_feature (enable = "avx512bw" )] |
| 10870 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10871 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
| 10872 | pub fn _mm512_cvtusepi16_epi8(a: __m512i) -> __m256i { |
| 10873 | unsafe { |
| 10874 | transmute(src:vpmovuswb( |
| 10875 | a.as_u16x32(), |
| 10876 | src:u8x32::ZERO, |
| 10877 | mask:0b11111111_11111111_11111111_11111111, |
| 10878 | )) |
| 10879 | } |
| 10880 | } |
| 10881 | |
| 10882 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 10883 | /// |
| 10884 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtusepi16_epi8&expand=2043) |
| 10885 | #[inline ] |
| 10886 | #[target_feature (enable = "avx512bw" )] |
| 10887 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10888 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
| 10889 | pub fn _mm512_mask_cvtusepi16_epi8(src: __m256i, k: __mmask32, a: __m512i) -> __m256i { |
| 10890 | unsafe { transmute(src:vpmovuswb(a.as_u16x32(), src.as_u8x32(), mask:k)) } |
| 10891 | } |
| 10892 | |
| 10893 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 10894 | /// |
| 10895 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_cvtusepi16_epi8&expand=2044) |
| 10896 | #[inline ] |
| 10897 | #[target_feature (enable = "avx512bw" )] |
| 10898 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10899 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
| 10900 | pub fn _mm512_maskz_cvtusepi16_epi8(k: __mmask32, a: __m512i) -> __m256i { |
| 10901 | unsafe { transmute(src:vpmovuswb(a.as_u16x32(), src:u8x32::ZERO, mask:k)) } |
| 10902 | } |
| 10903 | |
| 10904 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst. |
| 10905 | /// |
| 10906 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cvtusepi16_epi8&expand=2039) |
| 10907 | #[inline ] |
| 10908 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10909 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10910 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
| 10911 | pub fn _mm256_cvtusepi16_epi8(a: __m256i) -> __m128i { |
| 10912 | unsafe { |
| 10913 | transmute(src:vpmovuswb256( |
| 10914 | a.as_u16x16(), |
| 10915 | src:u8x16::ZERO, |
| 10916 | mask:0b11111111_11111111, |
| 10917 | )) |
| 10918 | } |
| 10919 | } |
| 10920 | |
| 10921 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 10922 | /// |
| 10923 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtusepi16_epi8&expand=2040) |
| 10924 | #[inline ] |
| 10925 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10926 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10927 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
| 10928 | pub fn _mm256_mask_cvtusepi16_epi8(src: __m128i, k: __mmask16, a: __m256i) -> __m128i { |
| 10929 | unsafe { transmute(src:vpmovuswb256(a.as_u16x16(), src.as_u8x16(), mask:k)) } |
| 10930 | } |
| 10931 | |
| 10932 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 10933 | /// |
| 10934 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_cvtusepi16_epi8&expand=2041) |
| 10935 | #[inline ] |
| 10936 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10937 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10938 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
| 10939 | pub fn _mm256_maskz_cvtusepi16_epi8(k: __mmask16, a: __m256i) -> __m128i { |
| 10940 | unsafe { transmute(src:vpmovuswb256(a.as_u16x16(), src:u8x16::ZERO, mask:k)) } |
| 10941 | } |
| 10942 | |
| 10943 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst. |
| 10944 | /// |
| 10945 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtusepi16_epi8&expand=2036) |
| 10946 | #[inline ] |
| 10947 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10948 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10949 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
| 10950 | pub fn _mm_cvtusepi16_epi8(a: __m128i) -> __m128i { |
| 10951 | unsafe { transmute(src:vpmovuswb128(a.as_u16x8(), src:u8x16::ZERO, mask:0b11111111)) } |
| 10952 | } |
| 10953 | |
| 10954 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 10955 | /// |
| 10956 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtusepi16_epi8&expand=2037) |
| 10957 | #[inline ] |
| 10958 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10959 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10960 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
| 10961 | pub fn _mm_mask_cvtusepi16_epi8(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 10962 | unsafe { transmute(src:vpmovuswb128(a.as_u16x8(), src.as_u8x16(), mask:k)) } |
| 10963 | } |
| 10964 | |
| 10965 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 10966 | /// |
| 10967 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_cvtusepi16_epi8&expand=2038) |
| 10968 | #[inline ] |
| 10969 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 10970 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10971 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
| 10972 | pub fn _mm_maskz_cvtusepi16_epi8(k: __mmask8, a: __m128i) -> __m128i { |
| 10973 | unsafe { transmute(src:vpmovuswb128(a.as_u16x8(), src:u8x16::ZERO, mask:k)) } |
| 10974 | } |
| 10975 | |
| 10976 | /// Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst. |
| 10977 | /// |
| 10978 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cvtepi8_epi16&expand=1526) |
| 10979 | #[inline ] |
| 10980 | #[target_feature (enable = "avx512bw" )] |
| 10981 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10982 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
| 10983 | pub fn _mm512_cvtepi8_epi16(a: __m256i) -> __m512i { |
| 10984 | unsafe { |
| 10985 | let a: i8x32 = a.as_i8x32(); |
| 10986 | transmute::<i16x32, _>(src:simd_cast(a)) |
| 10987 | } |
| 10988 | } |
| 10989 | |
| 10990 | /// Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 10991 | /// |
| 10992 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtepi8_epi16&expand=1527) |
| 10993 | #[inline ] |
| 10994 | #[target_feature (enable = "avx512bw" )] |
| 10995 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 10996 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
| 10997 | pub fn _mm512_mask_cvtepi8_epi16(src: __m512i, k: __mmask32, a: __m256i) -> __m512i { |
| 10998 | unsafe { |
| 10999 | let convert: i16x32 = _mm512_cvtepi8_epi16(a).as_i16x32(); |
| 11000 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i16x32())) |
| 11001 | } |
| 11002 | } |
| 11003 | |
| 11004 | /// Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 11005 | /// |
| 11006 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_cvtepi8_epi16&expand=1528) |
| 11007 | #[inline ] |
| 11008 | #[target_feature (enable = "avx512bw" )] |
| 11009 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11010 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
| 11011 | pub fn _mm512_maskz_cvtepi8_epi16(k: __mmask32, a: __m256i) -> __m512i { |
| 11012 | unsafe { |
| 11013 | let convert: i16x32 = _mm512_cvtepi8_epi16(a).as_i16x32(); |
| 11014 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:i16x32::ZERO)) |
| 11015 | } |
| 11016 | } |
| 11017 | |
| 11018 | /// Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 11019 | /// |
| 11020 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtepi8_epi16&expand=1524) |
| 11021 | #[inline ] |
| 11022 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11023 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11024 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
| 11025 | pub fn _mm256_mask_cvtepi8_epi16(src: __m256i, k: __mmask16, a: __m128i) -> __m256i { |
| 11026 | unsafe { |
| 11027 | let convert: i16x16 = _mm256_cvtepi8_epi16(a).as_i16x16(); |
| 11028 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i16x16())) |
| 11029 | } |
| 11030 | } |
| 11031 | |
| 11032 | /// Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 11033 | /// |
| 11034 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_cvtepi8_epi16&expand=1525) |
| 11035 | #[inline ] |
| 11036 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11037 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11038 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
| 11039 | pub fn _mm256_maskz_cvtepi8_epi16(k: __mmask16, a: __m128i) -> __m256i { |
| 11040 | unsafe { |
| 11041 | let convert: i16x16 = _mm256_cvtepi8_epi16(a).as_i16x16(); |
| 11042 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:i16x16::ZERO)) |
| 11043 | } |
| 11044 | } |
| 11045 | |
| 11046 | /// Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 11047 | /// |
| 11048 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtepi8_epi16&expand=1521) |
| 11049 | #[inline ] |
| 11050 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11051 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11052 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
| 11053 | pub fn _mm_mask_cvtepi8_epi16(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 11054 | unsafe { |
| 11055 | let convert: i16x8 = _mm_cvtepi8_epi16(a).as_i16x8(); |
| 11056 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i16x8())) |
| 11057 | } |
| 11058 | } |
| 11059 | |
| 11060 | /// Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 11061 | /// |
| 11062 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_cvtepi8_epi16&expand=1522) |
| 11063 | #[inline ] |
| 11064 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11065 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11066 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
| 11067 | pub fn _mm_maskz_cvtepi8_epi16(k: __mmask8, a: __m128i) -> __m128i { |
| 11068 | unsafe { |
| 11069 | let convert: i16x8 = _mm_cvtepi8_epi16(a).as_i16x8(); |
| 11070 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:i16x8::ZERO)) |
| 11071 | } |
| 11072 | } |
| 11073 | |
| 11074 | /// Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst. |
| 11075 | /// |
| 11076 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cvtepu8_epi16&expand=1612) |
| 11077 | #[inline ] |
| 11078 | #[target_feature (enable = "avx512bw" )] |
| 11079 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11080 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
| 11081 | pub fn _mm512_cvtepu8_epi16(a: __m256i) -> __m512i { |
| 11082 | unsafe { |
| 11083 | let a: u8x32 = a.as_u8x32(); |
| 11084 | transmute::<i16x32, _>(src:simd_cast(a)) |
| 11085 | } |
| 11086 | } |
| 11087 | |
| 11088 | /// Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 11089 | /// |
| 11090 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtepu8_epi16&expand=1613) |
| 11091 | #[inline ] |
| 11092 | #[target_feature (enable = "avx512bw" )] |
| 11093 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11094 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
| 11095 | pub fn _mm512_mask_cvtepu8_epi16(src: __m512i, k: __mmask32, a: __m256i) -> __m512i { |
| 11096 | unsafe { |
| 11097 | let convert: i16x32 = _mm512_cvtepu8_epi16(a).as_i16x32(); |
| 11098 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i16x32())) |
| 11099 | } |
| 11100 | } |
| 11101 | |
| 11102 | /// Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 11103 | /// |
| 11104 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_cvtepu8_epi16&expand=1614) |
| 11105 | #[inline ] |
| 11106 | #[target_feature (enable = "avx512bw" )] |
| 11107 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11108 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
| 11109 | pub fn _mm512_maskz_cvtepu8_epi16(k: __mmask32, a: __m256i) -> __m512i { |
| 11110 | unsafe { |
| 11111 | let convert: i16x32 = _mm512_cvtepu8_epi16(a).as_i16x32(); |
| 11112 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:i16x32::ZERO)) |
| 11113 | } |
| 11114 | } |
| 11115 | |
| 11116 | /// Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 11117 | /// |
| 11118 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtepu8_epi16&expand=1610) |
| 11119 | #[inline ] |
| 11120 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11121 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11122 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
| 11123 | pub fn _mm256_mask_cvtepu8_epi16(src: __m256i, k: __mmask16, a: __m128i) -> __m256i { |
| 11124 | unsafe { |
| 11125 | let convert: i16x16 = _mm256_cvtepu8_epi16(a).as_i16x16(); |
| 11126 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i16x16())) |
| 11127 | } |
| 11128 | } |
| 11129 | |
| 11130 | /// Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 11131 | /// |
| 11132 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_cvtepu8_epi16&expand=1611) |
| 11133 | #[inline ] |
| 11134 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11135 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11136 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
| 11137 | pub fn _mm256_maskz_cvtepu8_epi16(k: __mmask16, a: __m128i) -> __m256i { |
| 11138 | unsafe { |
| 11139 | let convert: i16x16 = _mm256_cvtepu8_epi16(a).as_i16x16(); |
| 11140 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:i16x16::ZERO)) |
| 11141 | } |
| 11142 | } |
| 11143 | |
| 11144 | /// Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 11145 | /// |
| 11146 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtepu8_epi16&expand=1607) |
| 11147 | #[inline ] |
| 11148 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11149 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11150 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
| 11151 | pub fn _mm_mask_cvtepu8_epi16(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
| 11152 | unsafe { |
| 11153 | let convert: i16x8 = _mm_cvtepu8_epi16(a).as_i16x8(); |
| 11154 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i16x8())) |
| 11155 | } |
| 11156 | } |
| 11157 | |
| 11158 | /// Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 11159 | /// |
| 11160 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_cvtepu8_epi16&expand=1608) |
| 11161 | #[inline ] |
| 11162 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11163 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11164 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
| 11165 | pub fn _mm_maskz_cvtepu8_epi16(k: __mmask8, a: __m128i) -> __m128i { |
| 11166 | unsafe { |
| 11167 | let convert: i16x8 = _mm_cvtepu8_epi16(a).as_i16x8(); |
| 11168 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:i16x8::ZERO)) |
| 11169 | } |
| 11170 | } |
| 11171 | |
| 11172 | /// Shift 128-bit lanes in a left by imm8 bytes while shifting in zeros, and store the results in dst. |
| 11173 | /// |
| 11174 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_bslli_epi128&expand=591) |
| 11175 | #[inline ] |
| 11176 | #[target_feature (enable = "avx512bw" )] |
| 11177 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11178 | #[cfg_attr (test, assert_instr(vpslldq, IMM8 = 3))] |
| 11179 | #[rustc_legacy_const_generics (1)] |
| 11180 | pub fn _mm512_bslli_epi128<const IMM8: i32>(a: __m512i) -> __m512i { |
| 11181 | unsafe { |
| 11182 | static_assert_uimm_bits!(IMM8, 8); |
| 11183 | const fn mask(shift: i32, i: u32) -> u32 { |
| 11184 | let shift = shift as u32 & 0xff; |
| 11185 | if shift > 15 || i % 16 < shift { |
| 11186 | 0 |
| 11187 | } else { |
| 11188 | 64 + (i - shift) |
| 11189 | } |
| 11190 | } |
| 11191 | let a = a.as_i8x64(); |
| 11192 | let zero = i8x64::ZERO; |
| 11193 | let r: i8x64 = simd_shuffle!( |
| 11194 | zero, |
| 11195 | a, |
| 11196 | [ |
| 11197 | mask(IMM8, 0), |
| 11198 | mask(IMM8, 1), |
| 11199 | mask(IMM8, 2), |
| 11200 | mask(IMM8, 3), |
| 11201 | mask(IMM8, 4), |
| 11202 | mask(IMM8, 5), |
| 11203 | mask(IMM8, 6), |
| 11204 | mask(IMM8, 7), |
| 11205 | mask(IMM8, 8), |
| 11206 | mask(IMM8, 9), |
| 11207 | mask(IMM8, 10), |
| 11208 | mask(IMM8, 11), |
| 11209 | mask(IMM8, 12), |
| 11210 | mask(IMM8, 13), |
| 11211 | mask(IMM8, 14), |
| 11212 | mask(IMM8, 15), |
| 11213 | mask(IMM8, 16), |
| 11214 | mask(IMM8, 17), |
| 11215 | mask(IMM8, 18), |
| 11216 | mask(IMM8, 19), |
| 11217 | mask(IMM8, 20), |
| 11218 | mask(IMM8, 21), |
| 11219 | mask(IMM8, 22), |
| 11220 | mask(IMM8, 23), |
| 11221 | mask(IMM8, 24), |
| 11222 | mask(IMM8, 25), |
| 11223 | mask(IMM8, 26), |
| 11224 | mask(IMM8, 27), |
| 11225 | mask(IMM8, 28), |
| 11226 | mask(IMM8, 29), |
| 11227 | mask(IMM8, 30), |
| 11228 | mask(IMM8, 31), |
| 11229 | mask(IMM8, 32), |
| 11230 | mask(IMM8, 33), |
| 11231 | mask(IMM8, 34), |
| 11232 | mask(IMM8, 35), |
| 11233 | mask(IMM8, 36), |
| 11234 | mask(IMM8, 37), |
| 11235 | mask(IMM8, 38), |
| 11236 | mask(IMM8, 39), |
| 11237 | mask(IMM8, 40), |
| 11238 | mask(IMM8, 41), |
| 11239 | mask(IMM8, 42), |
| 11240 | mask(IMM8, 43), |
| 11241 | mask(IMM8, 44), |
| 11242 | mask(IMM8, 45), |
| 11243 | mask(IMM8, 46), |
| 11244 | mask(IMM8, 47), |
| 11245 | mask(IMM8, 48), |
| 11246 | mask(IMM8, 49), |
| 11247 | mask(IMM8, 50), |
| 11248 | mask(IMM8, 51), |
| 11249 | mask(IMM8, 52), |
| 11250 | mask(IMM8, 53), |
| 11251 | mask(IMM8, 54), |
| 11252 | mask(IMM8, 55), |
| 11253 | mask(IMM8, 56), |
| 11254 | mask(IMM8, 57), |
| 11255 | mask(IMM8, 58), |
| 11256 | mask(IMM8, 59), |
| 11257 | mask(IMM8, 60), |
| 11258 | mask(IMM8, 61), |
| 11259 | mask(IMM8, 62), |
| 11260 | mask(IMM8, 63), |
| 11261 | ], |
| 11262 | ); |
| 11263 | transmute(r) |
| 11264 | } |
| 11265 | } |
| 11266 | |
| 11267 | /// Shift 128-bit lanes in a right by imm8 bytes while shifting in zeros, and store the results in dst. |
| 11268 | /// |
| 11269 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_bsrli_epi128&expand=594) |
| 11270 | #[inline ] |
| 11271 | #[target_feature (enable = "avx512bw" )] |
| 11272 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11273 | #[cfg_attr (test, assert_instr(vpsrldq, IMM8 = 3))] |
| 11274 | #[rustc_legacy_const_generics (1)] |
| 11275 | pub fn _mm512_bsrli_epi128<const IMM8: i32>(a: __m512i) -> __m512i { |
| 11276 | unsafe { |
| 11277 | static_assert_uimm_bits!(IMM8, 8); |
| 11278 | const fn mask(shift: i32, i: u32) -> u32 { |
| 11279 | let shift = shift as u32 & 0xff; |
| 11280 | if shift > 15 || (15 - (i % 16)) < shift { |
| 11281 | 0 |
| 11282 | } else { |
| 11283 | 64 + (i + shift) |
| 11284 | } |
| 11285 | } |
| 11286 | let a = a.as_i8x64(); |
| 11287 | let zero = i8x64::ZERO; |
| 11288 | let r: i8x64 = simd_shuffle!( |
| 11289 | zero, |
| 11290 | a, |
| 11291 | [ |
| 11292 | mask(IMM8, 0), |
| 11293 | mask(IMM8, 1), |
| 11294 | mask(IMM8, 2), |
| 11295 | mask(IMM8, 3), |
| 11296 | mask(IMM8, 4), |
| 11297 | mask(IMM8, 5), |
| 11298 | mask(IMM8, 6), |
| 11299 | mask(IMM8, 7), |
| 11300 | mask(IMM8, 8), |
| 11301 | mask(IMM8, 9), |
| 11302 | mask(IMM8, 10), |
| 11303 | mask(IMM8, 11), |
| 11304 | mask(IMM8, 12), |
| 11305 | mask(IMM8, 13), |
| 11306 | mask(IMM8, 14), |
| 11307 | mask(IMM8, 15), |
| 11308 | mask(IMM8, 16), |
| 11309 | mask(IMM8, 17), |
| 11310 | mask(IMM8, 18), |
| 11311 | mask(IMM8, 19), |
| 11312 | mask(IMM8, 20), |
| 11313 | mask(IMM8, 21), |
| 11314 | mask(IMM8, 22), |
| 11315 | mask(IMM8, 23), |
| 11316 | mask(IMM8, 24), |
| 11317 | mask(IMM8, 25), |
| 11318 | mask(IMM8, 26), |
| 11319 | mask(IMM8, 27), |
| 11320 | mask(IMM8, 28), |
| 11321 | mask(IMM8, 29), |
| 11322 | mask(IMM8, 30), |
| 11323 | mask(IMM8, 31), |
| 11324 | mask(IMM8, 32), |
| 11325 | mask(IMM8, 33), |
| 11326 | mask(IMM8, 34), |
| 11327 | mask(IMM8, 35), |
| 11328 | mask(IMM8, 36), |
| 11329 | mask(IMM8, 37), |
| 11330 | mask(IMM8, 38), |
| 11331 | mask(IMM8, 39), |
| 11332 | mask(IMM8, 40), |
| 11333 | mask(IMM8, 41), |
| 11334 | mask(IMM8, 42), |
| 11335 | mask(IMM8, 43), |
| 11336 | mask(IMM8, 44), |
| 11337 | mask(IMM8, 45), |
| 11338 | mask(IMM8, 46), |
| 11339 | mask(IMM8, 47), |
| 11340 | mask(IMM8, 48), |
| 11341 | mask(IMM8, 49), |
| 11342 | mask(IMM8, 50), |
| 11343 | mask(IMM8, 51), |
| 11344 | mask(IMM8, 52), |
| 11345 | mask(IMM8, 53), |
| 11346 | mask(IMM8, 54), |
| 11347 | mask(IMM8, 55), |
| 11348 | mask(IMM8, 56), |
| 11349 | mask(IMM8, 57), |
| 11350 | mask(IMM8, 58), |
| 11351 | mask(IMM8, 59), |
| 11352 | mask(IMM8, 60), |
| 11353 | mask(IMM8, 61), |
| 11354 | mask(IMM8, 62), |
| 11355 | mask(IMM8, 63), |
| 11356 | ], |
| 11357 | ); |
| 11358 | transmute(r) |
| 11359 | } |
| 11360 | } |
| 11361 | |
| 11362 | /// Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst. |
| 11363 | /// Unlike [`_mm_alignr_epi8`], [`_mm256_alignr_epi8`] functions, where the entire input vectors are concatenated to the temporary result, |
| 11364 | /// this concatenation happens in 4 steps, where each step builds 32-byte temporary result. |
| 11365 | /// |
| 11366 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_alignr_epi8&expand=263) |
| 11367 | #[inline ] |
| 11368 | #[target_feature (enable = "avx512bw" )] |
| 11369 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11370 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 1))] |
| 11371 | #[rustc_legacy_const_generics (2)] |
| 11372 | pub fn _mm512_alignr_epi8<const IMM8: i32>(a: __m512i, b: __m512i) -> __m512i { |
| 11373 | const fn mask(shift: u32, i: u32) -> u32 { |
| 11374 | let shift = shift % 16; |
| 11375 | let mod_i = i % 16; |
| 11376 | if mod_i < (16 - shift) { |
| 11377 | i + shift |
| 11378 | } else { |
| 11379 | i + 48 + shift |
| 11380 | } |
| 11381 | } |
| 11382 | |
| 11383 | // If palignr is shifting the pair of vectors more than the size of two |
| 11384 | // lanes, emit zero. |
| 11385 | if IMM8 >= 32 { |
| 11386 | return _mm512_setzero_si512(); |
| 11387 | } |
| 11388 | // If palignr is shifting the pair of input vectors more than one lane, |
| 11389 | // but less than two lanes, convert to shifting in zeroes. |
| 11390 | let (a, b) = if IMM8 > 16 { |
| 11391 | (_mm512_setzero_si512(), a) |
| 11392 | } else { |
| 11393 | (a, b) |
| 11394 | }; |
| 11395 | unsafe { |
| 11396 | if IMM8 == 16 { |
| 11397 | return transmute(a); |
| 11398 | } |
| 11399 | |
| 11400 | let r: i8x64 = simd_shuffle!( |
| 11401 | b.as_i8x64(), |
| 11402 | a.as_i8x64(), |
| 11403 | [ |
| 11404 | mask(IMM8 as u32, 0), |
| 11405 | mask(IMM8 as u32, 1), |
| 11406 | mask(IMM8 as u32, 2), |
| 11407 | mask(IMM8 as u32, 3), |
| 11408 | mask(IMM8 as u32, 4), |
| 11409 | mask(IMM8 as u32, 5), |
| 11410 | mask(IMM8 as u32, 6), |
| 11411 | mask(IMM8 as u32, 7), |
| 11412 | mask(IMM8 as u32, 8), |
| 11413 | mask(IMM8 as u32, 9), |
| 11414 | mask(IMM8 as u32, 10), |
| 11415 | mask(IMM8 as u32, 11), |
| 11416 | mask(IMM8 as u32, 12), |
| 11417 | mask(IMM8 as u32, 13), |
| 11418 | mask(IMM8 as u32, 14), |
| 11419 | mask(IMM8 as u32, 15), |
| 11420 | mask(IMM8 as u32, 16), |
| 11421 | mask(IMM8 as u32, 17), |
| 11422 | mask(IMM8 as u32, 18), |
| 11423 | mask(IMM8 as u32, 19), |
| 11424 | mask(IMM8 as u32, 20), |
| 11425 | mask(IMM8 as u32, 21), |
| 11426 | mask(IMM8 as u32, 22), |
| 11427 | mask(IMM8 as u32, 23), |
| 11428 | mask(IMM8 as u32, 24), |
| 11429 | mask(IMM8 as u32, 25), |
| 11430 | mask(IMM8 as u32, 26), |
| 11431 | mask(IMM8 as u32, 27), |
| 11432 | mask(IMM8 as u32, 28), |
| 11433 | mask(IMM8 as u32, 29), |
| 11434 | mask(IMM8 as u32, 30), |
| 11435 | mask(IMM8 as u32, 31), |
| 11436 | mask(IMM8 as u32, 32), |
| 11437 | mask(IMM8 as u32, 33), |
| 11438 | mask(IMM8 as u32, 34), |
| 11439 | mask(IMM8 as u32, 35), |
| 11440 | mask(IMM8 as u32, 36), |
| 11441 | mask(IMM8 as u32, 37), |
| 11442 | mask(IMM8 as u32, 38), |
| 11443 | mask(IMM8 as u32, 39), |
| 11444 | mask(IMM8 as u32, 40), |
| 11445 | mask(IMM8 as u32, 41), |
| 11446 | mask(IMM8 as u32, 42), |
| 11447 | mask(IMM8 as u32, 43), |
| 11448 | mask(IMM8 as u32, 44), |
| 11449 | mask(IMM8 as u32, 45), |
| 11450 | mask(IMM8 as u32, 46), |
| 11451 | mask(IMM8 as u32, 47), |
| 11452 | mask(IMM8 as u32, 48), |
| 11453 | mask(IMM8 as u32, 49), |
| 11454 | mask(IMM8 as u32, 50), |
| 11455 | mask(IMM8 as u32, 51), |
| 11456 | mask(IMM8 as u32, 52), |
| 11457 | mask(IMM8 as u32, 53), |
| 11458 | mask(IMM8 as u32, 54), |
| 11459 | mask(IMM8 as u32, 55), |
| 11460 | mask(IMM8 as u32, 56), |
| 11461 | mask(IMM8 as u32, 57), |
| 11462 | mask(IMM8 as u32, 58), |
| 11463 | mask(IMM8 as u32, 59), |
| 11464 | mask(IMM8 as u32, 60), |
| 11465 | mask(IMM8 as u32, 61), |
| 11466 | mask(IMM8 as u32, 62), |
| 11467 | mask(IMM8 as u32, 63), |
| 11468 | ], |
| 11469 | ); |
| 11470 | transmute(r) |
| 11471 | } |
| 11472 | } |
| 11473 | |
| 11474 | /// Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 11475 | /// |
| 11476 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_alignr_epi8&expand=264) |
| 11477 | #[inline ] |
| 11478 | #[target_feature (enable = "avx512bw" )] |
| 11479 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11480 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 1))] |
| 11481 | #[rustc_legacy_const_generics (4)] |
| 11482 | pub fn _mm512_mask_alignr_epi8<const IMM8: i32>( |
| 11483 | src: __m512i, |
| 11484 | k: __mmask64, |
| 11485 | a: __m512i, |
| 11486 | b: __m512i, |
| 11487 | ) -> __m512i { |
| 11488 | unsafe { |
| 11489 | static_assert_uimm_bits!(IMM8, 8); |
| 11490 | let r: __m512i = _mm512_alignr_epi8::<IMM8>(a, b); |
| 11491 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i8x64(), no:src.as_i8x64())) |
| 11492 | } |
| 11493 | } |
| 11494 | |
| 11495 | /// Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 11496 | /// |
| 11497 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_alignr_epi8&expand=265) |
| 11498 | #[inline ] |
| 11499 | #[target_feature (enable = "avx512bw" )] |
| 11500 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11501 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 1))] |
| 11502 | #[rustc_legacy_const_generics (3)] |
| 11503 | pub fn _mm512_maskz_alignr_epi8<const IMM8: i32>(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
| 11504 | unsafe { |
| 11505 | static_assert_uimm_bits!(IMM8, 8); |
| 11506 | let r: __m512i = _mm512_alignr_epi8::<IMM8>(a, b); |
| 11507 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i8x64(), no:i8x64::ZERO)) |
| 11508 | } |
| 11509 | } |
| 11510 | |
| 11511 | /// Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 11512 | /// |
| 11513 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_alignr_epi8&expand=261) |
| 11514 | #[inline ] |
| 11515 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11516 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11517 | #[rustc_legacy_const_generics (4)] |
| 11518 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 5))] |
| 11519 | pub fn _mm256_mask_alignr_epi8<const IMM8: i32>( |
| 11520 | src: __m256i, |
| 11521 | k: __mmask32, |
| 11522 | a: __m256i, |
| 11523 | b: __m256i, |
| 11524 | ) -> __m256i { |
| 11525 | unsafe { |
| 11526 | static_assert_uimm_bits!(IMM8, 8); |
| 11527 | let r: __m256i = _mm256_alignr_epi8::<IMM8>(a, b); |
| 11528 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i8x32(), no:src.as_i8x32())) |
| 11529 | } |
| 11530 | } |
| 11531 | |
| 11532 | /// Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 11533 | /// |
| 11534 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_alignr_epi8&expand=262) |
| 11535 | #[inline ] |
| 11536 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11537 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11538 | #[rustc_legacy_const_generics (3)] |
| 11539 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 5))] |
| 11540 | pub fn _mm256_maskz_alignr_epi8<const IMM8: i32>(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
| 11541 | unsafe { |
| 11542 | static_assert_uimm_bits!(IMM8, 8); |
| 11543 | let r: __m256i = _mm256_alignr_epi8::<IMM8>(a, b); |
| 11544 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i8x32(), no:i8x32::ZERO)) |
| 11545 | } |
| 11546 | } |
| 11547 | |
| 11548 | /// Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst using writemask k (elements are copied from src when the corresponding mask bit is not set). |
| 11549 | /// |
| 11550 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_alignr_epi8&expand=258) |
| 11551 | #[inline ] |
| 11552 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11553 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11554 | #[rustc_legacy_const_generics (4)] |
| 11555 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 5))] |
| 11556 | pub fn _mm_mask_alignr_epi8<const IMM8: i32>( |
| 11557 | src: __m128i, |
| 11558 | k: __mmask16, |
| 11559 | a: __m128i, |
| 11560 | b: __m128i, |
| 11561 | ) -> __m128i { |
| 11562 | unsafe { |
| 11563 | static_assert_uimm_bits!(IMM8, 8); |
| 11564 | let r: __m128i = _mm_alignr_epi8::<IMM8>(a, b); |
| 11565 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i8x16(), no:src.as_i8x16())) |
| 11566 | } |
| 11567 | } |
| 11568 | |
| 11569 | /// Concatenate pairs of 16-byte blocks in a and b into a 32-byte temporary result, shift the result right by imm8 bytes, and store the low 16 bytes in dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
| 11570 | /// |
| 11571 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_alignr_epi8&expand=259) |
| 11572 | #[inline ] |
| 11573 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11574 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11575 | #[rustc_legacy_const_generics (3)] |
| 11576 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 5))] |
| 11577 | pub fn _mm_maskz_alignr_epi8<const IMM8: i32>(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
| 11578 | unsafe { |
| 11579 | static_assert_uimm_bits!(IMM8, 8); |
| 11580 | let r: __m128i = _mm_alignr_epi8::<IMM8>(a, b); |
| 11581 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i8x16(), no:i8x16::ZERO)) |
| 11582 | } |
| 11583 | } |
| 11584 | |
| 11585 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr. |
| 11586 | /// |
| 11587 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtsepi16_storeu_epi8&expand=1812) |
| 11588 | #[inline ] |
| 11589 | #[target_feature (enable = "avx512bw" )] |
| 11590 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11591 | #[cfg_attr (test, assert_instr(vpmovswb))] |
| 11592 | pub unsafe fn _mm512_mask_cvtsepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask32, a: __m512i) { |
| 11593 | vpmovswbmem(mem_addr, a.as_i16x32(), mask:k); |
| 11594 | } |
| 11595 | |
| 11596 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr. |
| 11597 | /// |
| 11598 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtsepi16_storeu_epi8&expand=1811) |
| 11599 | #[inline ] |
| 11600 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11601 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11602 | #[cfg_attr (test, assert_instr(vpmovswb))] |
| 11603 | pub unsafe fn _mm256_mask_cvtsepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask16, a: __m256i) { |
| 11604 | vpmovswbmem256(mem_addr, a.as_i16x16(), mask:k); |
| 11605 | } |
| 11606 | |
| 11607 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr. |
| 11608 | /// |
| 11609 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtsepi16_storeu_epi8&expand=1810) |
| 11610 | #[inline ] |
| 11611 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11612 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11613 | #[cfg_attr (test, assert_instr(vpmovswb))] |
| 11614 | pub unsafe fn _mm_mask_cvtsepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask8, a: __m128i) { |
| 11615 | vpmovswbmem128(mem_addr, a.as_i16x8(), mask:k); |
| 11616 | } |
| 11617 | |
| 11618 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr. |
| 11619 | /// |
| 11620 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtepi16_storeu_epi8&expand=1412) |
| 11621 | #[inline ] |
| 11622 | #[target_feature (enable = "avx512bw" )] |
| 11623 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11624 | #[cfg_attr (test, assert_instr(vpmovwb))] |
| 11625 | pub unsafe fn _mm512_mask_cvtepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask32, a: __m512i) { |
| 11626 | vpmovwbmem(mem_addr, a.as_i16x32(), mask:k); |
| 11627 | } |
| 11628 | |
| 11629 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr. |
| 11630 | /// |
| 11631 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtepi16_storeu_epi8&expand=1411) |
| 11632 | #[inline ] |
| 11633 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11634 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11635 | #[cfg_attr (test, assert_instr(vpmovwb))] |
| 11636 | pub unsafe fn _mm256_mask_cvtepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask16, a: __m256i) { |
| 11637 | vpmovwbmem256(mem_addr, a.as_i16x16(), mask:k); |
| 11638 | } |
| 11639 | |
| 11640 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr. |
| 11641 | /// |
| 11642 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtepi16_storeu_epi8&expand=1410) |
| 11643 | #[inline ] |
| 11644 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11645 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11646 | #[cfg_attr (test, assert_instr(vpmovwb))] |
| 11647 | pub unsafe fn _mm_mask_cvtepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask8, a: __m128i) { |
| 11648 | vpmovwbmem128(mem_addr, a.as_i16x8(), mask:k); |
| 11649 | } |
| 11650 | |
| 11651 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr. |
| 11652 | /// |
| 11653 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtusepi16_storeu_epi8&expand=2047) |
| 11654 | #[inline ] |
| 11655 | #[target_feature (enable = "avx512bw" )] |
| 11656 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11657 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
| 11658 | pub unsafe fn _mm512_mask_cvtusepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask32, a: __m512i) { |
| 11659 | vpmovuswbmem(mem_addr, a.as_i16x32(), mask:k); |
| 11660 | } |
| 11661 | |
| 11662 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr. |
| 11663 | /// |
| 11664 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtusepi16_storeu_epi8&expand=2046) |
| 11665 | #[inline ] |
| 11666 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11667 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11668 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
| 11669 | pub unsafe fn _mm256_mask_cvtusepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask16, a: __m256i) { |
| 11670 | vpmovuswbmem256(mem_addr, a.as_i16x16(), mask:k); |
| 11671 | } |
| 11672 | |
| 11673 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the active results (those with their respective bit set in writemask k) to unaligned memory at base_addr. |
| 11674 | /// |
| 11675 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtusepi16_storeu_epi8&expand=2045) |
| 11676 | #[inline ] |
| 11677 | #[target_feature (enable = "avx512bw,avx512vl" )] |
| 11678 | #[stable (feature = "stdarch_x86_avx512" , since = "1.89" )] |
| 11679 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
| 11680 | pub unsafe fn _mm_mask_cvtusepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask8, a: __m128i) { |
| 11681 | vpmovuswbmem128(mem_addr, a.as_i16x8(), mask:k); |
| 11682 | } |
| 11683 | |
| 11684 | #[allow (improper_ctypes)] |
| 11685 | unsafe extern "C" { |
| 11686 | #[link_name = "llvm.x86.avx512.pmul.hr.sw.512" ] |
| 11687 | unsafefn vpmulhrsw(a: i16x32, b: i16x32) -> i16x32; |
| 11688 | |
| 11689 | #[link_name = "llvm.x86.avx512.pmaddw.d.512" ] |
| 11690 | unsafefn vpmaddwd(a: i16x32, b: i16x32) -> i32x16; |
| 11691 | #[link_name = "llvm.x86.avx512.pmaddubs.w.512" ] |
| 11692 | unsafefn vpmaddubsw(a: u8x64, b: i8x64) -> i16x32; |
| 11693 | |
| 11694 | #[link_name = "llvm.x86.avx512.packssdw.512" ] |
| 11695 | unsafefn vpackssdw(a: i32x16, b: i32x16) -> i16x32; |
| 11696 | #[link_name = "llvm.x86.avx512.packsswb.512" ] |
| 11697 | unsafefn vpacksswb(a: i16x32, b: i16x32) -> i8x64; |
| 11698 | #[link_name = "llvm.x86.avx512.packusdw.512" ] |
| 11699 | unsafefn vpackusdw(a: i32x16, b: i32x16) -> u16x32; |
| 11700 | #[link_name = "llvm.x86.avx512.packuswb.512" ] |
| 11701 | unsafefn vpackuswb(a: i16x32, b: i16x32) -> u8x64; |
| 11702 | |
| 11703 | #[link_name = "llvm.x86.avx512.psll.w.512" ] |
| 11704 | unsafefn vpsllw(a: i16x32, count: i16x8) -> i16x32; |
| 11705 | |
| 11706 | #[link_name = "llvm.x86.avx512.psrl.w.512" ] |
| 11707 | unsafefn vpsrlw(a: i16x32, count: i16x8) -> i16x32; |
| 11708 | |
| 11709 | #[link_name = "llvm.x86.avx512.psra.w.512" ] |
| 11710 | unsafefn vpsraw(a: i16x32, count: i16x8) -> i16x32; |
| 11711 | |
| 11712 | #[link_name = "llvm.x86.avx512.vpermi2var.hi.512" ] |
| 11713 | unsafefn vpermi2w(a: i16x32, idx: i16x32, b: i16x32) -> i16x32; |
| 11714 | #[link_name = "llvm.x86.avx512.vpermi2var.hi.256" ] |
| 11715 | unsafefn vpermi2w256(a: i16x16, idx: i16x16, b: i16x16) -> i16x16; |
| 11716 | #[link_name = "llvm.x86.avx512.vpermi2var.hi.128" ] |
| 11717 | unsafefn vpermi2w128(a: i16x8, idx: i16x8, b: i16x8) -> i16x8; |
| 11718 | |
| 11719 | #[link_name = "llvm.x86.avx512.permvar.hi.512" ] |
| 11720 | unsafefn vpermw(a: i16x32, idx: i16x32) -> i16x32; |
| 11721 | #[link_name = "llvm.x86.avx512.permvar.hi.256" ] |
| 11722 | unsafefn vpermw256(a: i16x16, idx: i16x16) -> i16x16; |
| 11723 | #[link_name = "llvm.x86.avx512.permvar.hi.128" ] |
| 11724 | unsafefn vpermw128(a: i16x8, idx: i16x8) -> i16x8; |
| 11725 | |
| 11726 | #[link_name = "llvm.x86.avx512.pshuf.b.512" ] |
| 11727 | unsafefn vpshufb(a: i8x64, b: i8x64) -> i8x64; |
| 11728 | |
| 11729 | #[link_name = "llvm.x86.avx512.psad.bw.512" ] |
| 11730 | unsafefn vpsadbw(a: u8x64, b: u8x64) -> u64x8; |
| 11731 | |
| 11732 | #[link_name = "llvm.x86.avx512.dbpsadbw.512" ] |
| 11733 | unsafefn vdbpsadbw(a: u8x64, b: u8x64, imm8: i32) -> u16x32; |
| 11734 | #[link_name = "llvm.x86.avx512.dbpsadbw.256" ] |
| 11735 | unsafefn vdbpsadbw256(a: u8x32, b: u8x32, imm8: i32) -> u16x16; |
| 11736 | #[link_name = "llvm.x86.avx512.dbpsadbw.128" ] |
| 11737 | unsafefn vdbpsadbw128(a: u8x16, b: u8x16, imm8: i32) -> u16x8; |
| 11738 | |
| 11739 | #[link_name = "llvm.x86.avx512.mask.pmovs.wb.512" ] |
| 11740 | unsafefn vpmovswb(a: i16x32, src: i8x32, mask: u32) -> i8x32; |
| 11741 | #[link_name = "llvm.x86.avx512.mask.pmovs.wb.256" ] |
| 11742 | unsafefn vpmovswb256(a: i16x16, src: i8x16, mask: u16) -> i8x16; |
| 11743 | #[link_name = "llvm.x86.avx512.mask.pmovs.wb.128" ] |
| 11744 | unsafefn vpmovswb128(a: i16x8, src: i8x16, mask: u8) -> i8x16; |
| 11745 | |
| 11746 | #[link_name = "llvm.x86.avx512.mask.pmovus.wb.512" ] |
| 11747 | unsafefn vpmovuswb(a: u16x32, src: u8x32, mask: u32) -> u8x32; |
| 11748 | #[link_name = "llvm.x86.avx512.mask.pmovus.wb.256" ] |
| 11749 | unsafefn vpmovuswb256(a: u16x16, src: u8x16, mask: u16) -> u8x16; |
| 11750 | #[link_name = "llvm.x86.avx512.mask.pmovus.wb.128" ] |
| 11751 | unsafefn vpmovuswb128(a: u16x8, src: u8x16, mask: u8) -> u8x16; |
| 11752 | |
| 11753 | #[link_name = "llvm.x86.avx512.mask.pmovs.wb.mem.512" ] |
| 11754 | unsafefn vpmovswbmem(mem_addr: *mut i8, a: i16x32, mask: u32); |
| 11755 | #[link_name = "llvm.x86.avx512.mask.pmovs.wb.mem.256" ] |
| 11756 | unsafefn vpmovswbmem256(mem_addr: *mut i8, a: i16x16, mask: u16); |
| 11757 | #[link_name = "llvm.x86.avx512.mask.pmovs.wb.mem.128" ] |
| 11758 | unsafefn vpmovswbmem128(mem_addr: *mut i8, a: i16x8, mask: u8); |
| 11759 | |
| 11760 | #[link_name = "llvm.x86.avx512.mask.pmov.wb.mem.512" ] |
| 11761 | unsafefn vpmovwbmem(mem_addr: *mut i8, a: i16x32, mask: u32); |
| 11762 | #[link_name = "llvm.x86.avx512.mask.pmov.wb.mem.256" ] |
| 11763 | unsafefn vpmovwbmem256(mem_addr: *mut i8, a: i16x16, mask: u16); |
| 11764 | #[link_name = "llvm.x86.avx512.mask.pmov.wb.mem.128" ] |
| 11765 | unsafefn vpmovwbmem128(mem_addr: *mut i8, a: i16x8, mask: u8); |
| 11766 | |
| 11767 | #[link_name = "llvm.x86.avx512.mask.pmovus.wb.mem.512" ] |
| 11768 | unsafefn vpmovuswbmem(mem_addr: *mut i8, a: i16x32, mask: u32); |
| 11769 | #[link_name = "llvm.x86.avx512.mask.pmovus.wb.mem.256" ] |
| 11770 | unsafefn vpmovuswbmem256(mem_addr: *mut i8, a: i16x16, mask: u16); |
| 11771 | #[link_name = "llvm.x86.avx512.mask.pmovus.wb.mem.128" ] |
| 11772 | unsafefn vpmovuswbmem128(mem_addr: *mut i8, a: i16x8, mask: u8); |
| 11773 | } |
| 11774 | |
| 11775 | #[cfg (test)] |
| 11776 | mod tests { |
| 11777 | |
| 11778 | use stdarch_test::simd_test; |
| 11779 | |
| 11780 | use crate::core_arch::x86::*; |
| 11781 | use crate::hint::black_box; |
| 11782 | use crate::mem::{self}; |
| 11783 | |
| 11784 | #[simd_test(enable = "avx512bw" )] |
| 11785 | unsafe fn test_mm512_abs_epi16() { |
| 11786 | let a = _mm512_set1_epi16(-1); |
| 11787 | let r = _mm512_abs_epi16(a); |
| 11788 | let e = _mm512_set1_epi16(1); |
| 11789 | assert_eq_m512i(r, e); |
| 11790 | } |
| 11791 | |
| 11792 | #[simd_test(enable = "avx512bw" )] |
| 11793 | unsafe fn test_mm512_mask_abs_epi16() { |
| 11794 | let a = _mm512_set1_epi16(-1); |
| 11795 | let r = _mm512_mask_abs_epi16(a, 0, a); |
| 11796 | assert_eq_m512i(r, a); |
| 11797 | let r = _mm512_mask_abs_epi16(a, 0b00000000_11111111_00000000_11111111, a); |
| 11798 | #[rustfmt::skip] |
| 11799 | let e = _mm512_set_epi16(-1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 11800 | -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1); |
| 11801 | assert_eq_m512i(r, e); |
| 11802 | } |
| 11803 | |
| 11804 | #[simd_test(enable = "avx512bw" )] |
| 11805 | unsafe fn test_mm512_maskz_abs_epi16() { |
| 11806 | let a = _mm512_set1_epi16(-1); |
| 11807 | let r = _mm512_maskz_abs_epi16(0, a); |
| 11808 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 11809 | let r = _mm512_maskz_abs_epi16(0b00000000_11111111_00000000_11111111, a); |
| 11810 | #[rustfmt::skip] |
| 11811 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, |
| 11812 | 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
| 11813 | assert_eq_m512i(r, e); |
| 11814 | } |
| 11815 | |
| 11816 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 11817 | unsafe fn test_mm256_mask_abs_epi16() { |
| 11818 | let a = _mm256_set1_epi16(-1); |
| 11819 | let r = _mm256_mask_abs_epi16(a, 0, a); |
| 11820 | assert_eq_m256i(r, a); |
| 11821 | let r = _mm256_mask_abs_epi16(a, 0b00000000_11111111, a); |
| 11822 | let e = _mm256_set_epi16(-1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1); |
| 11823 | assert_eq_m256i(r, e); |
| 11824 | } |
| 11825 | |
| 11826 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 11827 | unsafe fn test_mm256_maskz_abs_epi16() { |
| 11828 | let a = _mm256_set1_epi16(-1); |
| 11829 | let r = _mm256_maskz_abs_epi16(0, a); |
| 11830 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 11831 | let r = _mm256_maskz_abs_epi16(0b00000000_11111111, a); |
| 11832 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
| 11833 | assert_eq_m256i(r, e); |
| 11834 | } |
| 11835 | |
| 11836 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 11837 | unsafe fn test_mm_mask_abs_epi16() { |
| 11838 | let a = _mm_set1_epi16(-1); |
| 11839 | let r = _mm_mask_abs_epi16(a, 0, a); |
| 11840 | assert_eq_m128i(r, a); |
| 11841 | let r = _mm_mask_abs_epi16(a, 0b00001111, a); |
| 11842 | let e = _mm_set_epi16(-1, -1, -1, -1, 1, 1, 1, 1); |
| 11843 | assert_eq_m128i(r, e); |
| 11844 | } |
| 11845 | |
| 11846 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 11847 | unsafe fn test_mm_maskz_abs_epi16() { |
| 11848 | let a = _mm_set1_epi16(-1); |
| 11849 | let r = _mm_maskz_abs_epi16(0, a); |
| 11850 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 11851 | let r = _mm_maskz_abs_epi16(0b00001111, a); |
| 11852 | let e = _mm_set_epi16(0, 0, 0, 0, 1, 1, 1, 1); |
| 11853 | assert_eq_m128i(r, e); |
| 11854 | } |
| 11855 | |
| 11856 | #[simd_test(enable = "avx512bw" )] |
| 11857 | unsafe fn test_mm512_abs_epi8() { |
| 11858 | let a = _mm512_set1_epi8(-1); |
| 11859 | let r = _mm512_abs_epi8(a); |
| 11860 | let e = _mm512_set1_epi8(1); |
| 11861 | assert_eq_m512i(r, e); |
| 11862 | } |
| 11863 | |
| 11864 | #[simd_test(enable = "avx512bw" )] |
| 11865 | unsafe fn test_mm512_mask_abs_epi8() { |
| 11866 | let a = _mm512_set1_epi8(-1); |
| 11867 | let r = _mm512_mask_abs_epi8(a, 0, a); |
| 11868 | assert_eq_m512i(r, a); |
| 11869 | let r = _mm512_mask_abs_epi8( |
| 11870 | a, |
| 11871 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 11872 | a, |
| 11873 | ); |
| 11874 | #[rustfmt::skip] |
| 11875 | let e = _mm512_set_epi8(-1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 11876 | -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 11877 | -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 11878 | -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1); |
| 11879 | assert_eq_m512i(r, e); |
| 11880 | } |
| 11881 | |
| 11882 | #[simd_test(enable = "avx512bw" )] |
| 11883 | unsafe fn test_mm512_maskz_abs_epi8() { |
| 11884 | let a = _mm512_set1_epi8(-1); |
| 11885 | let r = _mm512_maskz_abs_epi8(0, a); |
| 11886 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 11887 | let r = _mm512_maskz_abs_epi8( |
| 11888 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 11889 | a, |
| 11890 | ); |
| 11891 | #[rustfmt::skip] |
| 11892 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, |
| 11893 | 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, |
| 11894 | 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, |
| 11895 | 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
| 11896 | assert_eq_m512i(r, e); |
| 11897 | } |
| 11898 | |
| 11899 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 11900 | unsafe fn test_mm256_mask_abs_epi8() { |
| 11901 | let a = _mm256_set1_epi8(-1); |
| 11902 | let r = _mm256_mask_abs_epi8(a, 0, a); |
| 11903 | assert_eq_m256i(r, a); |
| 11904 | let r = _mm256_mask_abs_epi8(a, 0b00000000_11111111_00000000_11111111, a); |
| 11905 | #[rustfmt::skip] |
| 11906 | let e = _mm256_set_epi8(-1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 11907 | -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1); |
| 11908 | assert_eq_m256i(r, e); |
| 11909 | } |
| 11910 | |
| 11911 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 11912 | unsafe fn test_mm256_maskz_abs_epi8() { |
| 11913 | let a = _mm256_set1_epi8(-1); |
| 11914 | let r = _mm256_maskz_abs_epi8(0, a); |
| 11915 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 11916 | let r = _mm256_maskz_abs_epi8(0b00000000_11111111_00000000_11111111, a); |
| 11917 | #[rustfmt::skip] |
| 11918 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, |
| 11919 | 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
| 11920 | assert_eq_m256i(r, e); |
| 11921 | } |
| 11922 | |
| 11923 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 11924 | unsafe fn test_mm_mask_abs_epi8() { |
| 11925 | let a = _mm_set1_epi8(-1); |
| 11926 | let r = _mm_mask_abs_epi8(a, 0, a); |
| 11927 | assert_eq_m128i(r, a); |
| 11928 | let r = _mm_mask_abs_epi8(a, 0b00000000_11111111, a); |
| 11929 | let e = _mm_set_epi8(-1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1); |
| 11930 | assert_eq_m128i(r, e); |
| 11931 | } |
| 11932 | |
| 11933 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 11934 | unsafe fn test_mm_maskz_abs_epi8() { |
| 11935 | let a = _mm_set1_epi8(-1); |
| 11936 | let r = _mm_maskz_abs_epi8(0, a); |
| 11937 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 11938 | let r = _mm_maskz_abs_epi8(0b00000000_11111111, a); |
| 11939 | #[rustfmt::skip] |
| 11940 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
| 11941 | assert_eq_m128i(r, e); |
| 11942 | } |
| 11943 | |
| 11944 | #[simd_test(enable = "avx512bw" )] |
| 11945 | unsafe fn test_mm512_add_epi16() { |
| 11946 | let a = _mm512_set1_epi16(1); |
| 11947 | let b = _mm512_set1_epi16(2); |
| 11948 | let r = _mm512_add_epi16(a, b); |
| 11949 | let e = _mm512_set1_epi16(3); |
| 11950 | assert_eq_m512i(r, e); |
| 11951 | } |
| 11952 | |
| 11953 | #[simd_test(enable = "avx512bw" )] |
| 11954 | unsafe fn test_mm512_mask_add_epi16() { |
| 11955 | let a = _mm512_set1_epi16(1); |
| 11956 | let b = _mm512_set1_epi16(2); |
| 11957 | let r = _mm512_mask_add_epi16(a, 0, a, b); |
| 11958 | assert_eq_m512i(r, a); |
| 11959 | let r = _mm512_mask_add_epi16(a, 0b00000000_11111111_00000000_11111111, a, b); |
| 11960 | #[rustfmt::skip] |
| 11961 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, |
| 11962 | 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3); |
| 11963 | assert_eq_m512i(r, e); |
| 11964 | } |
| 11965 | |
| 11966 | #[simd_test(enable = "avx512bw" )] |
| 11967 | unsafe fn test_mm512_maskz_add_epi16() { |
| 11968 | let a = _mm512_set1_epi16(1); |
| 11969 | let b = _mm512_set1_epi16(2); |
| 11970 | let r = _mm512_maskz_add_epi16(0, a, b); |
| 11971 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 11972 | let r = _mm512_maskz_add_epi16(0b00000000_11111111_00000000_11111111, a, b); |
| 11973 | #[rustfmt::skip] |
| 11974 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3, |
| 11975 | 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3); |
| 11976 | assert_eq_m512i(r, e); |
| 11977 | } |
| 11978 | |
| 11979 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 11980 | unsafe fn test_mm256_mask_add_epi16() { |
| 11981 | let a = _mm256_set1_epi16(1); |
| 11982 | let b = _mm256_set1_epi16(2); |
| 11983 | let r = _mm256_mask_add_epi16(a, 0, a, b); |
| 11984 | assert_eq_m256i(r, a); |
| 11985 | let r = _mm256_mask_add_epi16(a, 0b00000000_11111111, a, b); |
| 11986 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3); |
| 11987 | assert_eq_m256i(r, e); |
| 11988 | } |
| 11989 | |
| 11990 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 11991 | unsafe fn test_mm256_maskz_add_epi16() { |
| 11992 | let a = _mm256_set1_epi16(1); |
| 11993 | let b = _mm256_set1_epi16(2); |
| 11994 | let r = _mm256_maskz_add_epi16(0, a, b); |
| 11995 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 11996 | let r = _mm256_maskz_add_epi16(0b00000000_11111111, a, b); |
| 11997 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3); |
| 11998 | assert_eq_m256i(r, e); |
| 11999 | } |
| 12000 | |
| 12001 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12002 | unsafe fn test_mm_mask_add_epi16() { |
| 12003 | let a = _mm_set1_epi16(1); |
| 12004 | let b = _mm_set1_epi16(2); |
| 12005 | let r = _mm_mask_add_epi16(a, 0, a, b); |
| 12006 | assert_eq_m128i(r, a); |
| 12007 | let r = _mm_mask_add_epi16(a, 0b00001111, a, b); |
| 12008 | let e = _mm_set_epi16(1, 1, 1, 1, 3, 3, 3, 3); |
| 12009 | assert_eq_m128i(r, e); |
| 12010 | } |
| 12011 | |
| 12012 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12013 | unsafe fn test_mm_maskz_add_epi16() { |
| 12014 | let a = _mm_set1_epi16(1); |
| 12015 | let b = _mm_set1_epi16(2); |
| 12016 | let r = _mm_maskz_add_epi16(0, a, b); |
| 12017 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 12018 | let r = _mm_maskz_add_epi16(0b00001111, a, b); |
| 12019 | let e = _mm_set_epi16(0, 0, 0, 0, 3, 3, 3, 3); |
| 12020 | assert_eq_m128i(r, e); |
| 12021 | } |
| 12022 | |
| 12023 | #[simd_test(enable = "avx512bw" )] |
| 12024 | unsafe fn test_mm512_add_epi8() { |
| 12025 | let a = _mm512_set1_epi8(1); |
| 12026 | let b = _mm512_set1_epi8(2); |
| 12027 | let r = _mm512_add_epi8(a, b); |
| 12028 | let e = _mm512_set1_epi8(3); |
| 12029 | assert_eq_m512i(r, e); |
| 12030 | } |
| 12031 | |
| 12032 | #[simd_test(enable = "avx512bw" )] |
| 12033 | unsafe fn test_mm512_mask_add_epi8() { |
| 12034 | let a = _mm512_set1_epi8(1); |
| 12035 | let b = _mm512_set1_epi8(2); |
| 12036 | let r = _mm512_mask_add_epi8(a, 0, a, b); |
| 12037 | assert_eq_m512i(r, a); |
| 12038 | let r = _mm512_mask_add_epi8( |
| 12039 | a, |
| 12040 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 12041 | a, |
| 12042 | b, |
| 12043 | ); |
| 12044 | #[rustfmt::skip] |
| 12045 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, |
| 12046 | 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, |
| 12047 | 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, |
| 12048 | 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3); |
| 12049 | assert_eq_m512i(r, e); |
| 12050 | } |
| 12051 | |
| 12052 | #[simd_test(enable = "avx512bw" )] |
| 12053 | unsafe fn test_mm512_maskz_add_epi8() { |
| 12054 | let a = _mm512_set1_epi8(1); |
| 12055 | let b = _mm512_set1_epi8(2); |
| 12056 | let r = _mm512_maskz_add_epi8(0, a, b); |
| 12057 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 12058 | let r = _mm512_maskz_add_epi8( |
| 12059 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 12060 | a, |
| 12061 | b, |
| 12062 | ); |
| 12063 | #[rustfmt::skip] |
| 12064 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3, |
| 12065 | 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3, |
| 12066 | 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3, |
| 12067 | 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3); |
| 12068 | assert_eq_m512i(r, e); |
| 12069 | } |
| 12070 | |
| 12071 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12072 | unsafe fn test_mm256_mask_add_epi8() { |
| 12073 | let a = _mm256_set1_epi8(1); |
| 12074 | let b = _mm256_set1_epi8(2); |
| 12075 | let r = _mm256_mask_add_epi8(a, 0, a, b); |
| 12076 | assert_eq_m256i(r, a); |
| 12077 | let r = _mm256_mask_add_epi8(a, 0b00000000_11111111_00000000_11111111, a, b); |
| 12078 | #[rustfmt::skip] |
| 12079 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, |
| 12080 | 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3); |
| 12081 | assert_eq_m256i(r, e); |
| 12082 | } |
| 12083 | |
| 12084 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12085 | unsafe fn test_mm256_maskz_add_epi8() { |
| 12086 | let a = _mm256_set1_epi8(1); |
| 12087 | let b = _mm256_set1_epi8(2); |
| 12088 | let r = _mm256_maskz_add_epi8(0, a, b); |
| 12089 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 12090 | let r = _mm256_maskz_add_epi8(0b00000000_11111111_00000000_11111111, a, b); |
| 12091 | #[rustfmt::skip] |
| 12092 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3, |
| 12093 | 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3); |
| 12094 | assert_eq_m256i(r, e); |
| 12095 | } |
| 12096 | |
| 12097 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12098 | unsafe fn test_mm_mask_add_epi8() { |
| 12099 | let a = _mm_set1_epi8(1); |
| 12100 | let b = _mm_set1_epi8(2); |
| 12101 | let r = _mm_mask_add_epi8(a, 0, a, b); |
| 12102 | assert_eq_m128i(r, a); |
| 12103 | let r = _mm_mask_add_epi8(a, 0b00000000_11111111, a, b); |
| 12104 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3); |
| 12105 | assert_eq_m128i(r, e); |
| 12106 | } |
| 12107 | |
| 12108 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12109 | unsafe fn test_mm_maskz_add_epi8() { |
| 12110 | let a = _mm_set1_epi8(1); |
| 12111 | let b = _mm_set1_epi8(2); |
| 12112 | let r = _mm_maskz_add_epi8(0, a, b); |
| 12113 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 12114 | let r = _mm_maskz_add_epi8(0b00000000_11111111, a, b); |
| 12115 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3); |
| 12116 | assert_eq_m128i(r, e); |
| 12117 | } |
| 12118 | |
| 12119 | #[simd_test(enable = "avx512bw" )] |
| 12120 | unsafe fn test_mm512_adds_epu16() { |
| 12121 | let a = _mm512_set1_epi16(1); |
| 12122 | let b = _mm512_set1_epi16(u16::MAX as i16); |
| 12123 | let r = _mm512_adds_epu16(a, b); |
| 12124 | let e = _mm512_set1_epi16(u16::MAX as i16); |
| 12125 | assert_eq_m512i(r, e); |
| 12126 | } |
| 12127 | |
| 12128 | #[simd_test(enable = "avx512bw" )] |
| 12129 | unsafe fn test_mm512_mask_adds_epu16() { |
| 12130 | let a = _mm512_set1_epi16(1); |
| 12131 | let b = _mm512_set1_epi16(u16::MAX as i16); |
| 12132 | let r = _mm512_mask_adds_epu16(a, 0, a, b); |
| 12133 | assert_eq_m512i(r, a); |
| 12134 | let r = _mm512_mask_adds_epu16(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 12135 | #[rustfmt::skip] |
| 12136 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12137 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, u16::MAX as i16, u16::MAX as i16, u16::MAX as i16, u16::MAX as i16); |
| 12138 | assert_eq_m512i(r, e); |
| 12139 | } |
| 12140 | |
| 12141 | #[simd_test(enable = "avx512bw" )] |
| 12142 | unsafe fn test_mm512_maskz_adds_epu16() { |
| 12143 | let a = _mm512_set1_epi16(1); |
| 12144 | let b = _mm512_set1_epi16(u16::MAX as i16); |
| 12145 | let r = _mm512_maskz_adds_epu16(0, a, b); |
| 12146 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 12147 | let r = _mm512_maskz_adds_epu16(0b00000000_00000000_00000000_00001111, a, b); |
| 12148 | #[rustfmt::skip] |
| 12149 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12150 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, u16::MAX as i16, u16::MAX as i16, u16::MAX as i16, u16::MAX as i16); |
| 12151 | assert_eq_m512i(r, e); |
| 12152 | } |
| 12153 | |
| 12154 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12155 | unsafe fn test_mm256_mask_adds_epu16() { |
| 12156 | let a = _mm256_set1_epi16(1); |
| 12157 | let b = _mm256_set1_epi16(u16::MAX as i16); |
| 12158 | let r = _mm256_mask_adds_epu16(a, 0, a, b); |
| 12159 | assert_eq_m256i(r, a); |
| 12160 | let r = _mm256_mask_adds_epu16(a, 0b00000000_00001111, a, b); |
| 12161 | #[rustfmt::skip] |
| 12162 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, u16::MAX as i16, u16::MAX as i16, u16::MAX as i16, u16::MAX as i16); |
| 12163 | assert_eq_m256i(r, e); |
| 12164 | } |
| 12165 | |
| 12166 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12167 | unsafe fn test_mm256_maskz_adds_epu16() { |
| 12168 | let a = _mm256_set1_epi16(1); |
| 12169 | let b = _mm256_set1_epi16(u16::MAX as i16); |
| 12170 | let r = _mm256_maskz_adds_epu16(0, a, b); |
| 12171 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 12172 | let r = _mm256_maskz_adds_epu16(0b00000000_00001111, a, b); |
| 12173 | #[rustfmt::skip] |
| 12174 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, u16::MAX as i16, u16::MAX as i16, u16::MAX as i16, u16::MAX as i16); |
| 12175 | assert_eq_m256i(r, e); |
| 12176 | } |
| 12177 | |
| 12178 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12179 | unsafe fn test_mm_mask_adds_epu16() { |
| 12180 | let a = _mm_set1_epi16(1); |
| 12181 | let b = _mm_set1_epi16(u16::MAX as i16); |
| 12182 | let r = _mm_mask_adds_epu16(a, 0, a, b); |
| 12183 | assert_eq_m128i(r, a); |
| 12184 | let r = _mm_mask_adds_epu16(a, 0b00001111, a, b); |
| 12185 | #[rustfmt::skip] |
| 12186 | let e = _mm_set_epi16(1, 1, 1, 1, u16::MAX as i16, u16::MAX as i16, u16::MAX as i16, u16::MAX as i16); |
| 12187 | assert_eq_m128i(r, e); |
| 12188 | } |
| 12189 | |
| 12190 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12191 | unsafe fn test_mm_maskz_adds_epu16() { |
| 12192 | let a = _mm_set1_epi16(1); |
| 12193 | let b = _mm_set1_epi16(u16::MAX as i16); |
| 12194 | let r = _mm_maskz_adds_epu16(0, a, b); |
| 12195 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 12196 | let r = _mm_maskz_adds_epu16(0b00001111, a, b); |
| 12197 | #[rustfmt::skip] |
| 12198 | let e = _mm_set_epi16(0, 0, 0, 0, u16::MAX as i16, u16::MAX as i16, u16::MAX as i16, u16::MAX as i16); |
| 12199 | assert_eq_m128i(r, e); |
| 12200 | } |
| 12201 | |
| 12202 | #[simd_test(enable = "avx512bw" )] |
| 12203 | unsafe fn test_mm512_adds_epu8() { |
| 12204 | let a = _mm512_set1_epi8(1); |
| 12205 | let b = _mm512_set1_epi8(u8::MAX as i8); |
| 12206 | let r = _mm512_adds_epu8(a, b); |
| 12207 | let e = _mm512_set1_epi8(u8::MAX as i8); |
| 12208 | assert_eq_m512i(r, e); |
| 12209 | } |
| 12210 | |
| 12211 | #[simd_test(enable = "avx512bw" )] |
| 12212 | unsafe fn test_mm512_mask_adds_epu8() { |
| 12213 | let a = _mm512_set1_epi8(1); |
| 12214 | let b = _mm512_set1_epi8(u8::MAX as i8); |
| 12215 | let r = _mm512_mask_adds_epu8(a, 0, a, b); |
| 12216 | assert_eq_m512i(r, a); |
| 12217 | let r = _mm512_mask_adds_epu8( |
| 12218 | a, |
| 12219 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 12220 | a, |
| 12221 | b, |
| 12222 | ); |
| 12223 | #[rustfmt::skip] |
| 12224 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12225 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12226 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12227 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8); |
| 12228 | assert_eq_m512i(r, e); |
| 12229 | } |
| 12230 | |
| 12231 | #[simd_test(enable = "avx512bw" )] |
| 12232 | unsafe fn test_mm512_maskz_adds_epu8() { |
| 12233 | let a = _mm512_set1_epi8(1); |
| 12234 | let b = _mm512_set1_epi8(u8::MAX as i8); |
| 12235 | let r = _mm512_maskz_adds_epu8(0, a, b); |
| 12236 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 12237 | let r = _mm512_maskz_adds_epu8( |
| 12238 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 12239 | a, |
| 12240 | b, |
| 12241 | ); |
| 12242 | #[rustfmt::skip] |
| 12243 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12244 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12245 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12246 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8); |
| 12247 | assert_eq_m512i(r, e); |
| 12248 | } |
| 12249 | |
| 12250 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12251 | unsafe fn test_mm256_mask_adds_epu8() { |
| 12252 | let a = _mm256_set1_epi8(1); |
| 12253 | let b = _mm256_set1_epi8(u8::MAX as i8); |
| 12254 | let r = _mm256_mask_adds_epu8(a, 0, a, b); |
| 12255 | assert_eq_m256i(r, a); |
| 12256 | let r = _mm256_mask_adds_epu8(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 12257 | #[rustfmt::skip] |
| 12258 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12259 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8); |
| 12260 | assert_eq_m256i(r, e); |
| 12261 | } |
| 12262 | |
| 12263 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12264 | unsafe fn test_mm256_maskz_adds_epu8() { |
| 12265 | let a = _mm256_set1_epi8(1); |
| 12266 | let b = _mm256_set1_epi8(u8::MAX as i8); |
| 12267 | let r = _mm256_maskz_adds_epu8(0, a, b); |
| 12268 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 12269 | let r = _mm256_maskz_adds_epu8(0b00000000_00000000_00000000_00001111, a, b); |
| 12270 | #[rustfmt::skip] |
| 12271 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12272 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8); |
| 12273 | assert_eq_m256i(r, e); |
| 12274 | } |
| 12275 | |
| 12276 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12277 | unsafe fn test_mm_mask_adds_epu8() { |
| 12278 | let a = _mm_set1_epi8(1); |
| 12279 | let b = _mm_set1_epi8(u8::MAX as i8); |
| 12280 | let r = _mm_mask_adds_epu8(a, 0, a, b); |
| 12281 | assert_eq_m128i(r, a); |
| 12282 | let r = _mm_mask_adds_epu8(a, 0b00000000_00001111, a, b); |
| 12283 | #[rustfmt::skip] |
| 12284 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8); |
| 12285 | assert_eq_m128i(r, e); |
| 12286 | } |
| 12287 | |
| 12288 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12289 | unsafe fn test_mm_maskz_adds_epu8() { |
| 12290 | let a = _mm_set1_epi8(1); |
| 12291 | let b = _mm_set1_epi8(u8::MAX as i8); |
| 12292 | let r = _mm_maskz_adds_epu8(0, a, b); |
| 12293 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 12294 | let r = _mm_maskz_adds_epu8(0b00000000_00001111, a, b); |
| 12295 | #[rustfmt::skip] |
| 12296 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8); |
| 12297 | assert_eq_m128i(r, e); |
| 12298 | } |
| 12299 | |
| 12300 | #[simd_test(enable = "avx512bw" )] |
| 12301 | unsafe fn test_mm512_adds_epi16() { |
| 12302 | let a = _mm512_set1_epi16(1); |
| 12303 | let b = _mm512_set1_epi16(i16::MAX); |
| 12304 | let r = _mm512_adds_epi16(a, b); |
| 12305 | let e = _mm512_set1_epi16(i16::MAX); |
| 12306 | assert_eq_m512i(r, e); |
| 12307 | } |
| 12308 | |
| 12309 | #[simd_test(enable = "avx512bw" )] |
| 12310 | unsafe fn test_mm512_mask_adds_epi16() { |
| 12311 | let a = _mm512_set1_epi16(1); |
| 12312 | let b = _mm512_set1_epi16(i16::MAX); |
| 12313 | let r = _mm512_mask_adds_epi16(a, 0, a, b); |
| 12314 | assert_eq_m512i(r, a); |
| 12315 | let r = _mm512_mask_adds_epi16(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 12316 | #[rustfmt::skip] |
| 12317 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12318 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 12319 | assert_eq_m512i(r, e); |
| 12320 | } |
| 12321 | |
| 12322 | #[simd_test(enable = "avx512bw" )] |
| 12323 | unsafe fn test_mm512_maskz_adds_epi16() { |
| 12324 | let a = _mm512_set1_epi16(1); |
| 12325 | let b = _mm512_set1_epi16(i16::MAX); |
| 12326 | let r = _mm512_maskz_adds_epi16(0, a, b); |
| 12327 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 12328 | let r = _mm512_maskz_adds_epi16(0b00000000_00000000_00000000_00001111, a, b); |
| 12329 | #[rustfmt::skip] |
| 12330 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12331 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 12332 | assert_eq_m512i(r, e); |
| 12333 | } |
| 12334 | |
| 12335 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12336 | unsafe fn test_mm256_mask_adds_epi16() { |
| 12337 | let a = _mm256_set1_epi16(1); |
| 12338 | let b = _mm256_set1_epi16(i16::MAX); |
| 12339 | let r = _mm256_mask_adds_epi16(a, 0, a, b); |
| 12340 | assert_eq_m256i(r, a); |
| 12341 | let r = _mm256_mask_adds_epi16(a, 0b00000000_00001111, a, b); |
| 12342 | #[rustfmt::skip] |
| 12343 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 12344 | assert_eq_m256i(r, e); |
| 12345 | } |
| 12346 | |
| 12347 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12348 | unsafe fn test_mm256_maskz_adds_epi16() { |
| 12349 | let a = _mm256_set1_epi16(1); |
| 12350 | let b = _mm256_set1_epi16(i16::MAX); |
| 12351 | let r = _mm256_maskz_adds_epi16(0, a, b); |
| 12352 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 12353 | let r = _mm256_maskz_adds_epi16(0b00000000_00001111, a, b); |
| 12354 | #[rustfmt::skip] |
| 12355 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 12356 | assert_eq_m256i(r, e); |
| 12357 | } |
| 12358 | |
| 12359 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12360 | unsafe fn test_mm_mask_adds_epi16() { |
| 12361 | let a = _mm_set1_epi16(1); |
| 12362 | let b = _mm_set1_epi16(i16::MAX); |
| 12363 | let r = _mm_mask_adds_epi16(a, 0, a, b); |
| 12364 | assert_eq_m128i(r, a); |
| 12365 | let r = _mm_mask_adds_epi16(a, 0b00001111, a, b); |
| 12366 | let e = _mm_set_epi16(1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 12367 | assert_eq_m128i(r, e); |
| 12368 | } |
| 12369 | |
| 12370 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12371 | unsafe fn test_mm_maskz_adds_epi16() { |
| 12372 | let a = _mm_set1_epi16(1); |
| 12373 | let b = _mm_set1_epi16(i16::MAX); |
| 12374 | let r = _mm_maskz_adds_epi16(0, a, b); |
| 12375 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 12376 | let r = _mm_maskz_adds_epi16(0b00001111, a, b); |
| 12377 | let e = _mm_set_epi16(0, 0, 0, 0, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 12378 | assert_eq_m128i(r, e); |
| 12379 | } |
| 12380 | |
| 12381 | #[simd_test(enable = "avx512bw" )] |
| 12382 | unsafe fn test_mm512_adds_epi8() { |
| 12383 | let a = _mm512_set1_epi8(1); |
| 12384 | let b = _mm512_set1_epi8(i8::MAX); |
| 12385 | let r = _mm512_adds_epi8(a, b); |
| 12386 | let e = _mm512_set1_epi8(i8::MAX); |
| 12387 | assert_eq_m512i(r, e); |
| 12388 | } |
| 12389 | |
| 12390 | #[simd_test(enable = "avx512bw" )] |
| 12391 | unsafe fn test_mm512_mask_adds_epi8() { |
| 12392 | let a = _mm512_set1_epi8(1); |
| 12393 | let b = _mm512_set1_epi8(i8::MAX); |
| 12394 | let r = _mm512_mask_adds_epi8(a, 0, a, b); |
| 12395 | assert_eq_m512i(r, a); |
| 12396 | let r = _mm512_mask_adds_epi8( |
| 12397 | a, |
| 12398 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 12399 | a, |
| 12400 | b, |
| 12401 | ); |
| 12402 | #[rustfmt::skip] |
| 12403 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12404 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12405 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12406 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 12407 | assert_eq_m512i(r, e); |
| 12408 | } |
| 12409 | |
| 12410 | #[simd_test(enable = "avx512bw" )] |
| 12411 | unsafe fn test_mm512_maskz_adds_epi8() { |
| 12412 | let a = _mm512_set1_epi8(1); |
| 12413 | let b = _mm512_set1_epi8(i8::MAX); |
| 12414 | let r = _mm512_maskz_adds_epi8(0, a, b); |
| 12415 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 12416 | let r = _mm512_maskz_adds_epi8( |
| 12417 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 12418 | a, |
| 12419 | b, |
| 12420 | ); |
| 12421 | #[rustfmt::skip] |
| 12422 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12423 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12424 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12425 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 12426 | assert_eq_m512i(r, e); |
| 12427 | } |
| 12428 | |
| 12429 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12430 | unsafe fn test_mm256_mask_adds_epi8() { |
| 12431 | let a = _mm256_set1_epi8(1); |
| 12432 | let b = _mm256_set1_epi8(i8::MAX); |
| 12433 | let r = _mm256_mask_adds_epi8(a, 0, a, b); |
| 12434 | assert_eq_m256i(r, a); |
| 12435 | let r = _mm256_mask_adds_epi8(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 12436 | #[rustfmt::skip] |
| 12437 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12438 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 12439 | assert_eq_m256i(r, e); |
| 12440 | } |
| 12441 | |
| 12442 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12443 | unsafe fn test_mm256_maskz_adds_epi8() { |
| 12444 | let a = _mm256_set1_epi8(1); |
| 12445 | let b = _mm256_set1_epi8(i8::MAX); |
| 12446 | let r = _mm256_maskz_adds_epi8(0, a, b); |
| 12447 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 12448 | let r = _mm256_maskz_adds_epi8(0b00000000_00000000_00000000_00001111, a, b); |
| 12449 | #[rustfmt::skip] |
| 12450 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12451 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 12452 | assert_eq_m256i(r, e); |
| 12453 | } |
| 12454 | |
| 12455 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12456 | unsafe fn test_mm_mask_adds_epi8() { |
| 12457 | let a = _mm_set1_epi8(1); |
| 12458 | let b = _mm_set1_epi8(i8::MAX); |
| 12459 | let r = _mm_mask_adds_epi8(a, 0, a, b); |
| 12460 | assert_eq_m128i(r, a); |
| 12461 | let r = _mm_mask_adds_epi8(a, 0b00000000_00001111, a, b); |
| 12462 | #[rustfmt::skip] |
| 12463 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 12464 | assert_eq_m128i(r, e); |
| 12465 | } |
| 12466 | |
| 12467 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12468 | unsafe fn test_mm_maskz_adds_epi8() { |
| 12469 | let a = _mm_set1_epi8(1); |
| 12470 | let b = _mm_set1_epi8(i8::MAX); |
| 12471 | let r = _mm_maskz_adds_epi8(0, a, b); |
| 12472 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 12473 | let r = _mm_maskz_adds_epi8(0b00000000_00001111, a, b); |
| 12474 | #[rustfmt::skip] |
| 12475 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 12476 | assert_eq_m128i(r, e); |
| 12477 | } |
| 12478 | |
| 12479 | #[simd_test(enable = "avx512bw" )] |
| 12480 | unsafe fn test_mm512_sub_epi16() { |
| 12481 | let a = _mm512_set1_epi16(1); |
| 12482 | let b = _mm512_set1_epi16(2); |
| 12483 | let r = _mm512_sub_epi16(a, b); |
| 12484 | let e = _mm512_set1_epi16(-1); |
| 12485 | assert_eq_m512i(r, e); |
| 12486 | } |
| 12487 | |
| 12488 | #[simd_test(enable = "avx512bw" )] |
| 12489 | unsafe fn test_mm512_mask_sub_epi16() { |
| 12490 | let a = _mm512_set1_epi16(1); |
| 12491 | let b = _mm512_set1_epi16(2); |
| 12492 | let r = _mm512_mask_sub_epi16(a, 0, a, b); |
| 12493 | assert_eq_m512i(r, a); |
| 12494 | let r = _mm512_mask_sub_epi16(a, 0b00000000_11111111_00000000_11111111, a, b); |
| 12495 | #[rustfmt::skip] |
| 12496 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12497 | 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1); |
| 12498 | assert_eq_m512i(r, e); |
| 12499 | } |
| 12500 | |
| 12501 | #[simd_test(enable = "avx512bw" )] |
| 12502 | unsafe fn test_mm512_maskz_sub_epi16() { |
| 12503 | let a = _mm512_set1_epi16(1); |
| 12504 | let b = _mm512_set1_epi16(2); |
| 12505 | let r = _mm512_maskz_sub_epi16(0, a, b); |
| 12506 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 12507 | let r = _mm512_maskz_sub_epi16(0b00000000_11111111_00000000_11111111, a, b); |
| 12508 | #[rustfmt::skip] |
| 12509 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12510 | 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
| 12511 | assert_eq_m512i(r, e); |
| 12512 | } |
| 12513 | |
| 12514 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12515 | unsafe fn test_mm256_mask_sub_epi16() { |
| 12516 | let a = _mm256_set1_epi16(1); |
| 12517 | let b = _mm256_set1_epi16(2); |
| 12518 | let r = _mm256_mask_sub_epi16(a, 0, a, b); |
| 12519 | assert_eq_m256i(r, a); |
| 12520 | let r = _mm256_mask_sub_epi16(a, 0b00000000_11111111, a, b); |
| 12521 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1); |
| 12522 | assert_eq_m256i(r, e); |
| 12523 | } |
| 12524 | |
| 12525 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12526 | unsafe fn test_mm256_maskz_sub_epi16() { |
| 12527 | let a = _mm256_set1_epi16(1); |
| 12528 | let b = _mm256_set1_epi16(2); |
| 12529 | let r = _mm256_maskz_sub_epi16(0, a, b); |
| 12530 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 12531 | let r = _mm256_maskz_sub_epi16(0b00000000_11111111, a, b); |
| 12532 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
| 12533 | assert_eq_m256i(r, e); |
| 12534 | } |
| 12535 | |
| 12536 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12537 | unsafe fn test_mm_mask_sub_epi16() { |
| 12538 | let a = _mm_set1_epi16(1); |
| 12539 | let b = _mm_set1_epi16(2); |
| 12540 | let r = _mm_mask_sub_epi16(a, 0, a, b); |
| 12541 | assert_eq_m128i(r, a); |
| 12542 | let r = _mm_mask_sub_epi16(a, 0b00001111, a, b); |
| 12543 | let e = _mm_set_epi16(1, 1, 1, 1, -1, -1, -1, -1); |
| 12544 | assert_eq_m128i(r, e); |
| 12545 | } |
| 12546 | |
| 12547 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12548 | unsafe fn test_mm_maskz_sub_epi16() { |
| 12549 | let a = _mm_set1_epi16(1); |
| 12550 | let b = _mm_set1_epi16(2); |
| 12551 | let r = _mm_maskz_sub_epi16(0, a, b); |
| 12552 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 12553 | let r = _mm_maskz_sub_epi16(0b00001111, a, b); |
| 12554 | let e = _mm_set_epi16(0, 0, 0, 0, -1, -1, -1, -1); |
| 12555 | assert_eq_m128i(r, e); |
| 12556 | } |
| 12557 | |
| 12558 | #[simd_test(enable = "avx512bw" )] |
| 12559 | unsafe fn test_mm512_sub_epi8() { |
| 12560 | let a = _mm512_set1_epi8(1); |
| 12561 | let b = _mm512_set1_epi8(2); |
| 12562 | let r = _mm512_sub_epi8(a, b); |
| 12563 | let e = _mm512_set1_epi8(-1); |
| 12564 | assert_eq_m512i(r, e); |
| 12565 | } |
| 12566 | |
| 12567 | #[simd_test(enable = "avx512bw" )] |
| 12568 | unsafe fn test_mm512_mask_sub_epi8() { |
| 12569 | let a = _mm512_set1_epi8(1); |
| 12570 | let b = _mm512_set1_epi8(2); |
| 12571 | let r = _mm512_mask_sub_epi8(a, 0, a, b); |
| 12572 | assert_eq_m512i(r, a); |
| 12573 | let r = _mm512_mask_sub_epi8( |
| 12574 | a, |
| 12575 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 12576 | a, |
| 12577 | b, |
| 12578 | ); |
| 12579 | #[rustfmt::skip] |
| 12580 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12581 | 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12582 | 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12583 | 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1); |
| 12584 | assert_eq_m512i(r, e); |
| 12585 | } |
| 12586 | |
| 12587 | #[simd_test(enable = "avx512bw" )] |
| 12588 | unsafe fn test_mm512_maskz_sub_epi8() { |
| 12589 | let a = _mm512_set1_epi8(1); |
| 12590 | let b = _mm512_set1_epi8(2); |
| 12591 | let r = _mm512_maskz_sub_epi8(0, a, b); |
| 12592 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 12593 | let r = _mm512_maskz_sub_epi8( |
| 12594 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 12595 | a, |
| 12596 | b, |
| 12597 | ); |
| 12598 | #[rustfmt::skip] |
| 12599 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12600 | 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12601 | 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12602 | 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
| 12603 | assert_eq_m512i(r, e); |
| 12604 | } |
| 12605 | |
| 12606 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12607 | unsafe fn test_mm256_mask_sub_epi8() { |
| 12608 | let a = _mm256_set1_epi8(1); |
| 12609 | let b = _mm256_set1_epi8(2); |
| 12610 | let r = _mm256_mask_sub_epi8(a, 0, a, b); |
| 12611 | assert_eq_m256i(r, a); |
| 12612 | let r = _mm256_mask_sub_epi8(a, 0b00000000_11111111_00000000_11111111, a, b); |
| 12613 | #[rustfmt::skip] |
| 12614 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12615 | 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1); |
| 12616 | assert_eq_m256i(r, e); |
| 12617 | } |
| 12618 | |
| 12619 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12620 | unsafe fn test_mm256_maskz_sub_epi8() { |
| 12621 | let a = _mm256_set1_epi8(1); |
| 12622 | let b = _mm256_set1_epi8(2); |
| 12623 | let r = _mm256_maskz_sub_epi8(0, a, b); |
| 12624 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 12625 | let r = _mm256_maskz_sub_epi8(0b00000000_11111111_00000000_11111111, a, b); |
| 12626 | #[rustfmt::skip] |
| 12627 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12628 | 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
| 12629 | assert_eq_m256i(r, e); |
| 12630 | } |
| 12631 | |
| 12632 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12633 | unsafe fn test_mm_mask_sub_epi8() { |
| 12634 | let a = _mm_set1_epi8(1); |
| 12635 | let b = _mm_set1_epi8(2); |
| 12636 | let r = _mm_mask_sub_epi8(a, 0, a, b); |
| 12637 | assert_eq_m128i(r, a); |
| 12638 | let r = _mm_mask_sub_epi8(a, 0b00000000_11111111, a, b); |
| 12639 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1); |
| 12640 | assert_eq_m128i(r, e); |
| 12641 | } |
| 12642 | |
| 12643 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12644 | unsafe fn test_mm_maskz_sub_epi8() { |
| 12645 | let a = _mm_set1_epi8(1); |
| 12646 | let b = _mm_set1_epi8(2); |
| 12647 | let r = _mm_maskz_sub_epi8(0, a, b); |
| 12648 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 12649 | let r = _mm_maskz_sub_epi8(0b00000000_11111111, a, b); |
| 12650 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
| 12651 | assert_eq_m128i(r, e); |
| 12652 | } |
| 12653 | |
| 12654 | #[simd_test(enable = "avx512bw" )] |
| 12655 | unsafe fn test_mm512_subs_epu16() { |
| 12656 | let a = _mm512_set1_epi16(1); |
| 12657 | let b = _mm512_set1_epi16(u16::MAX as i16); |
| 12658 | let r = _mm512_subs_epu16(a, b); |
| 12659 | let e = _mm512_set1_epi16(0); |
| 12660 | assert_eq_m512i(r, e); |
| 12661 | } |
| 12662 | |
| 12663 | #[simd_test(enable = "avx512bw" )] |
| 12664 | unsafe fn test_mm512_mask_subs_epu16() { |
| 12665 | let a = _mm512_set1_epi16(1); |
| 12666 | let b = _mm512_set1_epi16(u16::MAX as i16); |
| 12667 | let r = _mm512_mask_subs_epu16(a, 0, a, b); |
| 12668 | assert_eq_m512i(r, a); |
| 12669 | let r = _mm512_mask_subs_epu16(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 12670 | #[rustfmt::skip] |
| 12671 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12672 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 12673 | assert_eq_m512i(r, e); |
| 12674 | } |
| 12675 | |
| 12676 | #[simd_test(enable = "avx512bw" )] |
| 12677 | unsafe fn test_mm512_maskz_subs_epu16() { |
| 12678 | let a = _mm512_set1_epi16(1); |
| 12679 | let b = _mm512_set1_epi16(u16::MAX as i16); |
| 12680 | let r = _mm512_maskz_subs_epu16(0, a, b); |
| 12681 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 12682 | let r = _mm512_maskz_subs_epu16(0b00000000_00000000_00000000_00001111, a, b); |
| 12683 | #[rustfmt::skip] |
| 12684 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12685 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 12686 | assert_eq_m512i(r, e); |
| 12687 | } |
| 12688 | |
| 12689 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12690 | unsafe fn test_mm256_mask_subs_epu16() { |
| 12691 | let a = _mm256_set1_epi16(1); |
| 12692 | let b = _mm256_set1_epi16(u16::MAX as i16); |
| 12693 | let r = _mm256_mask_subs_epu16(a, 0, a, b); |
| 12694 | assert_eq_m256i(r, a); |
| 12695 | let r = _mm256_mask_subs_epu16(a, 0b00000000_00001111, a, b); |
| 12696 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 12697 | assert_eq_m256i(r, e); |
| 12698 | } |
| 12699 | |
| 12700 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12701 | unsafe fn test_mm256_maskz_subs_epu16() { |
| 12702 | let a = _mm256_set1_epi16(1); |
| 12703 | let b = _mm256_set1_epi16(u16::MAX as i16); |
| 12704 | let r = _mm256_maskz_subs_epu16(0, a, b); |
| 12705 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 12706 | let r = _mm256_maskz_subs_epu16(0b00000000_00001111, a, b); |
| 12707 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 12708 | assert_eq_m256i(r, e); |
| 12709 | } |
| 12710 | |
| 12711 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12712 | unsafe fn test_mm_mask_subs_epu16() { |
| 12713 | let a = _mm_set1_epi16(1); |
| 12714 | let b = _mm_set1_epi16(u16::MAX as i16); |
| 12715 | let r = _mm_mask_subs_epu16(a, 0, a, b); |
| 12716 | assert_eq_m128i(r, a); |
| 12717 | let r = _mm_mask_subs_epu16(a, 0b00001111, a, b); |
| 12718 | let e = _mm_set_epi16(1, 1, 1, 1, 0, 0, 0, 0); |
| 12719 | assert_eq_m128i(r, e); |
| 12720 | } |
| 12721 | |
| 12722 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12723 | unsafe fn test_mm_maskz_subs_epu16() { |
| 12724 | let a = _mm_set1_epi16(1); |
| 12725 | let b = _mm_set1_epi16(u16::MAX as i16); |
| 12726 | let r = _mm_maskz_subs_epu16(0, a, b); |
| 12727 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 12728 | let r = _mm_maskz_subs_epu16(0b00001111, a, b); |
| 12729 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 0, 0); |
| 12730 | assert_eq_m128i(r, e); |
| 12731 | } |
| 12732 | |
| 12733 | #[simd_test(enable = "avx512bw" )] |
| 12734 | unsafe fn test_mm512_subs_epu8() { |
| 12735 | let a = _mm512_set1_epi8(1); |
| 12736 | let b = _mm512_set1_epi8(u8::MAX as i8); |
| 12737 | let r = _mm512_subs_epu8(a, b); |
| 12738 | let e = _mm512_set1_epi8(0); |
| 12739 | assert_eq_m512i(r, e); |
| 12740 | } |
| 12741 | |
| 12742 | #[simd_test(enable = "avx512bw" )] |
| 12743 | unsafe fn test_mm512_mask_subs_epu8() { |
| 12744 | let a = _mm512_set1_epi8(1); |
| 12745 | let b = _mm512_set1_epi8(u8::MAX as i8); |
| 12746 | let r = _mm512_mask_subs_epu8(a, 0, a, b); |
| 12747 | assert_eq_m512i(r, a); |
| 12748 | let r = _mm512_mask_subs_epu8( |
| 12749 | a, |
| 12750 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 12751 | a, |
| 12752 | b, |
| 12753 | ); |
| 12754 | #[rustfmt::skip] |
| 12755 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12756 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12757 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12758 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 12759 | assert_eq_m512i(r, e); |
| 12760 | } |
| 12761 | |
| 12762 | #[simd_test(enable = "avx512bw" )] |
| 12763 | unsafe fn test_mm512_maskz_subs_epu8() { |
| 12764 | let a = _mm512_set1_epi8(1); |
| 12765 | let b = _mm512_set1_epi8(u8::MAX as i8); |
| 12766 | let r = _mm512_maskz_subs_epu8(0, a, b); |
| 12767 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 12768 | let r = _mm512_maskz_subs_epu8( |
| 12769 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 12770 | a, |
| 12771 | b, |
| 12772 | ); |
| 12773 | #[rustfmt::skip] |
| 12774 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12775 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12776 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12777 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 12778 | assert_eq_m512i(r, e); |
| 12779 | } |
| 12780 | |
| 12781 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12782 | unsafe fn test_mm256_mask_subs_epu8() { |
| 12783 | let a = _mm256_set1_epi8(1); |
| 12784 | let b = _mm256_set1_epi8(u8::MAX as i8); |
| 12785 | let r = _mm256_mask_subs_epu8(a, 0, a, b); |
| 12786 | assert_eq_m256i(r, a); |
| 12787 | let r = _mm256_mask_subs_epu8(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 12788 | #[rustfmt::skip] |
| 12789 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 12790 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 12791 | assert_eq_m256i(r, e); |
| 12792 | } |
| 12793 | |
| 12794 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12795 | unsafe fn test_mm256_maskz_subs_epu8() { |
| 12796 | let a = _mm256_set1_epi8(1); |
| 12797 | let b = _mm256_set1_epi8(u8::MAX as i8); |
| 12798 | let r = _mm256_maskz_subs_epu8(0, a, b); |
| 12799 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 12800 | let r = _mm256_maskz_subs_epu8(0b00000000_00000000_00000000_00001111, a, b); |
| 12801 | #[rustfmt::skip] |
| 12802 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12803 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 12804 | assert_eq_m256i(r, e); |
| 12805 | } |
| 12806 | |
| 12807 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12808 | unsafe fn test_mm_mask_subs_epu8() { |
| 12809 | let a = _mm_set1_epi8(1); |
| 12810 | let b = _mm_set1_epi8(u8::MAX as i8); |
| 12811 | let r = _mm_mask_subs_epu8(a, 0, a, b); |
| 12812 | assert_eq_m128i(r, a); |
| 12813 | let r = _mm_mask_subs_epu8(a, 0b00000000_00001111, a, b); |
| 12814 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 12815 | assert_eq_m128i(r, e); |
| 12816 | } |
| 12817 | |
| 12818 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12819 | unsafe fn test_mm_maskz_subs_epu8() { |
| 12820 | let a = _mm_set1_epi8(1); |
| 12821 | let b = _mm_set1_epi8(u8::MAX as i8); |
| 12822 | let r = _mm_maskz_subs_epu8(0, a, b); |
| 12823 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 12824 | let r = _mm_maskz_subs_epu8(0b00000000_00001111, a, b); |
| 12825 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 12826 | assert_eq_m128i(r, e); |
| 12827 | } |
| 12828 | |
| 12829 | #[simd_test(enable = "avx512bw" )] |
| 12830 | unsafe fn test_mm512_subs_epi16() { |
| 12831 | let a = _mm512_set1_epi16(-1); |
| 12832 | let b = _mm512_set1_epi16(i16::MAX); |
| 12833 | let r = _mm512_subs_epi16(a, b); |
| 12834 | let e = _mm512_set1_epi16(i16::MIN); |
| 12835 | assert_eq_m512i(r, e); |
| 12836 | } |
| 12837 | |
| 12838 | #[simd_test(enable = "avx512bw" )] |
| 12839 | unsafe fn test_mm512_mask_subs_epi16() { |
| 12840 | let a = _mm512_set1_epi16(-1); |
| 12841 | let b = _mm512_set1_epi16(i16::MAX); |
| 12842 | let r = _mm512_mask_subs_epi16(a, 0, a, b); |
| 12843 | assert_eq_m512i(r, a); |
| 12844 | let r = _mm512_mask_subs_epi16(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 12845 | #[rustfmt::skip] |
| 12846 | let e = _mm512_set_epi16(-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12847 | -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, i16::MIN, i16::MIN, i16::MIN, i16::MIN); |
| 12848 | assert_eq_m512i(r, e); |
| 12849 | } |
| 12850 | |
| 12851 | #[simd_test(enable = "avx512bw" )] |
| 12852 | unsafe fn test_mm512_maskz_subs_epi16() { |
| 12853 | let a = _mm512_set1_epi16(-1); |
| 12854 | let b = _mm512_set1_epi16(i16::MAX); |
| 12855 | let r = _mm512_maskz_subs_epi16(0, a, b); |
| 12856 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 12857 | let r = _mm512_maskz_subs_epi16(0b00000000_00000000_00000000_00001111, a, b); |
| 12858 | #[rustfmt::skip] |
| 12859 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12860 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i16::MIN, i16::MIN, i16::MIN, i16::MIN); |
| 12861 | assert_eq_m512i(r, e); |
| 12862 | } |
| 12863 | |
| 12864 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12865 | unsafe fn test_mm256_mask_subs_epi16() { |
| 12866 | let a = _mm256_set1_epi16(-1); |
| 12867 | let b = _mm256_set1_epi16(i16::MAX); |
| 12868 | let r = _mm256_mask_subs_epi16(a, 0, a, b); |
| 12869 | assert_eq_m256i(r, a); |
| 12870 | let r = _mm256_mask_subs_epi16(a, 0b00000000_00001111, a, b); |
| 12871 | #[rustfmt::skip] |
| 12872 | let e = _mm256_set_epi16(-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, i16::MIN, i16::MIN, i16::MIN, i16::MIN); |
| 12873 | assert_eq_m256i(r, e); |
| 12874 | } |
| 12875 | |
| 12876 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12877 | unsafe fn test_mm256_maskz_subs_epi16() { |
| 12878 | let a = _mm256_set1_epi16(-1); |
| 12879 | let b = _mm256_set1_epi16(i16::MAX); |
| 12880 | let r = _mm256_maskz_subs_epi16(0, a, b); |
| 12881 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 12882 | let r = _mm256_maskz_subs_epi16(0b00000000_00001111, a, b); |
| 12883 | #[rustfmt::skip] |
| 12884 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i16::MIN, i16::MIN, i16::MIN, i16::MIN); |
| 12885 | assert_eq_m256i(r, e); |
| 12886 | } |
| 12887 | |
| 12888 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12889 | unsafe fn test_mm_mask_subs_epi16() { |
| 12890 | let a = _mm_set1_epi16(-1); |
| 12891 | let b = _mm_set1_epi16(i16::MAX); |
| 12892 | let r = _mm_mask_subs_epi16(a, 0, a, b); |
| 12893 | assert_eq_m128i(r, a); |
| 12894 | let r = _mm_mask_subs_epi16(a, 0b00001111, a, b); |
| 12895 | let e = _mm_set_epi16(-1, -1, -1, -1, i16::MIN, i16::MIN, i16::MIN, i16::MIN); |
| 12896 | assert_eq_m128i(r, e); |
| 12897 | } |
| 12898 | |
| 12899 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12900 | unsafe fn test_mm_maskz_subs_epi16() { |
| 12901 | let a = _mm_set1_epi16(-1); |
| 12902 | let b = _mm_set1_epi16(i16::MAX); |
| 12903 | let r = _mm_maskz_subs_epi16(0, a, b); |
| 12904 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 12905 | let r = _mm_maskz_subs_epi16(0b00001111, a, b); |
| 12906 | let e = _mm_set_epi16(0, 0, 0, 0, i16::MIN, i16::MIN, i16::MIN, i16::MIN); |
| 12907 | assert_eq_m128i(r, e); |
| 12908 | } |
| 12909 | |
| 12910 | #[simd_test(enable = "avx512bw" )] |
| 12911 | unsafe fn test_mm512_subs_epi8() { |
| 12912 | let a = _mm512_set1_epi8(-1); |
| 12913 | let b = _mm512_set1_epi8(i8::MAX); |
| 12914 | let r = _mm512_subs_epi8(a, b); |
| 12915 | let e = _mm512_set1_epi8(i8::MIN); |
| 12916 | assert_eq_m512i(r, e); |
| 12917 | } |
| 12918 | |
| 12919 | #[simd_test(enable = "avx512bw" )] |
| 12920 | unsafe fn test_mm512_mask_subs_epi8() { |
| 12921 | let a = _mm512_set1_epi8(-1); |
| 12922 | let b = _mm512_set1_epi8(i8::MAX); |
| 12923 | let r = _mm512_mask_subs_epi8(a, 0, a, b); |
| 12924 | assert_eq_m512i(r, a); |
| 12925 | let r = _mm512_mask_subs_epi8( |
| 12926 | a, |
| 12927 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 12928 | a, |
| 12929 | b, |
| 12930 | ); |
| 12931 | #[rustfmt::skip] |
| 12932 | let e = _mm512_set_epi8(-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12933 | -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12934 | -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12935 | -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, i8::MIN, i8::MIN, i8::MIN, i8::MIN); |
| 12936 | assert_eq_m512i(r, e); |
| 12937 | } |
| 12938 | |
| 12939 | #[simd_test(enable = "avx512bw" )] |
| 12940 | unsafe fn test_mm512_maskz_subs_epi8() { |
| 12941 | let a = _mm512_set1_epi8(-1); |
| 12942 | let b = _mm512_set1_epi8(i8::MAX); |
| 12943 | let r = _mm512_maskz_subs_epi8(0, a, b); |
| 12944 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 12945 | let r = _mm512_maskz_subs_epi8( |
| 12946 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 12947 | a, |
| 12948 | b, |
| 12949 | ); |
| 12950 | #[rustfmt::skip] |
| 12951 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12952 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12953 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12954 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MIN, i8::MIN, i8::MIN, i8::MIN); |
| 12955 | assert_eq_m512i(r, e); |
| 12956 | } |
| 12957 | |
| 12958 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12959 | unsafe fn test_mm256_mask_subs_epi8() { |
| 12960 | let a = _mm256_set1_epi8(-1); |
| 12961 | let b = _mm256_set1_epi8(i8::MAX); |
| 12962 | let r = _mm256_mask_subs_epi8(a, 0, a, b); |
| 12963 | assert_eq_m256i(r, a); |
| 12964 | let r = _mm256_mask_subs_epi8(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 12965 | #[rustfmt::skip] |
| 12966 | let e = _mm256_set_epi8(-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, |
| 12967 | -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, i8::MIN, i8::MIN, i8::MIN, i8::MIN); |
| 12968 | assert_eq_m256i(r, e); |
| 12969 | } |
| 12970 | |
| 12971 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12972 | unsafe fn test_mm256_maskz_subs_epi8() { |
| 12973 | let a = _mm256_set1_epi8(-1); |
| 12974 | let b = _mm256_set1_epi8(i8::MAX); |
| 12975 | let r = _mm256_maskz_subs_epi8(0, a, b); |
| 12976 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 12977 | let r = _mm256_maskz_subs_epi8(0b00000000_00000000_00000000_00001111, a, b); |
| 12978 | #[rustfmt::skip] |
| 12979 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 12980 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MIN, i8::MIN, i8::MIN, i8::MIN); |
| 12981 | assert_eq_m256i(r, e); |
| 12982 | } |
| 12983 | |
| 12984 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12985 | unsafe fn test_mm_mask_subs_epi8() { |
| 12986 | let a = _mm_set1_epi8(-1); |
| 12987 | let b = _mm_set1_epi8(i8::MAX); |
| 12988 | let r = _mm_mask_subs_epi8(a, 0, a, b); |
| 12989 | assert_eq_m128i(r, a); |
| 12990 | let r = _mm_mask_subs_epi8(a, 0b00000000_00001111, a, b); |
| 12991 | #[rustfmt::skip] |
| 12992 | let e = _mm_set_epi8(-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, i8::MIN, i8::MIN, i8::MIN, i8::MIN); |
| 12993 | assert_eq_m128i(r, e); |
| 12994 | } |
| 12995 | |
| 12996 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 12997 | unsafe fn test_mm_maskz_subs_epi8() { |
| 12998 | let a = _mm_set1_epi8(-1); |
| 12999 | let b = _mm_set1_epi8(i8::MAX); |
| 13000 | let r = _mm_maskz_subs_epi8(0, a, b); |
| 13001 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 13002 | let r = _mm_maskz_subs_epi8(0b00000000_00001111, a, b); |
| 13003 | #[rustfmt::skip] |
| 13004 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MIN, i8::MIN, i8::MIN, i8::MIN); |
| 13005 | assert_eq_m128i(r, e); |
| 13006 | } |
| 13007 | |
| 13008 | #[simd_test(enable = "avx512bw" )] |
| 13009 | unsafe fn test_mm512_mulhi_epu16() { |
| 13010 | let a = _mm512_set1_epi16(1); |
| 13011 | let b = _mm512_set1_epi16(1); |
| 13012 | let r = _mm512_mulhi_epu16(a, b); |
| 13013 | let e = _mm512_set1_epi16(0); |
| 13014 | assert_eq_m512i(r, e); |
| 13015 | } |
| 13016 | |
| 13017 | #[simd_test(enable = "avx512bw" )] |
| 13018 | unsafe fn test_mm512_mask_mulhi_epu16() { |
| 13019 | let a = _mm512_set1_epi16(1); |
| 13020 | let b = _mm512_set1_epi16(1); |
| 13021 | let r = _mm512_mask_mulhi_epu16(a, 0, a, b); |
| 13022 | assert_eq_m512i(r, a); |
| 13023 | let r = _mm512_mask_mulhi_epu16(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 13024 | #[rustfmt::skip] |
| 13025 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 13026 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 13027 | assert_eq_m512i(r, e); |
| 13028 | } |
| 13029 | |
| 13030 | #[simd_test(enable = "avx512bw" )] |
| 13031 | unsafe fn test_mm512_maskz_mulhi_epu16() { |
| 13032 | let a = _mm512_set1_epi16(1); |
| 13033 | let b = _mm512_set1_epi16(1); |
| 13034 | let r = _mm512_maskz_mulhi_epu16(0, a, b); |
| 13035 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 13036 | let r = _mm512_maskz_mulhi_epu16(0b00000000_00000000_00000000_00001111, a, b); |
| 13037 | #[rustfmt::skip] |
| 13038 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 13039 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 13040 | assert_eq_m512i(r, e); |
| 13041 | } |
| 13042 | |
| 13043 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13044 | unsafe fn test_mm256_mask_mulhi_epu16() { |
| 13045 | let a = _mm256_set1_epi16(1); |
| 13046 | let b = _mm256_set1_epi16(1); |
| 13047 | let r = _mm256_mask_mulhi_epu16(a, 0, a, b); |
| 13048 | assert_eq_m256i(r, a); |
| 13049 | let r = _mm256_mask_mulhi_epu16(a, 0b00000000_00001111, a, b); |
| 13050 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 13051 | assert_eq_m256i(r, e); |
| 13052 | } |
| 13053 | |
| 13054 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13055 | unsafe fn test_mm256_maskz_mulhi_epu16() { |
| 13056 | let a = _mm256_set1_epi16(1); |
| 13057 | let b = _mm256_set1_epi16(1); |
| 13058 | let r = _mm256_maskz_mulhi_epu16(0, a, b); |
| 13059 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 13060 | let r = _mm256_maskz_mulhi_epu16(0b00000000_00001111, a, b); |
| 13061 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 13062 | assert_eq_m256i(r, e); |
| 13063 | } |
| 13064 | |
| 13065 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13066 | unsafe fn test_mm_mask_mulhi_epu16() { |
| 13067 | let a = _mm_set1_epi16(1); |
| 13068 | let b = _mm_set1_epi16(1); |
| 13069 | let r = _mm_mask_mulhi_epu16(a, 0, a, b); |
| 13070 | assert_eq_m128i(r, a); |
| 13071 | let r = _mm_mask_mulhi_epu16(a, 0b00001111, a, b); |
| 13072 | let e = _mm_set_epi16(1, 1, 1, 1, 0, 0, 0, 0); |
| 13073 | assert_eq_m128i(r, e); |
| 13074 | } |
| 13075 | |
| 13076 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13077 | unsafe fn test_mm_maskz_mulhi_epu16() { |
| 13078 | let a = _mm_set1_epi16(1); |
| 13079 | let b = _mm_set1_epi16(1); |
| 13080 | let r = _mm_maskz_mulhi_epu16(0, a, b); |
| 13081 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 13082 | let r = _mm_maskz_mulhi_epu16(0b00001111, a, b); |
| 13083 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 0, 0); |
| 13084 | assert_eq_m128i(r, e); |
| 13085 | } |
| 13086 | |
| 13087 | #[simd_test(enable = "avx512bw" )] |
| 13088 | unsafe fn test_mm512_mulhi_epi16() { |
| 13089 | let a = _mm512_set1_epi16(1); |
| 13090 | let b = _mm512_set1_epi16(1); |
| 13091 | let r = _mm512_mulhi_epi16(a, b); |
| 13092 | let e = _mm512_set1_epi16(0); |
| 13093 | assert_eq_m512i(r, e); |
| 13094 | } |
| 13095 | |
| 13096 | #[simd_test(enable = "avx512bw" )] |
| 13097 | unsafe fn test_mm512_mask_mulhi_epi16() { |
| 13098 | let a = _mm512_set1_epi16(1); |
| 13099 | let b = _mm512_set1_epi16(1); |
| 13100 | let r = _mm512_mask_mulhi_epi16(a, 0, a, b); |
| 13101 | assert_eq_m512i(r, a); |
| 13102 | let r = _mm512_mask_mulhi_epi16(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 13103 | #[rustfmt::skip] |
| 13104 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 13105 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 13106 | assert_eq_m512i(r, e); |
| 13107 | } |
| 13108 | |
| 13109 | #[simd_test(enable = "avx512bw" )] |
| 13110 | unsafe fn test_mm512_maskz_mulhi_epi16() { |
| 13111 | let a = _mm512_set1_epi16(1); |
| 13112 | let b = _mm512_set1_epi16(1); |
| 13113 | let r = _mm512_maskz_mulhi_epi16(0, a, b); |
| 13114 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 13115 | let r = _mm512_maskz_mulhi_epi16(0b00000000_00000000_00000000_00001111, a, b); |
| 13116 | #[rustfmt::skip] |
| 13117 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 13118 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 13119 | assert_eq_m512i(r, e); |
| 13120 | } |
| 13121 | |
| 13122 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13123 | unsafe fn test_mm256_mask_mulhi_epi16() { |
| 13124 | let a = _mm256_set1_epi16(1); |
| 13125 | let b = _mm256_set1_epi16(1); |
| 13126 | let r = _mm256_mask_mulhi_epi16(a, 0, a, b); |
| 13127 | assert_eq_m256i(r, a); |
| 13128 | let r = _mm256_mask_mulhi_epi16(a, 0b00000000_00001111, a, b); |
| 13129 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 13130 | assert_eq_m256i(r, e); |
| 13131 | } |
| 13132 | |
| 13133 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13134 | unsafe fn test_mm256_maskz_mulhi_epi16() { |
| 13135 | let a = _mm256_set1_epi16(1); |
| 13136 | let b = _mm256_set1_epi16(1); |
| 13137 | let r = _mm256_maskz_mulhi_epi16(0, a, b); |
| 13138 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 13139 | let r = _mm256_maskz_mulhi_epi16(0b00000000_00001111, a, b); |
| 13140 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 13141 | assert_eq_m256i(r, e); |
| 13142 | } |
| 13143 | |
| 13144 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13145 | unsafe fn test_mm_mask_mulhi_epi16() { |
| 13146 | let a = _mm_set1_epi16(1); |
| 13147 | let b = _mm_set1_epi16(1); |
| 13148 | let r = _mm_mask_mulhi_epi16(a, 0, a, b); |
| 13149 | assert_eq_m128i(r, a); |
| 13150 | let r = _mm_mask_mulhi_epi16(a, 0b00001111, a, b); |
| 13151 | let e = _mm_set_epi16(1, 1, 1, 1, 0, 0, 0, 0); |
| 13152 | assert_eq_m128i(r, e); |
| 13153 | } |
| 13154 | |
| 13155 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13156 | unsafe fn test_mm_maskz_mulhi_epi16() { |
| 13157 | let a = _mm_set1_epi16(1); |
| 13158 | let b = _mm_set1_epi16(1); |
| 13159 | let r = _mm_maskz_mulhi_epi16(0, a, b); |
| 13160 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 13161 | let r = _mm_maskz_mulhi_epi16(0b00001111, a, b); |
| 13162 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 0, 0); |
| 13163 | assert_eq_m128i(r, e); |
| 13164 | } |
| 13165 | |
| 13166 | #[simd_test(enable = "avx512bw" )] |
| 13167 | unsafe fn test_mm512_mulhrs_epi16() { |
| 13168 | let a = _mm512_set1_epi16(1); |
| 13169 | let b = _mm512_set1_epi16(1); |
| 13170 | let r = _mm512_mulhrs_epi16(a, b); |
| 13171 | let e = _mm512_set1_epi16(0); |
| 13172 | assert_eq_m512i(r, e); |
| 13173 | } |
| 13174 | |
| 13175 | #[simd_test(enable = "avx512bw" )] |
| 13176 | unsafe fn test_mm512_mask_mulhrs_epi16() { |
| 13177 | let a = _mm512_set1_epi16(1); |
| 13178 | let b = _mm512_set1_epi16(1); |
| 13179 | let r = _mm512_mask_mulhrs_epi16(a, 0, a, b); |
| 13180 | assert_eq_m512i(r, a); |
| 13181 | let r = _mm512_mask_mulhrs_epi16(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 13182 | #[rustfmt::skip] |
| 13183 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 13184 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 13185 | assert_eq_m512i(r, e); |
| 13186 | } |
| 13187 | |
| 13188 | #[simd_test(enable = "avx512bw" )] |
| 13189 | unsafe fn test_mm512_maskz_mulhrs_epi16() { |
| 13190 | let a = _mm512_set1_epi16(1); |
| 13191 | let b = _mm512_set1_epi16(1); |
| 13192 | let r = _mm512_maskz_mulhrs_epi16(0, a, b); |
| 13193 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 13194 | let r = _mm512_maskz_mulhrs_epi16(0b00000000_00000000_00000000_00001111, a, b); |
| 13195 | #[rustfmt::skip] |
| 13196 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 13197 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 13198 | assert_eq_m512i(r, e); |
| 13199 | } |
| 13200 | |
| 13201 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13202 | unsafe fn test_mm256_mask_mulhrs_epi16() { |
| 13203 | let a = _mm256_set1_epi16(1); |
| 13204 | let b = _mm256_set1_epi16(1); |
| 13205 | let r = _mm256_mask_mulhrs_epi16(a, 0, a, b); |
| 13206 | assert_eq_m256i(r, a); |
| 13207 | let r = _mm256_mask_mulhrs_epi16(a, 0b00000000_00001111, a, b); |
| 13208 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 13209 | assert_eq_m256i(r, e); |
| 13210 | } |
| 13211 | |
| 13212 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13213 | unsafe fn test_mm256_maskz_mulhrs_epi16() { |
| 13214 | let a = _mm256_set1_epi16(1); |
| 13215 | let b = _mm256_set1_epi16(1); |
| 13216 | let r = _mm256_maskz_mulhrs_epi16(0, a, b); |
| 13217 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 13218 | let r = _mm256_maskz_mulhrs_epi16(0b00000000_00001111, a, b); |
| 13219 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 13220 | assert_eq_m256i(r, e); |
| 13221 | } |
| 13222 | |
| 13223 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13224 | unsafe fn test_mm_mask_mulhrs_epi16() { |
| 13225 | let a = _mm_set1_epi16(1); |
| 13226 | let b = _mm_set1_epi16(1); |
| 13227 | let r = _mm_mask_mulhrs_epi16(a, 0, a, b); |
| 13228 | assert_eq_m128i(r, a); |
| 13229 | let r = _mm_mask_mulhrs_epi16(a, 0b00001111, a, b); |
| 13230 | let e = _mm_set_epi16(1, 1, 1, 1, 0, 0, 0, 0); |
| 13231 | assert_eq_m128i(r, e); |
| 13232 | } |
| 13233 | |
| 13234 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13235 | unsafe fn test_mm_maskz_mulhrs_epi16() { |
| 13236 | let a = _mm_set1_epi16(1); |
| 13237 | let b = _mm_set1_epi16(1); |
| 13238 | let r = _mm_maskz_mulhrs_epi16(0, a, b); |
| 13239 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 13240 | let r = _mm_maskz_mulhrs_epi16(0b00001111, a, b); |
| 13241 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 0, 0); |
| 13242 | assert_eq_m128i(r, e); |
| 13243 | } |
| 13244 | |
| 13245 | #[simd_test(enable = "avx512bw" )] |
| 13246 | unsafe fn test_mm512_mullo_epi16() { |
| 13247 | let a = _mm512_set1_epi16(1); |
| 13248 | let b = _mm512_set1_epi16(1); |
| 13249 | let r = _mm512_mullo_epi16(a, b); |
| 13250 | let e = _mm512_set1_epi16(1); |
| 13251 | assert_eq_m512i(r, e); |
| 13252 | } |
| 13253 | |
| 13254 | #[simd_test(enable = "avx512bw" )] |
| 13255 | unsafe fn test_mm512_mask_mullo_epi16() { |
| 13256 | let a = _mm512_set1_epi16(1); |
| 13257 | let b = _mm512_set1_epi16(1); |
| 13258 | let r = _mm512_mask_mullo_epi16(a, 0, a, b); |
| 13259 | assert_eq_m512i(r, a); |
| 13260 | let r = _mm512_mask_mullo_epi16(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 13261 | #[rustfmt::skip] |
| 13262 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 13263 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
| 13264 | assert_eq_m512i(r, e); |
| 13265 | } |
| 13266 | |
| 13267 | #[simd_test(enable = "avx512bw" )] |
| 13268 | unsafe fn test_mm512_maskz_mullo_epi16() { |
| 13269 | let a = _mm512_set1_epi16(1); |
| 13270 | let b = _mm512_set1_epi16(1); |
| 13271 | let r = _mm512_maskz_mullo_epi16(0, a, b); |
| 13272 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 13273 | let r = _mm512_maskz_mullo_epi16(0b00000000_00000000_00000000_00001111, a, b); |
| 13274 | #[rustfmt::skip] |
| 13275 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 13276 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
| 13277 | assert_eq_m512i(r, e); |
| 13278 | } |
| 13279 | |
| 13280 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13281 | unsafe fn test_mm256_mask_mullo_epi16() { |
| 13282 | let a = _mm256_set1_epi16(1); |
| 13283 | let b = _mm256_set1_epi16(1); |
| 13284 | let r = _mm256_mask_mullo_epi16(a, 0, a, b); |
| 13285 | assert_eq_m256i(r, a); |
| 13286 | let r = _mm256_mask_mullo_epi16(a, 0b00000000_00001111, a, b); |
| 13287 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
| 13288 | assert_eq_m256i(r, e); |
| 13289 | } |
| 13290 | |
| 13291 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13292 | unsafe fn test_mm256_maskz_mullo_epi16() { |
| 13293 | let a = _mm256_set1_epi16(1); |
| 13294 | let b = _mm256_set1_epi16(1); |
| 13295 | let r = _mm256_maskz_mullo_epi16(0, a, b); |
| 13296 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 13297 | let r = _mm256_maskz_mullo_epi16(0b00000000_00001111, a, b); |
| 13298 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
| 13299 | assert_eq_m256i(r, e); |
| 13300 | } |
| 13301 | |
| 13302 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13303 | unsafe fn test_mm_mask_mullo_epi16() { |
| 13304 | let a = _mm_set1_epi16(1); |
| 13305 | let b = _mm_set1_epi16(1); |
| 13306 | let r = _mm_mask_mullo_epi16(a, 0, a, b); |
| 13307 | assert_eq_m128i(r, a); |
| 13308 | let r = _mm_mask_mullo_epi16(a, 0b00001111, a, b); |
| 13309 | let e = _mm_set_epi16(1, 1, 1, 1, 1, 1, 1, 1); |
| 13310 | assert_eq_m128i(r, e); |
| 13311 | } |
| 13312 | |
| 13313 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13314 | unsafe fn test_mm_maskz_mullo_epi16() { |
| 13315 | let a = _mm_set1_epi16(1); |
| 13316 | let b = _mm_set1_epi16(1); |
| 13317 | let r = _mm_maskz_mullo_epi16(0, a, b); |
| 13318 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 13319 | let r = _mm_maskz_mullo_epi16(0b00001111, a, b); |
| 13320 | let e = _mm_set_epi16(0, 0, 0, 0, 1, 1, 1, 1); |
| 13321 | assert_eq_m128i(r, e); |
| 13322 | } |
| 13323 | |
| 13324 | #[simd_test(enable = "avx512bw" )] |
| 13325 | unsafe fn test_mm512_max_epu16() { |
| 13326 | #[rustfmt::skip] |
| 13327 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13328 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13329 | #[rustfmt::skip] |
| 13330 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13331 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13332 | let r = _mm512_max_epu16(a, b); |
| 13333 | #[rustfmt::skip] |
| 13334 | let e = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13335 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13336 | assert_eq_m512i(r, e); |
| 13337 | } |
| 13338 | |
| 13339 | #[simd_test(enable = "avx512bw" )] |
| 13340 | unsafe fn test_mm512_mask_max_epu16() { |
| 13341 | #[rustfmt::skip] |
| 13342 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13343 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13344 | #[rustfmt::skip] |
| 13345 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13346 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13347 | let r = _mm512_mask_max_epu16(a, 0, a, b); |
| 13348 | assert_eq_m512i(r, a); |
| 13349 | let r = _mm512_mask_max_epu16(a, 0b00000000_11111111_00000000_11111111, a, b); |
| 13350 | #[rustfmt::skip] |
| 13351 | let e = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13352 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13353 | assert_eq_m512i(r, e); |
| 13354 | } |
| 13355 | |
| 13356 | #[simd_test(enable = "avx512bw" )] |
| 13357 | unsafe fn test_mm512_maskz_max_epu16() { |
| 13358 | #[rustfmt::skip] |
| 13359 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13360 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13361 | #[rustfmt::skip] |
| 13362 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13363 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13364 | let r = _mm512_maskz_max_epu16(0, a, b); |
| 13365 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 13366 | let r = _mm512_maskz_max_epu16(0b00000000_11111111_00000000_11111111, a, b); |
| 13367 | #[rustfmt::skip] |
| 13368 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13369 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13370 | assert_eq_m512i(r, e); |
| 13371 | } |
| 13372 | |
| 13373 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13374 | unsafe fn test_mm256_mask_max_epu16() { |
| 13375 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13376 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13377 | let r = _mm256_mask_max_epu16(a, 0, a, b); |
| 13378 | assert_eq_m256i(r, a); |
| 13379 | let r = _mm256_mask_max_epu16(a, 0b00000000_11111111, a, b); |
| 13380 | let e = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13381 | assert_eq_m256i(r, e); |
| 13382 | } |
| 13383 | |
| 13384 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13385 | unsafe fn test_mm256_maskz_max_epu16() { |
| 13386 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13387 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13388 | let r = _mm256_maskz_max_epu16(0, a, b); |
| 13389 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 13390 | let r = _mm256_maskz_max_epu16(0b00000000_11111111, a, b); |
| 13391 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13392 | assert_eq_m256i(r, e); |
| 13393 | } |
| 13394 | |
| 13395 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13396 | unsafe fn test_mm_mask_max_epu16() { |
| 13397 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 13398 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
| 13399 | let r = _mm_mask_max_epu16(a, 0, a, b); |
| 13400 | assert_eq_m128i(r, a); |
| 13401 | let r = _mm_mask_max_epu16(a, 0b00001111, a, b); |
| 13402 | let e = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 13403 | assert_eq_m128i(r, e); |
| 13404 | } |
| 13405 | |
| 13406 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13407 | unsafe fn test_mm_maskz_max_epu16() { |
| 13408 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 13409 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
| 13410 | let r = _mm_maskz_max_epu16(0, a, b); |
| 13411 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 13412 | let r = _mm_maskz_max_epu16(0b00001111, a, b); |
| 13413 | let e = _mm_set_epi16(0, 0, 0, 0, 4, 5, 6, 7); |
| 13414 | assert_eq_m128i(r, e); |
| 13415 | } |
| 13416 | |
| 13417 | #[simd_test(enable = "avx512bw" )] |
| 13418 | unsafe fn test_mm512_max_epu8() { |
| 13419 | #[rustfmt::skip] |
| 13420 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13421 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13422 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13423 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13424 | #[rustfmt::skip] |
| 13425 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13426 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13427 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13428 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13429 | let r = _mm512_max_epu8(a, b); |
| 13430 | #[rustfmt::skip] |
| 13431 | let e = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13432 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13433 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13434 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13435 | assert_eq_m512i(r, e); |
| 13436 | } |
| 13437 | |
| 13438 | #[simd_test(enable = "avx512bw" )] |
| 13439 | unsafe fn test_mm512_mask_max_epu8() { |
| 13440 | #[rustfmt::skip] |
| 13441 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13442 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13443 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13444 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13445 | #[rustfmt::skip] |
| 13446 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13447 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13448 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13449 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13450 | let r = _mm512_mask_max_epu8(a, 0, a, b); |
| 13451 | assert_eq_m512i(r, a); |
| 13452 | let r = _mm512_mask_max_epu8( |
| 13453 | a, |
| 13454 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 13455 | a, |
| 13456 | b, |
| 13457 | ); |
| 13458 | #[rustfmt::skip] |
| 13459 | let e = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13460 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13461 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13462 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13463 | assert_eq_m512i(r, e); |
| 13464 | } |
| 13465 | |
| 13466 | #[simd_test(enable = "avx512bw" )] |
| 13467 | unsafe fn test_mm512_maskz_max_epu8() { |
| 13468 | #[rustfmt::skip] |
| 13469 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13470 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13471 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13472 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13473 | #[rustfmt::skip] |
| 13474 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13475 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13476 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13477 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13478 | let r = _mm512_maskz_max_epu8(0, a, b); |
| 13479 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 13480 | let r = _mm512_maskz_max_epu8( |
| 13481 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 13482 | a, |
| 13483 | b, |
| 13484 | ); |
| 13485 | #[rustfmt::skip] |
| 13486 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13487 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13488 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13489 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13490 | assert_eq_m512i(r, e); |
| 13491 | } |
| 13492 | |
| 13493 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13494 | unsafe fn test_mm256_mask_max_epu8() { |
| 13495 | #[rustfmt::skip] |
| 13496 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13497 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13498 | #[rustfmt::skip] |
| 13499 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13500 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13501 | let r = _mm256_mask_max_epu8(a, 0, a, b); |
| 13502 | assert_eq_m256i(r, a); |
| 13503 | let r = _mm256_mask_max_epu8(a, 0b00000000_11111111_00000000_11111111, a, b); |
| 13504 | #[rustfmt::skip] |
| 13505 | let e = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13506 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13507 | assert_eq_m256i(r, e); |
| 13508 | } |
| 13509 | |
| 13510 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13511 | unsafe fn test_mm256_maskz_max_epu8() { |
| 13512 | #[rustfmt::skip] |
| 13513 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13514 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13515 | #[rustfmt::skip] |
| 13516 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13517 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13518 | let r = _mm256_maskz_max_epu8(0, a, b); |
| 13519 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 13520 | let r = _mm256_maskz_max_epu8(0b00000000_11111111_00000000_11111111, a, b); |
| 13521 | #[rustfmt::skip] |
| 13522 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13523 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13524 | assert_eq_m256i(r, e); |
| 13525 | } |
| 13526 | |
| 13527 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13528 | unsafe fn test_mm_mask_max_epu8() { |
| 13529 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13530 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13531 | let r = _mm_mask_max_epu8(a, 0, a, b); |
| 13532 | assert_eq_m128i(r, a); |
| 13533 | let r = _mm_mask_max_epu8(a, 0b00000000_11111111, a, b); |
| 13534 | let e = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13535 | assert_eq_m128i(r, e); |
| 13536 | } |
| 13537 | |
| 13538 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13539 | unsafe fn test_mm_maskz_max_epu8() { |
| 13540 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13541 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13542 | let r = _mm_maskz_max_epu8(0, a, b); |
| 13543 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 13544 | let r = _mm_maskz_max_epu8(0b00000000_11111111, a, b); |
| 13545 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13546 | assert_eq_m128i(r, e); |
| 13547 | } |
| 13548 | |
| 13549 | #[simd_test(enable = "avx512bw" )] |
| 13550 | unsafe fn test_mm512_max_epi16() { |
| 13551 | #[rustfmt::skip] |
| 13552 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13553 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13554 | #[rustfmt::skip] |
| 13555 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13556 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13557 | let r = _mm512_max_epi16(a, b); |
| 13558 | #[rustfmt::skip] |
| 13559 | let e = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13560 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13561 | assert_eq_m512i(r, e); |
| 13562 | } |
| 13563 | |
| 13564 | #[simd_test(enable = "avx512bw" )] |
| 13565 | unsafe fn test_mm512_mask_max_epi16() { |
| 13566 | #[rustfmt::skip] |
| 13567 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13568 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13569 | #[rustfmt::skip] |
| 13570 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13571 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13572 | let r = _mm512_mask_max_epi16(a, 0, a, b); |
| 13573 | assert_eq_m512i(r, a); |
| 13574 | let r = _mm512_mask_max_epi16(a, 0b00000000_11111111_00000000_11111111, a, b); |
| 13575 | #[rustfmt::skip] |
| 13576 | let e = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13577 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13578 | assert_eq_m512i(r, e); |
| 13579 | } |
| 13580 | |
| 13581 | #[simd_test(enable = "avx512bw" )] |
| 13582 | unsafe fn test_mm512_maskz_max_epi16() { |
| 13583 | #[rustfmt::skip] |
| 13584 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13585 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13586 | #[rustfmt::skip] |
| 13587 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13588 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13589 | let r = _mm512_maskz_max_epi16(0, a, b); |
| 13590 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 13591 | let r = _mm512_maskz_max_epi16(0b00000000_11111111_00000000_11111111, a, b); |
| 13592 | #[rustfmt::skip] |
| 13593 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13594 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13595 | assert_eq_m512i(r, e); |
| 13596 | } |
| 13597 | |
| 13598 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13599 | unsafe fn test_mm256_mask_max_epi16() { |
| 13600 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13601 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13602 | let r = _mm256_mask_max_epi16(a, 0, a, b); |
| 13603 | assert_eq_m256i(r, a); |
| 13604 | let r = _mm256_mask_max_epi16(a, 0b00000000_11111111, a, b); |
| 13605 | let e = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13606 | assert_eq_m256i(r, e); |
| 13607 | } |
| 13608 | |
| 13609 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13610 | unsafe fn test_mm256_maskz_max_epi16() { |
| 13611 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13612 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13613 | let r = _mm256_maskz_max_epi16(0, a, b); |
| 13614 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 13615 | let r = _mm256_maskz_max_epi16(0b00000000_11111111, a, b); |
| 13616 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13617 | assert_eq_m256i(r, e); |
| 13618 | } |
| 13619 | |
| 13620 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13621 | unsafe fn test_mm_mask_max_epi16() { |
| 13622 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 13623 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
| 13624 | let r = _mm_mask_max_epi16(a, 0, a, b); |
| 13625 | assert_eq_m128i(r, a); |
| 13626 | let r = _mm_mask_max_epi16(a, 0b00001111, a, b); |
| 13627 | let e = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 13628 | assert_eq_m128i(r, e); |
| 13629 | } |
| 13630 | |
| 13631 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13632 | unsafe fn test_mm_maskz_max_epi16() { |
| 13633 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 13634 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
| 13635 | let r = _mm_maskz_max_epi16(0, a, b); |
| 13636 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 13637 | let r = _mm_maskz_max_epi16(0b00001111, a, b); |
| 13638 | let e = _mm_set_epi16(0, 0, 0, 0, 4, 5, 6, 7); |
| 13639 | assert_eq_m128i(r, e); |
| 13640 | } |
| 13641 | |
| 13642 | #[simd_test(enable = "avx512bw" )] |
| 13643 | unsafe fn test_mm512_max_epi8() { |
| 13644 | #[rustfmt::skip] |
| 13645 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13646 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13647 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13648 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13649 | #[rustfmt::skip] |
| 13650 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13651 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13652 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13653 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13654 | let r = _mm512_max_epi8(a, b); |
| 13655 | #[rustfmt::skip] |
| 13656 | let e = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13657 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13658 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13659 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13660 | assert_eq_m512i(r, e); |
| 13661 | } |
| 13662 | |
| 13663 | #[simd_test(enable = "avx512bw" )] |
| 13664 | unsafe fn test_mm512_mask_max_epi8() { |
| 13665 | #[rustfmt::skip] |
| 13666 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13667 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13668 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13669 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13670 | #[rustfmt::skip] |
| 13671 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13672 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13673 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13674 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13675 | let r = _mm512_mask_max_epi8(a, 0, a, b); |
| 13676 | assert_eq_m512i(r, a); |
| 13677 | let r = _mm512_mask_max_epi8( |
| 13678 | a, |
| 13679 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 13680 | a, |
| 13681 | b, |
| 13682 | ); |
| 13683 | #[rustfmt::skip] |
| 13684 | let e = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13685 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13686 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13687 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13688 | assert_eq_m512i(r, e); |
| 13689 | } |
| 13690 | |
| 13691 | #[simd_test(enable = "avx512bw" )] |
| 13692 | unsafe fn test_mm512_maskz_max_epi8() { |
| 13693 | #[rustfmt::skip] |
| 13694 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13695 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13696 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13697 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13698 | #[rustfmt::skip] |
| 13699 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13700 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13701 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13702 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13703 | let r = _mm512_maskz_max_epi8(0, a, b); |
| 13704 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 13705 | let r = _mm512_maskz_max_epi8( |
| 13706 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 13707 | a, |
| 13708 | b, |
| 13709 | ); |
| 13710 | #[rustfmt::skip] |
| 13711 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13712 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13713 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13714 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13715 | assert_eq_m512i(r, e); |
| 13716 | } |
| 13717 | |
| 13718 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13719 | unsafe fn test_mm256_mask_max_epi8() { |
| 13720 | #[rustfmt::skip] |
| 13721 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13722 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13723 | #[rustfmt::skip] |
| 13724 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13725 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13726 | let r = _mm256_mask_max_epi8(a, 0, a, b); |
| 13727 | assert_eq_m256i(r, a); |
| 13728 | let r = _mm256_mask_max_epi8(a, 0b00000000_11111111_00000000_11111111, a, b); |
| 13729 | #[rustfmt::skip] |
| 13730 | let e = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13731 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13732 | assert_eq_m256i(r, e); |
| 13733 | } |
| 13734 | |
| 13735 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13736 | unsafe fn test_mm256_maskz_max_epi8() { |
| 13737 | #[rustfmt::skip] |
| 13738 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13739 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13740 | #[rustfmt::skip] |
| 13741 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13742 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13743 | let r = _mm256_maskz_max_epi8(0, a, b); |
| 13744 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 13745 | let r = _mm256_maskz_max_epi8(0b00000000_11111111_00000000_11111111, a, b); |
| 13746 | #[rustfmt::skip] |
| 13747 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13748 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13749 | assert_eq_m256i(r, e); |
| 13750 | } |
| 13751 | |
| 13752 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13753 | unsafe fn test_mm_mask_max_epi8() { |
| 13754 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13755 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13756 | let r = _mm_mask_max_epi8(a, 0, a, b); |
| 13757 | assert_eq_m128i(r, a); |
| 13758 | let r = _mm_mask_max_epi8(a, 0b00000000_11111111, a, b); |
| 13759 | let e = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13760 | assert_eq_m128i(r, e); |
| 13761 | } |
| 13762 | |
| 13763 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13764 | unsafe fn test_mm_maskz_max_epi8() { |
| 13765 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13766 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13767 | let r = _mm_maskz_max_epi8(0, a, b); |
| 13768 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 13769 | let r = _mm_maskz_max_epi8(0b00000000_11111111, a, b); |
| 13770 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13771 | assert_eq_m128i(r, e); |
| 13772 | } |
| 13773 | |
| 13774 | #[simd_test(enable = "avx512bw" )] |
| 13775 | unsafe fn test_mm512_min_epu16() { |
| 13776 | #[rustfmt::skip] |
| 13777 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13778 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13779 | #[rustfmt::skip] |
| 13780 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13781 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13782 | let r = _mm512_min_epu16(a, b); |
| 13783 | #[rustfmt::skip] |
| 13784 | let e = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13785 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13786 | assert_eq_m512i(r, e); |
| 13787 | } |
| 13788 | |
| 13789 | #[simd_test(enable = "avx512bw" )] |
| 13790 | unsafe fn test_mm512_mask_min_epu16() { |
| 13791 | #[rustfmt::skip] |
| 13792 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13793 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13794 | #[rustfmt::skip] |
| 13795 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13796 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13797 | let r = _mm512_mask_min_epu16(a, 0, a, b); |
| 13798 | assert_eq_m512i(r, a); |
| 13799 | let r = _mm512_mask_min_epu16(a, 0b00000000_11111111_00000000_11111111, a, b); |
| 13800 | #[rustfmt::skip] |
| 13801 | let e = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13802 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13803 | assert_eq_m512i(r, e); |
| 13804 | } |
| 13805 | |
| 13806 | #[simd_test(enable = "avx512bw" )] |
| 13807 | unsafe fn test_mm512_maskz_min_epu16() { |
| 13808 | #[rustfmt::skip] |
| 13809 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13810 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13811 | #[rustfmt::skip] |
| 13812 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13813 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13814 | let r = _mm512_maskz_min_epu16(0, a, b); |
| 13815 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 13816 | let r = _mm512_maskz_min_epu16(0b00000000_11111111_00000000_11111111, a, b); |
| 13817 | #[rustfmt::skip] |
| 13818 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13819 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13820 | assert_eq_m512i(r, e); |
| 13821 | } |
| 13822 | |
| 13823 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13824 | unsafe fn test_mm256_mask_min_epu16() { |
| 13825 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13826 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13827 | let r = _mm256_mask_min_epu16(a, 0, a, b); |
| 13828 | assert_eq_m256i(r, a); |
| 13829 | let r = _mm256_mask_min_epu16(a, 0b00000000_11111111, a, b); |
| 13830 | let e = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13831 | assert_eq_m256i(r, e); |
| 13832 | } |
| 13833 | |
| 13834 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13835 | unsafe fn test_mm256_maskz_min_epu16() { |
| 13836 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13837 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13838 | let r = _mm256_maskz_min_epu16(0, a, b); |
| 13839 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 13840 | let r = _mm256_maskz_min_epu16(0b00000000_11111111, a, b); |
| 13841 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13842 | assert_eq_m256i(r, e); |
| 13843 | } |
| 13844 | |
| 13845 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13846 | unsafe fn test_mm_mask_min_epu16() { |
| 13847 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 13848 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
| 13849 | let r = _mm_mask_min_epu16(a, 0, a, b); |
| 13850 | assert_eq_m128i(r, a); |
| 13851 | let r = _mm_mask_min_epu16(a, 0b00001111, a, b); |
| 13852 | let e = _mm_set_epi16(0, 1, 2, 3, 3, 2, 1, 0); |
| 13853 | assert_eq_m128i(r, e); |
| 13854 | } |
| 13855 | |
| 13856 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13857 | unsafe fn test_mm_maskz_min_epu16() { |
| 13858 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 13859 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
| 13860 | let r = _mm_maskz_min_epu16(0, a, b); |
| 13861 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 13862 | let r = _mm_maskz_min_epu16(0b00001111, a, b); |
| 13863 | let e = _mm_set_epi16(0, 0, 0, 0, 3, 2, 1, 0); |
| 13864 | assert_eq_m128i(r, e); |
| 13865 | } |
| 13866 | |
| 13867 | #[simd_test(enable = "avx512bw" )] |
| 13868 | unsafe fn test_mm512_min_epu8() { |
| 13869 | #[rustfmt::skip] |
| 13870 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13871 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13872 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13873 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13874 | #[rustfmt::skip] |
| 13875 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13876 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13877 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13878 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13879 | let r = _mm512_min_epu8(a, b); |
| 13880 | #[rustfmt::skip] |
| 13881 | let e = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13882 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13883 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13884 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13885 | assert_eq_m512i(r, e); |
| 13886 | } |
| 13887 | |
| 13888 | #[simd_test(enable = "avx512bw" )] |
| 13889 | unsafe fn test_mm512_mask_min_epu8() { |
| 13890 | #[rustfmt::skip] |
| 13891 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13892 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13893 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13894 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13895 | #[rustfmt::skip] |
| 13896 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13897 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13898 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13899 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13900 | let r = _mm512_mask_min_epu8(a, 0, a, b); |
| 13901 | assert_eq_m512i(r, a); |
| 13902 | let r = _mm512_mask_min_epu8( |
| 13903 | a, |
| 13904 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 13905 | a, |
| 13906 | b, |
| 13907 | ); |
| 13908 | #[rustfmt::skip] |
| 13909 | let e = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13910 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13911 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13912 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13913 | assert_eq_m512i(r, e); |
| 13914 | } |
| 13915 | |
| 13916 | #[simd_test(enable = "avx512bw" )] |
| 13917 | unsafe fn test_mm512_maskz_min_epu8() { |
| 13918 | #[rustfmt::skip] |
| 13919 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13920 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13921 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13922 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13923 | #[rustfmt::skip] |
| 13924 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13925 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13926 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13927 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13928 | let r = _mm512_maskz_min_epu8(0, a, b); |
| 13929 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 13930 | let r = _mm512_maskz_min_epu8( |
| 13931 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 13932 | a, |
| 13933 | b, |
| 13934 | ); |
| 13935 | #[rustfmt::skip] |
| 13936 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13937 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13938 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13939 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13940 | assert_eq_m512i(r, e); |
| 13941 | } |
| 13942 | |
| 13943 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13944 | unsafe fn test_mm256_mask_min_epu8() { |
| 13945 | #[rustfmt::skip] |
| 13946 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13947 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13948 | #[rustfmt::skip] |
| 13949 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13950 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13951 | let r = _mm256_mask_min_epu8(a, 0, a, b); |
| 13952 | assert_eq_m256i(r, a); |
| 13953 | let r = _mm256_mask_min_epu8(a, 0b00000000_11111111_00000000_11111111, a, b); |
| 13954 | #[rustfmt::skip] |
| 13955 | let e = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13956 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13957 | assert_eq_m256i(r, e); |
| 13958 | } |
| 13959 | |
| 13960 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13961 | unsafe fn test_mm256_maskz_min_epu8() { |
| 13962 | #[rustfmt::skip] |
| 13963 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 13964 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13965 | #[rustfmt::skip] |
| 13966 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13967 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13968 | let r = _mm256_maskz_min_epu8(0, a, b); |
| 13969 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 13970 | let r = _mm256_maskz_min_epu8(0b00000000_11111111_00000000_11111111, a, b); |
| 13971 | #[rustfmt::skip] |
| 13972 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
| 13973 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13974 | assert_eq_m256i(r, e); |
| 13975 | } |
| 13976 | |
| 13977 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13978 | unsafe fn test_mm_mask_min_epu8() { |
| 13979 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13980 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13981 | let r = _mm_mask_min_epu8(a, 0, a, b); |
| 13982 | assert_eq_m128i(r, a); |
| 13983 | let r = _mm_mask_min_epu8(a, 0b00000000_11111111, a, b); |
| 13984 | let e = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13985 | assert_eq_m128i(r, e); |
| 13986 | } |
| 13987 | |
| 13988 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 13989 | unsafe fn test_mm_maskz_min_epu8() { |
| 13990 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 13991 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13992 | let r = _mm_maskz_min_epu8(0, a, b); |
| 13993 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 13994 | let r = _mm_maskz_min_epu8(0b00000000_11111111, a, b); |
| 13995 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
| 13996 | assert_eq_m128i(r, e); |
| 13997 | } |
| 13998 | |
| 13999 | #[simd_test(enable = "avx512bw" )] |
| 14000 | unsafe fn test_mm512_min_epi16() { |
| 14001 | #[rustfmt::skip] |
| 14002 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14003 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 14004 | #[rustfmt::skip] |
| 14005 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14006 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14007 | let r = _mm512_min_epi16(a, b); |
| 14008 | #[rustfmt::skip] |
| 14009 | let e = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14010 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14011 | assert_eq_m512i(r, e); |
| 14012 | } |
| 14013 | |
| 14014 | #[simd_test(enable = "avx512bw" )] |
| 14015 | unsafe fn test_mm512_mask_min_epi16() { |
| 14016 | #[rustfmt::skip] |
| 14017 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14018 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 14019 | #[rustfmt::skip] |
| 14020 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14021 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14022 | let r = _mm512_mask_min_epi16(a, 0, a, b); |
| 14023 | assert_eq_m512i(r, a); |
| 14024 | let r = _mm512_mask_min_epi16(a, 0b00000000_11111111_00000000_11111111, a, b); |
| 14025 | #[rustfmt::skip] |
| 14026 | let e = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14027 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14028 | assert_eq_m512i(r, e); |
| 14029 | } |
| 14030 | |
| 14031 | #[simd_test(enable = "avx512bw" )] |
| 14032 | unsafe fn test_mm512_maskz_min_epi16() { |
| 14033 | #[rustfmt::skip] |
| 14034 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14035 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 14036 | #[rustfmt::skip] |
| 14037 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14038 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14039 | let r = _mm512_maskz_min_epi16(0, a, b); |
| 14040 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 14041 | let r = _mm512_maskz_min_epi16(0b00000000_11111111_00000000_11111111, a, b); |
| 14042 | #[rustfmt::skip] |
| 14043 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14044 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14045 | assert_eq_m512i(r, e); |
| 14046 | } |
| 14047 | |
| 14048 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14049 | unsafe fn test_mm256_mask_min_epi16() { |
| 14050 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 14051 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14052 | let r = _mm256_mask_min_epi16(a, 0, a, b); |
| 14053 | assert_eq_m256i(r, a); |
| 14054 | let r = _mm256_mask_min_epi16(a, 0b00000000_11111111, a, b); |
| 14055 | let e = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14056 | assert_eq_m256i(r, e); |
| 14057 | } |
| 14058 | |
| 14059 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14060 | unsafe fn test_mm256_maskz_min_epi16() { |
| 14061 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 14062 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14063 | let r = _mm256_maskz_min_epi16(0, a, b); |
| 14064 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 14065 | let r = _mm256_maskz_min_epi16(0b00000000_11111111, a, b); |
| 14066 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14067 | assert_eq_m256i(r, e); |
| 14068 | } |
| 14069 | |
| 14070 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14071 | unsafe fn test_mm_mask_min_epi16() { |
| 14072 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 14073 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
| 14074 | let r = _mm_mask_min_epi16(a, 0, a, b); |
| 14075 | assert_eq_m128i(r, a); |
| 14076 | let r = _mm_mask_min_epi16(a, 0b00001111, a, b); |
| 14077 | let e = _mm_set_epi16(0, 1, 2, 3, 3, 2, 1, 0); |
| 14078 | assert_eq_m128i(r, e); |
| 14079 | } |
| 14080 | |
| 14081 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14082 | unsafe fn test_mm_maskz_min_epi16() { |
| 14083 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 14084 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
| 14085 | let r = _mm_maskz_min_epi16(0, a, b); |
| 14086 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 14087 | let r = _mm_maskz_min_epi16(0b00001111, a, b); |
| 14088 | let e = _mm_set_epi16(0, 0, 0, 0, 3, 2, 1, 0); |
| 14089 | assert_eq_m128i(r, e); |
| 14090 | } |
| 14091 | |
| 14092 | #[simd_test(enable = "avx512bw" )] |
| 14093 | unsafe fn test_mm512_min_epi8() { |
| 14094 | #[rustfmt::skip] |
| 14095 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14096 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14097 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14098 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 14099 | #[rustfmt::skip] |
| 14100 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14101 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14102 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14103 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14104 | let r = _mm512_min_epi8(a, b); |
| 14105 | #[rustfmt::skip] |
| 14106 | let e = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14107 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14108 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14109 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14110 | assert_eq_m512i(r, e); |
| 14111 | } |
| 14112 | |
| 14113 | #[simd_test(enable = "avx512bw" )] |
| 14114 | unsafe fn test_mm512_mask_min_epi8() { |
| 14115 | #[rustfmt::skip] |
| 14116 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14117 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14118 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14119 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 14120 | #[rustfmt::skip] |
| 14121 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14122 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14123 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14124 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14125 | let r = _mm512_mask_min_epi8(a, 0, a, b); |
| 14126 | assert_eq_m512i(r, a); |
| 14127 | let r = _mm512_mask_min_epi8( |
| 14128 | a, |
| 14129 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 14130 | a, |
| 14131 | b, |
| 14132 | ); |
| 14133 | #[rustfmt::skip] |
| 14134 | let e = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14135 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14136 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14137 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14138 | assert_eq_m512i(r, e); |
| 14139 | } |
| 14140 | |
| 14141 | #[simd_test(enable = "avx512bw" )] |
| 14142 | unsafe fn test_mm512_maskz_min_epi8() { |
| 14143 | #[rustfmt::skip] |
| 14144 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14145 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14146 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14147 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 14148 | #[rustfmt::skip] |
| 14149 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14150 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14151 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14152 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14153 | let r = _mm512_maskz_min_epi8(0, a, b); |
| 14154 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 14155 | let r = _mm512_maskz_min_epi8( |
| 14156 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
| 14157 | a, |
| 14158 | b, |
| 14159 | ); |
| 14160 | #[rustfmt::skip] |
| 14161 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14162 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14163 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14164 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14165 | assert_eq_m512i(r, e); |
| 14166 | } |
| 14167 | |
| 14168 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14169 | unsafe fn test_mm256_mask_min_epi8() { |
| 14170 | #[rustfmt::skip] |
| 14171 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14172 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 14173 | #[rustfmt::skip] |
| 14174 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14175 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14176 | let r = _mm256_mask_min_epi8(a, 0, a, b); |
| 14177 | assert_eq_m256i(r, a); |
| 14178 | let r = _mm256_mask_min_epi8(a, 0b00000000_11111111_00000000_11111111, a, b); |
| 14179 | #[rustfmt::skip] |
| 14180 | let e = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14181 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14182 | assert_eq_m256i(r, e); |
| 14183 | } |
| 14184 | |
| 14185 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14186 | unsafe fn test_mm256_maskz_min_epi8() { |
| 14187 | #[rustfmt::skip] |
| 14188 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 14189 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 14190 | #[rustfmt::skip] |
| 14191 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14192 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14193 | let r = _mm256_maskz_min_epi8(0, a, b); |
| 14194 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 14195 | let r = _mm256_maskz_min_epi8(0b00000000_11111111_00000000_11111111, a, b); |
| 14196 | #[rustfmt::skip] |
| 14197 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
| 14198 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14199 | assert_eq_m256i(r, e); |
| 14200 | } |
| 14201 | |
| 14202 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14203 | unsafe fn test_mm_mask_min_epi8() { |
| 14204 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 14205 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14206 | let r = _mm_mask_min_epi8(a, 0, a, b); |
| 14207 | assert_eq_m128i(r, a); |
| 14208 | let r = _mm_mask_min_epi8(a, 0b00000000_11111111, a, b); |
| 14209 | let e = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14210 | assert_eq_m128i(r, e); |
| 14211 | } |
| 14212 | |
| 14213 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14214 | unsafe fn test_mm_maskz_min_epi8() { |
| 14215 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 14216 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14217 | let r = _mm_maskz_min_epi8(0, a, b); |
| 14218 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 14219 | let r = _mm_maskz_min_epi8(0b00000000_11111111, a, b); |
| 14220 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
| 14221 | assert_eq_m128i(r, e); |
| 14222 | } |
| 14223 | |
| 14224 | #[simd_test(enable = "avx512bw" )] |
| 14225 | unsafe fn test_mm512_cmplt_epu16_mask() { |
| 14226 | let a = _mm512_set1_epi16(-2); |
| 14227 | let b = _mm512_set1_epi16(-1); |
| 14228 | let m = _mm512_cmplt_epu16_mask(a, b); |
| 14229 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14230 | } |
| 14231 | |
| 14232 | #[simd_test(enable = "avx512bw" )] |
| 14233 | unsafe fn test_mm512_mask_cmplt_epu16_mask() { |
| 14234 | let a = _mm512_set1_epi16(-2); |
| 14235 | let b = _mm512_set1_epi16(-1); |
| 14236 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14237 | let r = _mm512_mask_cmplt_epu16_mask(mask, a, b); |
| 14238 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14239 | } |
| 14240 | |
| 14241 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14242 | unsafe fn test_mm256_cmplt_epu16_mask() { |
| 14243 | let a = _mm256_set1_epi16(-2); |
| 14244 | let b = _mm256_set1_epi16(-1); |
| 14245 | let m = _mm256_cmplt_epu16_mask(a, b); |
| 14246 | assert_eq!(m, 0b11111111_11111111); |
| 14247 | } |
| 14248 | |
| 14249 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14250 | unsafe fn test_mm256_mask_cmplt_epu16_mask() { |
| 14251 | let a = _mm256_set1_epi16(-2); |
| 14252 | let b = _mm256_set1_epi16(-1); |
| 14253 | let mask = 0b01010101_01010101; |
| 14254 | let r = _mm256_mask_cmplt_epu16_mask(mask, a, b); |
| 14255 | assert_eq!(r, 0b01010101_01010101); |
| 14256 | } |
| 14257 | |
| 14258 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14259 | unsafe fn test_mm_cmplt_epu16_mask() { |
| 14260 | let a = _mm_set1_epi16(-2); |
| 14261 | let b = _mm_set1_epi16(-1); |
| 14262 | let m = _mm_cmplt_epu16_mask(a, b); |
| 14263 | assert_eq!(m, 0b11111111); |
| 14264 | } |
| 14265 | |
| 14266 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14267 | unsafe fn test_mm_mask_cmplt_epu16_mask() { |
| 14268 | let a = _mm_set1_epi16(-2); |
| 14269 | let b = _mm_set1_epi16(-1); |
| 14270 | let mask = 0b01010101; |
| 14271 | let r = _mm_mask_cmplt_epu16_mask(mask, a, b); |
| 14272 | assert_eq!(r, 0b01010101); |
| 14273 | } |
| 14274 | |
| 14275 | #[simd_test(enable = "avx512bw" )] |
| 14276 | unsafe fn test_mm512_cmplt_epu8_mask() { |
| 14277 | let a = _mm512_set1_epi8(-2); |
| 14278 | let b = _mm512_set1_epi8(-1); |
| 14279 | let m = _mm512_cmplt_epu8_mask(a, b); |
| 14280 | assert_eq!( |
| 14281 | m, |
| 14282 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 14283 | ); |
| 14284 | } |
| 14285 | |
| 14286 | #[simd_test(enable = "avx512bw" )] |
| 14287 | unsafe fn test_mm512_mask_cmplt_epu8_mask() { |
| 14288 | let a = _mm512_set1_epi8(-2); |
| 14289 | let b = _mm512_set1_epi8(-1); |
| 14290 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 14291 | let r = _mm512_mask_cmplt_epu8_mask(mask, a, b); |
| 14292 | assert_eq!( |
| 14293 | r, |
| 14294 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 14295 | ); |
| 14296 | } |
| 14297 | |
| 14298 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14299 | unsafe fn test_mm256_cmplt_epu8_mask() { |
| 14300 | let a = _mm256_set1_epi8(-2); |
| 14301 | let b = _mm256_set1_epi8(-1); |
| 14302 | let m = _mm256_cmplt_epu8_mask(a, b); |
| 14303 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14304 | } |
| 14305 | |
| 14306 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14307 | unsafe fn test_mm256_mask_cmplt_epu8_mask() { |
| 14308 | let a = _mm256_set1_epi8(-2); |
| 14309 | let b = _mm256_set1_epi8(-1); |
| 14310 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14311 | let r = _mm256_mask_cmplt_epu8_mask(mask, a, b); |
| 14312 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14313 | } |
| 14314 | |
| 14315 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14316 | unsafe fn test_mm_cmplt_epu8_mask() { |
| 14317 | let a = _mm_set1_epi8(-2); |
| 14318 | let b = _mm_set1_epi8(-1); |
| 14319 | let m = _mm_cmplt_epu8_mask(a, b); |
| 14320 | assert_eq!(m, 0b11111111_11111111); |
| 14321 | } |
| 14322 | |
| 14323 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14324 | unsafe fn test_mm_mask_cmplt_epu8_mask() { |
| 14325 | let a = _mm_set1_epi8(-2); |
| 14326 | let b = _mm_set1_epi8(-1); |
| 14327 | let mask = 0b01010101_01010101; |
| 14328 | let r = _mm_mask_cmplt_epu8_mask(mask, a, b); |
| 14329 | assert_eq!(r, 0b01010101_01010101); |
| 14330 | } |
| 14331 | |
| 14332 | #[simd_test(enable = "avx512bw" )] |
| 14333 | unsafe fn test_mm512_cmplt_epi16_mask() { |
| 14334 | let a = _mm512_set1_epi16(-2); |
| 14335 | let b = _mm512_set1_epi16(-1); |
| 14336 | let m = _mm512_cmplt_epi16_mask(a, b); |
| 14337 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14338 | } |
| 14339 | |
| 14340 | #[simd_test(enable = "avx512bw" )] |
| 14341 | unsafe fn test_mm512_mask_cmplt_epi16_mask() { |
| 14342 | let a = _mm512_set1_epi16(-2); |
| 14343 | let b = _mm512_set1_epi16(-1); |
| 14344 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14345 | let r = _mm512_mask_cmplt_epi16_mask(mask, a, b); |
| 14346 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14347 | } |
| 14348 | |
| 14349 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14350 | unsafe fn test_mm256_cmplt_epi16_mask() { |
| 14351 | let a = _mm256_set1_epi16(-2); |
| 14352 | let b = _mm256_set1_epi16(-1); |
| 14353 | let m = _mm256_cmplt_epi16_mask(a, b); |
| 14354 | assert_eq!(m, 0b11111111_11111111); |
| 14355 | } |
| 14356 | |
| 14357 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14358 | unsafe fn test_mm256_mask_cmplt_epi16_mask() { |
| 14359 | let a = _mm256_set1_epi16(-2); |
| 14360 | let b = _mm256_set1_epi16(-1); |
| 14361 | let mask = 0b01010101_01010101; |
| 14362 | let r = _mm256_mask_cmplt_epi16_mask(mask, a, b); |
| 14363 | assert_eq!(r, 0b01010101_01010101); |
| 14364 | } |
| 14365 | |
| 14366 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14367 | unsafe fn test_mm_cmplt_epi16_mask() { |
| 14368 | let a = _mm_set1_epi16(-2); |
| 14369 | let b = _mm_set1_epi16(-1); |
| 14370 | let m = _mm_cmplt_epi16_mask(a, b); |
| 14371 | assert_eq!(m, 0b11111111); |
| 14372 | } |
| 14373 | |
| 14374 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14375 | unsafe fn test_mm_mask_cmplt_epi16_mask() { |
| 14376 | let a = _mm_set1_epi16(-2); |
| 14377 | let b = _mm_set1_epi16(-1); |
| 14378 | let mask = 0b01010101; |
| 14379 | let r = _mm_mask_cmplt_epi16_mask(mask, a, b); |
| 14380 | assert_eq!(r, 0b01010101); |
| 14381 | } |
| 14382 | |
| 14383 | #[simd_test(enable = "avx512bw" )] |
| 14384 | unsafe fn test_mm512_cmplt_epi8_mask() { |
| 14385 | let a = _mm512_set1_epi8(-2); |
| 14386 | let b = _mm512_set1_epi8(-1); |
| 14387 | let m = _mm512_cmplt_epi8_mask(a, b); |
| 14388 | assert_eq!( |
| 14389 | m, |
| 14390 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 14391 | ); |
| 14392 | } |
| 14393 | |
| 14394 | #[simd_test(enable = "avx512bw" )] |
| 14395 | unsafe fn test_mm512_mask_cmplt_epi8_mask() { |
| 14396 | let a = _mm512_set1_epi8(-2); |
| 14397 | let b = _mm512_set1_epi8(-1); |
| 14398 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 14399 | let r = _mm512_mask_cmplt_epi8_mask(mask, a, b); |
| 14400 | assert_eq!( |
| 14401 | r, |
| 14402 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 14403 | ); |
| 14404 | } |
| 14405 | |
| 14406 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14407 | unsafe fn test_mm256_cmplt_epi8_mask() { |
| 14408 | let a = _mm256_set1_epi8(-2); |
| 14409 | let b = _mm256_set1_epi8(-1); |
| 14410 | let m = _mm256_cmplt_epi8_mask(a, b); |
| 14411 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14412 | } |
| 14413 | |
| 14414 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14415 | unsafe fn test_mm256_mask_cmplt_epi8_mask() { |
| 14416 | let a = _mm256_set1_epi8(-2); |
| 14417 | let b = _mm256_set1_epi8(-1); |
| 14418 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14419 | let r = _mm256_mask_cmplt_epi8_mask(mask, a, b); |
| 14420 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14421 | } |
| 14422 | |
| 14423 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14424 | unsafe fn test_mm_cmplt_epi8_mask() { |
| 14425 | let a = _mm_set1_epi8(-2); |
| 14426 | let b = _mm_set1_epi8(-1); |
| 14427 | let m = _mm_cmplt_epi8_mask(a, b); |
| 14428 | assert_eq!(m, 0b11111111_11111111); |
| 14429 | } |
| 14430 | |
| 14431 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14432 | unsafe fn test_mm_mask_cmplt_epi8_mask() { |
| 14433 | let a = _mm_set1_epi8(-2); |
| 14434 | let b = _mm_set1_epi8(-1); |
| 14435 | let mask = 0b01010101_01010101; |
| 14436 | let r = _mm_mask_cmplt_epi8_mask(mask, a, b); |
| 14437 | assert_eq!(r, 0b01010101_01010101); |
| 14438 | } |
| 14439 | |
| 14440 | #[simd_test(enable = "avx512bw" )] |
| 14441 | unsafe fn test_mm512_cmpgt_epu16_mask() { |
| 14442 | let a = _mm512_set1_epi16(2); |
| 14443 | let b = _mm512_set1_epi16(1); |
| 14444 | let m = _mm512_cmpgt_epu16_mask(a, b); |
| 14445 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14446 | } |
| 14447 | |
| 14448 | #[simd_test(enable = "avx512bw" )] |
| 14449 | unsafe fn test_mm512_mask_cmpgt_epu16_mask() { |
| 14450 | let a = _mm512_set1_epi16(2); |
| 14451 | let b = _mm512_set1_epi16(1); |
| 14452 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14453 | let r = _mm512_mask_cmpgt_epu16_mask(mask, a, b); |
| 14454 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14455 | } |
| 14456 | |
| 14457 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14458 | unsafe fn test_mm256_cmpgt_epu16_mask() { |
| 14459 | let a = _mm256_set1_epi16(2); |
| 14460 | let b = _mm256_set1_epi16(1); |
| 14461 | let m = _mm256_cmpgt_epu16_mask(a, b); |
| 14462 | assert_eq!(m, 0b11111111_11111111); |
| 14463 | } |
| 14464 | |
| 14465 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14466 | unsafe fn test_mm256_mask_cmpgt_epu16_mask() { |
| 14467 | let a = _mm256_set1_epi16(2); |
| 14468 | let b = _mm256_set1_epi16(1); |
| 14469 | let mask = 0b01010101_01010101; |
| 14470 | let r = _mm256_mask_cmpgt_epu16_mask(mask, a, b); |
| 14471 | assert_eq!(r, 0b01010101_01010101); |
| 14472 | } |
| 14473 | |
| 14474 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14475 | unsafe fn test_mm_cmpgt_epu16_mask() { |
| 14476 | let a = _mm_set1_epi16(2); |
| 14477 | let b = _mm_set1_epi16(1); |
| 14478 | let m = _mm_cmpgt_epu16_mask(a, b); |
| 14479 | assert_eq!(m, 0b11111111); |
| 14480 | } |
| 14481 | |
| 14482 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14483 | unsafe fn test_mm_mask_cmpgt_epu16_mask() { |
| 14484 | let a = _mm_set1_epi16(2); |
| 14485 | let b = _mm_set1_epi16(1); |
| 14486 | let mask = 0b01010101; |
| 14487 | let r = _mm_mask_cmpgt_epu16_mask(mask, a, b); |
| 14488 | assert_eq!(r, 0b01010101); |
| 14489 | } |
| 14490 | |
| 14491 | #[simd_test(enable = "avx512bw" )] |
| 14492 | unsafe fn test_mm512_cmpgt_epu8_mask() { |
| 14493 | let a = _mm512_set1_epi8(2); |
| 14494 | let b = _mm512_set1_epi8(1); |
| 14495 | let m = _mm512_cmpgt_epu8_mask(a, b); |
| 14496 | assert_eq!( |
| 14497 | m, |
| 14498 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 14499 | ); |
| 14500 | } |
| 14501 | |
| 14502 | #[simd_test(enable = "avx512bw" )] |
| 14503 | unsafe fn test_mm512_mask_cmpgt_epu8_mask() { |
| 14504 | let a = _mm512_set1_epi8(2); |
| 14505 | let b = _mm512_set1_epi8(1); |
| 14506 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 14507 | let r = _mm512_mask_cmpgt_epu8_mask(mask, a, b); |
| 14508 | assert_eq!( |
| 14509 | r, |
| 14510 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 14511 | ); |
| 14512 | } |
| 14513 | |
| 14514 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14515 | unsafe fn test_mm256_cmpgt_epu8_mask() { |
| 14516 | let a = _mm256_set1_epi8(2); |
| 14517 | let b = _mm256_set1_epi8(1); |
| 14518 | let m = _mm256_cmpgt_epu8_mask(a, b); |
| 14519 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14520 | } |
| 14521 | |
| 14522 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14523 | unsafe fn test_mm256_mask_cmpgt_epu8_mask() { |
| 14524 | let a = _mm256_set1_epi8(2); |
| 14525 | let b = _mm256_set1_epi8(1); |
| 14526 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14527 | let r = _mm256_mask_cmpgt_epu8_mask(mask, a, b); |
| 14528 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14529 | } |
| 14530 | |
| 14531 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14532 | unsafe fn test_mm_cmpgt_epu8_mask() { |
| 14533 | let a = _mm_set1_epi8(2); |
| 14534 | let b = _mm_set1_epi8(1); |
| 14535 | let m = _mm_cmpgt_epu8_mask(a, b); |
| 14536 | assert_eq!(m, 0b11111111_11111111); |
| 14537 | } |
| 14538 | |
| 14539 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14540 | unsafe fn test_mm_mask_cmpgt_epu8_mask() { |
| 14541 | let a = _mm_set1_epi8(2); |
| 14542 | let b = _mm_set1_epi8(1); |
| 14543 | let mask = 0b01010101_01010101; |
| 14544 | let r = _mm_mask_cmpgt_epu8_mask(mask, a, b); |
| 14545 | assert_eq!(r, 0b01010101_01010101); |
| 14546 | } |
| 14547 | |
| 14548 | #[simd_test(enable = "avx512bw" )] |
| 14549 | unsafe fn test_mm512_cmpgt_epi16_mask() { |
| 14550 | let a = _mm512_set1_epi16(2); |
| 14551 | let b = _mm512_set1_epi16(-1); |
| 14552 | let m = _mm512_cmpgt_epi16_mask(a, b); |
| 14553 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14554 | } |
| 14555 | |
| 14556 | #[simd_test(enable = "avx512bw" )] |
| 14557 | unsafe fn test_mm512_mask_cmpgt_epi16_mask() { |
| 14558 | let a = _mm512_set1_epi16(2); |
| 14559 | let b = _mm512_set1_epi16(-1); |
| 14560 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14561 | let r = _mm512_mask_cmpgt_epi16_mask(mask, a, b); |
| 14562 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14563 | } |
| 14564 | |
| 14565 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14566 | unsafe fn test_mm256_cmpgt_epi16_mask() { |
| 14567 | let a = _mm256_set1_epi16(2); |
| 14568 | let b = _mm256_set1_epi16(-1); |
| 14569 | let m = _mm256_cmpgt_epi16_mask(a, b); |
| 14570 | assert_eq!(m, 0b11111111_11111111); |
| 14571 | } |
| 14572 | |
| 14573 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14574 | unsafe fn test_mm256_mask_cmpgt_epi16_mask() { |
| 14575 | let a = _mm256_set1_epi16(2); |
| 14576 | let b = _mm256_set1_epi16(-1); |
| 14577 | let mask = 0b001010101_01010101; |
| 14578 | let r = _mm256_mask_cmpgt_epi16_mask(mask, a, b); |
| 14579 | assert_eq!(r, 0b01010101_01010101); |
| 14580 | } |
| 14581 | |
| 14582 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14583 | unsafe fn test_mm_cmpgt_epi16_mask() { |
| 14584 | let a = _mm_set1_epi16(2); |
| 14585 | let b = _mm_set1_epi16(-1); |
| 14586 | let m = _mm_cmpgt_epi16_mask(a, b); |
| 14587 | assert_eq!(m, 0b11111111); |
| 14588 | } |
| 14589 | |
| 14590 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14591 | unsafe fn test_mm_mask_cmpgt_epi16_mask() { |
| 14592 | let a = _mm_set1_epi16(2); |
| 14593 | let b = _mm_set1_epi16(-1); |
| 14594 | let mask = 0b01010101; |
| 14595 | let r = _mm_mask_cmpgt_epi16_mask(mask, a, b); |
| 14596 | assert_eq!(r, 0b01010101); |
| 14597 | } |
| 14598 | |
| 14599 | #[simd_test(enable = "avx512bw" )] |
| 14600 | unsafe fn test_mm512_cmpgt_epi8_mask() { |
| 14601 | let a = _mm512_set1_epi8(2); |
| 14602 | let b = _mm512_set1_epi8(-1); |
| 14603 | let m = _mm512_cmpgt_epi8_mask(a, b); |
| 14604 | assert_eq!( |
| 14605 | m, |
| 14606 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 14607 | ); |
| 14608 | } |
| 14609 | |
| 14610 | #[simd_test(enable = "avx512bw" )] |
| 14611 | unsafe fn test_mm512_mask_cmpgt_epi8_mask() { |
| 14612 | let a = _mm512_set1_epi8(2); |
| 14613 | let b = _mm512_set1_epi8(-1); |
| 14614 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 14615 | let r = _mm512_mask_cmpgt_epi8_mask(mask, a, b); |
| 14616 | assert_eq!( |
| 14617 | r, |
| 14618 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 14619 | ); |
| 14620 | } |
| 14621 | |
| 14622 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14623 | unsafe fn test_mm256_cmpgt_epi8_mask() { |
| 14624 | let a = _mm256_set1_epi8(2); |
| 14625 | let b = _mm256_set1_epi8(-1); |
| 14626 | let m = _mm256_cmpgt_epi8_mask(a, b); |
| 14627 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14628 | } |
| 14629 | |
| 14630 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14631 | unsafe fn test_mm256_mask_cmpgt_epi8_mask() { |
| 14632 | let a = _mm256_set1_epi8(2); |
| 14633 | let b = _mm256_set1_epi8(-1); |
| 14634 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14635 | let r = _mm256_mask_cmpgt_epi8_mask(mask, a, b); |
| 14636 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14637 | } |
| 14638 | |
| 14639 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14640 | unsafe fn test_mm_cmpgt_epi8_mask() { |
| 14641 | let a = _mm_set1_epi8(2); |
| 14642 | let b = _mm_set1_epi8(-1); |
| 14643 | let m = _mm_cmpgt_epi8_mask(a, b); |
| 14644 | assert_eq!(m, 0b11111111_11111111); |
| 14645 | } |
| 14646 | |
| 14647 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14648 | unsafe fn test_mm_mask_cmpgt_epi8_mask() { |
| 14649 | let a = _mm_set1_epi8(2); |
| 14650 | let b = _mm_set1_epi8(-1); |
| 14651 | let mask = 0b01010101_01010101; |
| 14652 | let r = _mm_mask_cmpgt_epi8_mask(mask, a, b); |
| 14653 | assert_eq!(r, 0b01010101_01010101); |
| 14654 | } |
| 14655 | |
| 14656 | #[simd_test(enable = "avx512bw" )] |
| 14657 | unsafe fn test_mm512_cmple_epu16_mask() { |
| 14658 | let a = _mm512_set1_epi16(-1); |
| 14659 | let b = _mm512_set1_epi16(-1); |
| 14660 | let m = _mm512_cmple_epu16_mask(a, b); |
| 14661 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14662 | } |
| 14663 | |
| 14664 | #[simd_test(enable = "avx512bw" )] |
| 14665 | unsafe fn test_mm512_mask_cmple_epu16_mask() { |
| 14666 | let a = _mm512_set1_epi16(-1); |
| 14667 | let b = _mm512_set1_epi16(-1); |
| 14668 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14669 | let r = _mm512_mask_cmple_epu16_mask(mask, a, b); |
| 14670 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14671 | } |
| 14672 | |
| 14673 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14674 | unsafe fn test_mm256_cmple_epu16_mask() { |
| 14675 | let a = _mm256_set1_epi16(-1); |
| 14676 | let b = _mm256_set1_epi16(-1); |
| 14677 | let m = _mm256_cmple_epu16_mask(a, b); |
| 14678 | assert_eq!(m, 0b11111111_11111111); |
| 14679 | } |
| 14680 | |
| 14681 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14682 | unsafe fn test_mm256_mask_cmple_epu16_mask() { |
| 14683 | let a = _mm256_set1_epi16(-1); |
| 14684 | let b = _mm256_set1_epi16(-1); |
| 14685 | let mask = 0b01010101_01010101; |
| 14686 | let r = _mm256_mask_cmple_epu16_mask(mask, a, b); |
| 14687 | assert_eq!(r, 0b01010101_01010101); |
| 14688 | } |
| 14689 | |
| 14690 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14691 | unsafe fn test_mm_cmple_epu16_mask() { |
| 14692 | let a = _mm_set1_epi16(-1); |
| 14693 | let b = _mm_set1_epi16(-1); |
| 14694 | let m = _mm_cmple_epu16_mask(a, b); |
| 14695 | assert_eq!(m, 0b11111111); |
| 14696 | } |
| 14697 | |
| 14698 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14699 | unsafe fn test_mm_mask_cmple_epu16_mask() { |
| 14700 | let a = _mm_set1_epi16(-1); |
| 14701 | let b = _mm_set1_epi16(-1); |
| 14702 | let mask = 0b01010101; |
| 14703 | let r = _mm_mask_cmple_epu16_mask(mask, a, b); |
| 14704 | assert_eq!(r, 0b01010101); |
| 14705 | } |
| 14706 | |
| 14707 | #[simd_test(enable = "avx512bw" )] |
| 14708 | unsafe fn test_mm512_cmple_epu8_mask() { |
| 14709 | let a = _mm512_set1_epi8(-1); |
| 14710 | let b = _mm512_set1_epi8(-1); |
| 14711 | let m = _mm512_cmple_epu8_mask(a, b); |
| 14712 | assert_eq!( |
| 14713 | m, |
| 14714 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 14715 | ); |
| 14716 | } |
| 14717 | |
| 14718 | #[simd_test(enable = "avx512bw" )] |
| 14719 | unsafe fn test_mm512_mask_cmple_epu8_mask() { |
| 14720 | let a = _mm512_set1_epi8(-1); |
| 14721 | let b = _mm512_set1_epi8(-1); |
| 14722 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 14723 | let r = _mm512_mask_cmple_epu8_mask(mask, a, b); |
| 14724 | assert_eq!( |
| 14725 | r, |
| 14726 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 14727 | ); |
| 14728 | } |
| 14729 | |
| 14730 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14731 | unsafe fn test_mm256_cmple_epu8_mask() { |
| 14732 | let a = _mm256_set1_epi8(-1); |
| 14733 | let b = _mm256_set1_epi8(-1); |
| 14734 | let m = _mm256_cmple_epu8_mask(a, b); |
| 14735 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14736 | } |
| 14737 | |
| 14738 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14739 | unsafe fn test_mm256_mask_cmple_epu8_mask() { |
| 14740 | let a = _mm256_set1_epi8(-1); |
| 14741 | let b = _mm256_set1_epi8(-1); |
| 14742 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14743 | let r = _mm256_mask_cmple_epu8_mask(mask, a, b); |
| 14744 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14745 | } |
| 14746 | |
| 14747 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14748 | unsafe fn test_mm_cmple_epu8_mask() { |
| 14749 | let a = _mm_set1_epi8(-1); |
| 14750 | let b = _mm_set1_epi8(-1); |
| 14751 | let m = _mm_cmple_epu8_mask(a, b); |
| 14752 | assert_eq!(m, 0b11111111_11111111); |
| 14753 | } |
| 14754 | |
| 14755 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14756 | unsafe fn test_mm_mask_cmple_epu8_mask() { |
| 14757 | let a = _mm_set1_epi8(-1); |
| 14758 | let b = _mm_set1_epi8(-1); |
| 14759 | let mask = 0b01010101_01010101; |
| 14760 | let r = _mm_mask_cmple_epu8_mask(mask, a, b); |
| 14761 | assert_eq!(r, 0b01010101_01010101); |
| 14762 | } |
| 14763 | |
| 14764 | #[simd_test(enable = "avx512bw" )] |
| 14765 | unsafe fn test_mm512_cmple_epi16_mask() { |
| 14766 | let a = _mm512_set1_epi16(-1); |
| 14767 | let b = _mm512_set1_epi16(-1); |
| 14768 | let m = _mm512_cmple_epi16_mask(a, b); |
| 14769 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14770 | } |
| 14771 | |
| 14772 | #[simd_test(enable = "avx512bw" )] |
| 14773 | unsafe fn test_mm512_mask_cmple_epi16_mask() { |
| 14774 | let a = _mm512_set1_epi16(-1); |
| 14775 | let b = _mm512_set1_epi16(-1); |
| 14776 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14777 | let r = _mm512_mask_cmple_epi16_mask(mask, a, b); |
| 14778 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14779 | } |
| 14780 | |
| 14781 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14782 | unsafe fn test_mm256_cmple_epi16_mask() { |
| 14783 | let a = _mm256_set1_epi16(-1); |
| 14784 | let b = _mm256_set1_epi16(-1); |
| 14785 | let m = _mm256_cmple_epi16_mask(a, b); |
| 14786 | assert_eq!(m, 0b11111111_11111111); |
| 14787 | } |
| 14788 | |
| 14789 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14790 | unsafe fn test_mm256_mask_cmple_epi16_mask() { |
| 14791 | let a = _mm256_set1_epi16(-1); |
| 14792 | let b = _mm256_set1_epi16(-1); |
| 14793 | let mask = 0b01010101_01010101; |
| 14794 | let r = _mm256_mask_cmple_epi16_mask(mask, a, b); |
| 14795 | assert_eq!(r, 0b01010101_01010101); |
| 14796 | } |
| 14797 | |
| 14798 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14799 | unsafe fn test_mm_cmple_epi16_mask() { |
| 14800 | let a = _mm_set1_epi16(-1); |
| 14801 | let b = _mm_set1_epi16(-1); |
| 14802 | let m = _mm_cmple_epi16_mask(a, b); |
| 14803 | assert_eq!(m, 0b11111111); |
| 14804 | } |
| 14805 | |
| 14806 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14807 | unsafe fn test_mm_mask_cmple_epi16_mask() { |
| 14808 | let a = _mm_set1_epi16(-1); |
| 14809 | let b = _mm_set1_epi16(-1); |
| 14810 | let mask = 0b01010101; |
| 14811 | let r = _mm_mask_cmple_epi16_mask(mask, a, b); |
| 14812 | assert_eq!(r, 0b01010101); |
| 14813 | } |
| 14814 | |
| 14815 | #[simd_test(enable = "avx512bw" )] |
| 14816 | unsafe fn test_mm512_cmple_epi8_mask() { |
| 14817 | let a = _mm512_set1_epi8(-1); |
| 14818 | let b = _mm512_set1_epi8(-1); |
| 14819 | let m = _mm512_cmple_epi8_mask(a, b); |
| 14820 | assert_eq!( |
| 14821 | m, |
| 14822 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 14823 | ); |
| 14824 | } |
| 14825 | |
| 14826 | #[simd_test(enable = "avx512bw" )] |
| 14827 | unsafe fn test_mm512_mask_cmple_epi8_mask() { |
| 14828 | let a = _mm512_set1_epi8(-1); |
| 14829 | let b = _mm512_set1_epi8(-1); |
| 14830 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 14831 | let r = _mm512_mask_cmple_epi8_mask(mask, a, b); |
| 14832 | assert_eq!( |
| 14833 | r, |
| 14834 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 14835 | ); |
| 14836 | } |
| 14837 | |
| 14838 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14839 | unsafe fn test_mm256_cmple_epi8_mask() { |
| 14840 | let a = _mm256_set1_epi8(-1); |
| 14841 | let b = _mm256_set1_epi8(-1); |
| 14842 | let m = _mm256_cmple_epi8_mask(a, b); |
| 14843 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14844 | } |
| 14845 | |
| 14846 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14847 | unsafe fn test_mm256_mask_cmple_epi8_mask() { |
| 14848 | let a = _mm256_set1_epi8(-1); |
| 14849 | let b = _mm256_set1_epi8(-1); |
| 14850 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14851 | let r = _mm256_mask_cmple_epi8_mask(mask, a, b); |
| 14852 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14853 | } |
| 14854 | |
| 14855 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14856 | unsafe fn test_mm_cmple_epi8_mask() { |
| 14857 | let a = _mm_set1_epi8(-1); |
| 14858 | let b = _mm_set1_epi8(-1); |
| 14859 | let m = _mm_cmple_epi8_mask(a, b); |
| 14860 | assert_eq!(m, 0b11111111_11111111); |
| 14861 | } |
| 14862 | |
| 14863 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14864 | unsafe fn test_mm_mask_cmple_epi8_mask() { |
| 14865 | let a = _mm_set1_epi8(-1); |
| 14866 | let b = _mm_set1_epi8(-1); |
| 14867 | let mask = 0b01010101_01010101; |
| 14868 | let r = _mm_mask_cmple_epi8_mask(mask, a, b); |
| 14869 | assert_eq!(r, 0b01010101_01010101); |
| 14870 | } |
| 14871 | |
| 14872 | #[simd_test(enable = "avx512bw" )] |
| 14873 | unsafe fn test_mm512_cmpge_epu16_mask() { |
| 14874 | let a = _mm512_set1_epi16(1); |
| 14875 | let b = _mm512_set1_epi16(1); |
| 14876 | let m = _mm512_cmpge_epu16_mask(a, b); |
| 14877 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14878 | } |
| 14879 | |
| 14880 | #[simd_test(enable = "avx512bw" )] |
| 14881 | unsafe fn test_mm512_mask_cmpge_epu16_mask() { |
| 14882 | let a = _mm512_set1_epi16(1); |
| 14883 | let b = _mm512_set1_epi16(1); |
| 14884 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14885 | let r = _mm512_mask_cmpge_epu16_mask(mask, a, b); |
| 14886 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14887 | } |
| 14888 | |
| 14889 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14890 | unsafe fn test_mm256_cmpge_epu16_mask() { |
| 14891 | let a = _mm256_set1_epi16(1); |
| 14892 | let b = _mm256_set1_epi16(1); |
| 14893 | let m = _mm256_cmpge_epu16_mask(a, b); |
| 14894 | assert_eq!(m, 0b11111111_11111111); |
| 14895 | } |
| 14896 | |
| 14897 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14898 | unsafe fn test_mm256_mask_cmpge_epu16_mask() { |
| 14899 | let a = _mm256_set1_epi16(1); |
| 14900 | let b = _mm256_set1_epi16(1); |
| 14901 | let mask = 0b01010101_01010101; |
| 14902 | let r = _mm256_mask_cmpge_epu16_mask(mask, a, b); |
| 14903 | assert_eq!(r, 0b01010101_01010101); |
| 14904 | } |
| 14905 | |
| 14906 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14907 | unsafe fn test_mm_cmpge_epu16_mask() { |
| 14908 | let a = _mm_set1_epi16(1); |
| 14909 | let b = _mm_set1_epi16(1); |
| 14910 | let m = _mm_cmpge_epu16_mask(a, b); |
| 14911 | assert_eq!(m, 0b11111111); |
| 14912 | } |
| 14913 | |
| 14914 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14915 | unsafe fn test_mm_mask_cmpge_epu16_mask() { |
| 14916 | let a = _mm_set1_epi16(1); |
| 14917 | let b = _mm_set1_epi16(1); |
| 14918 | let mask = 0b01010101; |
| 14919 | let r = _mm_mask_cmpge_epu16_mask(mask, a, b); |
| 14920 | assert_eq!(r, 0b01010101); |
| 14921 | } |
| 14922 | |
| 14923 | #[simd_test(enable = "avx512bw" )] |
| 14924 | unsafe fn test_mm512_cmpge_epu8_mask() { |
| 14925 | let a = _mm512_set1_epi8(1); |
| 14926 | let b = _mm512_set1_epi8(1); |
| 14927 | let m = _mm512_cmpge_epu8_mask(a, b); |
| 14928 | assert_eq!( |
| 14929 | m, |
| 14930 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 14931 | ); |
| 14932 | } |
| 14933 | |
| 14934 | #[simd_test(enable = "avx512bw" )] |
| 14935 | unsafe fn test_mm512_mask_cmpge_epu8_mask() { |
| 14936 | let a = _mm512_set1_epi8(1); |
| 14937 | let b = _mm512_set1_epi8(1); |
| 14938 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 14939 | let r = _mm512_mask_cmpge_epu8_mask(mask, a, b); |
| 14940 | assert_eq!( |
| 14941 | r, |
| 14942 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 14943 | ); |
| 14944 | } |
| 14945 | |
| 14946 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14947 | unsafe fn test_mm256_cmpge_epu8_mask() { |
| 14948 | let a = _mm256_set1_epi8(1); |
| 14949 | let b = _mm256_set1_epi8(1); |
| 14950 | let m = _mm256_cmpge_epu8_mask(a, b); |
| 14951 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14952 | } |
| 14953 | |
| 14954 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14955 | unsafe fn test_mm256_mask_cmpge_epu8_mask() { |
| 14956 | let a = _mm256_set1_epi8(1); |
| 14957 | let b = _mm256_set1_epi8(1); |
| 14958 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14959 | let r = _mm256_mask_cmpge_epu8_mask(mask, a, b); |
| 14960 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14961 | } |
| 14962 | |
| 14963 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14964 | unsafe fn test_mm_cmpge_epu8_mask() { |
| 14965 | let a = _mm_set1_epi8(1); |
| 14966 | let b = _mm_set1_epi8(1); |
| 14967 | let m = _mm_cmpge_epu8_mask(a, b); |
| 14968 | assert_eq!(m, 0b11111111_11111111); |
| 14969 | } |
| 14970 | |
| 14971 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14972 | unsafe fn test_mm_mask_cmpge_epu8_mask() { |
| 14973 | let a = _mm_set1_epi8(1); |
| 14974 | let b = _mm_set1_epi8(1); |
| 14975 | let mask = 0b01010101_01010101; |
| 14976 | let r = _mm_mask_cmpge_epu8_mask(mask, a, b); |
| 14977 | assert_eq!(r, 0b01010101_01010101); |
| 14978 | } |
| 14979 | |
| 14980 | #[simd_test(enable = "avx512bw" )] |
| 14981 | unsafe fn test_mm512_cmpge_epi16_mask() { |
| 14982 | let a = _mm512_set1_epi16(-1); |
| 14983 | let b = _mm512_set1_epi16(-1); |
| 14984 | let m = _mm512_cmpge_epi16_mask(a, b); |
| 14985 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 14986 | } |
| 14987 | |
| 14988 | #[simd_test(enable = "avx512bw" )] |
| 14989 | unsafe fn test_mm512_mask_cmpge_epi16_mask() { |
| 14990 | let a = _mm512_set1_epi16(-1); |
| 14991 | let b = _mm512_set1_epi16(-1); |
| 14992 | let mask = 0b01010101_01010101_01010101_01010101; |
| 14993 | let r = _mm512_mask_cmpge_epi16_mask(mask, a, b); |
| 14994 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 14995 | } |
| 14996 | |
| 14997 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 14998 | unsafe fn test_mm256_cmpge_epi16_mask() { |
| 14999 | let a = _mm256_set1_epi16(-1); |
| 15000 | let b = _mm256_set1_epi16(-1); |
| 15001 | let m = _mm256_cmpge_epi16_mask(a, b); |
| 15002 | assert_eq!(m, 0b11111111_11111111); |
| 15003 | } |
| 15004 | |
| 15005 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15006 | unsafe fn test_mm256_mask_cmpge_epi16_mask() { |
| 15007 | let a = _mm256_set1_epi16(-1); |
| 15008 | let b = _mm256_set1_epi16(-1); |
| 15009 | let mask = 0b01010101_01010101; |
| 15010 | let r = _mm256_mask_cmpge_epi16_mask(mask, a, b); |
| 15011 | assert_eq!(r, 0b01010101_01010101); |
| 15012 | } |
| 15013 | |
| 15014 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15015 | unsafe fn test_mm_cmpge_epi16_mask() { |
| 15016 | let a = _mm_set1_epi16(-1); |
| 15017 | let b = _mm_set1_epi16(-1); |
| 15018 | let m = _mm_cmpge_epi16_mask(a, b); |
| 15019 | assert_eq!(m, 0b11111111); |
| 15020 | } |
| 15021 | |
| 15022 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15023 | unsafe fn test_mm_mask_cmpge_epi16_mask() { |
| 15024 | let a = _mm_set1_epi16(-1); |
| 15025 | let b = _mm_set1_epi16(-1); |
| 15026 | let mask = 0b01010101; |
| 15027 | let r = _mm_mask_cmpge_epi16_mask(mask, a, b); |
| 15028 | assert_eq!(r, 0b01010101); |
| 15029 | } |
| 15030 | |
| 15031 | #[simd_test(enable = "avx512bw" )] |
| 15032 | unsafe fn test_mm512_cmpge_epi8_mask() { |
| 15033 | let a = _mm512_set1_epi8(-1); |
| 15034 | let b = _mm512_set1_epi8(-1); |
| 15035 | let m = _mm512_cmpge_epi8_mask(a, b); |
| 15036 | assert_eq!( |
| 15037 | m, |
| 15038 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 15039 | ); |
| 15040 | } |
| 15041 | |
| 15042 | #[simd_test(enable = "avx512bw" )] |
| 15043 | unsafe fn test_mm512_mask_cmpge_epi8_mask() { |
| 15044 | let a = _mm512_set1_epi8(-1); |
| 15045 | let b = _mm512_set1_epi8(-1); |
| 15046 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 15047 | let r = _mm512_mask_cmpge_epi8_mask(mask, a, b); |
| 15048 | assert_eq!( |
| 15049 | r, |
| 15050 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 15051 | ); |
| 15052 | } |
| 15053 | |
| 15054 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15055 | unsafe fn test_mm256_cmpge_epi8_mask() { |
| 15056 | let a = _mm256_set1_epi8(-1); |
| 15057 | let b = _mm256_set1_epi8(-1); |
| 15058 | let m = _mm256_cmpge_epi8_mask(a, b); |
| 15059 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15060 | } |
| 15061 | |
| 15062 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15063 | unsafe fn test_mm256_mask_cmpge_epi8_mask() { |
| 15064 | let a = _mm256_set1_epi8(-1); |
| 15065 | let b = _mm256_set1_epi8(-1); |
| 15066 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15067 | let r = _mm256_mask_cmpge_epi8_mask(mask, a, b); |
| 15068 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15069 | } |
| 15070 | |
| 15071 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15072 | unsafe fn test_mm_cmpge_epi8_mask() { |
| 15073 | let a = _mm_set1_epi8(-1); |
| 15074 | let b = _mm_set1_epi8(-1); |
| 15075 | let m = _mm_cmpge_epi8_mask(a, b); |
| 15076 | assert_eq!(m, 0b11111111_11111111); |
| 15077 | } |
| 15078 | |
| 15079 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15080 | unsafe fn test_mm_mask_cmpge_epi8_mask() { |
| 15081 | let a = _mm_set1_epi8(-1); |
| 15082 | let b = _mm_set1_epi8(-1); |
| 15083 | let mask = 0b01010101_01010101; |
| 15084 | let r = _mm_mask_cmpge_epi8_mask(mask, a, b); |
| 15085 | assert_eq!(r, 0b01010101_01010101); |
| 15086 | } |
| 15087 | |
| 15088 | #[simd_test(enable = "avx512bw" )] |
| 15089 | unsafe fn test_mm512_cmpeq_epu16_mask() { |
| 15090 | let a = _mm512_set1_epi16(1); |
| 15091 | let b = _mm512_set1_epi16(1); |
| 15092 | let m = _mm512_cmpeq_epu16_mask(a, b); |
| 15093 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15094 | } |
| 15095 | |
| 15096 | #[simd_test(enable = "avx512bw" )] |
| 15097 | unsafe fn test_mm512_mask_cmpeq_epu16_mask() { |
| 15098 | let a = _mm512_set1_epi16(1); |
| 15099 | let b = _mm512_set1_epi16(1); |
| 15100 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15101 | let r = _mm512_mask_cmpeq_epu16_mask(mask, a, b); |
| 15102 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15103 | } |
| 15104 | |
| 15105 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15106 | unsafe fn test_mm256_cmpeq_epu16_mask() { |
| 15107 | let a = _mm256_set1_epi16(1); |
| 15108 | let b = _mm256_set1_epi16(1); |
| 15109 | let m = _mm256_cmpeq_epu16_mask(a, b); |
| 15110 | assert_eq!(m, 0b11111111_11111111); |
| 15111 | } |
| 15112 | |
| 15113 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15114 | unsafe fn test_mm256_mask_cmpeq_epu16_mask() { |
| 15115 | let a = _mm256_set1_epi16(1); |
| 15116 | let b = _mm256_set1_epi16(1); |
| 15117 | let mask = 0b01010101_01010101; |
| 15118 | let r = _mm256_mask_cmpeq_epu16_mask(mask, a, b); |
| 15119 | assert_eq!(r, 0b01010101_01010101); |
| 15120 | } |
| 15121 | |
| 15122 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15123 | unsafe fn test_mm_cmpeq_epu16_mask() { |
| 15124 | let a = _mm_set1_epi16(1); |
| 15125 | let b = _mm_set1_epi16(1); |
| 15126 | let m = _mm_cmpeq_epu16_mask(a, b); |
| 15127 | assert_eq!(m, 0b11111111); |
| 15128 | } |
| 15129 | |
| 15130 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15131 | unsafe fn test_mm_mask_cmpeq_epu16_mask() { |
| 15132 | let a = _mm_set1_epi16(1); |
| 15133 | let b = _mm_set1_epi16(1); |
| 15134 | let mask = 0b01010101; |
| 15135 | let r = _mm_mask_cmpeq_epu16_mask(mask, a, b); |
| 15136 | assert_eq!(r, 0b01010101); |
| 15137 | } |
| 15138 | |
| 15139 | #[simd_test(enable = "avx512bw" )] |
| 15140 | unsafe fn test_mm512_cmpeq_epu8_mask() { |
| 15141 | let a = _mm512_set1_epi8(1); |
| 15142 | let b = _mm512_set1_epi8(1); |
| 15143 | let m = _mm512_cmpeq_epu8_mask(a, b); |
| 15144 | assert_eq!( |
| 15145 | m, |
| 15146 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 15147 | ); |
| 15148 | } |
| 15149 | |
| 15150 | #[simd_test(enable = "avx512bw" )] |
| 15151 | unsafe fn test_mm512_mask_cmpeq_epu8_mask() { |
| 15152 | let a = _mm512_set1_epi8(1); |
| 15153 | let b = _mm512_set1_epi8(1); |
| 15154 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 15155 | let r = _mm512_mask_cmpeq_epu8_mask(mask, a, b); |
| 15156 | assert_eq!( |
| 15157 | r, |
| 15158 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 15159 | ); |
| 15160 | } |
| 15161 | |
| 15162 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15163 | unsafe fn test_mm256_cmpeq_epu8_mask() { |
| 15164 | let a = _mm256_set1_epi8(1); |
| 15165 | let b = _mm256_set1_epi8(1); |
| 15166 | let m = _mm256_cmpeq_epu8_mask(a, b); |
| 15167 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15168 | } |
| 15169 | |
| 15170 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15171 | unsafe fn test_mm256_mask_cmpeq_epu8_mask() { |
| 15172 | let a = _mm256_set1_epi8(1); |
| 15173 | let b = _mm256_set1_epi8(1); |
| 15174 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15175 | let r = _mm256_mask_cmpeq_epu8_mask(mask, a, b); |
| 15176 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15177 | } |
| 15178 | |
| 15179 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15180 | unsafe fn test_mm_cmpeq_epu8_mask() { |
| 15181 | let a = _mm_set1_epi8(1); |
| 15182 | let b = _mm_set1_epi8(1); |
| 15183 | let m = _mm_cmpeq_epu8_mask(a, b); |
| 15184 | assert_eq!(m, 0b11111111_11111111); |
| 15185 | } |
| 15186 | |
| 15187 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15188 | unsafe fn test_mm_mask_cmpeq_epu8_mask() { |
| 15189 | let a = _mm_set1_epi8(1); |
| 15190 | let b = _mm_set1_epi8(1); |
| 15191 | let mask = 0b01010101_01010101; |
| 15192 | let r = _mm_mask_cmpeq_epu8_mask(mask, a, b); |
| 15193 | assert_eq!(r, 0b01010101_01010101); |
| 15194 | } |
| 15195 | |
| 15196 | #[simd_test(enable = "avx512bw" )] |
| 15197 | unsafe fn test_mm512_cmpeq_epi16_mask() { |
| 15198 | let a = _mm512_set1_epi16(-1); |
| 15199 | let b = _mm512_set1_epi16(-1); |
| 15200 | let m = _mm512_cmpeq_epi16_mask(a, b); |
| 15201 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15202 | } |
| 15203 | |
| 15204 | #[simd_test(enable = "avx512bw" )] |
| 15205 | unsafe fn test_mm512_mask_cmpeq_epi16_mask() { |
| 15206 | let a = _mm512_set1_epi16(-1); |
| 15207 | let b = _mm512_set1_epi16(-1); |
| 15208 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15209 | let r = _mm512_mask_cmpeq_epi16_mask(mask, a, b); |
| 15210 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15211 | } |
| 15212 | |
| 15213 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15214 | unsafe fn test_mm256_cmpeq_epi16_mask() { |
| 15215 | let a = _mm256_set1_epi16(-1); |
| 15216 | let b = _mm256_set1_epi16(-1); |
| 15217 | let m = _mm256_cmpeq_epi16_mask(a, b); |
| 15218 | assert_eq!(m, 0b11111111_11111111); |
| 15219 | } |
| 15220 | |
| 15221 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15222 | unsafe fn test_mm256_mask_cmpeq_epi16_mask() { |
| 15223 | let a = _mm256_set1_epi16(-1); |
| 15224 | let b = _mm256_set1_epi16(-1); |
| 15225 | let mask = 0b01010101_01010101; |
| 15226 | let r = _mm256_mask_cmpeq_epi16_mask(mask, a, b); |
| 15227 | assert_eq!(r, 0b01010101_01010101); |
| 15228 | } |
| 15229 | |
| 15230 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15231 | unsafe fn test_mm_cmpeq_epi16_mask() { |
| 15232 | let a = _mm_set1_epi16(-1); |
| 15233 | let b = _mm_set1_epi16(-1); |
| 15234 | let m = _mm_cmpeq_epi16_mask(a, b); |
| 15235 | assert_eq!(m, 0b11111111); |
| 15236 | } |
| 15237 | |
| 15238 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15239 | unsafe fn test_mm_mask_cmpeq_epi16_mask() { |
| 15240 | let a = _mm_set1_epi16(-1); |
| 15241 | let b = _mm_set1_epi16(-1); |
| 15242 | let mask = 0b01010101; |
| 15243 | let r = _mm_mask_cmpeq_epi16_mask(mask, a, b); |
| 15244 | assert_eq!(r, 0b01010101); |
| 15245 | } |
| 15246 | |
| 15247 | #[simd_test(enable = "avx512bw" )] |
| 15248 | unsafe fn test_mm512_cmpeq_epi8_mask() { |
| 15249 | let a = _mm512_set1_epi8(-1); |
| 15250 | let b = _mm512_set1_epi8(-1); |
| 15251 | let m = _mm512_cmpeq_epi8_mask(a, b); |
| 15252 | assert_eq!( |
| 15253 | m, |
| 15254 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 15255 | ); |
| 15256 | } |
| 15257 | |
| 15258 | #[simd_test(enable = "avx512bw" )] |
| 15259 | unsafe fn test_mm512_mask_cmpeq_epi8_mask() { |
| 15260 | let a = _mm512_set1_epi8(-1); |
| 15261 | let b = _mm512_set1_epi8(-1); |
| 15262 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 15263 | let r = _mm512_mask_cmpeq_epi8_mask(mask, a, b); |
| 15264 | assert_eq!( |
| 15265 | r, |
| 15266 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 15267 | ); |
| 15268 | } |
| 15269 | |
| 15270 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15271 | unsafe fn test_mm256_cmpeq_epi8_mask() { |
| 15272 | let a = _mm256_set1_epi8(-1); |
| 15273 | let b = _mm256_set1_epi8(-1); |
| 15274 | let m = _mm256_cmpeq_epi8_mask(a, b); |
| 15275 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15276 | } |
| 15277 | |
| 15278 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15279 | unsafe fn test_mm256_mask_cmpeq_epi8_mask() { |
| 15280 | let a = _mm256_set1_epi8(-1); |
| 15281 | let b = _mm256_set1_epi8(-1); |
| 15282 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15283 | let r = _mm256_mask_cmpeq_epi8_mask(mask, a, b); |
| 15284 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15285 | } |
| 15286 | |
| 15287 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15288 | unsafe fn test_mm_cmpeq_epi8_mask() { |
| 15289 | let a = _mm_set1_epi8(-1); |
| 15290 | let b = _mm_set1_epi8(-1); |
| 15291 | let m = _mm_cmpeq_epi8_mask(a, b); |
| 15292 | assert_eq!(m, 0b11111111_11111111); |
| 15293 | } |
| 15294 | |
| 15295 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15296 | unsafe fn test_mm_mask_cmpeq_epi8_mask() { |
| 15297 | let a = _mm_set1_epi8(-1); |
| 15298 | let b = _mm_set1_epi8(-1); |
| 15299 | let mask = 0b01010101_01010101; |
| 15300 | let r = _mm_mask_cmpeq_epi8_mask(mask, a, b); |
| 15301 | assert_eq!(r, 0b01010101_01010101); |
| 15302 | } |
| 15303 | |
| 15304 | #[simd_test(enable = "avx512bw" )] |
| 15305 | unsafe fn test_mm512_cmpneq_epu16_mask() { |
| 15306 | let a = _mm512_set1_epi16(2); |
| 15307 | let b = _mm512_set1_epi16(1); |
| 15308 | let m = _mm512_cmpneq_epu16_mask(a, b); |
| 15309 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15310 | } |
| 15311 | |
| 15312 | #[simd_test(enable = "avx512bw" )] |
| 15313 | unsafe fn test_mm512_mask_cmpneq_epu16_mask() { |
| 15314 | let a = _mm512_set1_epi16(2); |
| 15315 | let b = _mm512_set1_epi16(1); |
| 15316 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15317 | let r = _mm512_mask_cmpneq_epu16_mask(mask, a, b); |
| 15318 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15319 | } |
| 15320 | |
| 15321 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15322 | unsafe fn test_mm256_cmpneq_epu16_mask() { |
| 15323 | let a = _mm256_set1_epi16(2); |
| 15324 | let b = _mm256_set1_epi16(1); |
| 15325 | let m = _mm256_cmpneq_epu16_mask(a, b); |
| 15326 | assert_eq!(m, 0b11111111_11111111); |
| 15327 | } |
| 15328 | |
| 15329 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15330 | unsafe fn test_mm256_mask_cmpneq_epu16_mask() { |
| 15331 | let a = _mm256_set1_epi16(2); |
| 15332 | let b = _mm256_set1_epi16(1); |
| 15333 | let mask = 0b01010101_01010101; |
| 15334 | let r = _mm256_mask_cmpneq_epu16_mask(mask, a, b); |
| 15335 | assert_eq!(r, 0b01010101_01010101); |
| 15336 | } |
| 15337 | |
| 15338 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15339 | unsafe fn test_mm_cmpneq_epu16_mask() { |
| 15340 | let a = _mm_set1_epi16(2); |
| 15341 | let b = _mm_set1_epi16(1); |
| 15342 | let m = _mm_cmpneq_epu16_mask(a, b); |
| 15343 | assert_eq!(m, 0b11111111); |
| 15344 | } |
| 15345 | |
| 15346 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15347 | unsafe fn test_mm_mask_cmpneq_epu16_mask() { |
| 15348 | let a = _mm_set1_epi16(2); |
| 15349 | let b = _mm_set1_epi16(1); |
| 15350 | let mask = 0b01010101; |
| 15351 | let r = _mm_mask_cmpneq_epu16_mask(mask, a, b); |
| 15352 | assert_eq!(r, 0b01010101); |
| 15353 | } |
| 15354 | |
| 15355 | #[simd_test(enable = "avx512bw" )] |
| 15356 | unsafe fn test_mm512_cmpneq_epu8_mask() { |
| 15357 | let a = _mm512_set1_epi8(2); |
| 15358 | let b = _mm512_set1_epi8(1); |
| 15359 | let m = _mm512_cmpneq_epu8_mask(a, b); |
| 15360 | assert_eq!( |
| 15361 | m, |
| 15362 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 15363 | ); |
| 15364 | } |
| 15365 | |
| 15366 | #[simd_test(enable = "avx512bw" )] |
| 15367 | unsafe fn test_mm512_mask_cmpneq_epu8_mask() { |
| 15368 | let a = _mm512_set1_epi8(2); |
| 15369 | let b = _mm512_set1_epi8(1); |
| 15370 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 15371 | let r = _mm512_mask_cmpneq_epu8_mask(mask, a, b); |
| 15372 | assert_eq!( |
| 15373 | r, |
| 15374 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 15375 | ); |
| 15376 | } |
| 15377 | |
| 15378 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15379 | unsafe fn test_mm256_cmpneq_epu8_mask() { |
| 15380 | let a = _mm256_set1_epi8(2); |
| 15381 | let b = _mm256_set1_epi8(1); |
| 15382 | let m = _mm256_cmpneq_epu8_mask(a, b); |
| 15383 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15384 | } |
| 15385 | |
| 15386 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15387 | unsafe fn test_mm256_mask_cmpneq_epu8_mask() { |
| 15388 | let a = _mm256_set1_epi8(2); |
| 15389 | let b = _mm256_set1_epi8(1); |
| 15390 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15391 | let r = _mm256_mask_cmpneq_epu8_mask(mask, a, b); |
| 15392 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15393 | } |
| 15394 | |
| 15395 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15396 | unsafe fn test_mm_cmpneq_epu8_mask() { |
| 15397 | let a = _mm_set1_epi8(2); |
| 15398 | let b = _mm_set1_epi8(1); |
| 15399 | let m = _mm_cmpneq_epu8_mask(a, b); |
| 15400 | assert_eq!(m, 0b11111111_11111111); |
| 15401 | } |
| 15402 | |
| 15403 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15404 | unsafe fn test_mm_mask_cmpneq_epu8_mask() { |
| 15405 | let a = _mm_set1_epi8(2); |
| 15406 | let b = _mm_set1_epi8(1); |
| 15407 | let mask = 0b01010101_01010101; |
| 15408 | let r = _mm_mask_cmpneq_epu8_mask(mask, a, b); |
| 15409 | assert_eq!(r, 0b01010101_01010101); |
| 15410 | } |
| 15411 | |
| 15412 | #[simd_test(enable = "avx512bw" )] |
| 15413 | unsafe fn test_mm512_cmpneq_epi16_mask() { |
| 15414 | let a = _mm512_set1_epi16(1); |
| 15415 | let b = _mm512_set1_epi16(-1); |
| 15416 | let m = _mm512_cmpneq_epi16_mask(a, b); |
| 15417 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15418 | } |
| 15419 | |
| 15420 | #[simd_test(enable = "avx512bw" )] |
| 15421 | unsafe fn test_mm512_mask_cmpneq_epi16_mask() { |
| 15422 | let a = _mm512_set1_epi16(1); |
| 15423 | let b = _mm512_set1_epi16(-1); |
| 15424 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15425 | let r = _mm512_mask_cmpneq_epi16_mask(mask, a, b); |
| 15426 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15427 | } |
| 15428 | |
| 15429 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15430 | unsafe fn test_mm256_cmpneq_epi16_mask() { |
| 15431 | let a = _mm256_set1_epi16(1); |
| 15432 | let b = _mm256_set1_epi16(-1); |
| 15433 | let m = _mm256_cmpneq_epi16_mask(a, b); |
| 15434 | assert_eq!(m, 0b11111111_11111111); |
| 15435 | } |
| 15436 | |
| 15437 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15438 | unsafe fn test_mm256_mask_cmpneq_epi16_mask() { |
| 15439 | let a = _mm256_set1_epi16(1); |
| 15440 | let b = _mm256_set1_epi16(-1); |
| 15441 | let mask = 0b01010101_01010101; |
| 15442 | let r = _mm256_mask_cmpneq_epi16_mask(mask, a, b); |
| 15443 | assert_eq!(r, 0b01010101_01010101); |
| 15444 | } |
| 15445 | |
| 15446 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15447 | unsafe fn test_mm_cmpneq_epi16_mask() { |
| 15448 | let a = _mm_set1_epi16(1); |
| 15449 | let b = _mm_set1_epi16(-1); |
| 15450 | let m = _mm_cmpneq_epi16_mask(a, b); |
| 15451 | assert_eq!(m, 0b11111111); |
| 15452 | } |
| 15453 | |
| 15454 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15455 | unsafe fn test_mm_mask_cmpneq_epi16_mask() { |
| 15456 | let a = _mm_set1_epi16(1); |
| 15457 | let b = _mm_set1_epi16(-1); |
| 15458 | let mask = 0b01010101; |
| 15459 | let r = _mm_mask_cmpneq_epi16_mask(mask, a, b); |
| 15460 | assert_eq!(r, 0b01010101); |
| 15461 | } |
| 15462 | |
| 15463 | #[simd_test(enable = "avx512bw" )] |
| 15464 | unsafe fn test_mm512_cmpneq_epi8_mask() { |
| 15465 | let a = _mm512_set1_epi8(1); |
| 15466 | let b = _mm512_set1_epi8(-1); |
| 15467 | let m = _mm512_cmpneq_epi8_mask(a, b); |
| 15468 | assert_eq!( |
| 15469 | m, |
| 15470 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 15471 | ); |
| 15472 | } |
| 15473 | |
| 15474 | #[simd_test(enable = "avx512bw" )] |
| 15475 | unsafe fn test_mm512_mask_cmpneq_epi8_mask() { |
| 15476 | let a = _mm512_set1_epi8(1); |
| 15477 | let b = _mm512_set1_epi8(-1); |
| 15478 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 15479 | let r = _mm512_mask_cmpneq_epi8_mask(mask, a, b); |
| 15480 | assert_eq!( |
| 15481 | r, |
| 15482 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 15483 | ); |
| 15484 | } |
| 15485 | |
| 15486 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15487 | unsafe fn test_mm256_cmpneq_epi8_mask() { |
| 15488 | let a = _mm256_set1_epi8(1); |
| 15489 | let b = _mm256_set1_epi8(-1); |
| 15490 | let m = _mm256_cmpneq_epi8_mask(a, b); |
| 15491 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15492 | } |
| 15493 | |
| 15494 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15495 | unsafe fn test_mm256_mask_cmpneq_epi8_mask() { |
| 15496 | let a = _mm256_set1_epi8(1); |
| 15497 | let b = _mm256_set1_epi8(-1); |
| 15498 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15499 | let r = _mm256_mask_cmpneq_epi8_mask(mask, a, b); |
| 15500 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15501 | } |
| 15502 | |
| 15503 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15504 | unsafe fn test_mm_cmpneq_epi8_mask() { |
| 15505 | let a = _mm_set1_epi8(1); |
| 15506 | let b = _mm_set1_epi8(-1); |
| 15507 | let m = _mm_cmpneq_epi8_mask(a, b); |
| 15508 | assert_eq!(m, 0b11111111_11111111); |
| 15509 | } |
| 15510 | |
| 15511 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15512 | unsafe fn test_mm_mask_cmpneq_epi8_mask() { |
| 15513 | let a = _mm_set1_epi8(1); |
| 15514 | let b = _mm_set1_epi8(-1); |
| 15515 | let mask = 0b01010101_01010101; |
| 15516 | let r = _mm_mask_cmpneq_epi8_mask(mask, a, b); |
| 15517 | assert_eq!(r, 0b01010101_01010101); |
| 15518 | } |
| 15519 | |
| 15520 | #[simd_test(enable = "avx512bw" )] |
| 15521 | unsafe fn test_mm512_cmp_epu16_mask() { |
| 15522 | let a = _mm512_set1_epi16(0); |
| 15523 | let b = _mm512_set1_epi16(1); |
| 15524 | let m = _mm512_cmp_epu16_mask::<_MM_CMPINT_LT>(a, b); |
| 15525 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15526 | } |
| 15527 | |
| 15528 | #[simd_test(enable = "avx512bw" )] |
| 15529 | unsafe fn test_mm512_mask_cmp_epu16_mask() { |
| 15530 | let a = _mm512_set1_epi16(0); |
| 15531 | let b = _mm512_set1_epi16(1); |
| 15532 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15533 | let r = _mm512_mask_cmp_epu16_mask::<_MM_CMPINT_LT>(mask, a, b); |
| 15534 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15535 | } |
| 15536 | |
| 15537 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15538 | unsafe fn test_mm256_cmp_epu16_mask() { |
| 15539 | let a = _mm256_set1_epi16(0); |
| 15540 | let b = _mm256_set1_epi16(1); |
| 15541 | let m = _mm256_cmp_epu16_mask::<_MM_CMPINT_LT>(a, b); |
| 15542 | assert_eq!(m, 0b11111111_11111111); |
| 15543 | } |
| 15544 | |
| 15545 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15546 | unsafe fn test_mm256_mask_cmp_epu16_mask() { |
| 15547 | let a = _mm256_set1_epi16(0); |
| 15548 | let b = _mm256_set1_epi16(1); |
| 15549 | let mask = 0b01010101_01010101; |
| 15550 | let r = _mm256_mask_cmp_epu16_mask::<_MM_CMPINT_LT>(mask, a, b); |
| 15551 | assert_eq!(r, 0b01010101_01010101); |
| 15552 | } |
| 15553 | |
| 15554 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15555 | unsafe fn test_mm_cmp_epu16_mask() { |
| 15556 | let a = _mm_set1_epi16(0); |
| 15557 | let b = _mm_set1_epi16(1); |
| 15558 | let m = _mm_cmp_epu16_mask::<_MM_CMPINT_LT>(a, b); |
| 15559 | assert_eq!(m, 0b11111111); |
| 15560 | } |
| 15561 | |
| 15562 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15563 | unsafe fn test_mm_mask_cmp_epu16_mask() { |
| 15564 | let a = _mm_set1_epi16(0); |
| 15565 | let b = _mm_set1_epi16(1); |
| 15566 | let mask = 0b01010101; |
| 15567 | let r = _mm_mask_cmp_epu16_mask::<_MM_CMPINT_LT>(mask, a, b); |
| 15568 | assert_eq!(r, 0b01010101); |
| 15569 | } |
| 15570 | |
| 15571 | #[simd_test(enable = "avx512bw" )] |
| 15572 | unsafe fn test_mm512_cmp_epu8_mask() { |
| 15573 | let a = _mm512_set1_epi8(0); |
| 15574 | let b = _mm512_set1_epi8(1); |
| 15575 | let m = _mm512_cmp_epu8_mask::<_MM_CMPINT_LT>(a, b); |
| 15576 | assert_eq!( |
| 15577 | m, |
| 15578 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 15579 | ); |
| 15580 | } |
| 15581 | |
| 15582 | #[simd_test(enable = "avx512bw" )] |
| 15583 | unsafe fn test_mm512_mask_cmp_epu8_mask() { |
| 15584 | let a = _mm512_set1_epi8(0); |
| 15585 | let b = _mm512_set1_epi8(1); |
| 15586 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 15587 | let r = _mm512_mask_cmp_epu8_mask::<_MM_CMPINT_LT>(mask, a, b); |
| 15588 | assert_eq!( |
| 15589 | r, |
| 15590 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 15591 | ); |
| 15592 | } |
| 15593 | |
| 15594 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15595 | unsafe fn test_mm256_cmp_epu8_mask() { |
| 15596 | let a = _mm256_set1_epi8(0); |
| 15597 | let b = _mm256_set1_epi8(1); |
| 15598 | let m = _mm256_cmp_epu8_mask::<_MM_CMPINT_LT>(a, b); |
| 15599 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15600 | } |
| 15601 | |
| 15602 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15603 | unsafe fn test_mm256_mask_cmp_epu8_mask() { |
| 15604 | let a = _mm256_set1_epi8(0); |
| 15605 | let b = _mm256_set1_epi8(1); |
| 15606 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15607 | let r = _mm256_mask_cmp_epu8_mask::<_MM_CMPINT_LT>(mask, a, b); |
| 15608 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15609 | } |
| 15610 | |
| 15611 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15612 | unsafe fn test_mm_cmp_epu8_mask() { |
| 15613 | let a = _mm_set1_epi8(0); |
| 15614 | let b = _mm_set1_epi8(1); |
| 15615 | let m = _mm_cmp_epu8_mask::<_MM_CMPINT_LT>(a, b); |
| 15616 | assert_eq!(m, 0b11111111_11111111); |
| 15617 | } |
| 15618 | |
| 15619 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15620 | unsafe fn test_mm_mask_cmp_epu8_mask() { |
| 15621 | let a = _mm_set1_epi8(0); |
| 15622 | let b = _mm_set1_epi8(1); |
| 15623 | let mask = 0b01010101_01010101; |
| 15624 | let r = _mm_mask_cmp_epu8_mask::<_MM_CMPINT_LT>(mask, a, b); |
| 15625 | assert_eq!(r, 0b01010101_01010101); |
| 15626 | } |
| 15627 | |
| 15628 | #[simd_test(enable = "avx512bw" )] |
| 15629 | unsafe fn test_mm512_cmp_epi16_mask() { |
| 15630 | let a = _mm512_set1_epi16(0); |
| 15631 | let b = _mm512_set1_epi16(1); |
| 15632 | let m = _mm512_cmp_epi16_mask::<_MM_CMPINT_LT>(a, b); |
| 15633 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15634 | } |
| 15635 | |
| 15636 | #[simd_test(enable = "avx512bw" )] |
| 15637 | unsafe fn test_mm512_mask_cmp_epi16_mask() { |
| 15638 | let a = _mm512_set1_epi16(0); |
| 15639 | let b = _mm512_set1_epi16(1); |
| 15640 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15641 | let r = _mm512_mask_cmp_epi16_mask::<_MM_CMPINT_LT>(mask, a, b); |
| 15642 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15643 | } |
| 15644 | |
| 15645 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15646 | unsafe fn test_mm256_cmp_epi16_mask() { |
| 15647 | let a = _mm256_set1_epi16(0); |
| 15648 | let b = _mm256_set1_epi16(1); |
| 15649 | let m = _mm256_cmp_epi16_mask::<_MM_CMPINT_LT>(a, b); |
| 15650 | assert_eq!(m, 0b11111111_11111111); |
| 15651 | } |
| 15652 | |
| 15653 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15654 | unsafe fn test_mm256_mask_cmp_epi16_mask() { |
| 15655 | let a = _mm256_set1_epi16(0); |
| 15656 | let b = _mm256_set1_epi16(1); |
| 15657 | let mask = 0b01010101_01010101; |
| 15658 | let r = _mm256_mask_cmp_epi16_mask::<_MM_CMPINT_LT>(mask, a, b); |
| 15659 | assert_eq!(r, 0b01010101_01010101); |
| 15660 | } |
| 15661 | |
| 15662 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15663 | unsafe fn test_mm_cmp_epi16_mask() { |
| 15664 | let a = _mm_set1_epi16(0); |
| 15665 | let b = _mm_set1_epi16(1); |
| 15666 | let m = _mm_cmp_epi16_mask::<_MM_CMPINT_LT>(a, b); |
| 15667 | assert_eq!(m, 0b11111111); |
| 15668 | } |
| 15669 | |
| 15670 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15671 | unsafe fn test_mm_mask_cmp_epi16_mask() { |
| 15672 | let a = _mm_set1_epi16(0); |
| 15673 | let b = _mm_set1_epi16(1); |
| 15674 | let mask = 0b01010101; |
| 15675 | let r = _mm_mask_cmp_epi16_mask::<_MM_CMPINT_LT>(mask, a, b); |
| 15676 | assert_eq!(r, 0b01010101); |
| 15677 | } |
| 15678 | |
| 15679 | #[simd_test(enable = "avx512bw" )] |
| 15680 | unsafe fn test_mm512_cmp_epi8_mask() { |
| 15681 | let a = _mm512_set1_epi8(0); |
| 15682 | let b = _mm512_set1_epi8(1); |
| 15683 | let m = _mm512_cmp_epi8_mask::<_MM_CMPINT_LT>(a, b); |
| 15684 | assert_eq!( |
| 15685 | m, |
| 15686 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
| 15687 | ); |
| 15688 | } |
| 15689 | |
| 15690 | #[simd_test(enable = "avx512bw" )] |
| 15691 | unsafe fn test_mm512_mask_cmp_epi8_mask() { |
| 15692 | let a = _mm512_set1_epi8(0); |
| 15693 | let b = _mm512_set1_epi8(1); |
| 15694 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
| 15695 | let r = _mm512_mask_cmp_epi8_mask::<_MM_CMPINT_LT>(mask, a, b); |
| 15696 | assert_eq!( |
| 15697 | r, |
| 15698 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
| 15699 | ); |
| 15700 | } |
| 15701 | |
| 15702 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15703 | unsafe fn test_mm256_cmp_epi8_mask() { |
| 15704 | let a = _mm256_set1_epi8(0); |
| 15705 | let b = _mm256_set1_epi8(1); |
| 15706 | let m = _mm256_cmp_epi8_mask::<_MM_CMPINT_LT>(a, b); |
| 15707 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
| 15708 | } |
| 15709 | |
| 15710 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15711 | unsafe fn test_mm256_mask_cmp_epi8_mask() { |
| 15712 | let a = _mm256_set1_epi8(0); |
| 15713 | let b = _mm256_set1_epi8(1); |
| 15714 | let mask = 0b01010101_01010101_01010101_01010101; |
| 15715 | let r = _mm256_mask_cmp_epi8_mask::<_MM_CMPINT_LT>(mask, a, b); |
| 15716 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
| 15717 | } |
| 15718 | |
| 15719 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15720 | unsafe fn test_mm_cmp_epi8_mask() { |
| 15721 | let a = _mm_set1_epi8(0); |
| 15722 | let b = _mm_set1_epi8(1); |
| 15723 | let m = _mm_cmp_epi8_mask::<_MM_CMPINT_LT>(a, b); |
| 15724 | assert_eq!(m, 0b11111111_11111111); |
| 15725 | } |
| 15726 | |
| 15727 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15728 | unsafe fn test_mm_mask_cmp_epi8_mask() { |
| 15729 | let a = _mm_set1_epi8(0); |
| 15730 | let b = _mm_set1_epi8(1); |
| 15731 | let mask = 0b01010101_01010101; |
| 15732 | let r = _mm_mask_cmp_epi8_mask::<_MM_CMPINT_LT>(mask, a, b); |
| 15733 | assert_eq!(r, 0b01010101_01010101); |
| 15734 | } |
| 15735 | |
| 15736 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15737 | unsafe fn test_mm256_reduce_add_epi16() { |
| 15738 | let a = _mm256_set1_epi16(1); |
| 15739 | let e = _mm256_reduce_add_epi16(a); |
| 15740 | assert_eq!(16, e); |
| 15741 | } |
| 15742 | |
| 15743 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15744 | unsafe fn test_mm256_mask_reduce_add_epi16() { |
| 15745 | let a = _mm256_set1_epi16(1); |
| 15746 | let e = _mm256_mask_reduce_add_epi16(0b11111111_00000000, a); |
| 15747 | assert_eq!(8, e); |
| 15748 | } |
| 15749 | |
| 15750 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15751 | unsafe fn test_mm_reduce_add_epi16() { |
| 15752 | let a = _mm_set1_epi16(1); |
| 15753 | let e = _mm_reduce_add_epi16(a); |
| 15754 | assert_eq!(8, e); |
| 15755 | } |
| 15756 | |
| 15757 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15758 | unsafe fn test_mm_mask_reduce_add_epi16() { |
| 15759 | let a = _mm_set1_epi16(1); |
| 15760 | let e = _mm_mask_reduce_add_epi16(0b11110000, a); |
| 15761 | assert_eq!(4, e); |
| 15762 | } |
| 15763 | |
| 15764 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15765 | unsafe fn test_mm256_reduce_add_epi8() { |
| 15766 | let a = _mm256_set1_epi8(1); |
| 15767 | let e = _mm256_reduce_add_epi8(a); |
| 15768 | assert_eq!(32, e); |
| 15769 | } |
| 15770 | |
| 15771 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15772 | unsafe fn test_mm256_mask_reduce_add_epi8() { |
| 15773 | let a = _mm256_set1_epi8(1); |
| 15774 | let e = _mm256_mask_reduce_add_epi8(0b11111111_00000000_11111111_00000000, a); |
| 15775 | assert_eq!(16, e); |
| 15776 | } |
| 15777 | |
| 15778 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15779 | unsafe fn test_mm_reduce_add_epi8() { |
| 15780 | let a = _mm_set1_epi8(1); |
| 15781 | let e = _mm_reduce_add_epi8(a); |
| 15782 | assert_eq!(16, e); |
| 15783 | } |
| 15784 | |
| 15785 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15786 | unsafe fn test_mm_mask_reduce_add_epi8() { |
| 15787 | let a = _mm_set1_epi8(1); |
| 15788 | let e = _mm_mask_reduce_add_epi8(0b11111111_00000000, a); |
| 15789 | assert_eq!(8, e); |
| 15790 | } |
| 15791 | |
| 15792 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15793 | unsafe fn test_mm256_reduce_and_epi16() { |
| 15794 | let a = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2); |
| 15795 | let e = _mm256_reduce_and_epi16(a); |
| 15796 | assert_eq!(0, e); |
| 15797 | } |
| 15798 | |
| 15799 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15800 | unsafe fn test_mm256_mask_reduce_and_epi16() { |
| 15801 | let a = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2); |
| 15802 | let e = _mm256_mask_reduce_and_epi16(0b11111111_00000000, a); |
| 15803 | assert_eq!(1, e); |
| 15804 | } |
| 15805 | |
| 15806 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15807 | unsafe fn test_mm_reduce_and_epi16() { |
| 15808 | let a = _mm_set_epi16(1, 1, 1, 1, 2, 2, 2, 2); |
| 15809 | let e = _mm_reduce_and_epi16(a); |
| 15810 | assert_eq!(0, e); |
| 15811 | } |
| 15812 | |
| 15813 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15814 | unsafe fn test_mm_mask_reduce_and_epi16() { |
| 15815 | let a = _mm_set_epi16(1, 1, 1, 1, 2, 2, 2, 2); |
| 15816 | let e = _mm_mask_reduce_and_epi16(0b11110000, a); |
| 15817 | assert_eq!(1, e); |
| 15818 | } |
| 15819 | |
| 15820 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15821 | unsafe fn test_mm256_reduce_and_epi8() { |
| 15822 | let a = _mm256_set_epi8( |
| 15823 | 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, |
| 15824 | 2, 2, 2, |
| 15825 | ); |
| 15826 | let e = _mm256_reduce_and_epi8(a); |
| 15827 | assert_eq!(0, e); |
| 15828 | } |
| 15829 | |
| 15830 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15831 | unsafe fn test_mm256_mask_reduce_and_epi8() { |
| 15832 | let a = _mm256_set_epi8( |
| 15833 | 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, |
| 15834 | 2, 2, 2, |
| 15835 | ); |
| 15836 | let e = _mm256_mask_reduce_and_epi8(0b11111111_00000000_11111111_00000000, a); |
| 15837 | assert_eq!(1, e); |
| 15838 | } |
| 15839 | |
| 15840 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15841 | unsafe fn test_mm_reduce_and_epi8() { |
| 15842 | let a = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2); |
| 15843 | let e = _mm_reduce_and_epi8(a); |
| 15844 | assert_eq!(0, e); |
| 15845 | } |
| 15846 | |
| 15847 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15848 | unsafe fn test_mm_mask_reduce_and_epi8() { |
| 15849 | let a = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2); |
| 15850 | let e = _mm_mask_reduce_and_epi8(0b11111111_00000000, a); |
| 15851 | assert_eq!(1, e); |
| 15852 | } |
| 15853 | |
| 15854 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15855 | unsafe fn test_mm256_reduce_mul_epi16() { |
| 15856 | let a = _mm256_set_epi16(2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1); |
| 15857 | let e = _mm256_reduce_mul_epi16(a); |
| 15858 | assert_eq!(256, e); |
| 15859 | } |
| 15860 | |
| 15861 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15862 | unsafe fn test_mm256_mask_reduce_mul_epi16() { |
| 15863 | let a = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2); |
| 15864 | let e = _mm256_mask_reduce_mul_epi16(0b11111111_00000000, a); |
| 15865 | assert_eq!(1, e); |
| 15866 | } |
| 15867 | |
| 15868 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15869 | unsafe fn test_mm_reduce_mul_epi16() { |
| 15870 | let a = _mm_set_epi16(2, 2, 2, 2, 1, 1, 1, 1); |
| 15871 | let e = _mm_reduce_mul_epi16(a); |
| 15872 | assert_eq!(16, e); |
| 15873 | } |
| 15874 | |
| 15875 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15876 | unsafe fn test_mm_mask_reduce_mul_epi16() { |
| 15877 | let a = _mm_set_epi16(1, 1, 1, 1, 2, 2, 2, 2); |
| 15878 | let e = _mm_mask_reduce_mul_epi16(0b11110000, a); |
| 15879 | assert_eq!(1, e); |
| 15880 | } |
| 15881 | |
| 15882 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15883 | unsafe fn test_mm256_reduce_mul_epi8() { |
| 15884 | let a = _mm256_set_epi8( |
| 15885 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 15886 | 2, 2, 2, |
| 15887 | ); |
| 15888 | let e = _mm256_reduce_mul_epi8(a); |
| 15889 | assert_eq!(64, e); |
| 15890 | } |
| 15891 | |
| 15892 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15893 | unsafe fn test_mm256_mask_reduce_mul_epi8() { |
| 15894 | let a = _mm256_set_epi8( |
| 15895 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 15896 | 2, 2, 2, |
| 15897 | ); |
| 15898 | let e = _mm256_mask_reduce_mul_epi8(0b11111111_00000000_11111111_00000000, a); |
| 15899 | assert_eq!(1, e); |
| 15900 | } |
| 15901 | |
| 15902 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15903 | unsafe fn test_mm_reduce_mul_epi8() { |
| 15904 | let a = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2); |
| 15905 | let e = _mm_reduce_mul_epi8(a); |
| 15906 | assert_eq!(8, e); |
| 15907 | } |
| 15908 | |
| 15909 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15910 | unsafe fn test_mm_mask_reduce_mul_epi8() { |
| 15911 | let a = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2); |
| 15912 | let e = _mm_mask_reduce_mul_epi8(0b11111111_00000000, a); |
| 15913 | assert_eq!(1, e); |
| 15914 | } |
| 15915 | |
| 15916 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15917 | unsafe fn test_mm256_reduce_max_epi16() { |
| 15918 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 15919 | let e: i16 = _mm256_reduce_max_epi16(a); |
| 15920 | assert_eq!(15, e); |
| 15921 | } |
| 15922 | |
| 15923 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15924 | unsafe fn test_mm256_mask_reduce_max_epi16() { |
| 15925 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 15926 | let e: i16 = _mm256_mask_reduce_max_epi16(0b11111111_00000000, a); |
| 15927 | assert_eq!(7, e); |
| 15928 | } |
| 15929 | |
| 15930 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15931 | unsafe fn test_mm_reduce_max_epi16() { |
| 15932 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 15933 | let e: i16 = _mm_reduce_max_epi16(a); |
| 15934 | assert_eq!(7, e); |
| 15935 | } |
| 15936 | |
| 15937 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15938 | unsafe fn test_mm_mask_reduce_max_epi16() { |
| 15939 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 15940 | let e: i16 = _mm_mask_reduce_max_epi16(0b11110000, a); |
| 15941 | assert_eq!(3, e); |
| 15942 | } |
| 15943 | |
| 15944 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15945 | unsafe fn test_mm256_reduce_max_epi8() { |
| 15946 | let a = _mm256_set_epi8( |
| 15947 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 15948 | 24, 25, 26, 27, 28, 29, 30, 31, |
| 15949 | ); |
| 15950 | let e: i8 = _mm256_reduce_max_epi8(a); |
| 15951 | assert_eq!(31, e); |
| 15952 | } |
| 15953 | |
| 15954 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15955 | unsafe fn test_mm256_mask_reduce_max_epi8() { |
| 15956 | let a = _mm256_set_epi8( |
| 15957 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 15958 | 24, 25, 26, 27, 28, 29, 30, 31, |
| 15959 | ); |
| 15960 | let e: i8 = _mm256_mask_reduce_max_epi8(0b1111111111111111_0000000000000000, a); |
| 15961 | assert_eq!(15, e); |
| 15962 | } |
| 15963 | |
| 15964 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15965 | unsafe fn test_mm_reduce_max_epi8() { |
| 15966 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 15967 | let e: i8 = _mm_reduce_max_epi8(a); |
| 15968 | assert_eq!(15, e); |
| 15969 | } |
| 15970 | |
| 15971 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15972 | unsafe fn test_mm_mask_reduce_max_epi8() { |
| 15973 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 15974 | let e: i8 = _mm_mask_reduce_max_epi8(0b11111111_00000000, a); |
| 15975 | assert_eq!(7, e); |
| 15976 | } |
| 15977 | |
| 15978 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15979 | unsafe fn test_mm256_reduce_max_epu16() { |
| 15980 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 15981 | let e: u16 = _mm256_reduce_max_epu16(a); |
| 15982 | assert_eq!(15, e); |
| 15983 | } |
| 15984 | |
| 15985 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15986 | unsafe fn test_mm256_mask_reduce_max_epu16() { |
| 15987 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 15988 | let e: u16 = _mm256_mask_reduce_max_epu16(0b11111111_00000000, a); |
| 15989 | assert_eq!(7, e); |
| 15990 | } |
| 15991 | |
| 15992 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 15993 | unsafe fn test_mm_reduce_max_epu16() { |
| 15994 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 15995 | let e: u16 = _mm_reduce_max_epu16(a); |
| 15996 | assert_eq!(7, e); |
| 15997 | } |
| 15998 | |
| 15999 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16000 | unsafe fn test_mm_mask_reduce_max_epu16() { |
| 16001 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 16002 | let e: u16 = _mm_mask_reduce_max_epu16(0b11110000, a); |
| 16003 | assert_eq!(3, e); |
| 16004 | } |
| 16005 | |
| 16006 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16007 | unsafe fn test_mm256_reduce_max_epu8() { |
| 16008 | let a = _mm256_set_epi8( |
| 16009 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16010 | 24, 25, 26, 27, 28, 29, 30, 31, |
| 16011 | ); |
| 16012 | let e: u8 = _mm256_reduce_max_epu8(a); |
| 16013 | assert_eq!(31, e); |
| 16014 | } |
| 16015 | |
| 16016 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16017 | unsafe fn test_mm256_mask_reduce_max_epu8() { |
| 16018 | let a = _mm256_set_epi8( |
| 16019 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16020 | 24, 25, 26, 27, 28, 29, 30, 31, |
| 16021 | ); |
| 16022 | let e: u8 = _mm256_mask_reduce_max_epu8(0b1111111111111111_0000000000000000, a); |
| 16023 | assert_eq!(15, e); |
| 16024 | } |
| 16025 | |
| 16026 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16027 | unsafe fn test_mm_reduce_max_epu8() { |
| 16028 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 16029 | let e: u8 = _mm_reduce_max_epu8(a); |
| 16030 | assert_eq!(15, e); |
| 16031 | } |
| 16032 | |
| 16033 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16034 | unsafe fn test_mm_mask_reduce_max_epu8() { |
| 16035 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 16036 | let e: u8 = _mm_mask_reduce_max_epu8(0b11111111_00000000, a); |
| 16037 | assert_eq!(7, e); |
| 16038 | } |
| 16039 | |
| 16040 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16041 | unsafe fn test_mm256_reduce_min_epi16() { |
| 16042 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 16043 | let e: i16 = _mm256_reduce_min_epi16(a); |
| 16044 | assert_eq!(0, e); |
| 16045 | } |
| 16046 | |
| 16047 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16048 | unsafe fn test_mm256_mask_reduce_min_epi16() { |
| 16049 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 16050 | let e: i16 = _mm256_mask_reduce_min_epi16(0b11111111_00000000, a); |
| 16051 | assert_eq!(0, e); |
| 16052 | } |
| 16053 | |
| 16054 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16055 | unsafe fn test_mm_reduce_min_epi16() { |
| 16056 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 16057 | let e: i16 = _mm_reduce_min_epi16(a); |
| 16058 | assert_eq!(0, e); |
| 16059 | } |
| 16060 | |
| 16061 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16062 | unsafe fn test_mm_mask_reduce_min_epi16() { |
| 16063 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 16064 | let e: i16 = _mm_mask_reduce_min_epi16(0b11110000, a); |
| 16065 | assert_eq!(0, e); |
| 16066 | } |
| 16067 | |
| 16068 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16069 | unsafe fn test_mm256_reduce_min_epi8() { |
| 16070 | let a = _mm256_set_epi8( |
| 16071 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16072 | 24, 25, 26, 27, 28, 29, 30, 31, |
| 16073 | ); |
| 16074 | let e: i8 = _mm256_reduce_min_epi8(a); |
| 16075 | assert_eq!(0, e); |
| 16076 | } |
| 16077 | |
| 16078 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16079 | unsafe fn test_mm256_mask_reduce_min_epi8() { |
| 16080 | let a = _mm256_set_epi8( |
| 16081 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16082 | 24, 25, 26, 27, 28, 29, 30, 31, |
| 16083 | ); |
| 16084 | let e: i8 = _mm256_mask_reduce_min_epi8(0b1111111111111111_0000000000000000, a); |
| 16085 | assert_eq!(0, e); |
| 16086 | } |
| 16087 | |
| 16088 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16089 | unsafe fn test_mm_reduce_min_epi8() { |
| 16090 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 16091 | let e: i8 = _mm_reduce_min_epi8(a); |
| 16092 | assert_eq!(0, e); |
| 16093 | } |
| 16094 | |
| 16095 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16096 | unsafe fn test_mm_mask_reduce_min_epi8() { |
| 16097 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 16098 | let e: i8 = _mm_mask_reduce_min_epi8(0b11111111_00000000, a); |
| 16099 | assert_eq!(0, e); |
| 16100 | } |
| 16101 | |
| 16102 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16103 | unsafe fn test_mm256_reduce_min_epu16() { |
| 16104 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 16105 | let e: u16 = _mm256_reduce_min_epu16(a); |
| 16106 | assert_eq!(0, e); |
| 16107 | } |
| 16108 | |
| 16109 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16110 | unsafe fn test_mm256_mask_reduce_min_epu16() { |
| 16111 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 16112 | let e: u16 = _mm256_mask_reduce_min_epu16(0b11111111_00000000, a); |
| 16113 | assert_eq!(0, e); |
| 16114 | } |
| 16115 | |
| 16116 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16117 | unsafe fn test_mm_reduce_min_epu16() { |
| 16118 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 16119 | let e: u16 = _mm_reduce_min_epu16(a); |
| 16120 | assert_eq!(0, e); |
| 16121 | } |
| 16122 | |
| 16123 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16124 | unsafe fn test_mm_mask_reduce_min_epu16() { |
| 16125 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 16126 | let e: u16 = _mm_mask_reduce_min_epu16(0b11110000, a); |
| 16127 | assert_eq!(0, e); |
| 16128 | } |
| 16129 | |
| 16130 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16131 | unsafe fn test_mm256_reduce_min_epu8() { |
| 16132 | let a = _mm256_set_epi8( |
| 16133 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16134 | 24, 25, 26, 27, 28, 29, 30, 31, |
| 16135 | ); |
| 16136 | let e: u8 = _mm256_reduce_min_epu8(a); |
| 16137 | assert_eq!(0, e); |
| 16138 | } |
| 16139 | |
| 16140 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16141 | unsafe fn test_mm256_mask_reduce_min_epu8() { |
| 16142 | let a = _mm256_set_epi8( |
| 16143 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16144 | 24, 25, 26, 27, 28, 29, 30, 31, |
| 16145 | ); |
| 16146 | let e: u8 = _mm256_mask_reduce_min_epu8(0b1111111111111111_0000000000000000, a); |
| 16147 | assert_eq!(0, e); |
| 16148 | } |
| 16149 | |
| 16150 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16151 | unsafe fn test_mm_reduce_min_epu8() { |
| 16152 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 16153 | let e: u8 = _mm_reduce_min_epu8(a); |
| 16154 | assert_eq!(0, e); |
| 16155 | } |
| 16156 | |
| 16157 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16158 | unsafe fn test_mm_mask_reduce_min_epu8() { |
| 16159 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 16160 | let e: u8 = _mm_mask_reduce_min_epu8(0b11111111_00000000, a); |
| 16161 | assert_eq!(0, e); |
| 16162 | } |
| 16163 | |
| 16164 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16165 | unsafe fn test_mm256_reduce_or_epi16() { |
| 16166 | let a = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2); |
| 16167 | let e = _mm256_reduce_or_epi16(a); |
| 16168 | assert_eq!(3, e); |
| 16169 | } |
| 16170 | |
| 16171 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16172 | unsafe fn test_mm256_mask_reduce_or_epi16() { |
| 16173 | let a = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2); |
| 16174 | let e = _mm256_mask_reduce_or_epi16(0b11111111_00000000, a); |
| 16175 | assert_eq!(1, e); |
| 16176 | } |
| 16177 | |
| 16178 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16179 | unsafe fn test_mm_reduce_or_epi16() { |
| 16180 | let a = _mm_set_epi16(1, 1, 1, 1, 2, 2, 2, 2); |
| 16181 | let e = _mm_reduce_or_epi16(a); |
| 16182 | assert_eq!(3, e); |
| 16183 | } |
| 16184 | |
| 16185 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16186 | unsafe fn test_mm_mask_reduce_or_epi16() { |
| 16187 | let a = _mm_set_epi16(1, 1, 1, 1, 2, 2, 2, 2); |
| 16188 | let e = _mm_mask_reduce_or_epi16(0b11110000, a); |
| 16189 | assert_eq!(1, e); |
| 16190 | } |
| 16191 | |
| 16192 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16193 | unsafe fn test_mm256_reduce_or_epi8() { |
| 16194 | let a = _mm256_set_epi8( |
| 16195 | 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, |
| 16196 | 2, 2, 2, |
| 16197 | ); |
| 16198 | let e = _mm256_reduce_or_epi8(a); |
| 16199 | assert_eq!(3, e); |
| 16200 | } |
| 16201 | |
| 16202 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16203 | unsafe fn test_mm256_mask_reduce_or_epi8() { |
| 16204 | let a = _mm256_set_epi8( |
| 16205 | 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, |
| 16206 | 2, 2, 2, |
| 16207 | ); |
| 16208 | let e = _mm256_mask_reduce_or_epi8(0b11111111_00000000_11111111_00000000, a); |
| 16209 | assert_eq!(1, e); |
| 16210 | } |
| 16211 | |
| 16212 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16213 | unsafe fn test_mm_reduce_or_epi8() { |
| 16214 | let a = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2); |
| 16215 | let e = _mm_reduce_or_epi8(a); |
| 16216 | assert_eq!(3, e); |
| 16217 | } |
| 16218 | |
| 16219 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16220 | unsafe fn test_mm_mask_reduce_or_epi8() { |
| 16221 | let a = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2); |
| 16222 | let e = _mm_mask_reduce_or_epi8(0b11111111_00000000, a); |
| 16223 | assert_eq!(1, e); |
| 16224 | } |
| 16225 | |
| 16226 | #[simd_test(enable = "avx512bw" )] |
| 16227 | unsafe fn test_mm512_loadu_epi16() { |
| 16228 | #[rustfmt::skip] |
| 16229 | let a: [i16; 32] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32]; |
| 16230 | let r = _mm512_loadu_epi16(&a[0]); |
| 16231 | #[rustfmt::skip] |
| 16232 | let e = _mm512_set_epi16(32, 31, 30, 29, 28, 27, 26, 25, 24, 23, 22, 21, 20, 19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1); |
| 16233 | assert_eq_m512i(r, e); |
| 16234 | } |
| 16235 | |
| 16236 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16237 | unsafe fn test_mm256_loadu_epi16() { |
| 16238 | let a: [i16; 16] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
| 16239 | let r = _mm256_loadu_epi16(&a[0]); |
| 16240 | let e = _mm256_set_epi16(16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1); |
| 16241 | assert_eq_m256i(r, e); |
| 16242 | } |
| 16243 | |
| 16244 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16245 | unsafe fn test_mm_loadu_epi16() { |
| 16246 | let a: [i16; 8] = [1, 2, 3, 4, 5, 6, 7, 8]; |
| 16247 | let r = _mm_loadu_epi16(&a[0]); |
| 16248 | let e = _mm_set_epi16(8, 7, 6, 5, 4, 3, 2, 1); |
| 16249 | assert_eq_m128i(r, e); |
| 16250 | } |
| 16251 | |
| 16252 | #[simd_test(enable = "avx512bw" )] |
| 16253 | unsafe fn test_mm512_loadu_epi8() { |
| 16254 | #[rustfmt::skip] |
| 16255 | let a: [i8; 64] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 16256 | 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32]; |
| 16257 | let r = _mm512_loadu_epi8(&a[0]); |
| 16258 | #[rustfmt::skip] |
| 16259 | let e = _mm512_set_epi8(32, 31, 30, 29, 28, 27, 26, 25, 24, 23, 22, 21, 20, 19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, |
| 16260 | 32, 31, 30, 29, 28, 27, 26, 25, 24, 23, 22, 21, 20, 19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1); |
| 16261 | assert_eq_m512i(r, e); |
| 16262 | } |
| 16263 | |
| 16264 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16265 | unsafe fn test_mm256_loadu_epi8() { |
| 16266 | #[rustfmt::skip] |
| 16267 | let a: [i8; 32] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32]; |
| 16268 | let r = _mm256_loadu_epi8(&a[0]); |
| 16269 | #[rustfmt::skip] |
| 16270 | let e = _mm256_set_epi8(32, 31, 30, 29, 28, 27, 26, 25, 24, 23, 22, 21, 20, 19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1); |
| 16271 | assert_eq_m256i(r, e); |
| 16272 | } |
| 16273 | |
| 16274 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16275 | unsafe fn test_mm_loadu_epi8() { |
| 16276 | let a: [i8; 16] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
| 16277 | let r = _mm_loadu_epi8(&a[0]); |
| 16278 | let e = _mm_set_epi8(16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1); |
| 16279 | assert_eq_m128i(r, e); |
| 16280 | } |
| 16281 | |
| 16282 | #[simd_test(enable = "avx512bw" )] |
| 16283 | unsafe fn test_mm512_storeu_epi16() { |
| 16284 | let a = _mm512_set1_epi16(9); |
| 16285 | let mut r = _mm512_undefined_epi32(); |
| 16286 | _mm512_storeu_epi16(&mut r as *mut _ as *mut i16, a); |
| 16287 | assert_eq_m512i(r, a); |
| 16288 | } |
| 16289 | |
| 16290 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16291 | unsafe fn test_mm256_storeu_epi16() { |
| 16292 | let a = _mm256_set1_epi16(9); |
| 16293 | let mut r = _mm256_set1_epi32(0); |
| 16294 | _mm256_storeu_epi16(&mut r as *mut _ as *mut i16, a); |
| 16295 | assert_eq_m256i(r, a); |
| 16296 | } |
| 16297 | |
| 16298 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16299 | unsafe fn test_mm_storeu_epi16() { |
| 16300 | let a = _mm_set1_epi16(9); |
| 16301 | let mut r = _mm_set1_epi32(0); |
| 16302 | _mm_storeu_epi16(&mut r as *mut _ as *mut i16, a); |
| 16303 | assert_eq_m128i(r, a); |
| 16304 | } |
| 16305 | |
| 16306 | #[simd_test(enable = "avx512bw" )] |
| 16307 | unsafe fn test_mm512_storeu_epi8() { |
| 16308 | let a = _mm512_set1_epi8(9); |
| 16309 | let mut r = _mm512_undefined_epi32(); |
| 16310 | _mm512_storeu_epi8(&mut r as *mut _ as *mut i8, a); |
| 16311 | assert_eq_m512i(r, a); |
| 16312 | } |
| 16313 | |
| 16314 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16315 | unsafe fn test_mm256_storeu_epi8() { |
| 16316 | let a = _mm256_set1_epi8(9); |
| 16317 | let mut r = _mm256_set1_epi32(0); |
| 16318 | _mm256_storeu_epi8(&mut r as *mut _ as *mut i8, a); |
| 16319 | assert_eq_m256i(r, a); |
| 16320 | } |
| 16321 | |
| 16322 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16323 | unsafe fn test_mm_storeu_epi8() { |
| 16324 | let a = _mm_set1_epi8(9); |
| 16325 | let mut r = _mm_set1_epi32(0); |
| 16326 | _mm_storeu_epi8(&mut r as *mut _ as *mut i8, a); |
| 16327 | assert_eq_m128i(r, a); |
| 16328 | } |
| 16329 | |
| 16330 | #[simd_test(enable = "avx512bw" )] |
| 16331 | unsafe fn test_mm512_mask_loadu_epi16() { |
| 16332 | let src = _mm512_set1_epi16(42); |
| 16333 | let a = &[ |
| 16334 | 1_i16, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16335 | 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 16336 | ]; |
| 16337 | let p = a.as_ptr(); |
| 16338 | let m = 0b10101010_11001100_11101000_11001010; |
| 16339 | let r = _mm512_mask_loadu_epi16(src, m, black_box(p)); |
| 16340 | let e = &[ |
| 16341 | 42_i16, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, 42, 42, 19, 20, 42, 42, |
| 16342 | 23, 24, 42, 26, 42, 28, 42, 30, 42, 32, |
| 16343 | ]; |
| 16344 | let e = _mm512_loadu_epi16(e.as_ptr()); |
| 16345 | assert_eq_m512i(r, e); |
| 16346 | } |
| 16347 | |
| 16348 | #[simd_test(enable = "avx512bw" )] |
| 16349 | unsafe fn test_mm512_maskz_loadu_epi16() { |
| 16350 | let a = &[ |
| 16351 | 1_i16, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16352 | 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 16353 | ]; |
| 16354 | let p = a.as_ptr(); |
| 16355 | let m = 0b10101010_11001100_11101000_11001010; |
| 16356 | let r = _mm512_maskz_loadu_epi16(m, black_box(p)); |
| 16357 | let e = &[ |
| 16358 | 0_i16, 2, 0, 4, 0, 0, 7, 8, 0, 0, 0, 12, 0, 14, 15, 16, 0, 0, 19, 20, 0, 0, 23, 24, 0, |
| 16359 | 26, 0, 28, 0, 30, 0, 32, |
| 16360 | ]; |
| 16361 | let e = _mm512_loadu_epi16(e.as_ptr()); |
| 16362 | assert_eq_m512i(r, e); |
| 16363 | } |
| 16364 | |
| 16365 | #[simd_test(enable = "avx512bw" )] |
| 16366 | unsafe fn test_mm512_mask_storeu_epi16() { |
| 16367 | let mut r = [42_i16; 32]; |
| 16368 | let a = &[ |
| 16369 | 1_i16, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16370 | 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 16371 | ]; |
| 16372 | let a = _mm512_loadu_epi16(a.as_ptr()); |
| 16373 | let m = 0b10101010_11001100_11101000_11001010; |
| 16374 | _mm512_mask_storeu_epi16(r.as_mut_ptr(), m, a); |
| 16375 | let e = &[ |
| 16376 | 42_i16, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, 42, 42, 19, 20, 42, 42, |
| 16377 | 23, 24, 42, 26, 42, 28, 42, 30, 42, 32, |
| 16378 | ]; |
| 16379 | let e = _mm512_loadu_epi16(e.as_ptr()); |
| 16380 | assert_eq_m512i(_mm512_loadu_epi16(r.as_ptr()), e); |
| 16381 | } |
| 16382 | |
| 16383 | #[simd_test(enable = "avx512bw" )] |
| 16384 | unsafe fn test_mm512_mask_loadu_epi8() { |
| 16385 | let src = _mm512_set1_epi8(42); |
| 16386 | let a = &[ |
| 16387 | 1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16388 | 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, |
| 16389 | 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, |
| 16390 | ]; |
| 16391 | let p = a.as_ptr(); |
| 16392 | let m = 0b00000000_11111111_11111111_00000000_10101010_11001100_11101000_11001010; |
| 16393 | let r = _mm512_mask_loadu_epi8(src, m, black_box(p)); |
| 16394 | let e = &[ |
| 16395 | 42_i8, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, 42, 42, 19, 20, 42, 42, |
| 16396 | 23, 24, 42, 26, 42, 28, 42, 30, 42, 32, 42, 42, 42, 42, 42, 42, 42, 42, 41, 42, 43, 44, |
| 16397 | 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 42, 42, 42, 42, 42, 42, 42, 42, |
| 16398 | ]; |
| 16399 | let e = _mm512_loadu_epi8(e.as_ptr()); |
| 16400 | assert_eq_m512i(r, e); |
| 16401 | } |
| 16402 | |
| 16403 | #[simd_test(enable = "avx512bw" )] |
| 16404 | unsafe fn test_mm512_maskz_loadu_epi8() { |
| 16405 | let a = &[ |
| 16406 | 1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16407 | 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, |
| 16408 | 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, |
| 16409 | ]; |
| 16410 | let p = a.as_ptr(); |
| 16411 | let m = 0b00000000_11111111_11111111_00000000_10101010_11001100_11101000_11001010; |
| 16412 | let r = _mm512_maskz_loadu_epi8(m, black_box(p)); |
| 16413 | let e = &[ |
| 16414 | 0_i8, 2, 0, 4, 0, 0, 7, 8, 0, 0, 0, 12, 0, 14, 15, 16, 0, 0, 19, 20, 0, 0, 23, 24, 0, |
| 16415 | 26, 0, 28, 0, 30, 0, 32, 0, 0, 0, 0, 0, 0, 0, 0, 41, 42, 43, 44, 45, 46, 47, 48, 49, |
| 16416 | 50, 51, 52, 53, 54, 55, 56, 0, 0, 0, 0, 0, 0, 0, 0, |
| 16417 | ]; |
| 16418 | let e = _mm512_loadu_epi8(e.as_ptr()); |
| 16419 | assert_eq_m512i(r, e); |
| 16420 | } |
| 16421 | |
| 16422 | #[simd_test(enable = "avx512bw" )] |
| 16423 | unsafe fn test_mm512_mask_storeu_epi8() { |
| 16424 | let mut r = [42_i8; 64]; |
| 16425 | let a = &[ |
| 16426 | 1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16427 | 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, |
| 16428 | 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, |
| 16429 | ]; |
| 16430 | let a = _mm512_loadu_epi8(a.as_ptr()); |
| 16431 | let m = 0b00000000_11111111_11111111_00000000_10101010_11001100_11101000_11001010; |
| 16432 | _mm512_mask_storeu_epi8(r.as_mut_ptr(), m, a); |
| 16433 | let e = &[ |
| 16434 | 42_i8, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, 42, 42, 19, 20, 42, 42, |
| 16435 | 23, 24, 42, 26, 42, 28, 42, 30, 42, 32, 42, 42, 42, 42, 42, 42, 42, 42, 41, 42, 43, 44, |
| 16436 | 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 42, 42, 42, 42, 42, 42, 42, 42, |
| 16437 | ]; |
| 16438 | let e = _mm512_loadu_epi8(e.as_ptr()); |
| 16439 | assert_eq_m512i(_mm512_loadu_epi8(r.as_ptr()), e); |
| 16440 | } |
| 16441 | |
| 16442 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16443 | unsafe fn test_mm256_mask_loadu_epi16() { |
| 16444 | let src = _mm256_set1_epi16(42); |
| 16445 | let a = &[1_i16, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
| 16446 | let p = a.as_ptr(); |
| 16447 | let m = 0b11101000_11001010; |
| 16448 | let r = _mm256_mask_loadu_epi16(src, m, black_box(p)); |
| 16449 | let e = &[ |
| 16450 | 42_i16, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, |
| 16451 | ]; |
| 16452 | let e = _mm256_loadu_epi16(e.as_ptr()); |
| 16453 | assert_eq_m256i(r, e); |
| 16454 | } |
| 16455 | |
| 16456 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16457 | unsafe fn test_mm256_maskz_loadu_epi16() { |
| 16458 | let a = &[1_i16, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
| 16459 | let p = a.as_ptr(); |
| 16460 | let m = 0b11101000_11001010; |
| 16461 | let r = _mm256_maskz_loadu_epi16(m, black_box(p)); |
| 16462 | let e = &[0_i16, 2, 0, 4, 0, 0, 7, 8, 0, 0, 0, 12, 0, 14, 15, 16]; |
| 16463 | let e = _mm256_loadu_epi16(e.as_ptr()); |
| 16464 | assert_eq_m256i(r, e); |
| 16465 | } |
| 16466 | |
| 16467 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16468 | unsafe fn test_mm256_mask_storeu_epi16() { |
| 16469 | let mut r = [42_i16; 16]; |
| 16470 | let a = &[1_i16, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
| 16471 | let a = _mm256_loadu_epi16(a.as_ptr()); |
| 16472 | let m = 0b11101000_11001010; |
| 16473 | _mm256_mask_storeu_epi16(r.as_mut_ptr(), m, a); |
| 16474 | let e = &[ |
| 16475 | 42_i16, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, |
| 16476 | ]; |
| 16477 | let e = _mm256_loadu_epi16(e.as_ptr()); |
| 16478 | assert_eq_m256i(_mm256_loadu_epi16(r.as_ptr()), e); |
| 16479 | } |
| 16480 | |
| 16481 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16482 | unsafe fn test_mm256_mask_loadu_epi8() { |
| 16483 | let src = _mm256_set1_epi8(42); |
| 16484 | let a = &[ |
| 16485 | 1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16486 | 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 16487 | ]; |
| 16488 | let p = a.as_ptr(); |
| 16489 | let m = 0b10101010_11001100_11101000_11001010; |
| 16490 | let r = _mm256_mask_loadu_epi8(src, m, black_box(p)); |
| 16491 | let e = &[ |
| 16492 | 42_i8, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, 42, 42, 19, 20, 42, 42, |
| 16493 | 23, 24, 42, 26, 42, 28, 42, 30, 42, 32, |
| 16494 | ]; |
| 16495 | let e = _mm256_loadu_epi8(e.as_ptr()); |
| 16496 | assert_eq_m256i(r, e); |
| 16497 | } |
| 16498 | |
| 16499 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16500 | unsafe fn test_mm256_maskz_loadu_epi8() { |
| 16501 | let a = &[ |
| 16502 | 1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16503 | 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 16504 | ]; |
| 16505 | let p = a.as_ptr(); |
| 16506 | let m = 0b10101010_11001100_11101000_11001010; |
| 16507 | let r = _mm256_maskz_loadu_epi8(m, black_box(p)); |
| 16508 | let e = &[ |
| 16509 | 0_i8, 2, 0, 4, 0, 0, 7, 8, 0, 0, 0, 12, 0, 14, 15, 16, 0, 0, 19, 20, 0, 0, 23, 24, 0, |
| 16510 | 26, 0, 28, 0, 30, 0, 32, |
| 16511 | ]; |
| 16512 | let e = _mm256_loadu_epi8(e.as_ptr()); |
| 16513 | assert_eq_m256i(r, e); |
| 16514 | } |
| 16515 | |
| 16516 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16517 | unsafe fn test_mm256_mask_storeu_epi8() { |
| 16518 | let mut r = [42_i8; 32]; |
| 16519 | let a = &[ |
| 16520 | 1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
| 16521 | 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 16522 | ]; |
| 16523 | let a = _mm256_loadu_epi8(a.as_ptr()); |
| 16524 | let m = 0b10101010_11001100_11101000_11001010; |
| 16525 | _mm256_mask_storeu_epi8(r.as_mut_ptr(), m, a); |
| 16526 | let e = &[ |
| 16527 | 42_i8, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, 42, 42, 19, 20, 42, 42, |
| 16528 | 23, 24, 42, 26, 42, 28, 42, 30, 42, 32, |
| 16529 | ]; |
| 16530 | let e = _mm256_loadu_epi8(e.as_ptr()); |
| 16531 | assert_eq_m256i(_mm256_loadu_epi8(r.as_ptr()), e); |
| 16532 | } |
| 16533 | |
| 16534 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16535 | unsafe fn test_mm_mask_loadu_epi16() { |
| 16536 | let src = _mm_set1_epi16(42); |
| 16537 | let a = &[1_i16, 2, 3, 4, 5, 6, 7, 8]; |
| 16538 | let p = a.as_ptr(); |
| 16539 | let m = 0b11001010; |
| 16540 | let r = _mm_mask_loadu_epi16(src, m, black_box(p)); |
| 16541 | let e = &[42_i16, 2, 42, 4, 42, 42, 7, 8]; |
| 16542 | let e = _mm_loadu_epi16(e.as_ptr()); |
| 16543 | assert_eq_m128i(r, e); |
| 16544 | } |
| 16545 | |
| 16546 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16547 | unsafe fn test_mm_maskz_loadu_epi16() { |
| 16548 | let a = &[1_i16, 2, 3, 4, 5, 6, 7, 8]; |
| 16549 | let p = a.as_ptr(); |
| 16550 | let m = 0b11001010; |
| 16551 | let r = _mm_maskz_loadu_epi16(m, black_box(p)); |
| 16552 | let e = &[0_i16, 2, 0, 4, 0, 0, 7, 8]; |
| 16553 | let e = _mm_loadu_epi16(e.as_ptr()); |
| 16554 | assert_eq_m128i(r, e); |
| 16555 | } |
| 16556 | |
| 16557 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16558 | unsafe fn test_mm_mask_storeu_epi16() { |
| 16559 | let mut r = [42_i16; 8]; |
| 16560 | let a = &[1_i16, 2, 3, 4, 5, 6, 7, 8]; |
| 16561 | let a = _mm_loadu_epi16(a.as_ptr()); |
| 16562 | let m = 0b11001010; |
| 16563 | _mm_mask_storeu_epi16(r.as_mut_ptr(), m, a); |
| 16564 | let e = &[42_i16, 2, 42, 4, 42, 42, 7, 8]; |
| 16565 | let e = _mm_loadu_epi16(e.as_ptr()); |
| 16566 | assert_eq_m128i(_mm_loadu_epi16(r.as_ptr()), e); |
| 16567 | } |
| 16568 | |
| 16569 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16570 | unsafe fn test_mm_mask_loadu_epi8() { |
| 16571 | let src = _mm_set1_epi8(42); |
| 16572 | let a = &[1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
| 16573 | let p = a.as_ptr(); |
| 16574 | let m = 0b11101000_11001010; |
| 16575 | let r = _mm_mask_loadu_epi8(src, m, black_box(p)); |
| 16576 | let e = &[ |
| 16577 | 42_i8, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, |
| 16578 | ]; |
| 16579 | let e = _mm_loadu_epi8(e.as_ptr()); |
| 16580 | assert_eq_m128i(r, e); |
| 16581 | } |
| 16582 | |
| 16583 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16584 | unsafe fn test_mm_maskz_loadu_epi8() { |
| 16585 | let a = &[1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
| 16586 | let p = a.as_ptr(); |
| 16587 | let m = 0b11101000_11001010; |
| 16588 | let r = _mm_maskz_loadu_epi8(m, black_box(p)); |
| 16589 | let e = &[0_i8, 2, 0, 4, 0, 0, 7, 8, 0, 0, 0, 12, 0, 14, 15, 16]; |
| 16590 | let e = _mm_loadu_epi8(e.as_ptr()); |
| 16591 | assert_eq_m128i(r, e); |
| 16592 | } |
| 16593 | |
| 16594 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16595 | unsafe fn test_mm_mask_storeu_epi8() { |
| 16596 | let mut r = [42_i8; 16]; |
| 16597 | let a = &[1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
| 16598 | let a = _mm_loadu_epi8(a.as_ptr()); |
| 16599 | let m = 0b11101000_11001010; |
| 16600 | _mm_mask_storeu_epi8(r.as_mut_ptr(), m, a); |
| 16601 | let e = &[ |
| 16602 | 42_i8, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, |
| 16603 | ]; |
| 16604 | let e = _mm_loadu_epi8(e.as_ptr()); |
| 16605 | assert_eq_m128i(_mm_loadu_epi8(r.as_ptr()), e); |
| 16606 | } |
| 16607 | |
| 16608 | #[simd_test(enable = "avx512bw" )] |
| 16609 | unsafe fn test_mm512_madd_epi16() { |
| 16610 | let a = _mm512_set1_epi16(1); |
| 16611 | let b = _mm512_set1_epi16(1); |
| 16612 | let r = _mm512_madd_epi16(a, b); |
| 16613 | let e = _mm512_set1_epi32(2); |
| 16614 | assert_eq_m512i(r, e); |
| 16615 | } |
| 16616 | |
| 16617 | #[simd_test(enable = "avx512bw" )] |
| 16618 | unsafe fn test_mm512_mask_madd_epi16() { |
| 16619 | let a = _mm512_set1_epi16(1); |
| 16620 | let b = _mm512_set1_epi16(1); |
| 16621 | let r = _mm512_mask_madd_epi16(a, 0, a, b); |
| 16622 | assert_eq_m512i(r, a); |
| 16623 | let r = _mm512_mask_madd_epi16(a, 0b00000000_00001111, a, b); |
| 16624 | let e = _mm512_set_epi32( |
| 16625 | 1 << 16 | 1, |
| 16626 | 1 << 16 | 1, |
| 16627 | 1 << 16 | 1, |
| 16628 | 1 << 16 | 1, |
| 16629 | 1 << 16 | 1, |
| 16630 | 1 << 16 | 1, |
| 16631 | 1 << 16 | 1, |
| 16632 | 1 << 16 | 1, |
| 16633 | 1 << 16 | 1, |
| 16634 | 1 << 16 | 1, |
| 16635 | 1 << 16 | 1, |
| 16636 | 1 << 16 | 1, |
| 16637 | 2, |
| 16638 | 2, |
| 16639 | 2, |
| 16640 | 2, |
| 16641 | ); |
| 16642 | assert_eq_m512i(r, e); |
| 16643 | } |
| 16644 | |
| 16645 | #[simd_test(enable = "avx512bw" )] |
| 16646 | unsafe fn test_mm512_maskz_madd_epi16() { |
| 16647 | let a = _mm512_set1_epi16(1); |
| 16648 | let b = _mm512_set1_epi16(1); |
| 16649 | let r = _mm512_maskz_madd_epi16(0, a, b); |
| 16650 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 16651 | let r = _mm512_maskz_madd_epi16(0b00000000_00001111, a, b); |
| 16652 | let e = _mm512_set_epi32(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2); |
| 16653 | assert_eq_m512i(r, e); |
| 16654 | } |
| 16655 | |
| 16656 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16657 | unsafe fn test_mm256_mask_madd_epi16() { |
| 16658 | let a = _mm256_set1_epi16(1); |
| 16659 | let b = _mm256_set1_epi16(1); |
| 16660 | let r = _mm256_mask_madd_epi16(a, 0, a, b); |
| 16661 | assert_eq_m256i(r, a); |
| 16662 | let r = _mm256_mask_madd_epi16(a, 0b00001111, a, b); |
| 16663 | let e = _mm256_set_epi32( |
| 16664 | 1 << 16 | 1, |
| 16665 | 1 << 16 | 1, |
| 16666 | 1 << 16 | 1, |
| 16667 | 1 << 16 | 1, |
| 16668 | 2, |
| 16669 | 2, |
| 16670 | 2, |
| 16671 | 2, |
| 16672 | ); |
| 16673 | assert_eq_m256i(r, e); |
| 16674 | } |
| 16675 | |
| 16676 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16677 | unsafe fn test_mm256_maskz_madd_epi16() { |
| 16678 | let a = _mm256_set1_epi16(1); |
| 16679 | let b = _mm256_set1_epi16(1); |
| 16680 | let r = _mm256_maskz_madd_epi16(0, a, b); |
| 16681 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 16682 | let r = _mm256_maskz_madd_epi16(0b00001111, a, b); |
| 16683 | let e = _mm256_set_epi32(0, 0, 0, 0, 2, 2, 2, 2); |
| 16684 | assert_eq_m256i(r, e); |
| 16685 | } |
| 16686 | |
| 16687 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16688 | unsafe fn test_mm_mask_madd_epi16() { |
| 16689 | let a = _mm_set1_epi16(1); |
| 16690 | let b = _mm_set1_epi16(1); |
| 16691 | let r = _mm_mask_madd_epi16(a, 0, a, b); |
| 16692 | assert_eq_m128i(r, a); |
| 16693 | let r = _mm_mask_madd_epi16(a, 0b00001111, a, b); |
| 16694 | let e = _mm_set_epi32(2, 2, 2, 2); |
| 16695 | assert_eq_m128i(r, e); |
| 16696 | } |
| 16697 | |
| 16698 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16699 | unsafe fn test_mm_maskz_madd_epi16() { |
| 16700 | let a = _mm_set1_epi16(1); |
| 16701 | let b = _mm_set1_epi16(1); |
| 16702 | let r = _mm_maskz_madd_epi16(0, a, b); |
| 16703 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 16704 | let r = _mm_maskz_madd_epi16(0b00001111, a, b); |
| 16705 | let e = _mm_set_epi32(2, 2, 2, 2); |
| 16706 | assert_eq_m128i(r, e); |
| 16707 | } |
| 16708 | |
| 16709 | #[simd_test(enable = "avx512bw" )] |
| 16710 | unsafe fn test_mm512_maddubs_epi16() { |
| 16711 | let a = _mm512_set1_epi8(1); |
| 16712 | let b = _mm512_set1_epi8(1); |
| 16713 | let r = _mm512_maddubs_epi16(a, b); |
| 16714 | let e = _mm512_set1_epi16(2); |
| 16715 | assert_eq_m512i(r, e); |
| 16716 | } |
| 16717 | |
| 16718 | #[simd_test(enable = "avx512bw" )] |
| 16719 | unsafe fn test_mm512_mask_maddubs_epi16() { |
| 16720 | let a = _mm512_set1_epi8(1); |
| 16721 | let b = _mm512_set1_epi8(1); |
| 16722 | let src = _mm512_set1_epi16(1); |
| 16723 | let r = _mm512_mask_maddubs_epi16(src, 0, a, b); |
| 16724 | assert_eq_m512i(r, src); |
| 16725 | let r = _mm512_mask_add_epi16(src, 0b00000000_00000000_00000000_00000001, a, b); |
| 16726 | #[rustfmt::skip] |
| 16727 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 16728 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1<<9|2); |
| 16729 | assert_eq_m512i(r, e); |
| 16730 | } |
| 16731 | |
| 16732 | #[simd_test(enable = "avx512bw" )] |
| 16733 | unsafe fn test_mm512_maskz_maddubs_epi16() { |
| 16734 | let a = _mm512_set1_epi8(1); |
| 16735 | let b = _mm512_set1_epi8(1); |
| 16736 | let r = _mm512_maskz_maddubs_epi16(0, a, b); |
| 16737 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 16738 | let r = _mm512_maskz_maddubs_epi16(0b00000000_11111111_00000000_11111111, a, b); |
| 16739 | #[rustfmt::skip] |
| 16740 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, |
| 16741 | 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2); |
| 16742 | assert_eq_m512i(r, e); |
| 16743 | } |
| 16744 | |
| 16745 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16746 | unsafe fn test_mm256_mask_maddubs_epi16() { |
| 16747 | let a = _mm256_set1_epi8(1); |
| 16748 | let b = _mm256_set1_epi8(1); |
| 16749 | let src = _mm256_set1_epi16(1); |
| 16750 | let r = _mm256_mask_maddubs_epi16(src, 0, a, b); |
| 16751 | assert_eq_m256i(r, src); |
| 16752 | let r = _mm256_mask_add_epi16(src, 0b00000000_00000001, a, b); |
| 16753 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 << 9 | 2); |
| 16754 | assert_eq_m256i(r, e); |
| 16755 | } |
| 16756 | |
| 16757 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16758 | unsafe fn test_mm256_maskz_maddubs_epi16() { |
| 16759 | let a = _mm256_set1_epi8(1); |
| 16760 | let b = _mm256_set1_epi8(1); |
| 16761 | let r = _mm256_maskz_maddubs_epi16(0, a, b); |
| 16762 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 16763 | let r = _mm256_maskz_maddubs_epi16(0b00000000_11111111, a, b); |
| 16764 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2); |
| 16765 | assert_eq_m256i(r, e); |
| 16766 | } |
| 16767 | |
| 16768 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16769 | unsafe fn test_mm_mask_maddubs_epi16() { |
| 16770 | let a = _mm_set1_epi8(1); |
| 16771 | let b = _mm_set1_epi8(1); |
| 16772 | let src = _mm_set1_epi16(1); |
| 16773 | let r = _mm_mask_maddubs_epi16(src, 0, a, b); |
| 16774 | assert_eq_m128i(r, src); |
| 16775 | let r = _mm_mask_add_epi16(src, 0b00000001, a, b); |
| 16776 | let e = _mm_set_epi16(1, 1, 1, 1, 1, 1, 1, 1 << 9 | 2); |
| 16777 | assert_eq_m128i(r, e); |
| 16778 | } |
| 16779 | |
| 16780 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16781 | unsafe fn test_mm_maskz_maddubs_epi16() { |
| 16782 | let a = _mm_set1_epi8(1); |
| 16783 | let b = _mm_set1_epi8(1); |
| 16784 | let r = _mm_maskz_maddubs_epi16(0, a, b); |
| 16785 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 16786 | let r = _mm_maskz_maddubs_epi16(0b00001111, a, b); |
| 16787 | let e = _mm_set_epi16(0, 0, 0, 0, 2, 2, 2, 2); |
| 16788 | assert_eq_m128i(r, e); |
| 16789 | } |
| 16790 | |
| 16791 | #[simd_test(enable = "avx512bw" )] |
| 16792 | unsafe fn test_mm512_packs_epi32() { |
| 16793 | let a = _mm512_set1_epi32(i32::MAX); |
| 16794 | let b = _mm512_set1_epi32(1); |
| 16795 | let r = _mm512_packs_epi32(a, b); |
| 16796 | #[rustfmt::skip] |
| 16797 | let e = _mm512_set_epi16(1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX, 1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX, |
| 16798 | 1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX, 1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 16799 | assert_eq_m512i(r, e); |
| 16800 | } |
| 16801 | |
| 16802 | #[simd_test(enable = "avx512bw" )] |
| 16803 | unsafe fn test_mm512_mask_packs_epi32() { |
| 16804 | let a = _mm512_set1_epi32(i32::MAX); |
| 16805 | let b = _mm512_set1_epi32(1 << 16 | 1); |
| 16806 | let r = _mm512_mask_packs_epi32(a, 0, a, b); |
| 16807 | assert_eq_m512i(r, a); |
| 16808 | let r = _mm512_mask_packs_epi32(b, 0b00000000_00000000_00000000_00001111, a, b); |
| 16809 | #[rustfmt::skip] |
| 16810 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 16811 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 16812 | assert_eq_m512i(r, e); |
| 16813 | } |
| 16814 | |
| 16815 | #[simd_test(enable = "avx512bw" )] |
| 16816 | unsafe fn test_mm512_maskz_packs_epi32() { |
| 16817 | let a = _mm512_set1_epi32(i32::MAX); |
| 16818 | let b = _mm512_set1_epi32(1); |
| 16819 | let r = _mm512_maskz_packs_epi32(0, a, b); |
| 16820 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 16821 | let r = _mm512_maskz_packs_epi32(0b00000000_00000000_00000000_00001111, a, b); |
| 16822 | #[rustfmt::skip] |
| 16823 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 16824 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 16825 | assert_eq_m512i(r, e); |
| 16826 | } |
| 16827 | |
| 16828 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16829 | unsafe fn test_mm256_mask_packs_epi32() { |
| 16830 | let a = _mm256_set1_epi32(i32::MAX); |
| 16831 | let b = _mm256_set1_epi32(1 << 16 | 1); |
| 16832 | let r = _mm256_mask_packs_epi32(a, 0, a, b); |
| 16833 | assert_eq_m256i(r, a); |
| 16834 | let r = _mm256_mask_packs_epi32(b, 0b00000000_00001111, a, b); |
| 16835 | #[rustfmt::skip] |
| 16836 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 16837 | assert_eq_m256i(r, e); |
| 16838 | } |
| 16839 | |
| 16840 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16841 | unsafe fn test_mm256_maskz_packs_epi32() { |
| 16842 | let a = _mm256_set1_epi32(i32::MAX); |
| 16843 | let b = _mm256_set1_epi32(1); |
| 16844 | let r = _mm256_maskz_packs_epi32(0, a, b); |
| 16845 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 16846 | let r = _mm256_maskz_packs_epi32(0b00000000_00001111, a, b); |
| 16847 | #[rustfmt::skip] |
| 16848 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 16849 | assert_eq_m256i(r, e); |
| 16850 | } |
| 16851 | |
| 16852 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16853 | unsafe fn test_mm_mask_packs_epi32() { |
| 16854 | let a = _mm_set1_epi32(i32::MAX); |
| 16855 | let b = _mm_set1_epi32(1 << 16 | 1); |
| 16856 | let r = _mm_mask_packs_epi32(a, 0, a, b); |
| 16857 | assert_eq_m128i(r, a); |
| 16858 | let r = _mm_mask_packs_epi32(b, 0b00001111, a, b); |
| 16859 | let e = _mm_set_epi16(1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 16860 | assert_eq_m128i(r, e); |
| 16861 | } |
| 16862 | |
| 16863 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16864 | unsafe fn test_mm_maskz_packs_epi32() { |
| 16865 | let a = _mm_set1_epi32(i32::MAX); |
| 16866 | let b = _mm_set1_epi32(1); |
| 16867 | let r = _mm_maskz_packs_epi32(0, a, b); |
| 16868 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 16869 | let r = _mm_maskz_packs_epi32(0b00001111, a, b); |
| 16870 | let e = _mm_set_epi16(0, 0, 0, 0, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
| 16871 | assert_eq_m128i(r, e); |
| 16872 | } |
| 16873 | |
| 16874 | #[simd_test(enable = "avx512bw" )] |
| 16875 | unsafe fn test_mm512_packs_epi16() { |
| 16876 | let a = _mm512_set1_epi16(i16::MAX); |
| 16877 | let b = _mm512_set1_epi16(1); |
| 16878 | let r = _mm512_packs_epi16(a, b); |
| 16879 | #[rustfmt::skip] |
| 16880 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, |
| 16881 | 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, |
| 16882 | 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, |
| 16883 | 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 16884 | assert_eq_m512i(r, e); |
| 16885 | } |
| 16886 | |
| 16887 | #[simd_test(enable = "avx512bw" )] |
| 16888 | unsafe fn test_mm512_mask_packs_epi16() { |
| 16889 | let a = _mm512_set1_epi16(i16::MAX); |
| 16890 | let b = _mm512_set1_epi16(1 << 8 | 1); |
| 16891 | let r = _mm512_mask_packs_epi16(a, 0, a, b); |
| 16892 | assert_eq_m512i(r, a); |
| 16893 | let r = _mm512_mask_packs_epi16( |
| 16894 | b, |
| 16895 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 16896 | a, |
| 16897 | b, |
| 16898 | ); |
| 16899 | #[rustfmt::skip] |
| 16900 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 16901 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 16902 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 16903 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 16904 | assert_eq_m512i(r, e); |
| 16905 | } |
| 16906 | |
| 16907 | #[simd_test(enable = "avx512bw" )] |
| 16908 | unsafe fn test_mm512_maskz_packs_epi16() { |
| 16909 | let a = _mm512_set1_epi16(i16::MAX); |
| 16910 | let b = _mm512_set1_epi16(1); |
| 16911 | let r = _mm512_maskz_packs_epi16(0, a, b); |
| 16912 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 16913 | let r = _mm512_maskz_packs_epi16( |
| 16914 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 16915 | a, |
| 16916 | b, |
| 16917 | ); |
| 16918 | #[rustfmt::skip] |
| 16919 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 16920 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 16921 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 16922 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 16923 | assert_eq_m512i(r, e); |
| 16924 | } |
| 16925 | |
| 16926 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16927 | unsafe fn test_mm256_mask_packs_epi16() { |
| 16928 | let a = _mm256_set1_epi16(i16::MAX); |
| 16929 | let b = _mm256_set1_epi16(1 << 8 | 1); |
| 16930 | let r = _mm256_mask_packs_epi16(a, 0, a, b); |
| 16931 | assert_eq_m256i(r, a); |
| 16932 | let r = _mm256_mask_packs_epi16(b, 0b00000000_00000000_00000000_00001111, a, b); |
| 16933 | #[rustfmt::skip] |
| 16934 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 16935 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 16936 | assert_eq_m256i(r, e); |
| 16937 | } |
| 16938 | |
| 16939 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16940 | unsafe fn test_mm256_maskz_packs_epi16() { |
| 16941 | let a = _mm256_set1_epi16(i16::MAX); |
| 16942 | let b = _mm256_set1_epi16(1); |
| 16943 | let r = _mm256_maskz_packs_epi16(0, a, b); |
| 16944 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 16945 | let r = _mm256_maskz_packs_epi16(0b00000000_00000000_00000000_00001111, a, b); |
| 16946 | #[rustfmt::skip] |
| 16947 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 16948 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 16949 | assert_eq_m256i(r, e); |
| 16950 | } |
| 16951 | |
| 16952 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16953 | unsafe fn test_mm_mask_packs_epi16() { |
| 16954 | let a = _mm_set1_epi16(i16::MAX); |
| 16955 | let b = _mm_set1_epi16(1 << 8 | 1); |
| 16956 | let r = _mm_mask_packs_epi16(a, 0, a, b); |
| 16957 | assert_eq_m128i(r, a); |
| 16958 | let r = _mm_mask_packs_epi16(b, 0b00000000_00001111, a, b); |
| 16959 | #[rustfmt::skip] |
| 16960 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 16961 | assert_eq_m128i(r, e); |
| 16962 | } |
| 16963 | |
| 16964 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 16965 | unsafe fn test_mm_maskz_packs_epi16() { |
| 16966 | let a = _mm_set1_epi16(i16::MAX); |
| 16967 | let b = _mm_set1_epi16(1); |
| 16968 | let r = _mm_maskz_packs_epi16(0, a, b); |
| 16969 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 16970 | let r = _mm_maskz_packs_epi16(0b00000000_00001111, a, b); |
| 16971 | #[rustfmt::skip] |
| 16972 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 16973 | assert_eq_m128i(r, e); |
| 16974 | } |
| 16975 | |
| 16976 | #[simd_test(enable = "avx512bw" )] |
| 16977 | unsafe fn test_mm512_packus_epi32() { |
| 16978 | let a = _mm512_set1_epi32(-1); |
| 16979 | let b = _mm512_set1_epi32(1); |
| 16980 | let r = _mm512_packus_epi32(a, b); |
| 16981 | #[rustfmt::skip] |
| 16982 | let e = _mm512_set_epi16(1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, |
| 16983 | 1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0); |
| 16984 | assert_eq_m512i(r, e); |
| 16985 | } |
| 16986 | |
| 16987 | #[simd_test(enable = "avx512bw" )] |
| 16988 | unsafe fn test_mm512_mask_packus_epi32() { |
| 16989 | let a = _mm512_set1_epi32(-1); |
| 16990 | let b = _mm512_set1_epi32(1 << 16 | 1); |
| 16991 | let r = _mm512_mask_packus_epi32(a, 0, a, b); |
| 16992 | assert_eq_m512i(r, a); |
| 16993 | let r = _mm512_mask_packus_epi32(b, 0b00000000_00000000_00000000_00001111, a, b); |
| 16994 | #[rustfmt::skip] |
| 16995 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 16996 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 16997 | assert_eq_m512i(r, e); |
| 16998 | } |
| 16999 | |
| 17000 | #[simd_test(enable = "avx512bw" )] |
| 17001 | unsafe fn test_mm512_maskz_packus_epi32() { |
| 17002 | let a = _mm512_set1_epi32(-1); |
| 17003 | let b = _mm512_set1_epi32(1); |
| 17004 | let r = _mm512_maskz_packus_epi32(0, a, b); |
| 17005 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17006 | let r = _mm512_maskz_packus_epi32(0b00000000_00000000_00000000_00001111, a, b); |
| 17007 | #[rustfmt::skip] |
| 17008 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17009 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 17010 | assert_eq_m512i(r, e); |
| 17011 | } |
| 17012 | |
| 17013 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17014 | unsafe fn test_mm256_mask_packus_epi32() { |
| 17015 | let a = _mm256_set1_epi32(-1); |
| 17016 | let b = _mm256_set1_epi32(1 << 16 | 1); |
| 17017 | let r = _mm256_mask_packus_epi32(a, 0, a, b); |
| 17018 | assert_eq_m256i(r, a); |
| 17019 | let r = _mm256_mask_packus_epi32(b, 0b00000000_00001111, a, b); |
| 17020 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 17021 | assert_eq_m256i(r, e); |
| 17022 | } |
| 17023 | |
| 17024 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17025 | unsafe fn test_mm256_maskz_packus_epi32() { |
| 17026 | let a = _mm256_set1_epi32(-1); |
| 17027 | let b = _mm256_set1_epi32(1); |
| 17028 | let r = _mm256_maskz_packus_epi32(0, a, b); |
| 17029 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 17030 | let r = _mm256_maskz_packus_epi32(0b00000000_00001111, a, b); |
| 17031 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 17032 | assert_eq_m256i(r, e); |
| 17033 | } |
| 17034 | |
| 17035 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17036 | unsafe fn test_mm_mask_packus_epi32() { |
| 17037 | let a = _mm_set1_epi32(-1); |
| 17038 | let b = _mm_set1_epi32(1 << 16 | 1); |
| 17039 | let r = _mm_mask_packus_epi32(a, 0, a, b); |
| 17040 | assert_eq_m128i(r, a); |
| 17041 | let r = _mm_mask_packus_epi32(b, 0b00001111, a, b); |
| 17042 | let e = _mm_set_epi16(1, 1, 1, 1, 0, 0, 0, 0); |
| 17043 | assert_eq_m128i(r, e); |
| 17044 | } |
| 17045 | |
| 17046 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17047 | unsafe fn test_mm_maskz_packus_epi32() { |
| 17048 | let a = _mm_set1_epi32(-1); |
| 17049 | let b = _mm_set1_epi32(1); |
| 17050 | let r = _mm_maskz_packus_epi32(0, a, b); |
| 17051 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 17052 | let r = _mm_maskz_packus_epi32(0b00001111, a, b); |
| 17053 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 0, 0); |
| 17054 | assert_eq_m128i(r, e); |
| 17055 | } |
| 17056 | |
| 17057 | #[simd_test(enable = "avx512bw" )] |
| 17058 | unsafe fn test_mm512_packus_epi16() { |
| 17059 | let a = _mm512_set1_epi16(-1); |
| 17060 | let b = _mm512_set1_epi16(1); |
| 17061 | let r = _mm512_packus_epi16(a, b); |
| 17062 | #[rustfmt::skip] |
| 17063 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17064 | 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17065 | 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17066 | 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0); |
| 17067 | assert_eq_m512i(r, e); |
| 17068 | } |
| 17069 | |
| 17070 | #[simd_test(enable = "avx512bw" )] |
| 17071 | unsafe fn test_mm512_mask_packus_epi16() { |
| 17072 | let a = _mm512_set1_epi16(-1); |
| 17073 | let b = _mm512_set1_epi16(1 << 8 | 1); |
| 17074 | let r = _mm512_mask_packus_epi16(a, 0, a, b); |
| 17075 | assert_eq_m512i(r, a); |
| 17076 | let r = _mm512_mask_packus_epi16( |
| 17077 | b, |
| 17078 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 17079 | a, |
| 17080 | b, |
| 17081 | ); |
| 17082 | #[rustfmt::skip] |
| 17083 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 17084 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 17085 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 17086 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 17087 | assert_eq_m512i(r, e); |
| 17088 | } |
| 17089 | |
| 17090 | #[simd_test(enable = "avx512bw" )] |
| 17091 | unsafe fn test_mm512_maskz_packus_epi16() { |
| 17092 | let a = _mm512_set1_epi16(-1); |
| 17093 | let b = _mm512_set1_epi16(1); |
| 17094 | let r = _mm512_maskz_packus_epi16(0, a, b); |
| 17095 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17096 | let r = _mm512_maskz_packus_epi16( |
| 17097 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 17098 | a, |
| 17099 | b, |
| 17100 | ); |
| 17101 | #[rustfmt::skip] |
| 17102 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17103 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17104 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17105 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 17106 | assert_eq_m512i(r, e); |
| 17107 | } |
| 17108 | |
| 17109 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17110 | unsafe fn test_mm256_mask_packus_epi16() { |
| 17111 | let a = _mm256_set1_epi16(-1); |
| 17112 | let b = _mm256_set1_epi16(1 << 8 | 1); |
| 17113 | let r = _mm256_mask_packus_epi16(a, 0, a, b); |
| 17114 | assert_eq_m256i(r, a); |
| 17115 | let r = _mm256_mask_packus_epi16(b, 0b00000000_00000000_00000000_00001111, a, b); |
| 17116 | #[rustfmt::skip] |
| 17117 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 17118 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 17119 | assert_eq_m256i(r, e); |
| 17120 | } |
| 17121 | |
| 17122 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17123 | unsafe fn test_mm256_maskz_packus_epi16() { |
| 17124 | let a = _mm256_set1_epi16(-1); |
| 17125 | let b = _mm256_set1_epi16(1); |
| 17126 | let r = _mm256_maskz_packus_epi16(0, a, b); |
| 17127 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 17128 | let r = _mm256_maskz_packus_epi16(0b00000000_00000000_00000000_00001111, a, b); |
| 17129 | #[rustfmt::skip] |
| 17130 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17131 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 17132 | assert_eq_m256i(r, e); |
| 17133 | } |
| 17134 | |
| 17135 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17136 | unsafe fn test_mm_mask_packus_epi16() { |
| 17137 | let a = _mm_set1_epi16(-1); |
| 17138 | let b = _mm_set1_epi16(1 << 8 | 1); |
| 17139 | let r = _mm_mask_packus_epi16(a, 0, a, b); |
| 17140 | assert_eq_m128i(r, a); |
| 17141 | let r = _mm_mask_packus_epi16(b, 0b00000000_00001111, a, b); |
| 17142 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
| 17143 | assert_eq_m128i(r, e); |
| 17144 | } |
| 17145 | |
| 17146 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17147 | unsafe fn test_mm_maskz_packus_epi16() { |
| 17148 | let a = _mm_set1_epi16(-1); |
| 17149 | let b = _mm_set1_epi16(1); |
| 17150 | let r = _mm_maskz_packus_epi16(0, a, b); |
| 17151 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 17152 | let r = _mm_maskz_packus_epi16(0b00000000_00001111, a, b); |
| 17153 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| 17154 | assert_eq_m128i(r, e); |
| 17155 | } |
| 17156 | |
| 17157 | #[simd_test(enable = "avx512bw" )] |
| 17158 | unsafe fn test_mm512_avg_epu16() { |
| 17159 | let a = _mm512_set1_epi16(1); |
| 17160 | let b = _mm512_set1_epi16(1); |
| 17161 | let r = _mm512_avg_epu16(a, b); |
| 17162 | let e = _mm512_set1_epi16(1); |
| 17163 | assert_eq_m512i(r, e); |
| 17164 | } |
| 17165 | |
| 17166 | #[simd_test(enable = "avx512bw" )] |
| 17167 | unsafe fn test_mm512_mask_avg_epu16() { |
| 17168 | let a = _mm512_set1_epi16(1); |
| 17169 | let b = _mm512_set1_epi16(1); |
| 17170 | let r = _mm512_mask_avg_epu16(a, 0, a, b); |
| 17171 | assert_eq_m512i(r, a); |
| 17172 | let r = _mm512_mask_avg_epu16(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 17173 | #[rustfmt::skip] |
| 17174 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 17175 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
| 17176 | assert_eq_m512i(r, e); |
| 17177 | } |
| 17178 | |
| 17179 | #[simd_test(enable = "avx512bw" )] |
| 17180 | unsafe fn test_mm512_maskz_avg_epu16() { |
| 17181 | let a = _mm512_set1_epi16(1); |
| 17182 | let b = _mm512_set1_epi16(1); |
| 17183 | let r = _mm512_maskz_avg_epu16(0, a, b); |
| 17184 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17185 | let r = _mm512_maskz_avg_epu16(0b00000000_00000000_00000000_00001111, a, b); |
| 17186 | #[rustfmt::skip] |
| 17187 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17188 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
| 17189 | assert_eq_m512i(r, e); |
| 17190 | } |
| 17191 | |
| 17192 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17193 | unsafe fn test_mm256_mask_avg_epu16() { |
| 17194 | let a = _mm256_set1_epi16(1); |
| 17195 | let b = _mm256_set1_epi16(1); |
| 17196 | let r = _mm256_mask_avg_epu16(a, 0, a, b); |
| 17197 | assert_eq_m256i(r, a); |
| 17198 | let r = _mm256_mask_avg_epu16(a, 0b00000000_00001111, a, b); |
| 17199 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
| 17200 | assert_eq_m256i(r, e); |
| 17201 | } |
| 17202 | |
| 17203 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17204 | unsafe fn test_mm256_maskz_avg_epu16() { |
| 17205 | let a = _mm256_set1_epi16(1); |
| 17206 | let b = _mm256_set1_epi16(1); |
| 17207 | let r = _mm256_maskz_avg_epu16(0, a, b); |
| 17208 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 17209 | let r = _mm256_maskz_avg_epu16(0b00000000_00001111, a, b); |
| 17210 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
| 17211 | assert_eq_m256i(r, e); |
| 17212 | } |
| 17213 | |
| 17214 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17215 | unsafe fn test_mm_mask_avg_epu16() { |
| 17216 | let a = _mm_set1_epi16(1); |
| 17217 | let b = _mm_set1_epi16(1); |
| 17218 | let r = _mm_mask_avg_epu16(a, 0, a, b); |
| 17219 | assert_eq_m128i(r, a); |
| 17220 | let r = _mm_mask_avg_epu16(a, 0b00001111, a, b); |
| 17221 | let e = _mm_set_epi16(1, 1, 1, 1, 1, 1, 1, 1); |
| 17222 | assert_eq_m128i(r, e); |
| 17223 | } |
| 17224 | |
| 17225 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17226 | unsafe fn test_mm_maskz_avg_epu16() { |
| 17227 | let a = _mm_set1_epi16(1); |
| 17228 | let b = _mm_set1_epi16(1); |
| 17229 | let r = _mm_maskz_avg_epu16(0, a, b); |
| 17230 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 17231 | let r = _mm_maskz_avg_epu16(0b00001111, a, b); |
| 17232 | let e = _mm_set_epi16(0, 0, 0, 0, 1, 1, 1, 1); |
| 17233 | assert_eq_m128i(r, e); |
| 17234 | } |
| 17235 | |
| 17236 | #[simd_test(enable = "avx512bw" )] |
| 17237 | unsafe fn test_mm512_avg_epu8() { |
| 17238 | let a = _mm512_set1_epi8(1); |
| 17239 | let b = _mm512_set1_epi8(1); |
| 17240 | let r = _mm512_avg_epu8(a, b); |
| 17241 | let e = _mm512_set1_epi8(1); |
| 17242 | assert_eq_m512i(r, e); |
| 17243 | } |
| 17244 | |
| 17245 | #[simd_test(enable = "avx512bw" )] |
| 17246 | unsafe fn test_mm512_mask_avg_epu8() { |
| 17247 | let a = _mm512_set1_epi8(1); |
| 17248 | let b = _mm512_set1_epi8(1); |
| 17249 | let r = _mm512_mask_avg_epu8(a, 0, a, b); |
| 17250 | assert_eq_m512i(r, a); |
| 17251 | let r = _mm512_mask_avg_epu8( |
| 17252 | a, |
| 17253 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
| 17254 | a, |
| 17255 | b, |
| 17256 | ); |
| 17257 | #[rustfmt::skip] |
| 17258 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 17259 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 17260 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 17261 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
| 17262 | assert_eq_m512i(r, e); |
| 17263 | } |
| 17264 | |
| 17265 | #[simd_test(enable = "avx512bw" )] |
| 17266 | unsafe fn test_mm512_maskz_avg_epu8() { |
| 17267 | let a = _mm512_set1_epi8(1); |
| 17268 | let b = _mm512_set1_epi8(1); |
| 17269 | let r = _mm512_maskz_avg_epu8(0, a, b); |
| 17270 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17271 | let r = _mm512_maskz_avg_epu8( |
| 17272 | 0b00000000_000000000_00000000_00000000_00000000_0000000_00000000_00001111, |
| 17273 | a, |
| 17274 | b, |
| 17275 | ); |
| 17276 | #[rustfmt::skip] |
| 17277 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17278 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17279 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17280 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
| 17281 | assert_eq_m512i(r, e); |
| 17282 | } |
| 17283 | |
| 17284 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17285 | unsafe fn test_mm256_mask_avg_epu8() { |
| 17286 | let a = _mm256_set1_epi8(1); |
| 17287 | let b = _mm256_set1_epi8(1); |
| 17288 | let r = _mm256_mask_avg_epu8(a, 0, a, b); |
| 17289 | assert_eq_m256i(r, a); |
| 17290 | let r = _mm256_mask_avg_epu8(a, 0b00000000_00000000_00000000_00001111, a, b); |
| 17291 | #[rustfmt::skip] |
| 17292 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
| 17293 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
| 17294 | assert_eq_m256i(r, e); |
| 17295 | } |
| 17296 | |
| 17297 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17298 | unsafe fn test_mm256_maskz_avg_epu8() { |
| 17299 | let a = _mm256_set1_epi8(1); |
| 17300 | let b = _mm256_set1_epi8(1); |
| 17301 | let r = _mm256_maskz_avg_epu8(0, a, b); |
| 17302 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 17303 | let r = _mm256_maskz_avg_epu8(0b00000000_0000000_00000000_00001111, a, b); |
| 17304 | #[rustfmt::skip] |
| 17305 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 17306 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
| 17307 | assert_eq_m256i(r, e); |
| 17308 | } |
| 17309 | |
| 17310 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17311 | unsafe fn test_mm_mask_avg_epu8() { |
| 17312 | let a = _mm_set1_epi8(1); |
| 17313 | let b = _mm_set1_epi8(1); |
| 17314 | let r = _mm_mask_avg_epu8(a, 0, a, b); |
| 17315 | assert_eq_m128i(r, a); |
| 17316 | let r = _mm_mask_avg_epu8(a, 0b00000000_00001111, a, b); |
| 17317 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
| 17318 | assert_eq_m128i(r, e); |
| 17319 | } |
| 17320 | |
| 17321 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17322 | unsafe fn test_mm_maskz_avg_epu8() { |
| 17323 | let a = _mm_set1_epi8(1); |
| 17324 | let b = _mm_set1_epi8(1); |
| 17325 | let r = _mm_maskz_avg_epu8(0, a, b); |
| 17326 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 17327 | let r = _mm_maskz_avg_epu8(0b00000000_00001111, a, b); |
| 17328 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
| 17329 | assert_eq_m128i(r, e); |
| 17330 | } |
| 17331 | |
| 17332 | #[simd_test(enable = "avx512bw" )] |
| 17333 | unsafe fn test_mm512_sll_epi16() { |
| 17334 | let a = _mm512_set1_epi16(1 << 15); |
| 17335 | let count = _mm_set1_epi16(2); |
| 17336 | let r = _mm512_sll_epi16(a, count); |
| 17337 | let e = _mm512_set1_epi16(0); |
| 17338 | assert_eq_m512i(r, e); |
| 17339 | } |
| 17340 | |
| 17341 | #[simd_test(enable = "avx512bw" )] |
| 17342 | unsafe fn test_mm512_mask_sll_epi16() { |
| 17343 | let a = _mm512_set1_epi16(1 << 15); |
| 17344 | let count = _mm_set1_epi16(2); |
| 17345 | let r = _mm512_mask_sll_epi16(a, 0, a, count); |
| 17346 | assert_eq_m512i(r, a); |
| 17347 | let r = _mm512_mask_sll_epi16(a, 0b11111111_11111111_11111111_11111111, a, count); |
| 17348 | let e = _mm512_set1_epi16(0); |
| 17349 | assert_eq_m512i(r, e); |
| 17350 | } |
| 17351 | |
| 17352 | #[simd_test(enable = "avx512bw" )] |
| 17353 | unsafe fn test_mm512_maskz_sll_epi16() { |
| 17354 | let a = _mm512_set1_epi16(1 << 15); |
| 17355 | let count = _mm_set1_epi16(2); |
| 17356 | let r = _mm512_maskz_sll_epi16(0, a, count); |
| 17357 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17358 | let r = _mm512_maskz_sll_epi16(0b11111111_11111111_11111111_11111111, a, count); |
| 17359 | let e = _mm512_set1_epi16(0); |
| 17360 | assert_eq_m512i(r, e); |
| 17361 | } |
| 17362 | |
| 17363 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17364 | unsafe fn test_mm256_mask_sll_epi16() { |
| 17365 | let a = _mm256_set1_epi16(1 << 15); |
| 17366 | let count = _mm_set1_epi16(2); |
| 17367 | let r = _mm256_mask_sll_epi16(a, 0, a, count); |
| 17368 | assert_eq_m256i(r, a); |
| 17369 | let r = _mm256_mask_sll_epi16(a, 0b11111111_11111111, a, count); |
| 17370 | let e = _mm256_set1_epi16(0); |
| 17371 | assert_eq_m256i(r, e); |
| 17372 | } |
| 17373 | |
| 17374 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17375 | unsafe fn test_mm256_maskz_sll_epi16() { |
| 17376 | let a = _mm256_set1_epi16(1 << 15); |
| 17377 | let count = _mm_set1_epi16(2); |
| 17378 | let r = _mm256_maskz_sll_epi16(0, a, count); |
| 17379 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 17380 | let r = _mm256_maskz_sll_epi16(0b11111111_11111111, a, count); |
| 17381 | let e = _mm256_set1_epi16(0); |
| 17382 | assert_eq_m256i(r, e); |
| 17383 | } |
| 17384 | |
| 17385 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17386 | unsafe fn test_mm_mask_sll_epi16() { |
| 17387 | let a = _mm_set1_epi16(1 << 15); |
| 17388 | let count = _mm_set1_epi16(2); |
| 17389 | let r = _mm_mask_sll_epi16(a, 0, a, count); |
| 17390 | assert_eq_m128i(r, a); |
| 17391 | let r = _mm_mask_sll_epi16(a, 0b11111111, a, count); |
| 17392 | let e = _mm_set1_epi16(0); |
| 17393 | assert_eq_m128i(r, e); |
| 17394 | } |
| 17395 | |
| 17396 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17397 | unsafe fn test_mm_maskz_sll_epi16() { |
| 17398 | let a = _mm_set1_epi16(1 << 15); |
| 17399 | let count = _mm_set1_epi16(2); |
| 17400 | let r = _mm_maskz_sll_epi16(0, a, count); |
| 17401 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 17402 | let r = _mm_maskz_sll_epi16(0b11111111, a, count); |
| 17403 | let e = _mm_set1_epi16(0); |
| 17404 | assert_eq_m128i(r, e); |
| 17405 | } |
| 17406 | |
| 17407 | #[simd_test(enable = "avx512bw" )] |
| 17408 | unsafe fn test_mm512_slli_epi16() { |
| 17409 | let a = _mm512_set1_epi16(1 << 15); |
| 17410 | let r = _mm512_slli_epi16::<1>(a); |
| 17411 | let e = _mm512_set1_epi16(0); |
| 17412 | assert_eq_m512i(r, e); |
| 17413 | } |
| 17414 | |
| 17415 | #[simd_test(enable = "avx512bw" )] |
| 17416 | unsafe fn test_mm512_mask_slli_epi16() { |
| 17417 | let a = _mm512_set1_epi16(1 << 15); |
| 17418 | let r = _mm512_mask_slli_epi16::<1>(a, 0, a); |
| 17419 | assert_eq_m512i(r, a); |
| 17420 | let r = _mm512_mask_slli_epi16::<1>(a, 0b11111111_11111111_11111111_11111111, a); |
| 17421 | let e = _mm512_set1_epi16(0); |
| 17422 | assert_eq_m512i(r, e); |
| 17423 | } |
| 17424 | |
| 17425 | #[simd_test(enable = "avx512bw" )] |
| 17426 | unsafe fn test_mm512_maskz_slli_epi16() { |
| 17427 | let a = _mm512_set1_epi16(1 << 15); |
| 17428 | let r = _mm512_maskz_slli_epi16::<1>(0, a); |
| 17429 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17430 | let r = _mm512_maskz_slli_epi16::<1>(0b11111111_11111111_11111111_11111111, a); |
| 17431 | let e = _mm512_set1_epi16(0); |
| 17432 | assert_eq_m512i(r, e); |
| 17433 | } |
| 17434 | |
| 17435 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17436 | unsafe fn test_mm256_mask_slli_epi16() { |
| 17437 | let a = _mm256_set1_epi16(1 << 15); |
| 17438 | let r = _mm256_mask_slli_epi16::<1>(a, 0, a); |
| 17439 | assert_eq_m256i(r, a); |
| 17440 | let r = _mm256_mask_slli_epi16::<1>(a, 0b11111111_11111111, a); |
| 17441 | let e = _mm256_set1_epi16(0); |
| 17442 | assert_eq_m256i(r, e); |
| 17443 | } |
| 17444 | |
| 17445 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17446 | unsafe fn test_mm256_maskz_slli_epi16() { |
| 17447 | let a = _mm256_set1_epi16(1 << 15); |
| 17448 | let r = _mm256_maskz_slli_epi16::<1>(0, a); |
| 17449 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 17450 | let r = _mm256_maskz_slli_epi16::<1>(0b11111111_11111111, a); |
| 17451 | let e = _mm256_set1_epi16(0); |
| 17452 | assert_eq_m256i(r, e); |
| 17453 | } |
| 17454 | |
| 17455 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17456 | unsafe fn test_mm_mask_slli_epi16() { |
| 17457 | let a = _mm_set1_epi16(1 << 15); |
| 17458 | let r = _mm_mask_slli_epi16::<1>(a, 0, a); |
| 17459 | assert_eq_m128i(r, a); |
| 17460 | let r = _mm_mask_slli_epi16::<1>(a, 0b11111111, a); |
| 17461 | let e = _mm_set1_epi16(0); |
| 17462 | assert_eq_m128i(r, e); |
| 17463 | } |
| 17464 | |
| 17465 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17466 | unsafe fn test_mm_maskz_slli_epi16() { |
| 17467 | let a = _mm_set1_epi16(1 << 15); |
| 17468 | let r = _mm_maskz_slli_epi16::<1>(0, a); |
| 17469 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 17470 | let r = _mm_maskz_slli_epi16::<1>(0b11111111, a); |
| 17471 | let e = _mm_set1_epi16(0); |
| 17472 | assert_eq_m128i(r, e); |
| 17473 | } |
| 17474 | |
| 17475 | #[simd_test(enable = "avx512bw" )] |
| 17476 | unsafe fn test_mm512_sllv_epi16() { |
| 17477 | let a = _mm512_set1_epi16(1 << 15); |
| 17478 | let count = _mm512_set1_epi16(2); |
| 17479 | let r = _mm512_sllv_epi16(a, count); |
| 17480 | let e = _mm512_set1_epi16(0); |
| 17481 | assert_eq_m512i(r, e); |
| 17482 | } |
| 17483 | |
| 17484 | #[simd_test(enable = "avx512bw" )] |
| 17485 | unsafe fn test_mm512_mask_sllv_epi16() { |
| 17486 | let a = _mm512_set1_epi16(1 << 15); |
| 17487 | let count = _mm512_set1_epi16(2); |
| 17488 | let r = _mm512_mask_sllv_epi16(a, 0, a, count); |
| 17489 | assert_eq_m512i(r, a); |
| 17490 | let r = _mm512_mask_sllv_epi16(a, 0b11111111_11111111_11111111_11111111, a, count); |
| 17491 | let e = _mm512_set1_epi16(0); |
| 17492 | assert_eq_m512i(r, e); |
| 17493 | } |
| 17494 | |
| 17495 | #[simd_test(enable = "avx512bw" )] |
| 17496 | unsafe fn test_mm512_maskz_sllv_epi16() { |
| 17497 | let a = _mm512_set1_epi16(1 << 15); |
| 17498 | let count = _mm512_set1_epi16(2); |
| 17499 | let r = _mm512_maskz_sllv_epi16(0, a, count); |
| 17500 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17501 | let r = _mm512_maskz_sllv_epi16(0b11111111_11111111_11111111_11111111, a, count); |
| 17502 | let e = _mm512_set1_epi16(0); |
| 17503 | assert_eq_m512i(r, e); |
| 17504 | } |
| 17505 | |
| 17506 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17507 | unsafe fn test_mm256_sllv_epi16() { |
| 17508 | let a = _mm256_set1_epi16(1 << 15); |
| 17509 | let count = _mm256_set1_epi16(2); |
| 17510 | let r = _mm256_sllv_epi16(a, count); |
| 17511 | let e = _mm256_set1_epi16(0); |
| 17512 | assert_eq_m256i(r, e); |
| 17513 | } |
| 17514 | |
| 17515 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17516 | unsafe fn test_mm256_mask_sllv_epi16() { |
| 17517 | let a = _mm256_set1_epi16(1 << 15); |
| 17518 | let count = _mm256_set1_epi16(2); |
| 17519 | let r = _mm256_mask_sllv_epi16(a, 0, a, count); |
| 17520 | assert_eq_m256i(r, a); |
| 17521 | let r = _mm256_mask_sllv_epi16(a, 0b11111111_11111111, a, count); |
| 17522 | let e = _mm256_set1_epi16(0); |
| 17523 | assert_eq_m256i(r, e); |
| 17524 | } |
| 17525 | |
| 17526 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17527 | unsafe fn test_mm256_maskz_sllv_epi16() { |
| 17528 | let a = _mm256_set1_epi16(1 << 15); |
| 17529 | let count = _mm256_set1_epi16(2); |
| 17530 | let r = _mm256_maskz_sllv_epi16(0, a, count); |
| 17531 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 17532 | let r = _mm256_maskz_sllv_epi16(0b11111111_11111111, a, count); |
| 17533 | let e = _mm256_set1_epi16(0); |
| 17534 | assert_eq_m256i(r, e); |
| 17535 | } |
| 17536 | |
| 17537 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17538 | unsafe fn test_mm_sllv_epi16() { |
| 17539 | let a = _mm_set1_epi16(1 << 15); |
| 17540 | let count = _mm_set1_epi16(2); |
| 17541 | let r = _mm_sllv_epi16(a, count); |
| 17542 | let e = _mm_set1_epi16(0); |
| 17543 | assert_eq_m128i(r, e); |
| 17544 | } |
| 17545 | |
| 17546 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17547 | unsafe fn test_mm_mask_sllv_epi16() { |
| 17548 | let a = _mm_set1_epi16(1 << 15); |
| 17549 | let count = _mm_set1_epi16(2); |
| 17550 | let r = _mm_mask_sllv_epi16(a, 0, a, count); |
| 17551 | assert_eq_m128i(r, a); |
| 17552 | let r = _mm_mask_sllv_epi16(a, 0b11111111, a, count); |
| 17553 | let e = _mm_set1_epi16(0); |
| 17554 | assert_eq_m128i(r, e); |
| 17555 | } |
| 17556 | |
| 17557 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17558 | unsafe fn test_mm_maskz_sllv_epi16() { |
| 17559 | let a = _mm_set1_epi16(1 << 15); |
| 17560 | let count = _mm_set1_epi16(2); |
| 17561 | let r = _mm_maskz_sllv_epi16(0, a, count); |
| 17562 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 17563 | let r = _mm_maskz_sllv_epi16(0b11111111, a, count); |
| 17564 | let e = _mm_set1_epi16(0); |
| 17565 | assert_eq_m128i(r, e); |
| 17566 | } |
| 17567 | |
| 17568 | #[simd_test(enable = "avx512bw" )] |
| 17569 | unsafe fn test_mm512_srl_epi16() { |
| 17570 | let a = _mm512_set1_epi16(1 << 1); |
| 17571 | let count = _mm_set1_epi16(2); |
| 17572 | let r = _mm512_srl_epi16(a, count); |
| 17573 | let e = _mm512_set1_epi16(0); |
| 17574 | assert_eq_m512i(r, e); |
| 17575 | } |
| 17576 | |
| 17577 | #[simd_test(enable = "avx512bw" )] |
| 17578 | unsafe fn test_mm512_mask_srl_epi16() { |
| 17579 | let a = _mm512_set1_epi16(1 << 1); |
| 17580 | let count = _mm_set1_epi16(2); |
| 17581 | let r = _mm512_mask_srl_epi16(a, 0, a, count); |
| 17582 | assert_eq_m512i(r, a); |
| 17583 | let r = _mm512_mask_srl_epi16(a, 0b11111111_11111111_11111111_11111111, a, count); |
| 17584 | let e = _mm512_set1_epi16(0); |
| 17585 | assert_eq_m512i(r, e); |
| 17586 | } |
| 17587 | |
| 17588 | #[simd_test(enable = "avx512bw" )] |
| 17589 | unsafe fn test_mm512_maskz_srl_epi16() { |
| 17590 | let a = _mm512_set1_epi16(1 << 1); |
| 17591 | let count = _mm_set1_epi16(2); |
| 17592 | let r = _mm512_maskz_srl_epi16(0, a, count); |
| 17593 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17594 | let r = _mm512_maskz_srl_epi16(0b11111111_11111111_11111111_11111111, a, count); |
| 17595 | let e = _mm512_set1_epi16(0); |
| 17596 | assert_eq_m512i(r, e); |
| 17597 | } |
| 17598 | |
| 17599 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17600 | unsafe fn test_mm256_mask_srl_epi16() { |
| 17601 | let a = _mm256_set1_epi16(1 << 1); |
| 17602 | let count = _mm_set1_epi16(2); |
| 17603 | let r = _mm256_mask_srl_epi16(a, 0, a, count); |
| 17604 | assert_eq_m256i(r, a); |
| 17605 | let r = _mm256_mask_srl_epi16(a, 0b11111111_11111111, a, count); |
| 17606 | let e = _mm256_set1_epi16(0); |
| 17607 | assert_eq_m256i(r, e); |
| 17608 | } |
| 17609 | |
| 17610 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17611 | unsafe fn test_mm256_maskz_srl_epi16() { |
| 17612 | let a = _mm256_set1_epi16(1 << 1); |
| 17613 | let count = _mm_set1_epi16(2); |
| 17614 | let r = _mm256_maskz_srl_epi16(0, a, count); |
| 17615 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 17616 | let r = _mm256_maskz_srl_epi16(0b11111111_11111111, a, count); |
| 17617 | let e = _mm256_set1_epi16(0); |
| 17618 | assert_eq_m256i(r, e); |
| 17619 | } |
| 17620 | |
| 17621 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17622 | unsafe fn test_mm_mask_srl_epi16() { |
| 17623 | let a = _mm_set1_epi16(1 << 1); |
| 17624 | let count = _mm_set1_epi16(2); |
| 17625 | let r = _mm_mask_srl_epi16(a, 0, a, count); |
| 17626 | assert_eq_m128i(r, a); |
| 17627 | let r = _mm_mask_srl_epi16(a, 0b11111111, a, count); |
| 17628 | let e = _mm_set1_epi16(0); |
| 17629 | assert_eq_m128i(r, e); |
| 17630 | } |
| 17631 | |
| 17632 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17633 | unsafe fn test_mm_maskz_srl_epi16() { |
| 17634 | let a = _mm_set1_epi16(1 << 1); |
| 17635 | let count = _mm_set1_epi16(2); |
| 17636 | let r = _mm_maskz_srl_epi16(0, a, count); |
| 17637 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 17638 | let r = _mm_maskz_srl_epi16(0b11111111, a, count); |
| 17639 | let e = _mm_set1_epi16(0); |
| 17640 | assert_eq_m128i(r, e); |
| 17641 | } |
| 17642 | |
| 17643 | #[simd_test(enable = "avx512bw" )] |
| 17644 | unsafe fn test_mm512_srli_epi16() { |
| 17645 | let a = _mm512_set1_epi16(1 << 1); |
| 17646 | let r = _mm512_srli_epi16::<2>(a); |
| 17647 | let e = _mm512_set1_epi16(0); |
| 17648 | assert_eq_m512i(r, e); |
| 17649 | } |
| 17650 | |
| 17651 | #[simd_test(enable = "avx512bw" )] |
| 17652 | unsafe fn test_mm512_mask_srli_epi16() { |
| 17653 | let a = _mm512_set1_epi16(1 << 1); |
| 17654 | let r = _mm512_mask_srli_epi16::<2>(a, 0, a); |
| 17655 | assert_eq_m512i(r, a); |
| 17656 | let r = _mm512_mask_srli_epi16::<2>(a, 0b11111111_11111111_11111111_11111111, a); |
| 17657 | let e = _mm512_set1_epi16(0); |
| 17658 | assert_eq_m512i(r, e); |
| 17659 | } |
| 17660 | |
| 17661 | #[simd_test(enable = "avx512bw" )] |
| 17662 | unsafe fn test_mm512_maskz_srli_epi16() { |
| 17663 | let a = _mm512_set1_epi16(1 << 1); |
| 17664 | let r = _mm512_maskz_srli_epi16::<2>(0, a); |
| 17665 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17666 | let r = _mm512_maskz_srli_epi16::<2>(0b11111111_11111111_11111111_11111111, a); |
| 17667 | let e = _mm512_set1_epi16(0); |
| 17668 | assert_eq_m512i(r, e); |
| 17669 | } |
| 17670 | |
| 17671 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17672 | unsafe fn test_mm256_mask_srli_epi16() { |
| 17673 | let a = _mm256_set1_epi16(1 << 1); |
| 17674 | let r = _mm256_mask_srli_epi16::<2>(a, 0, a); |
| 17675 | assert_eq_m256i(r, a); |
| 17676 | let r = _mm256_mask_srli_epi16::<2>(a, 0b11111111_11111111, a); |
| 17677 | let e = _mm256_set1_epi16(0); |
| 17678 | assert_eq_m256i(r, e); |
| 17679 | } |
| 17680 | |
| 17681 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17682 | unsafe fn test_mm256_maskz_srli_epi16() { |
| 17683 | let a = _mm256_set1_epi16(1 << 1); |
| 17684 | let r = _mm256_maskz_srli_epi16::<2>(0, a); |
| 17685 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 17686 | let r = _mm256_maskz_srli_epi16::<2>(0b11111111_11111111, a); |
| 17687 | let e = _mm256_set1_epi16(0); |
| 17688 | assert_eq_m256i(r, e); |
| 17689 | } |
| 17690 | |
| 17691 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17692 | unsafe fn test_mm_mask_srli_epi16() { |
| 17693 | let a = _mm_set1_epi16(1 << 1); |
| 17694 | let r = _mm_mask_srli_epi16::<2>(a, 0, a); |
| 17695 | assert_eq_m128i(r, a); |
| 17696 | let r = _mm_mask_srli_epi16::<2>(a, 0b11111111, a); |
| 17697 | let e = _mm_set1_epi16(0); |
| 17698 | assert_eq_m128i(r, e); |
| 17699 | } |
| 17700 | |
| 17701 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17702 | unsafe fn test_mm_maskz_srli_epi16() { |
| 17703 | let a = _mm_set1_epi16(1 << 1); |
| 17704 | let r = _mm_maskz_srli_epi16::<2>(0, a); |
| 17705 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 17706 | let r = _mm_maskz_srli_epi16::<2>(0b11111111, a); |
| 17707 | let e = _mm_set1_epi16(0); |
| 17708 | assert_eq_m128i(r, e); |
| 17709 | } |
| 17710 | |
| 17711 | #[simd_test(enable = "avx512bw" )] |
| 17712 | unsafe fn test_mm512_srlv_epi16() { |
| 17713 | let a = _mm512_set1_epi16(1 << 1); |
| 17714 | let count = _mm512_set1_epi16(2); |
| 17715 | let r = _mm512_srlv_epi16(a, count); |
| 17716 | let e = _mm512_set1_epi16(0); |
| 17717 | assert_eq_m512i(r, e); |
| 17718 | } |
| 17719 | |
| 17720 | #[simd_test(enable = "avx512bw" )] |
| 17721 | unsafe fn test_mm512_mask_srlv_epi16() { |
| 17722 | let a = _mm512_set1_epi16(1 << 1); |
| 17723 | let count = _mm512_set1_epi16(2); |
| 17724 | let r = _mm512_mask_srlv_epi16(a, 0, a, count); |
| 17725 | assert_eq_m512i(r, a); |
| 17726 | let r = _mm512_mask_srlv_epi16(a, 0b11111111_11111111_11111111_11111111, a, count); |
| 17727 | let e = _mm512_set1_epi16(0); |
| 17728 | assert_eq_m512i(r, e); |
| 17729 | } |
| 17730 | |
| 17731 | #[simd_test(enable = "avx512bw" )] |
| 17732 | unsafe fn test_mm512_maskz_srlv_epi16() { |
| 17733 | let a = _mm512_set1_epi16(1 << 1); |
| 17734 | let count = _mm512_set1_epi16(2); |
| 17735 | let r = _mm512_maskz_srlv_epi16(0, a, count); |
| 17736 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17737 | let r = _mm512_maskz_srlv_epi16(0b11111111_11111111_11111111_11111111, a, count); |
| 17738 | let e = _mm512_set1_epi16(0); |
| 17739 | assert_eq_m512i(r, e); |
| 17740 | } |
| 17741 | |
| 17742 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17743 | unsafe fn test_mm256_srlv_epi16() { |
| 17744 | let a = _mm256_set1_epi16(1 << 1); |
| 17745 | let count = _mm256_set1_epi16(2); |
| 17746 | let r = _mm256_srlv_epi16(a, count); |
| 17747 | let e = _mm256_set1_epi16(0); |
| 17748 | assert_eq_m256i(r, e); |
| 17749 | } |
| 17750 | |
| 17751 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17752 | unsafe fn test_mm256_mask_srlv_epi16() { |
| 17753 | let a = _mm256_set1_epi16(1 << 1); |
| 17754 | let count = _mm256_set1_epi16(2); |
| 17755 | let r = _mm256_mask_srlv_epi16(a, 0, a, count); |
| 17756 | assert_eq_m256i(r, a); |
| 17757 | let r = _mm256_mask_srlv_epi16(a, 0b11111111_11111111, a, count); |
| 17758 | let e = _mm256_set1_epi16(0); |
| 17759 | assert_eq_m256i(r, e); |
| 17760 | } |
| 17761 | |
| 17762 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17763 | unsafe fn test_mm256_maskz_srlv_epi16() { |
| 17764 | let a = _mm256_set1_epi16(1 << 1); |
| 17765 | let count = _mm256_set1_epi16(2); |
| 17766 | let r = _mm256_maskz_srlv_epi16(0, a, count); |
| 17767 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 17768 | let r = _mm256_maskz_srlv_epi16(0b11111111_11111111, a, count); |
| 17769 | let e = _mm256_set1_epi16(0); |
| 17770 | assert_eq_m256i(r, e); |
| 17771 | } |
| 17772 | |
| 17773 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17774 | unsafe fn test_mm_srlv_epi16() { |
| 17775 | let a = _mm_set1_epi16(1 << 1); |
| 17776 | let count = _mm_set1_epi16(2); |
| 17777 | let r = _mm_srlv_epi16(a, count); |
| 17778 | let e = _mm_set1_epi16(0); |
| 17779 | assert_eq_m128i(r, e); |
| 17780 | } |
| 17781 | |
| 17782 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17783 | unsafe fn test_mm_mask_srlv_epi16() { |
| 17784 | let a = _mm_set1_epi16(1 << 1); |
| 17785 | let count = _mm_set1_epi16(2); |
| 17786 | let r = _mm_mask_srlv_epi16(a, 0, a, count); |
| 17787 | assert_eq_m128i(r, a); |
| 17788 | let r = _mm_mask_srlv_epi16(a, 0b11111111, a, count); |
| 17789 | let e = _mm_set1_epi16(0); |
| 17790 | assert_eq_m128i(r, e); |
| 17791 | } |
| 17792 | |
| 17793 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17794 | unsafe fn test_mm_maskz_srlv_epi16() { |
| 17795 | let a = _mm_set1_epi16(1 << 1); |
| 17796 | let count = _mm_set1_epi16(2); |
| 17797 | let r = _mm_maskz_srlv_epi16(0, a, count); |
| 17798 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 17799 | let r = _mm_maskz_srlv_epi16(0b11111111, a, count); |
| 17800 | let e = _mm_set1_epi16(0); |
| 17801 | assert_eq_m128i(r, e); |
| 17802 | } |
| 17803 | |
| 17804 | #[simd_test(enable = "avx512bw" )] |
| 17805 | unsafe fn test_mm512_sra_epi16() { |
| 17806 | let a = _mm512_set1_epi16(8); |
| 17807 | let count = _mm_set1_epi16(1); |
| 17808 | let r = _mm512_sra_epi16(a, count); |
| 17809 | let e = _mm512_set1_epi16(0); |
| 17810 | assert_eq_m512i(r, e); |
| 17811 | } |
| 17812 | |
| 17813 | #[simd_test(enable = "avx512bw" )] |
| 17814 | unsafe fn test_mm512_mask_sra_epi16() { |
| 17815 | let a = _mm512_set1_epi16(8); |
| 17816 | let count = _mm_set1_epi16(1); |
| 17817 | let r = _mm512_mask_sra_epi16(a, 0, a, count); |
| 17818 | assert_eq_m512i(r, a); |
| 17819 | let r = _mm512_mask_sra_epi16(a, 0b11111111_11111111_11111111_11111111, a, count); |
| 17820 | let e = _mm512_set1_epi16(0); |
| 17821 | assert_eq_m512i(r, e); |
| 17822 | } |
| 17823 | |
| 17824 | #[simd_test(enable = "avx512bw" )] |
| 17825 | unsafe fn test_mm512_maskz_sra_epi16() { |
| 17826 | let a = _mm512_set1_epi16(8); |
| 17827 | let count = _mm_set1_epi16(1); |
| 17828 | let r = _mm512_maskz_sra_epi16(0, a, count); |
| 17829 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17830 | let r = _mm512_maskz_sra_epi16(0b11111111_11111111_11111111_11111111, a, count); |
| 17831 | let e = _mm512_set1_epi16(0); |
| 17832 | assert_eq_m512i(r, e); |
| 17833 | } |
| 17834 | |
| 17835 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17836 | unsafe fn test_mm256_mask_sra_epi16() { |
| 17837 | let a = _mm256_set1_epi16(8); |
| 17838 | let count = _mm_set1_epi16(1); |
| 17839 | let r = _mm256_mask_sra_epi16(a, 0, a, count); |
| 17840 | assert_eq_m256i(r, a); |
| 17841 | let r = _mm256_mask_sra_epi16(a, 0b11111111_11111111, a, count); |
| 17842 | let e = _mm256_set1_epi16(0); |
| 17843 | assert_eq_m256i(r, e); |
| 17844 | } |
| 17845 | |
| 17846 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17847 | unsafe fn test_mm256_maskz_sra_epi16() { |
| 17848 | let a = _mm256_set1_epi16(8); |
| 17849 | let count = _mm_set1_epi16(1); |
| 17850 | let r = _mm256_maskz_sra_epi16(0, a, count); |
| 17851 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 17852 | let r = _mm256_maskz_sra_epi16(0b11111111_11111111, a, count); |
| 17853 | let e = _mm256_set1_epi16(0); |
| 17854 | assert_eq_m256i(r, e); |
| 17855 | } |
| 17856 | |
| 17857 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17858 | unsafe fn test_mm_mask_sra_epi16() { |
| 17859 | let a = _mm_set1_epi16(8); |
| 17860 | let count = _mm_set1_epi16(1); |
| 17861 | let r = _mm_mask_sra_epi16(a, 0, a, count); |
| 17862 | assert_eq_m128i(r, a); |
| 17863 | let r = _mm_mask_sra_epi16(a, 0b11111111, a, count); |
| 17864 | let e = _mm_set1_epi16(0); |
| 17865 | assert_eq_m128i(r, e); |
| 17866 | } |
| 17867 | |
| 17868 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17869 | unsafe fn test_mm_maskz_sra_epi16() { |
| 17870 | let a = _mm_set1_epi16(8); |
| 17871 | let count = _mm_set1_epi16(1); |
| 17872 | let r = _mm_maskz_sra_epi16(0, a, count); |
| 17873 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 17874 | let r = _mm_maskz_sra_epi16(0b11111111, a, count); |
| 17875 | let e = _mm_set1_epi16(0); |
| 17876 | assert_eq_m128i(r, e); |
| 17877 | } |
| 17878 | |
| 17879 | #[simd_test(enable = "avx512bw" )] |
| 17880 | unsafe fn test_mm512_srai_epi16() { |
| 17881 | let a = _mm512_set1_epi16(8); |
| 17882 | let r = _mm512_srai_epi16::<2>(a); |
| 17883 | let e = _mm512_set1_epi16(2); |
| 17884 | assert_eq_m512i(r, e); |
| 17885 | } |
| 17886 | |
| 17887 | #[simd_test(enable = "avx512bw" )] |
| 17888 | unsafe fn test_mm512_mask_srai_epi16() { |
| 17889 | let a = _mm512_set1_epi16(8); |
| 17890 | let r = _mm512_mask_srai_epi16::<2>(a, 0, a); |
| 17891 | assert_eq_m512i(r, a); |
| 17892 | let r = _mm512_mask_srai_epi16::<2>(a, 0b11111111_11111111_11111111_11111111, a); |
| 17893 | let e = _mm512_set1_epi16(2); |
| 17894 | assert_eq_m512i(r, e); |
| 17895 | } |
| 17896 | |
| 17897 | #[simd_test(enable = "avx512bw" )] |
| 17898 | unsafe fn test_mm512_maskz_srai_epi16() { |
| 17899 | let a = _mm512_set1_epi16(8); |
| 17900 | let r = _mm512_maskz_srai_epi16::<2>(0, a); |
| 17901 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17902 | let r = _mm512_maskz_srai_epi16::<2>(0b11111111_11111111_11111111_11111111, a); |
| 17903 | let e = _mm512_set1_epi16(2); |
| 17904 | assert_eq_m512i(r, e); |
| 17905 | } |
| 17906 | |
| 17907 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17908 | unsafe fn test_mm256_mask_srai_epi16() { |
| 17909 | let a = _mm256_set1_epi16(8); |
| 17910 | let r = _mm256_mask_srai_epi16::<2>(a, 0, a); |
| 17911 | assert_eq_m256i(r, a); |
| 17912 | let r = _mm256_mask_srai_epi16::<2>(a, 0b11111111_11111111, a); |
| 17913 | let e = _mm256_set1_epi16(2); |
| 17914 | assert_eq_m256i(r, e); |
| 17915 | } |
| 17916 | |
| 17917 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17918 | unsafe fn test_mm256_maskz_srai_epi16() { |
| 17919 | let a = _mm256_set1_epi16(8); |
| 17920 | let r = _mm256_maskz_srai_epi16::<2>(0, a); |
| 17921 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 17922 | let r = _mm256_maskz_srai_epi16::<2>(0b11111111_11111111, a); |
| 17923 | let e = _mm256_set1_epi16(2); |
| 17924 | assert_eq_m256i(r, e); |
| 17925 | } |
| 17926 | |
| 17927 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17928 | unsafe fn test_mm_mask_srai_epi16() { |
| 17929 | let a = _mm_set1_epi16(8); |
| 17930 | let r = _mm_mask_srai_epi16::<2>(a, 0, a); |
| 17931 | assert_eq_m128i(r, a); |
| 17932 | let r = _mm_mask_srai_epi16::<2>(a, 0b11111111, a); |
| 17933 | let e = _mm_set1_epi16(2); |
| 17934 | assert_eq_m128i(r, e); |
| 17935 | } |
| 17936 | |
| 17937 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17938 | unsafe fn test_mm_maskz_srai_epi16() { |
| 17939 | let a = _mm_set1_epi16(8); |
| 17940 | let r = _mm_maskz_srai_epi16::<2>(0, a); |
| 17941 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 17942 | let r = _mm_maskz_srai_epi16::<2>(0b11111111, a); |
| 17943 | let e = _mm_set1_epi16(2); |
| 17944 | assert_eq_m128i(r, e); |
| 17945 | } |
| 17946 | |
| 17947 | #[simd_test(enable = "avx512bw" )] |
| 17948 | unsafe fn test_mm512_srav_epi16() { |
| 17949 | let a = _mm512_set1_epi16(8); |
| 17950 | let count = _mm512_set1_epi16(2); |
| 17951 | let r = _mm512_srav_epi16(a, count); |
| 17952 | let e = _mm512_set1_epi16(2); |
| 17953 | assert_eq_m512i(r, e); |
| 17954 | } |
| 17955 | |
| 17956 | #[simd_test(enable = "avx512bw" )] |
| 17957 | unsafe fn test_mm512_mask_srav_epi16() { |
| 17958 | let a = _mm512_set1_epi16(8); |
| 17959 | let count = _mm512_set1_epi16(2); |
| 17960 | let r = _mm512_mask_srav_epi16(a, 0, a, count); |
| 17961 | assert_eq_m512i(r, a); |
| 17962 | let r = _mm512_mask_srav_epi16(a, 0b11111111_11111111_11111111_11111111, a, count); |
| 17963 | let e = _mm512_set1_epi16(2); |
| 17964 | assert_eq_m512i(r, e); |
| 17965 | } |
| 17966 | |
| 17967 | #[simd_test(enable = "avx512bw" )] |
| 17968 | unsafe fn test_mm512_maskz_srav_epi16() { |
| 17969 | let a = _mm512_set1_epi16(8); |
| 17970 | let count = _mm512_set1_epi16(2); |
| 17971 | let r = _mm512_maskz_srav_epi16(0, a, count); |
| 17972 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 17973 | let r = _mm512_maskz_srav_epi16(0b11111111_11111111_11111111_11111111, a, count); |
| 17974 | let e = _mm512_set1_epi16(2); |
| 17975 | assert_eq_m512i(r, e); |
| 17976 | } |
| 17977 | |
| 17978 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17979 | unsafe fn test_mm256_srav_epi16() { |
| 17980 | let a = _mm256_set1_epi16(8); |
| 17981 | let count = _mm256_set1_epi16(2); |
| 17982 | let r = _mm256_srav_epi16(a, count); |
| 17983 | let e = _mm256_set1_epi16(2); |
| 17984 | assert_eq_m256i(r, e); |
| 17985 | } |
| 17986 | |
| 17987 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17988 | unsafe fn test_mm256_mask_srav_epi16() { |
| 17989 | let a = _mm256_set1_epi16(8); |
| 17990 | let count = _mm256_set1_epi16(2); |
| 17991 | let r = _mm256_mask_srav_epi16(a, 0, a, count); |
| 17992 | assert_eq_m256i(r, a); |
| 17993 | let r = _mm256_mask_srav_epi16(a, 0b11111111_11111111, a, count); |
| 17994 | let e = _mm256_set1_epi16(2); |
| 17995 | assert_eq_m256i(r, e); |
| 17996 | } |
| 17997 | |
| 17998 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 17999 | unsafe fn test_mm256_maskz_srav_epi16() { |
| 18000 | let a = _mm256_set1_epi16(8); |
| 18001 | let count = _mm256_set1_epi16(2); |
| 18002 | let r = _mm256_maskz_srav_epi16(0, a, count); |
| 18003 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 18004 | let r = _mm256_maskz_srav_epi16(0b11111111_11111111, a, count); |
| 18005 | let e = _mm256_set1_epi16(2); |
| 18006 | assert_eq_m256i(r, e); |
| 18007 | } |
| 18008 | |
| 18009 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18010 | unsafe fn test_mm_srav_epi16() { |
| 18011 | let a = _mm_set1_epi16(8); |
| 18012 | let count = _mm_set1_epi16(2); |
| 18013 | let r = _mm_srav_epi16(a, count); |
| 18014 | let e = _mm_set1_epi16(2); |
| 18015 | assert_eq_m128i(r, e); |
| 18016 | } |
| 18017 | |
| 18018 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18019 | unsafe fn test_mm_mask_srav_epi16() { |
| 18020 | let a = _mm_set1_epi16(8); |
| 18021 | let count = _mm_set1_epi16(2); |
| 18022 | let r = _mm_mask_srav_epi16(a, 0, a, count); |
| 18023 | assert_eq_m128i(r, a); |
| 18024 | let r = _mm_mask_srav_epi16(a, 0b11111111, a, count); |
| 18025 | let e = _mm_set1_epi16(2); |
| 18026 | assert_eq_m128i(r, e); |
| 18027 | } |
| 18028 | |
| 18029 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18030 | unsafe fn test_mm_maskz_srav_epi16() { |
| 18031 | let a = _mm_set1_epi16(8); |
| 18032 | let count = _mm_set1_epi16(2); |
| 18033 | let r = _mm_maskz_srav_epi16(0, a, count); |
| 18034 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 18035 | let r = _mm_maskz_srav_epi16(0b11111111, a, count); |
| 18036 | let e = _mm_set1_epi16(2); |
| 18037 | assert_eq_m128i(r, e); |
| 18038 | } |
| 18039 | |
| 18040 | #[simd_test(enable = "avx512bw" )] |
| 18041 | unsafe fn test_mm512_permutex2var_epi16() { |
| 18042 | #[rustfmt::skip] |
| 18043 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 18044 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
| 18045 | #[rustfmt::skip] |
| 18046 | let idx = _mm512_set_epi16(1, 1<<5, 2, 1<<5, 3, 1<<5, 4, 1<<5, 5, 1<<5, 6, 1<<5, 7, 1<<5, 8, 1<<5, |
| 18047 | 9, 1<<5, 10, 1<<5, 11, 1<<5, 12, 1<<5, 13, 1<<5, 14, 1<<5, 15, 1<<5, 16, 1<<5); |
| 18048 | let b = _mm512_set1_epi16(100); |
| 18049 | let r = _mm512_permutex2var_epi16(a, idx, b); |
| 18050 | #[rustfmt::skip] |
| 18051 | let e = _mm512_set_epi16( |
| 18052 | 30, 100, 29, 100, 28, 100, 27, 100, 26, 100, 25, 100, 24, 100, 23, 100, |
| 18053 | 22, 100, 21, 100, 20, 100, 19, 100, 18, 100, 17, 100, 16, 100, 15, 100, |
| 18054 | ); |
| 18055 | assert_eq_m512i(r, e); |
| 18056 | } |
| 18057 | |
| 18058 | #[simd_test(enable = "avx512bw" )] |
| 18059 | unsafe fn test_mm512_mask_permutex2var_epi16() { |
| 18060 | #[rustfmt::skip] |
| 18061 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 18062 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
| 18063 | #[rustfmt::skip] |
| 18064 | let idx = _mm512_set_epi16(1, 1<<5, 2, 1<<5, 3, 1<<5, 4, 1<<5, 5, 1<<5, 6, 1<<5, 7, 1<<5, 8, 1<<5, |
| 18065 | 9, 1<<5, 10, 1<<5, 11, 1<<5, 12, 1<<5, 13, 1<<5, 14, 1<<5, 15, 1<<5, 16, 1<<5); |
| 18066 | let b = _mm512_set1_epi16(100); |
| 18067 | let r = _mm512_mask_permutex2var_epi16(a, 0, idx, b); |
| 18068 | assert_eq_m512i(r, a); |
| 18069 | let r = _mm512_mask_permutex2var_epi16(a, 0b11111111_11111111_11111111_11111111, idx, b); |
| 18070 | #[rustfmt::skip] |
| 18071 | let e = _mm512_set_epi16( |
| 18072 | 30, 100, 29, 100, 28, 100, 27, 100, 26, 100, 25, 100, 24, 100, 23, 100, |
| 18073 | 22, 100, 21, 100, 20, 100, 19, 100, 18, 100, 17, 100, 16, 100, 15, 100, |
| 18074 | ); |
| 18075 | assert_eq_m512i(r, e); |
| 18076 | } |
| 18077 | |
| 18078 | #[simd_test(enable = "avx512bw" )] |
| 18079 | unsafe fn test_mm512_maskz_permutex2var_epi16() { |
| 18080 | #[rustfmt::skip] |
| 18081 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 18082 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
| 18083 | #[rustfmt::skip] |
| 18084 | let idx = _mm512_set_epi16(1, 1<<5, 2, 1<<5, 3, 1<<5, 4, 1<<5, 5, 1<<5, 6, 1<<5, 7, 1<<5, 8, 1<<5, |
| 18085 | 9, 1<<5, 10, 1<<5, 11, 1<<5, 12, 1<<5, 13, 1<<5, 14, 1<<5, 15, 1<<5, 16, 1<<5); |
| 18086 | let b = _mm512_set1_epi16(100); |
| 18087 | let r = _mm512_maskz_permutex2var_epi16(0, a, idx, b); |
| 18088 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 18089 | let r = _mm512_maskz_permutex2var_epi16(0b11111111_11111111_11111111_11111111, a, idx, b); |
| 18090 | #[rustfmt::skip] |
| 18091 | let e = _mm512_set_epi16( |
| 18092 | 30, 100, 29, 100, 28, 100, 27, 100, 26, 100, 25, 100, 24, 100, 23, 100, |
| 18093 | 22, 100, 21, 100, 20, 100, 19, 100, 18, 100, 17, 100, 16, 100, 15, 100, |
| 18094 | ); |
| 18095 | assert_eq_m512i(r, e); |
| 18096 | } |
| 18097 | |
| 18098 | #[simd_test(enable = "avx512bw" )] |
| 18099 | unsafe fn test_mm512_mask2_permutex2var_epi16() { |
| 18100 | #[rustfmt::skip] |
| 18101 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 18102 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
| 18103 | #[rustfmt::skip] |
| 18104 | let idx = _mm512_set_epi16(1, 1<<5, 2, 1<<5, 3, 1<<5, 4, 1<<5, 5, 1<<5, 6, 1<<5, 7, 1<<5, 8, 1<<5, |
| 18105 | 9, 1<<5, 10, 1<<5, 11, 1<<5, 12, 1<<5, 13, 1<<5, 14, 1<<5, 15, 1<<5, 16, 1<<5); |
| 18106 | let b = _mm512_set1_epi16(100); |
| 18107 | let r = _mm512_mask2_permutex2var_epi16(a, idx, 0, b); |
| 18108 | assert_eq_m512i(r, idx); |
| 18109 | let r = _mm512_mask2_permutex2var_epi16(a, idx, 0b11111111_11111111_11111111_11111111, b); |
| 18110 | #[rustfmt::skip] |
| 18111 | let e = _mm512_set_epi16( |
| 18112 | 30, 100, 29, 100, 28, 100, 27, 100, 26, 100, 25, 100, 24, 100, 23, 100, |
| 18113 | 22, 100, 21, 100, 20, 100, 19, 100, 18, 100, 17, 100, 16, 100, 15, 100, |
| 18114 | ); |
| 18115 | assert_eq_m512i(r, e); |
| 18116 | } |
| 18117 | |
| 18118 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18119 | unsafe fn test_mm256_permutex2var_epi16() { |
| 18120 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 18121 | #[rustfmt::skip] |
| 18122 | let idx = _mm256_set_epi16(1, 1<<4, 2, 1<<4, 3, 1<<4, 4, 1<<4, 5, 1<<4, 6, 1<<4, 7, 1<<4, 8, 1<<4); |
| 18123 | let b = _mm256_set1_epi16(100); |
| 18124 | let r = _mm256_permutex2var_epi16(a, idx, b); |
| 18125 | let e = _mm256_set_epi16( |
| 18126 | 14, 100, 13, 100, 12, 100, 11, 100, 10, 100, 9, 100, 8, 100, 7, 100, |
| 18127 | ); |
| 18128 | assert_eq_m256i(r, e); |
| 18129 | } |
| 18130 | |
| 18131 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18132 | unsafe fn test_mm256_mask_permutex2var_epi16() { |
| 18133 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 18134 | #[rustfmt::skip] |
| 18135 | let idx = _mm256_set_epi16(1, 1<<4, 2, 1<<4, 3, 1<<4, 4, 1<<4, 5, 1<<4, 6, 1<<4, 7, 1<<4, 8, 1<<4); |
| 18136 | let b = _mm256_set1_epi16(100); |
| 18137 | let r = _mm256_mask_permutex2var_epi16(a, 0, idx, b); |
| 18138 | assert_eq_m256i(r, a); |
| 18139 | let r = _mm256_mask_permutex2var_epi16(a, 0b11111111_11111111, idx, b); |
| 18140 | let e = _mm256_set_epi16( |
| 18141 | 14, 100, 13, 100, 12, 100, 11, 100, 10, 100, 9, 100, 8, 100, 7, 100, |
| 18142 | ); |
| 18143 | assert_eq_m256i(r, e); |
| 18144 | } |
| 18145 | |
| 18146 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18147 | unsafe fn test_mm256_maskz_permutex2var_epi16() { |
| 18148 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 18149 | #[rustfmt::skip] |
| 18150 | let idx = _mm256_set_epi16(1, 1<<4, 2, 1<<4, 3, 1<<4, 4, 1<<4, 5, 1<<4, 6, 1<<4, 7, 1<<4, 8, 1<<4); |
| 18151 | let b = _mm256_set1_epi16(100); |
| 18152 | let r = _mm256_maskz_permutex2var_epi16(0, a, idx, b); |
| 18153 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 18154 | let r = _mm256_maskz_permutex2var_epi16(0b11111111_11111111, a, idx, b); |
| 18155 | let e = _mm256_set_epi16( |
| 18156 | 14, 100, 13, 100, 12, 100, 11, 100, 10, 100, 9, 100, 8, 100, 7, 100, |
| 18157 | ); |
| 18158 | assert_eq_m256i(r, e); |
| 18159 | } |
| 18160 | |
| 18161 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18162 | unsafe fn test_mm256_mask2_permutex2var_epi16() { |
| 18163 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 18164 | #[rustfmt::skip] |
| 18165 | let idx = _mm256_set_epi16(1, 1<<4, 2, 1<<4, 3, 1<<4, 4, 1<<4, 5, 1<<4, 6, 1<<4, 7, 1<<4, 8, 1<<4); |
| 18166 | let b = _mm256_set1_epi16(100); |
| 18167 | let r = _mm256_mask2_permutex2var_epi16(a, idx, 0, b); |
| 18168 | assert_eq_m256i(r, idx); |
| 18169 | let r = _mm256_mask2_permutex2var_epi16(a, idx, 0b11111111_11111111, b); |
| 18170 | #[rustfmt::skip] |
| 18171 | let e = _mm256_set_epi16( |
| 18172 | 14, 100, 13, 100, 12, 100, 11, 100, 10, 100, 9, 100, 8, 100, 7, 100, |
| 18173 | ); |
| 18174 | assert_eq_m256i(r, e); |
| 18175 | } |
| 18176 | |
| 18177 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18178 | unsafe fn test_mm_permutex2var_epi16() { |
| 18179 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 18180 | let idx = _mm_set_epi16(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
| 18181 | let b = _mm_set1_epi16(100); |
| 18182 | let r = _mm_permutex2var_epi16(a, idx, b); |
| 18183 | let e = _mm_set_epi16(6, 100, 5, 100, 4, 100, 3, 100); |
| 18184 | assert_eq_m128i(r, e); |
| 18185 | } |
| 18186 | |
| 18187 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18188 | unsafe fn test_mm_mask_permutex2var_epi16() { |
| 18189 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 18190 | let idx = _mm_set_epi16(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
| 18191 | let b = _mm_set1_epi16(100); |
| 18192 | let r = _mm_mask_permutex2var_epi16(a, 0, idx, b); |
| 18193 | assert_eq_m128i(r, a); |
| 18194 | let r = _mm_mask_permutex2var_epi16(a, 0b11111111, idx, b); |
| 18195 | let e = _mm_set_epi16(6, 100, 5, 100, 4, 100, 3, 100); |
| 18196 | assert_eq_m128i(r, e); |
| 18197 | } |
| 18198 | |
| 18199 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18200 | unsafe fn test_mm_maskz_permutex2var_epi16() { |
| 18201 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 18202 | let idx = _mm_set_epi16(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
| 18203 | let b = _mm_set1_epi16(100); |
| 18204 | let r = _mm_maskz_permutex2var_epi16(0, a, idx, b); |
| 18205 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 18206 | let r = _mm_maskz_permutex2var_epi16(0b11111111, a, idx, b); |
| 18207 | let e = _mm_set_epi16(6, 100, 5, 100, 4, 100, 3, 100); |
| 18208 | assert_eq_m128i(r, e); |
| 18209 | } |
| 18210 | |
| 18211 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18212 | unsafe fn test_mm_mask2_permutex2var_epi16() { |
| 18213 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 18214 | let idx = _mm_set_epi16(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
| 18215 | let b = _mm_set1_epi16(100); |
| 18216 | let r = _mm_mask2_permutex2var_epi16(a, idx, 0, b); |
| 18217 | assert_eq_m128i(r, idx); |
| 18218 | let r = _mm_mask2_permutex2var_epi16(a, idx, 0b11111111, b); |
| 18219 | let e = _mm_set_epi16(6, 100, 5, 100, 4, 100, 3, 100); |
| 18220 | assert_eq_m128i(r, e); |
| 18221 | } |
| 18222 | |
| 18223 | #[simd_test(enable = "avx512bw" )] |
| 18224 | unsafe fn test_mm512_permutexvar_epi16() { |
| 18225 | let idx = _mm512_set1_epi16(1); |
| 18226 | #[rustfmt::skip] |
| 18227 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 18228 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
| 18229 | let r = _mm512_permutexvar_epi16(idx, a); |
| 18230 | let e = _mm512_set1_epi16(30); |
| 18231 | assert_eq_m512i(r, e); |
| 18232 | } |
| 18233 | |
| 18234 | #[simd_test(enable = "avx512bw" )] |
| 18235 | unsafe fn test_mm512_mask_permutexvar_epi16() { |
| 18236 | let idx = _mm512_set1_epi16(1); |
| 18237 | #[rustfmt::skip] |
| 18238 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 18239 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
| 18240 | let r = _mm512_mask_permutexvar_epi16(a, 0, idx, a); |
| 18241 | assert_eq_m512i(r, a); |
| 18242 | let r = _mm512_mask_permutexvar_epi16(a, 0b11111111_11111111_11111111_11111111, idx, a); |
| 18243 | let e = _mm512_set1_epi16(30); |
| 18244 | assert_eq_m512i(r, e); |
| 18245 | } |
| 18246 | |
| 18247 | #[simd_test(enable = "avx512bw" )] |
| 18248 | unsafe fn test_mm512_maskz_permutexvar_epi16() { |
| 18249 | let idx = _mm512_set1_epi16(1); |
| 18250 | #[rustfmt::skip] |
| 18251 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 18252 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
| 18253 | let r = _mm512_maskz_permutexvar_epi16(0, idx, a); |
| 18254 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 18255 | let r = _mm512_maskz_permutexvar_epi16(0b11111111_11111111_11111111_11111111, idx, a); |
| 18256 | let e = _mm512_set1_epi16(30); |
| 18257 | assert_eq_m512i(r, e); |
| 18258 | } |
| 18259 | |
| 18260 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18261 | unsafe fn test_mm256_permutexvar_epi16() { |
| 18262 | let idx = _mm256_set1_epi16(1); |
| 18263 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 18264 | let r = _mm256_permutexvar_epi16(idx, a); |
| 18265 | let e = _mm256_set1_epi16(14); |
| 18266 | assert_eq_m256i(r, e); |
| 18267 | } |
| 18268 | |
| 18269 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18270 | unsafe fn test_mm256_mask_permutexvar_epi16() { |
| 18271 | let idx = _mm256_set1_epi16(1); |
| 18272 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 18273 | let r = _mm256_mask_permutexvar_epi16(a, 0, idx, a); |
| 18274 | assert_eq_m256i(r, a); |
| 18275 | let r = _mm256_mask_permutexvar_epi16(a, 0b11111111_11111111, idx, a); |
| 18276 | let e = _mm256_set1_epi16(14); |
| 18277 | assert_eq_m256i(r, e); |
| 18278 | } |
| 18279 | |
| 18280 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18281 | unsafe fn test_mm256_maskz_permutexvar_epi16() { |
| 18282 | let idx = _mm256_set1_epi16(1); |
| 18283 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 18284 | let r = _mm256_maskz_permutexvar_epi16(0, idx, a); |
| 18285 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 18286 | let r = _mm256_maskz_permutexvar_epi16(0b11111111_11111111, idx, a); |
| 18287 | let e = _mm256_set1_epi16(14); |
| 18288 | assert_eq_m256i(r, e); |
| 18289 | } |
| 18290 | |
| 18291 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18292 | unsafe fn test_mm_permutexvar_epi16() { |
| 18293 | let idx = _mm_set1_epi16(1); |
| 18294 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 18295 | let r = _mm_permutexvar_epi16(idx, a); |
| 18296 | let e = _mm_set1_epi16(6); |
| 18297 | assert_eq_m128i(r, e); |
| 18298 | } |
| 18299 | |
| 18300 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18301 | unsafe fn test_mm_mask_permutexvar_epi16() { |
| 18302 | let idx = _mm_set1_epi16(1); |
| 18303 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 18304 | let r = _mm_mask_permutexvar_epi16(a, 0, idx, a); |
| 18305 | assert_eq_m128i(r, a); |
| 18306 | let r = _mm_mask_permutexvar_epi16(a, 0b11111111, idx, a); |
| 18307 | let e = _mm_set1_epi16(6); |
| 18308 | assert_eq_m128i(r, e); |
| 18309 | } |
| 18310 | |
| 18311 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18312 | unsafe fn test_mm_maskz_permutexvar_epi16() { |
| 18313 | let idx = _mm_set1_epi16(1); |
| 18314 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 18315 | let r = _mm_maskz_permutexvar_epi16(0, idx, a); |
| 18316 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 18317 | let r = _mm_maskz_permutexvar_epi16(0b11111111, idx, a); |
| 18318 | let e = _mm_set1_epi16(6); |
| 18319 | assert_eq_m128i(r, e); |
| 18320 | } |
| 18321 | |
| 18322 | #[simd_test(enable = "avx512bw" )] |
| 18323 | unsafe fn test_mm512_mask_blend_epi16() { |
| 18324 | let a = _mm512_set1_epi16(1); |
| 18325 | let b = _mm512_set1_epi16(2); |
| 18326 | let r = _mm512_mask_blend_epi16(0b11111111_00000000_11111111_00000000, a, b); |
| 18327 | #[rustfmt::skip] |
| 18328 | let e = _mm512_set_epi16(2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, |
| 18329 | 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1); |
| 18330 | assert_eq_m512i(r, e); |
| 18331 | } |
| 18332 | |
| 18333 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18334 | unsafe fn test_mm256_mask_blend_epi16() { |
| 18335 | let a = _mm256_set1_epi16(1); |
| 18336 | let b = _mm256_set1_epi16(2); |
| 18337 | let r = _mm256_mask_blend_epi16(0b11111111_00000000, a, b); |
| 18338 | let e = _mm256_set_epi16(2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1); |
| 18339 | assert_eq_m256i(r, e); |
| 18340 | } |
| 18341 | |
| 18342 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18343 | unsafe fn test_mm_mask_blend_epi16() { |
| 18344 | let a = _mm_set1_epi16(1); |
| 18345 | let b = _mm_set1_epi16(2); |
| 18346 | let r = _mm_mask_blend_epi16(0b11110000, a, b); |
| 18347 | let e = _mm_set_epi16(2, 2, 2, 2, 1, 1, 1, 1); |
| 18348 | assert_eq_m128i(r, e); |
| 18349 | } |
| 18350 | |
| 18351 | #[simd_test(enable = "avx512bw" )] |
| 18352 | unsafe fn test_mm512_mask_blend_epi8() { |
| 18353 | let a = _mm512_set1_epi8(1); |
| 18354 | let b = _mm512_set1_epi8(2); |
| 18355 | let r = _mm512_mask_blend_epi8( |
| 18356 | 0b11111111_00000000_11111111_00000000_11111111_00000000_11111111_00000000, |
| 18357 | a, |
| 18358 | b, |
| 18359 | ); |
| 18360 | #[rustfmt::skip] |
| 18361 | let e = _mm512_set_epi8(2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, |
| 18362 | 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, |
| 18363 | 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, |
| 18364 | 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1); |
| 18365 | assert_eq_m512i(r, e); |
| 18366 | } |
| 18367 | |
| 18368 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18369 | unsafe fn test_mm256_mask_blend_epi8() { |
| 18370 | let a = _mm256_set1_epi8(1); |
| 18371 | let b = _mm256_set1_epi8(2); |
| 18372 | let r = _mm256_mask_blend_epi8(0b11111111_00000000_11111111_00000000, a, b); |
| 18373 | #[rustfmt::skip] |
| 18374 | let e = _mm256_set_epi8(2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, |
| 18375 | 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1); |
| 18376 | assert_eq_m256i(r, e); |
| 18377 | } |
| 18378 | |
| 18379 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18380 | unsafe fn test_mm_mask_blend_epi8() { |
| 18381 | let a = _mm_set1_epi8(1); |
| 18382 | let b = _mm_set1_epi8(2); |
| 18383 | let r = _mm_mask_blend_epi8(0b11111111_00000000, a, b); |
| 18384 | let e = _mm_set_epi8(2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1); |
| 18385 | assert_eq_m128i(r, e); |
| 18386 | } |
| 18387 | |
| 18388 | #[simd_test(enable = "avx512bw" )] |
| 18389 | unsafe fn test_mm512_broadcastw_epi16() { |
| 18390 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
| 18391 | let r = _mm512_broadcastw_epi16(a); |
| 18392 | let e = _mm512_set1_epi16(24); |
| 18393 | assert_eq_m512i(r, e); |
| 18394 | } |
| 18395 | |
| 18396 | #[simd_test(enable = "avx512bw" )] |
| 18397 | unsafe fn test_mm512_mask_broadcastw_epi16() { |
| 18398 | let src = _mm512_set1_epi16(1); |
| 18399 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
| 18400 | let r = _mm512_mask_broadcastw_epi16(src, 0, a); |
| 18401 | assert_eq_m512i(r, src); |
| 18402 | let r = _mm512_mask_broadcastw_epi16(src, 0b11111111_11111111_11111111_11111111, a); |
| 18403 | let e = _mm512_set1_epi16(24); |
| 18404 | assert_eq_m512i(r, e); |
| 18405 | } |
| 18406 | |
| 18407 | #[simd_test(enable = "avx512bw" )] |
| 18408 | unsafe fn test_mm512_maskz_broadcastw_epi16() { |
| 18409 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
| 18410 | let r = _mm512_maskz_broadcastw_epi16(0, a); |
| 18411 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 18412 | let r = _mm512_maskz_broadcastw_epi16(0b11111111_11111111_11111111_11111111, a); |
| 18413 | let e = _mm512_set1_epi16(24); |
| 18414 | assert_eq_m512i(r, e); |
| 18415 | } |
| 18416 | |
| 18417 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18418 | unsafe fn test_mm256_mask_broadcastw_epi16() { |
| 18419 | let src = _mm256_set1_epi16(1); |
| 18420 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
| 18421 | let r = _mm256_mask_broadcastw_epi16(src, 0, a); |
| 18422 | assert_eq_m256i(r, src); |
| 18423 | let r = _mm256_mask_broadcastw_epi16(src, 0b11111111_11111111, a); |
| 18424 | let e = _mm256_set1_epi16(24); |
| 18425 | assert_eq_m256i(r, e); |
| 18426 | } |
| 18427 | |
| 18428 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18429 | unsafe fn test_mm256_maskz_broadcastw_epi16() { |
| 18430 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
| 18431 | let r = _mm256_maskz_broadcastw_epi16(0, a); |
| 18432 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 18433 | let r = _mm256_maskz_broadcastw_epi16(0b11111111_11111111, a); |
| 18434 | let e = _mm256_set1_epi16(24); |
| 18435 | assert_eq_m256i(r, e); |
| 18436 | } |
| 18437 | |
| 18438 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18439 | unsafe fn test_mm_mask_broadcastw_epi16() { |
| 18440 | let src = _mm_set1_epi16(1); |
| 18441 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
| 18442 | let r = _mm_mask_broadcastw_epi16(src, 0, a); |
| 18443 | assert_eq_m128i(r, src); |
| 18444 | let r = _mm_mask_broadcastw_epi16(src, 0b11111111, a); |
| 18445 | let e = _mm_set1_epi16(24); |
| 18446 | assert_eq_m128i(r, e); |
| 18447 | } |
| 18448 | |
| 18449 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18450 | unsafe fn test_mm_maskz_broadcastw_epi16() { |
| 18451 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
| 18452 | let r = _mm_maskz_broadcastw_epi16(0, a); |
| 18453 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 18454 | let r = _mm_maskz_broadcastw_epi16(0b11111111, a); |
| 18455 | let e = _mm_set1_epi16(24); |
| 18456 | assert_eq_m128i(r, e); |
| 18457 | } |
| 18458 | |
| 18459 | #[simd_test(enable = "avx512bw" )] |
| 18460 | unsafe fn test_mm512_broadcastb_epi8() { |
| 18461 | let a = _mm_set_epi8( |
| 18462 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18463 | ); |
| 18464 | let r = _mm512_broadcastb_epi8(a); |
| 18465 | let e = _mm512_set1_epi8(32); |
| 18466 | assert_eq_m512i(r, e); |
| 18467 | } |
| 18468 | |
| 18469 | #[simd_test(enable = "avx512bw" )] |
| 18470 | unsafe fn test_mm512_mask_broadcastb_epi8() { |
| 18471 | let src = _mm512_set1_epi8(1); |
| 18472 | let a = _mm_set_epi8( |
| 18473 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18474 | ); |
| 18475 | let r = _mm512_mask_broadcastb_epi8(src, 0, a); |
| 18476 | assert_eq_m512i(r, src); |
| 18477 | let r = _mm512_mask_broadcastb_epi8( |
| 18478 | src, |
| 18479 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 18480 | a, |
| 18481 | ); |
| 18482 | let e = _mm512_set1_epi8(32); |
| 18483 | assert_eq_m512i(r, e); |
| 18484 | } |
| 18485 | |
| 18486 | #[simd_test(enable = "avx512bw" )] |
| 18487 | unsafe fn test_mm512_maskz_broadcastb_epi8() { |
| 18488 | let a = _mm_set_epi8( |
| 18489 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18490 | ); |
| 18491 | let r = _mm512_maskz_broadcastb_epi8(0, a); |
| 18492 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 18493 | let r = _mm512_maskz_broadcastb_epi8( |
| 18494 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 18495 | a, |
| 18496 | ); |
| 18497 | let e = _mm512_set1_epi8(32); |
| 18498 | assert_eq_m512i(r, e); |
| 18499 | } |
| 18500 | |
| 18501 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18502 | unsafe fn test_mm256_mask_broadcastb_epi8() { |
| 18503 | let src = _mm256_set1_epi8(1); |
| 18504 | let a = _mm_set_epi8( |
| 18505 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18506 | ); |
| 18507 | let r = _mm256_mask_broadcastb_epi8(src, 0, a); |
| 18508 | assert_eq_m256i(r, src); |
| 18509 | let r = _mm256_mask_broadcastb_epi8(src, 0b11111111_11111111_11111111_11111111, a); |
| 18510 | let e = _mm256_set1_epi8(32); |
| 18511 | assert_eq_m256i(r, e); |
| 18512 | } |
| 18513 | |
| 18514 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18515 | unsafe fn test_mm256_maskz_broadcastb_epi8() { |
| 18516 | let a = _mm_set_epi8( |
| 18517 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18518 | ); |
| 18519 | let r = _mm256_maskz_broadcastb_epi8(0, a); |
| 18520 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 18521 | let r = _mm256_maskz_broadcastb_epi8(0b11111111_11111111_11111111_11111111, a); |
| 18522 | let e = _mm256_set1_epi8(32); |
| 18523 | assert_eq_m256i(r, e); |
| 18524 | } |
| 18525 | |
| 18526 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18527 | unsafe fn test_mm_mask_broadcastb_epi8() { |
| 18528 | let src = _mm_set1_epi8(1); |
| 18529 | let a = _mm_set_epi8( |
| 18530 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18531 | ); |
| 18532 | let r = _mm_mask_broadcastb_epi8(src, 0, a); |
| 18533 | assert_eq_m128i(r, src); |
| 18534 | let r = _mm_mask_broadcastb_epi8(src, 0b11111111_11111111, a); |
| 18535 | let e = _mm_set1_epi8(32); |
| 18536 | assert_eq_m128i(r, e); |
| 18537 | } |
| 18538 | |
| 18539 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18540 | unsafe fn test_mm_maskz_broadcastb_epi8() { |
| 18541 | let a = _mm_set_epi8( |
| 18542 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18543 | ); |
| 18544 | let r = _mm_maskz_broadcastb_epi8(0, a); |
| 18545 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 18546 | let r = _mm_maskz_broadcastb_epi8(0b11111111_11111111, a); |
| 18547 | let e = _mm_set1_epi8(32); |
| 18548 | assert_eq_m128i(r, e); |
| 18549 | } |
| 18550 | |
| 18551 | #[simd_test(enable = "avx512bw" )] |
| 18552 | unsafe fn test_mm512_unpackhi_epi16() { |
| 18553 | #[rustfmt::skip] |
| 18554 | let a = _mm512_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18555 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
| 18556 | #[rustfmt::skip] |
| 18557 | let b = _mm512_set_epi16(33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18558 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
| 18559 | let r = _mm512_unpackhi_epi16(a, b); |
| 18560 | #[rustfmt::skip] |
| 18561 | let e = _mm512_set_epi16(33, 1, 34, 2, 35, 3, 36, 4, 41, 9, 42, 10, 43, 11, 44, 12, |
| 18562 | 49, 17, 50, 18, 51, 19, 52, 20, 57, 25, 58, 26, 59, 27, 60, 28); |
| 18563 | assert_eq_m512i(r, e); |
| 18564 | } |
| 18565 | |
| 18566 | #[simd_test(enable = "avx512bw" )] |
| 18567 | unsafe fn test_mm512_mask_unpackhi_epi16() { |
| 18568 | #[rustfmt::skip] |
| 18569 | let a = _mm512_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18570 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
| 18571 | #[rustfmt::skip] |
| 18572 | let b = _mm512_set_epi16(33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18573 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
| 18574 | let r = _mm512_mask_unpackhi_epi16(a, 0, a, b); |
| 18575 | assert_eq_m512i(r, a); |
| 18576 | let r = _mm512_mask_unpackhi_epi16(a, 0b11111111_11111111_11111111_11111111, a, b); |
| 18577 | #[rustfmt::skip] |
| 18578 | let e = _mm512_set_epi16(33, 1, 34, 2, 35, 3, 36, 4, 41, 9, 42, 10, 43, 11, 44, 12, |
| 18579 | 49, 17, 50, 18, 51, 19, 52, 20, 57, 25, 58, 26, 59, 27, 60, 28); |
| 18580 | assert_eq_m512i(r, e); |
| 18581 | } |
| 18582 | |
| 18583 | #[simd_test(enable = "avx512bw" )] |
| 18584 | unsafe fn test_mm512_maskz_unpackhi_epi16() { |
| 18585 | #[rustfmt::skip] |
| 18586 | let a = _mm512_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18587 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
| 18588 | #[rustfmt::skip] |
| 18589 | let b = _mm512_set_epi16(33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18590 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
| 18591 | let r = _mm512_maskz_unpackhi_epi16(0, a, b); |
| 18592 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 18593 | let r = _mm512_maskz_unpackhi_epi16(0b11111111_11111111_11111111_11111111, a, b); |
| 18594 | #[rustfmt::skip] |
| 18595 | let e = _mm512_set_epi16(33, 1, 34, 2, 35, 3, 36, 4, 41, 9, 42, 10, 43, 11, 44, 12, |
| 18596 | 49, 17, 50, 18, 51, 19, 52, 20, 57, 25, 58, 26, 59, 27, 60, 28); |
| 18597 | assert_eq_m512i(r, e); |
| 18598 | } |
| 18599 | |
| 18600 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18601 | unsafe fn test_mm256_mask_unpackhi_epi16() { |
| 18602 | let a = _mm256_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
| 18603 | let b = _mm256_set_epi16( |
| 18604 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18605 | ); |
| 18606 | let r = _mm256_mask_unpackhi_epi16(a, 0, a, b); |
| 18607 | assert_eq_m256i(r, a); |
| 18608 | let r = _mm256_mask_unpackhi_epi16(a, 0b11111111_11111111, a, b); |
| 18609 | let e = _mm256_set_epi16(33, 1, 34, 2, 35, 3, 36, 4, 41, 9, 42, 10, 43, 11, 44, 12); |
| 18610 | assert_eq_m256i(r, e); |
| 18611 | } |
| 18612 | |
| 18613 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18614 | unsafe fn test_mm256_maskz_unpackhi_epi16() { |
| 18615 | let a = _mm256_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
| 18616 | let b = _mm256_set_epi16( |
| 18617 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18618 | ); |
| 18619 | let r = _mm256_maskz_unpackhi_epi16(0, a, b); |
| 18620 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 18621 | let r = _mm256_maskz_unpackhi_epi16(0b11111111_11111111, a, b); |
| 18622 | let e = _mm256_set_epi16(33, 1, 34, 2, 35, 3, 36, 4, 41, 9, 42, 10, 43, 11, 44, 12); |
| 18623 | assert_eq_m256i(r, e); |
| 18624 | } |
| 18625 | |
| 18626 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18627 | unsafe fn test_mm_mask_unpackhi_epi16() { |
| 18628 | let a = _mm_set_epi16(1, 2, 3, 4, 5, 6, 7, 8); |
| 18629 | let b = _mm_set_epi16(33, 34, 35, 36, 37, 38, 39, 40); |
| 18630 | let r = _mm_mask_unpackhi_epi16(a, 0, a, b); |
| 18631 | assert_eq_m128i(r, a); |
| 18632 | let r = _mm_mask_unpackhi_epi16(a, 0b11111111, a, b); |
| 18633 | let e = _mm_set_epi16(33, 1, 34, 2, 35, 3, 36, 4); |
| 18634 | assert_eq_m128i(r, e); |
| 18635 | } |
| 18636 | |
| 18637 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18638 | unsafe fn test_mm_maskz_unpackhi_epi16() { |
| 18639 | let a = _mm_set_epi16(1, 2, 3, 4, 5, 6, 7, 8); |
| 18640 | let b = _mm_set_epi16(33, 34, 35, 36, 37, 38, 39, 40); |
| 18641 | let r = _mm_maskz_unpackhi_epi16(0, a, b); |
| 18642 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 18643 | let r = _mm_maskz_unpackhi_epi16(0b11111111, a, b); |
| 18644 | let e = _mm_set_epi16(33, 1, 34, 2, 35, 3, 36, 4); |
| 18645 | assert_eq_m128i(r, e); |
| 18646 | } |
| 18647 | |
| 18648 | #[simd_test(enable = "avx512bw" )] |
| 18649 | unsafe fn test_mm512_unpackhi_epi8() { |
| 18650 | #[rustfmt::skip] |
| 18651 | let a = _mm512_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18652 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18653 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18654 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
| 18655 | #[rustfmt::skip] |
| 18656 | let b = _mm512_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18657 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, |
| 18658 | 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, |
| 18659 | 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 0); |
| 18660 | let r = _mm512_unpackhi_epi8(a, b); |
| 18661 | #[rustfmt::skip] |
| 18662 | let e = _mm512_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8, |
| 18663 | 81, 17, 82, 18, 83, 19, 84, 20, 85, 21, 86, 22, 87, 23, 88, 24, |
| 18664 | 97, 33, 98, 34, 99, 35, 100, 36, 101, 37, 102, 38, 103, 39, 104, 40, |
| 18665 | 113, 49, 114, 50, 115, 51, 116, 52, 117, 53, 118, 54, 119, 55, 120, 56); |
| 18666 | assert_eq_m512i(r, e); |
| 18667 | } |
| 18668 | |
| 18669 | #[simd_test(enable = "avx512bw" )] |
| 18670 | unsafe fn test_mm512_mask_unpackhi_epi8() { |
| 18671 | #[rustfmt::skip] |
| 18672 | let a = _mm512_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18673 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18674 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18675 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
| 18676 | #[rustfmt::skip] |
| 18677 | let b = _mm512_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18678 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, |
| 18679 | 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, |
| 18680 | 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 0); |
| 18681 | let r = _mm512_mask_unpackhi_epi8(a, 0, a, b); |
| 18682 | assert_eq_m512i(r, a); |
| 18683 | let r = _mm512_mask_unpackhi_epi8( |
| 18684 | a, |
| 18685 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 18686 | a, |
| 18687 | b, |
| 18688 | ); |
| 18689 | #[rustfmt::skip] |
| 18690 | let e = _mm512_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8, |
| 18691 | 81, 17, 82, 18, 83, 19, 84, 20, 85, 21, 86, 22, 87, 23, 88, 24, |
| 18692 | 97, 33, 98, 34, 99, 35, 100, 36, 101, 37, 102, 38, 103, 39, 104, 40, |
| 18693 | 113, 49, 114, 50, 115, 51, 116, 52, 117, 53, 118, 54, 119, 55, 120, 56); |
| 18694 | assert_eq_m512i(r, e); |
| 18695 | } |
| 18696 | |
| 18697 | #[simd_test(enable = "avx512bw" )] |
| 18698 | unsafe fn test_mm512_maskz_unpackhi_epi8() { |
| 18699 | #[rustfmt::skip] |
| 18700 | let a = _mm512_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18701 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18702 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18703 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
| 18704 | #[rustfmt::skip] |
| 18705 | let b = _mm512_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18706 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, |
| 18707 | 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, |
| 18708 | 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 0); |
| 18709 | let r = _mm512_maskz_unpackhi_epi8(0, a, b); |
| 18710 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 18711 | let r = _mm512_maskz_unpackhi_epi8( |
| 18712 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 18713 | a, |
| 18714 | b, |
| 18715 | ); |
| 18716 | #[rustfmt::skip] |
| 18717 | let e = _mm512_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8, |
| 18718 | 81, 17, 82, 18, 83, 19, 84, 20, 85, 21, 86, 22, 87, 23, 88, 24, |
| 18719 | 97, 33, 98, 34, 99, 35, 100, 36, 101, 37, 102, 38, 103, 39, 104, 40, |
| 18720 | 113, 49, 114, 50, 115, 51, 116, 52, 117, 53, 118, 54, 119, 55, 120, 56); |
| 18721 | assert_eq_m512i(r, e); |
| 18722 | } |
| 18723 | |
| 18724 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18725 | unsafe fn test_mm256_mask_unpackhi_epi8() { |
| 18726 | #[rustfmt::skip] |
| 18727 | let a = _mm256_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18728 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
| 18729 | #[rustfmt::skip] |
| 18730 | let b = _mm256_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18731 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96); |
| 18732 | let r = _mm256_mask_unpackhi_epi8(a, 0, a, b); |
| 18733 | assert_eq_m256i(r, a); |
| 18734 | let r = _mm256_mask_unpackhi_epi8(a, 0b11111111_11111111_11111111_11111111, a, b); |
| 18735 | #[rustfmt::skip] |
| 18736 | let e = _mm256_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8, |
| 18737 | 81, 17, 82, 18, 83, 19, 84, 20, 85, 21, 86, 22, 87, 23, 88, 24); |
| 18738 | assert_eq_m256i(r, e); |
| 18739 | } |
| 18740 | |
| 18741 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18742 | unsafe fn test_mm256_maskz_unpackhi_epi8() { |
| 18743 | #[rustfmt::skip] |
| 18744 | let a = _mm256_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18745 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
| 18746 | #[rustfmt::skip] |
| 18747 | let b = _mm256_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18748 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96); |
| 18749 | let r = _mm256_maskz_unpackhi_epi8(0, a, b); |
| 18750 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 18751 | let r = _mm256_maskz_unpackhi_epi8(0b11111111_11111111_11111111_11111111, a, b); |
| 18752 | #[rustfmt::skip] |
| 18753 | let e = _mm256_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8, |
| 18754 | 81, 17, 82, 18, 83, 19, 84, 20, 85, 21, 86, 22, 87, 23, 88, 24); |
| 18755 | assert_eq_m256i(r, e); |
| 18756 | } |
| 18757 | |
| 18758 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18759 | unsafe fn test_mm_mask_unpackhi_epi8() { |
| 18760 | let a = _mm_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
| 18761 | let b = _mm_set_epi8( |
| 18762 | 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18763 | ); |
| 18764 | let r = _mm_mask_unpackhi_epi8(a, 0, a, b); |
| 18765 | assert_eq_m128i(r, a); |
| 18766 | let r = _mm_mask_unpackhi_epi8(a, 0b11111111_11111111, a, b); |
| 18767 | let e = _mm_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8); |
| 18768 | assert_eq_m128i(r, e); |
| 18769 | } |
| 18770 | |
| 18771 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18772 | unsafe fn test_mm_maskz_unpackhi_epi8() { |
| 18773 | let a = _mm_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
| 18774 | let b = _mm_set_epi8( |
| 18775 | 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18776 | ); |
| 18777 | let r = _mm_maskz_unpackhi_epi8(0, a, b); |
| 18778 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 18779 | let r = _mm_maskz_unpackhi_epi8(0b11111111_11111111, a, b); |
| 18780 | let e = _mm_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8); |
| 18781 | assert_eq_m128i(r, e); |
| 18782 | } |
| 18783 | |
| 18784 | #[simd_test(enable = "avx512bw" )] |
| 18785 | unsafe fn test_mm512_unpacklo_epi16() { |
| 18786 | #[rustfmt::skip] |
| 18787 | let a = _mm512_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18788 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
| 18789 | #[rustfmt::skip] |
| 18790 | let b = _mm512_set_epi16(33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18791 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
| 18792 | let r = _mm512_unpacklo_epi16(a, b); |
| 18793 | #[rustfmt::skip] |
| 18794 | let e = _mm512_set_epi16(37, 5, 38, 6, 39, 7, 40, 8, 45, 13, 46, 14, 47, 15, 48, 16, |
| 18795 | 53, 21, 54, 22, 55, 23, 56, 24, 61, 29, 62, 30, 63, 31, 64, 32); |
| 18796 | assert_eq_m512i(r, e); |
| 18797 | } |
| 18798 | |
| 18799 | #[simd_test(enable = "avx512bw" )] |
| 18800 | unsafe fn test_mm512_mask_unpacklo_epi16() { |
| 18801 | #[rustfmt::skip] |
| 18802 | let a = _mm512_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18803 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
| 18804 | #[rustfmt::skip] |
| 18805 | let b = _mm512_set_epi16(33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18806 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
| 18807 | let r = _mm512_mask_unpacklo_epi16(a, 0, a, b); |
| 18808 | assert_eq_m512i(r, a); |
| 18809 | let r = _mm512_mask_unpacklo_epi16(a, 0b11111111_11111111_11111111_11111111, a, b); |
| 18810 | #[rustfmt::skip] |
| 18811 | let e = _mm512_set_epi16(37, 5, 38, 6, 39, 7, 40, 8, 45, 13, 46, 14, 47, 15, 48, 16, |
| 18812 | 53, 21, 54, 22, 55, 23, 56, 24, 61, 29, 62, 30, 63, 31, 64, 32); |
| 18813 | assert_eq_m512i(r, e); |
| 18814 | } |
| 18815 | |
| 18816 | #[simd_test(enable = "avx512bw" )] |
| 18817 | unsafe fn test_mm512_maskz_unpacklo_epi16() { |
| 18818 | #[rustfmt::skip] |
| 18819 | let a = _mm512_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18820 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
| 18821 | #[rustfmt::skip] |
| 18822 | let b = _mm512_set_epi16(33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18823 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
| 18824 | let r = _mm512_maskz_unpacklo_epi16(0, a, b); |
| 18825 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 18826 | let r = _mm512_maskz_unpacklo_epi16(0b11111111_11111111_11111111_11111111, a, b); |
| 18827 | #[rustfmt::skip] |
| 18828 | let e = _mm512_set_epi16(37, 5, 38, 6, 39, 7, 40, 8, 45, 13, 46, 14, 47, 15, 48, 16, |
| 18829 | 53, 21, 54, 22, 55, 23, 56, 24, 61, 29, 62, 30, 63, 31, 64, 32); |
| 18830 | assert_eq_m512i(r, e); |
| 18831 | } |
| 18832 | |
| 18833 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18834 | unsafe fn test_mm256_mask_unpacklo_epi16() { |
| 18835 | let a = _mm256_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
| 18836 | let b = _mm256_set_epi16( |
| 18837 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18838 | ); |
| 18839 | let r = _mm256_mask_unpacklo_epi16(a, 0, a, b); |
| 18840 | assert_eq_m256i(r, a); |
| 18841 | let r = _mm256_mask_unpacklo_epi16(a, 0b11111111_11111111, a, b); |
| 18842 | let e = _mm256_set_epi16(37, 5, 38, 6, 39, 7, 40, 8, 45, 13, 46, 14, 47, 15, 48, 16); |
| 18843 | assert_eq_m256i(r, e); |
| 18844 | } |
| 18845 | |
| 18846 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18847 | unsafe fn test_mm256_maskz_unpacklo_epi16() { |
| 18848 | let a = _mm256_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
| 18849 | let b = _mm256_set_epi16( |
| 18850 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18851 | ); |
| 18852 | let r = _mm256_maskz_unpacklo_epi16(0, a, b); |
| 18853 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 18854 | let r = _mm256_maskz_unpacklo_epi16(0b11111111_11111111, a, b); |
| 18855 | let e = _mm256_set_epi16(37, 5, 38, 6, 39, 7, 40, 8, 45, 13, 46, 14, 47, 15, 48, 16); |
| 18856 | assert_eq_m256i(r, e); |
| 18857 | } |
| 18858 | |
| 18859 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18860 | unsafe fn test_mm_mask_unpacklo_epi16() { |
| 18861 | let a = _mm_set_epi16(1, 2, 3, 4, 5, 6, 7, 8); |
| 18862 | let b = _mm_set_epi16(33, 34, 35, 36, 37, 38, 39, 40); |
| 18863 | let r = _mm_mask_unpacklo_epi16(a, 0, a, b); |
| 18864 | assert_eq_m128i(r, a); |
| 18865 | let r = _mm_mask_unpacklo_epi16(a, 0b11111111, a, b); |
| 18866 | let e = _mm_set_epi16(37, 5, 38, 6, 39, 7, 40, 8); |
| 18867 | assert_eq_m128i(r, e); |
| 18868 | } |
| 18869 | |
| 18870 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18871 | unsafe fn test_mm_maskz_unpacklo_epi16() { |
| 18872 | let a = _mm_set_epi16(1, 2, 3, 4, 5, 6, 7, 8); |
| 18873 | let b = _mm_set_epi16(33, 34, 35, 36, 37, 38, 39, 40); |
| 18874 | let r = _mm_maskz_unpacklo_epi16(0, a, b); |
| 18875 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 18876 | let r = _mm_maskz_unpacklo_epi16(0b11111111, a, b); |
| 18877 | let e = _mm_set_epi16(37, 5, 38, 6, 39, 7, 40, 8); |
| 18878 | assert_eq_m128i(r, e); |
| 18879 | } |
| 18880 | |
| 18881 | #[simd_test(enable = "avx512bw" )] |
| 18882 | unsafe fn test_mm512_unpacklo_epi8() { |
| 18883 | #[rustfmt::skip] |
| 18884 | let a = _mm512_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18885 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18886 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18887 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
| 18888 | #[rustfmt::skip] |
| 18889 | let b = _mm512_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18890 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, |
| 18891 | 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, |
| 18892 | 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 0); |
| 18893 | let r = _mm512_unpacklo_epi8(a, b); |
| 18894 | #[rustfmt::skip] |
| 18895 | let e = _mm512_set_epi8(73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
| 18896 | 89, 25, 90, 26, 91, 27, 92, 28, 93, 29, 94, 30, 95, 31, 96, 32, |
| 18897 | 105, 41, 106, 42, 107, 43, 108, 44, 109, 45, 110, 46, 111, 47, 112, 48, |
| 18898 | 121, 57, 122, 58, 123, 59, 124, 60, 125, 61, 126, 62, 127, 63, 0, 64); |
| 18899 | assert_eq_m512i(r, e); |
| 18900 | } |
| 18901 | |
| 18902 | #[simd_test(enable = "avx512bw" )] |
| 18903 | unsafe fn test_mm512_mask_unpacklo_epi8() { |
| 18904 | #[rustfmt::skip] |
| 18905 | let a = _mm512_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18906 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18907 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18908 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
| 18909 | #[rustfmt::skip] |
| 18910 | let b = _mm512_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18911 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, |
| 18912 | 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, |
| 18913 | 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 0); |
| 18914 | let r = _mm512_mask_unpacklo_epi8(a, 0, a, b); |
| 18915 | assert_eq_m512i(r, a); |
| 18916 | let r = _mm512_mask_unpacklo_epi8( |
| 18917 | a, |
| 18918 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 18919 | a, |
| 18920 | b, |
| 18921 | ); |
| 18922 | #[rustfmt::skip] |
| 18923 | let e = _mm512_set_epi8(73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
| 18924 | 89, 25, 90, 26, 91, 27, 92, 28, 93, 29, 94, 30, 95, 31, 96, 32, |
| 18925 | 105, 41, 106, 42, 107, 43, 108, 44, 109, 45, 110, 46, 111, 47, 112, 48, |
| 18926 | 121, 57, 122, 58, 123, 59, 124, 60, 125, 61, 126, 62, 127, 63, 0, 64); |
| 18927 | assert_eq_m512i(r, e); |
| 18928 | } |
| 18929 | |
| 18930 | #[simd_test(enable = "avx512bw" )] |
| 18931 | unsafe fn test_mm512_maskz_unpacklo_epi8() { |
| 18932 | #[rustfmt::skip] |
| 18933 | let a = _mm512_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18934 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 18935 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 18936 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
| 18937 | #[rustfmt::skip] |
| 18938 | let b = _mm512_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18939 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, |
| 18940 | 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, |
| 18941 | 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 0); |
| 18942 | let r = _mm512_maskz_unpacklo_epi8(0, a, b); |
| 18943 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 18944 | let r = _mm512_maskz_unpacklo_epi8( |
| 18945 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 18946 | a, |
| 18947 | b, |
| 18948 | ); |
| 18949 | #[rustfmt::skip] |
| 18950 | let e = _mm512_set_epi8(73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
| 18951 | 89, 25, 90, 26, 91, 27, 92, 28, 93, 29, 94, 30, 95, 31, 96, 32, |
| 18952 | 105, 41, 106, 42, 107, 43, 108, 44, 109, 45, 110, 46, 111, 47, 112, 48, |
| 18953 | 121, 57, 122, 58, 123, 59, 124, 60, 125, 61, 126, 62, 127, 63, 0, 64); |
| 18954 | assert_eq_m512i(r, e); |
| 18955 | } |
| 18956 | |
| 18957 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18958 | unsafe fn test_mm256_mask_unpacklo_epi8() { |
| 18959 | #[rustfmt::skip] |
| 18960 | let a = _mm256_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18961 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
| 18962 | #[rustfmt::skip] |
| 18963 | let b = _mm256_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18964 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96); |
| 18965 | let r = _mm256_mask_unpacklo_epi8(a, 0, a, b); |
| 18966 | assert_eq_m256i(r, a); |
| 18967 | let r = _mm256_mask_unpacklo_epi8(a, 0b11111111_11111111_11111111_11111111, a, b); |
| 18968 | #[rustfmt::skip] |
| 18969 | let e = _mm256_set_epi8(73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
| 18970 | 89, 25, 90, 26, 91, 27, 92, 28, 93, 29, 94, 30, 95, 31, 96, 32); |
| 18971 | assert_eq_m256i(r, e); |
| 18972 | } |
| 18973 | |
| 18974 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18975 | unsafe fn test_mm256_maskz_unpacklo_epi8() { |
| 18976 | #[rustfmt::skip] |
| 18977 | let a = _mm256_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 18978 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
| 18979 | #[rustfmt::skip] |
| 18980 | let b = _mm256_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18981 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96); |
| 18982 | let r = _mm256_maskz_unpacklo_epi8(0, a, b); |
| 18983 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 18984 | let r = _mm256_maskz_unpacklo_epi8(0b11111111_11111111_11111111_11111111, a, b); |
| 18985 | #[rustfmt::skip] |
| 18986 | let e = _mm256_set_epi8(73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
| 18987 | 89, 25, 90, 26, 91, 27, 92, 28, 93, 29, 94, 30, 95, 31, 96, 32); |
| 18988 | assert_eq_m256i(r, e); |
| 18989 | } |
| 18990 | |
| 18991 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 18992 | unsafe fn test_mm_mask_unpacklo_epi8() { |
| 18993 | let a = _mm_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
| 18994 | let b = _mm_set_epi8( |
| 18995 | 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 18996 | ); |
| 18997 | let r = _mm_mask_unpacklo_epi8(a, 0, a, b); |
| 18998 | assert_eq_m128i(r, a); |
| 18999 | let r = _mm_mask_unpacklo_epi8(a, 0b11111111_11111111, a, b); |
| 19000 | let e = _mm_set_epi8( |
| 19001 | 73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
| 19002 | ); |
| 19003 | assert_eq_m128i(r, e); |
| 19004 | } |
| 19005 | |
| 19006 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19007 | unsafe fn test_mm_maskz_unpacklo_epi8() { |
| 19008 | let a = _mm_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
| 19009 | let b = _mm_set_epi8( |
| 19010 | 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
| 19011 | ); |
| 19012 | let r = _mm_maskz_unpacklo_epi8(0, a, b); |
| 19013 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 19014 | let r = _mm_maskz_unpacklo_epi8(0b11111111_11111111, a, b); |
| 19015 | let e = _mm_set_epi8( |
| 19016 | 73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
| 19017 | ); |
| 19018 | assert_eq_m128i(r, e); |
| 19019 | } |
| 19020 | |
| 19021 | #[simd_test(enable = "avx512bw" )] |
| 19022 | unsafe fn test_mm512_mask_mov_epi16() { |
| 19023 | let src = _mm512_set1_epi16(1); |
| 19024 | let a = _mm512_set1_epi16(2); |
| 19025 | let r = _mm512_mask_mov_epi16(src, 0, a); |
| 19026 | assert_eq_m512i(r, src); |
| 19027 | let r = _mm512_mask_mov_epi16(src, 0b11111111_11111111_11111111_11111111, a); |
| 19028 | assert_eq_m512i(r, a); |
| 19029 | } |
| 19030 | |
| 19031 | #[simd_test(enable = "avx512bw" )] |
| 19032 | unsafe fn test_mm512_maskz_mov_epi16() { |
| 19033 | let a = _mm512_set1_epi16(2); |
| 19034 | let r = _mm512_maskz_mov_epi16(0, a); |
| 19035 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 19036 | let r = _mm512_maskz_mov_epi16(0b11111111_11111111_11111111_11111111, a); |
| 19037 | assert_eq_m512i(r, a); |
| 19038 | } |
| 19039 | |
| 19040 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19041 | unsafe fn test_mm256_mask_mov_epi16() { |
| 19042 | let src = _mm256_set1_epi16(1); |
| 19043 | let a = _mm256_set1_epi16(2); |
| 19044 | let r = _mm256_mask_mov_epi16(src, 0, a); |
| 19045 | assert_eq_m256i(r, src); |
| 19046 | let r = _mm256_mask_mov_epi16(src, 0b11111111_11111111, a); |
| 19047 | assert_eq_m256i(r, a); |
| 19048 | } |
| 19049 | |
| 19050 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19051 | unsafe fn test_mm256_maskz_mov_epi16() { |
| 19052 | let a = _mm256_set1_epi16(2); |
| 19053 | let r = _mm256_maskz_mov_epi16(0, a); |
| 19054 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 19055 | let r = _mm256_maskz_mov_epi16(0b11111111_11111111, a); |
| 19056 | assert_eq_m256i(r, a); |
| 19057 | } |
| 19058 | |
| 19059 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19060 | unsafe fn test_mm_mask_mov_epi16() { |
| 19061 | let src = _mm_set1_epi16(1); |
| 19062 | let a = _mm_set1_epi16(2); |
| 19063 | let r = _mm_mask_mov_epi16(src, 0, a); |
| 19064 | assert_eq_m128i(r, src); |
| 19065 | let r = _mm_mask_mov_epi16(src, 0b11111111, a); |
| 19066 | assert_eq_m128i(r, a); |
| 19067 | } |
| 19068 | |
| 19069 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19070 | unsafe fn test_mm_maskz_mov_epi16() { |
| 19071 | let a = _mm_set1_epi16(2); |
| 19072 | let r = _mm_maskz_mov_epi16(0, a); |
| 19073 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 19074 | let r = _mm_maskz_mov_epi16(0b11111111, a); |
| 19075 | assert_eq_m128i(r, a); |
| 19076 | } |
| 19077 | |
| 19078 | #[simd_test(enable = "avx512bw" )] |
| 19079 | unsafe fn test_mm512_mask_mov_epi8() { |
| 19080 | let src = _mm512_set1_epi8(1); |
| 19081 | let a = _mm512_set1_epi8(2); |
| 19082 | let r = _mm512_mask_mov_epi8(src, 0, a); |
| 19083 | assert_eq_m512i(r, src); |
| 19084 | let r = _mm512_mask_mov_epi8( |
| 19085 | src, |
| 19086 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 19087 | a, |
| 19088 | ); |
| 19089 | assert_eq_m512i(r, a); |
| 19090 | } |
| 19091 | |
| 19092 | #[simd_test(enable = "avx512bw" )] |
| 19093 | unsafe fn test_mm512_maskz_mov_epi8() { |
| 19094 | let a = _mm512_set1_epi8(2); |
| 19095 | let r = _mm512_maskz_mov_epi8(0, a); |
| 19096 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 19097 | let r = _mm512_maskz_mov_epi8( |
| 19098 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 19099 | a, |
| 19100 | ); |
| 19101 | assert_eq_m512i(r, a); |
| 19102 | } |
| 19103 | |
| 19104 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19105 | unsafe fn test_mm256_mask_mov_epi8() { |
| 19106 | let src = _mm256_set1_epi8(1); |
| 19107 | let a = _mm256_set1_epi8(2); |
| 19108 | let r = _mm256_mask_mov_epi8(src, 0, a); |
| 19109 | assert_eq_m256i(r, src); |
| 19110 | let r = _mm256_mask_mov_epi8(src, 0b11111111_11111111_11111111_11111111, a); |
| 19111 | assert_eq_m256i(r, a); |
| 19112 | } |
| 19113 | |
| 19114 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19115 | unsafe fn test_mm256_maskz_mov_epi8() { |
| 19116 | let a = _mm256_set1_epi8(2); |
| 19117 | let r = _mm256_maskz_mov_epi8(0, a); |
| 19118 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 19119 | let r = _mm256_maskz_mov_epi8(0b11111111_11111111_11111111_11111111, a); |
| 19120 | assert_eq_m256i(r, a); |
| 19121 | } |
| 19122 | |
| 19123 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19124 | unsafe fn test_mm_mask_mov_epi8() { |
| 19125 | let src = _mm_set1_epi8(1); |
| 19126 | let a = _mm_set1_epi8(2); |
| 19127 | let r = _mm_mask_mov_epi8(src, 0, a); |
| 19128 | assert_eq_m128i(r, src); |
| 19129 | let r = _mm_mask_mov_epi8(src, 0b11111111_11111111, a); |
| 19130 | assert_eq_m128i(r, a); |
| 19131 | } |
| 19132 | |
| 19133 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19134 | unsafe fn test_mm_maskz_mov_epi8() { |
| 19135 | let a = _mm_set1_epi8(2); |
| 19136 | let r = _mm_maskz_mov_epi8(0, a); |
| 19137 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 19138 | let r = _mm_maskz_mov_epi8(0b11111111_11111111, a); |
| 19139 | assert_eq_m128i(r, a); |
| 19140 | } |
| 19141 | |
| 19142 | #[simd_test(enable = "avx512bw" )] |
| 19143 | unsafe fn test_mm512_mask_set1_epi16() { |
| 19144 | let src = _mm512_set1_epi16(2); |
| 19145 | let a: i16 = 11; |
| 19146 | let r = _mm512_mask_set1_epi16(src, 0, a); |
| 19147 | assert_eq_m512i(r, src); |
| 19148 | let r = _mm512_mask_set1_epi16(src, 0b11111111_11111111_11111111_11111111, a); |
| 19149 | let e = _mm512_set1_epi16(11); |
| 19150 | assert_eq_m512i(r, e); |
| 19151 | } |
| 19152 | |
| 19153 | #[simd_test(enable = "avx512bw" )] |
| 19154 | unsafe fn test_mm512_maskz_set1_epi16() { |
| 19155 | let a: i16 = 11; |
| 19156 | let r = _mm512_maskz_set1_epi16(0, a); |
| 19157 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 19158 | let r = _mm512_maskz_set1_epi16(0b11111111_11111111_11111111_11111111, a); |
| 19159 | let e = _mm512_set1_epi16(11); |
| 19160 | assert_eq_m512i(r, e); |
| 19161 | } |
| 19162 | |
| 19163 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19164 | unsafe fn test_mm256_mask_set1_epi16() { |
| 19165 | let src = _mm256_set1_epi16(2); |
| 19166 | let a: i16 = 11; |
| 19167 | let r = _mm256_mask_set1_epi16(src, 0, a); |
| 19168 | assert_eq_m256i(r, src); |
| 19169 | let r = _mm256_mask_set1_epi16(src, 0b11111111_11111111, a); |
| 19170 | let e = _mm256_set1_epi16(11); |
| 19171 | assert_eq_m256i(r, e); |
| 19172 | } |
| 19173 | |
| 19174 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19175 | unsafe fn test_mm256_maskz_set1_epi16() { |
| 19176 | let a: i16 = 11; |
| 19177 | let r = _mm256_maskz_set1_epi16(0, a); |
| 19178 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 19179 | let r = _mm256_maskz_set1_epi16(0b11111111_11111111, a); |
| 19180 | let e = _mm256_set1_epi16(11); |
| 19181 | assert_eq_m256i(r, e); |
| 19182 | } |
| 19183 | |
| 19184 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19185 | unsafe fn test_mm_mask_set1_epi16() { |
| 19186 | let src = _mm_set1_epi16(2); |
| 19187 | let a: i16 = 11; |
| 19188 | let r = _mm_mask_set1_epi16(src, 0, a); |
| 19189 | assert_eq_m128i(r, src); |
| 19190 | let r = _mm_mask_set1_epi16(src, 0b11111111, a); |
| 19191 | let e = _mm_set1_epi16(11); |
| 19192 | assert_eq_m128i(r, e); |
| 19193 | } |
| 19194 | |
| 19195 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19196 | unsafe fn test_mm_maskz_set1_epi16() { |
| 19197 | let a: i16 = 11; |
| 19198 | let r = _mm_maskz_set1_epi16(0, a); |
| 19199 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 19200 | let r = _mm_maskz_set1_epi16(0b11111111, a); |
| 19201 | let e = _mm_set1_epi16(11); |
| 19202 | assert_eq_m128i(r, e); |
| 19203 | } |
| 19204 | |
| 19205 | #[simd_test(enable = "avx512bw" )] |
| 19206 | unsafe fn test_mm512_mask_set1_epi8() { |
| 19207 | let src = _mm512_set1_epi8(2); |
| 19208 | let a: i8 = 11; |
| 19209 | let r = _mm512_mask_set1_epi8(src, 0, a); |
| 19210 | assert_eq_m512i(r, src); |
| 19211 | let r = _mm512_mask_set1_epi8( |
| 19212 | src, |
| 19213 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 19214 | a, |
| 19215 | ); |
| 19216 | let e = _mm512_set1_epi8(11); |
| 19217 | assert_eq_m512i(r, e); |
| 19218 | } |
| 19219 | |
| 19220 | #[simd_test(enable = "avx512bw" )] |
| 19221 | unsafe fn test_mm512_maskz_set1_epi8() { |
| 19222 | let a: i8 = 11; |
| 19223 | let r = _mm512_maskz_set1_epi8(0, a); |
| 19224 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 19225 | let r = _mm512_maskz_set1_epi8( |
| 19226 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 19227 | a, |
| 19228 | ); |
| 19229 | let e = _mm512_set1_epi8(11); |
| 19230 | assert_eq_m512i(r, e); |
| 19231 | } |
| 19232 | |
| 19233 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19234 | unsafe fn test_mm256_mask_set1_epi8() { |
| 19235 | let src = _mm256_set1_epi8(2); |
| 19236 | let a: i8 = 11; |
| 19237 | let r = _mm256_mask_set1_epi8(src, 0, a); |
| 19238 | assert_eq_m256i(r, src); |
| 19239 | let r = _mm256_mask_set1_epi8(src, 0b11111111_11111111_11111111_11111111, a); |
| 19240 | let e = _mm256_set1_epi8(11); |
| 19241 | assert_eq_m256i(r, e); |
| 19242 | } |
| 19243 | |
| 19244 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19245 | unsafe fn test_mm256_maskz_set1_epi8() { |
| 19246 | let a: i8 = 11; |
| 19247 | let r = _mm256_maskz_set1_epi8(0, a); |
| 19248 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 19249 | let r = _mm256_maskz_set1_epi8(0b11111111_11111111_11111111_11111111, a); |
| 19250 | let e = _mm256_set1_epi8(11); |
| 19251 | assert_eq_m256i(r, e); |
| 19252 | } |
| 19253 | |
| 19254 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19255 | unsafe fn test_mm_mask_set1_epi8() { |
| 19256 | let src = _mm_set1_epi8(2); |
| 19257 | let a: i8 = 11; |
| 19258 | let r = _mm_mask_set1_epi8(src, 0, a); |
| 19259 | assert_eq_m128i(r, src); |
| 19260 | let r = _mm_mask_set1_epi8(src, 0b11111111_11111111, a); |
| 19261 | let e = _mm_set1_epi8(11); |
| 19262 | assert_eq_m128i(r, e); |
| 19263 | } |
| 19264 | |
| 19265 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19266 | unsafe fn test_mm_maskz_set1_epi8() { |
| 19267 | let a: i8 = 11; |
| 19268 | let r = _mm_maskz_set1_epi8(0, a); |
| 19269 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 19270 | let r = _mm_maskz_set1_epi8(0b11111111_11111111, a); |
| 19271 | let e = _mm_set1_epi8(11); |
| 19272 | assert_eq_m128i(r, e); |
| 19273 | } |
| 19274 | |
| 19275 | #[simd_test(enable = "avx512bw" )] |
| 19276 | unsafe fn test_mm512_shufflelo_epi16() { |
| 19277 | #[rustfmt::skip] |
| 19278 | let a = _mm512_set_epi16( |
| 19279 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 19280 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
| 19281 | ); |
| 19282 | #[rustfmt::skip] |
| 19283 | let e = _mm512_set_epi16( |
| 19284 | 0, 1, 2, 3, 7, 6, 6, 4, 8, 9, 10, 11, 15, 14, 14, 12, |
| 19285 | 16, 17, 18, 19, 23, 22, 22, 20, 24, 25, 26, 27, 31, 30, 30, 28, |
| 19286 | ); |
| 19287 | let r = _mm512_shufflelo_epi16::<0b00_01_01_11>(a); |
| 19288 | assert_eq_m512i(r, e); |
| 19289 | } |
| 19290 | |
| 19291 | #[simd_test(enable = "avx512bw" )] |
| 19292 | unsafe fn test_mm512_mask_shufflelo_epi16() { |
| 19293 | #[rustfmt::skip] |
| 19294 | let a = _mm512_set_epi16( |
| 19295 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 19296 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
| 19297 | ); |
| 19298 | let r = _mm512_mask_shufflelo_epi16::<0b00_01_01_11>(a, 0, a); |
| 19299 | assert_eq_m512i(r, a); |
| 19300 | let r = _mm512_mask_shufflelo_epi16::<0b00_01_01_11>( |
| 19301 | a, |
| 19302 | 0b11111111_11111111_11111111_11111111, |
| 19303 | a, |
| 19304 | ); |
| 19305 | #[rustfmt::skip] |
| 19306 | let e = _mm512_set_epi16( |
| 19307 | 0, 1, 2, 3, 7, 6, 6, 4, 8, 9, 10, 11, 15, 14, 14, 12, |
| 19308 | 16, 17, 18, 19, 23, 22, 22, 20, 24, 25, 26, 27, 31, 30, 30, 28, |
| 19309 | ); |
| 19310 | assert_eq_m512i(r, e); |
| 19311 | } |
| 19312 | |
| 19313 | #[simd_test(enable = "avx512bw" )] |
| 19314 | unsafe fn test_mm512_maskz_shufflelo_epi16() { |
| 19315 | #[rustfmt::skip] |
| 19316 | let a = _mm512_set_epi16( |
| 19317 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 19318 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
| 19319 | ); |
| 19320 | let r = _mm512_maskz_shufflelo_epi16::<0b00_01_01_11>(0, a); |
| 19321 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 19322 | let r = |
| 19323 | _mm512_maskz_shufflelo_epi16::<0b00_01_01_11>(0b11111111_11111111_11111111_11111111, a); |
| 19324 | #[rustfmt::skip] |
| 19325 | let e = _mm512_set_epi16( |
| 19326 | 0, 1, 2, 3, 7, 6, 6, 4, 8, 9, 10, 11, 15, 14, 14, 12, |
| 19327 | 16, 17, 18, 19, 23, 22, 22, 20, 24, 25, 26, 27, 31, 30, 30, 28, |
| 19328 | ); |
| 19329 | assert_eq_m512i(r, e); |
| 19330 | } |
| 19331 | |
| 19332 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19333 | unsafe fn test_mm256_mask_shufflelo_epi16() { |
| 19334 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 19335 | let r = _mm256_mask_shufflelo_epi16::<0b00_01_01_11>(a, 0, a); |
| 19336 | assert_eq_m256i(r, a); |
| 19337 | let r = _mm256_mask_shufflelo_epi16::<0b00_01_01_11>(a, 0b11111111_11111111, a); |
| 19338 | let e = _mm256_set_epi16(0, 1, 2, 3, 7, 6, 6, 4, 8, 9, 10, 11, 15, 14, 14, 12); |
| 19339 | assert_eq_m256i(r, e); |
| 19340 | } |
| 19341 | |
| 19342 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19343 | unsafe fn test_mm256_maskz_shufflelo_epi16() { |
| 19344 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 19345 | let r = _mm256_maskz_shufflelo_epi16::<0b00_01_01_11>(0, a); |
| 19346 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 19347 | let r = _mm256_maskz_shufflelo_epi16::<0b00_01_01_11>(0b11111111_11111111, a); |
| 19348 | let e = _mm256_set_epi16(0, 1, 2, 3, 7, 6, 6, 4, 8, 9, 10, 11, 15, 14, 14, 12); |
| 19349 | assert_eq_m256i(r, e); |
| 19350 | } |
| 19351 | |
| 19352 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19353 | unsafe fn test_mm_mask_shufflelo_epi16() { |
| 19354 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 19355 | let r = _mm_mask_shufflelo_epi16::<0b00_01_01_11>(a, 0, a); |
| 19356 | assert_eq_m128i(r, a); |
| 19357 | let r = _mm_mask_shufflelo_epi16::<0b00_01_01_11>(a, 0b11111111, a); |
| 19358 | let e = _mm_set_epi16(0, 1, 2, 3, 7, 6, 6, 4); |
| 19359 | assert_eq_m128i(r, e); |
| 19360 | } |
| 19361 | |
| 19362 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19363 | unsafe fn test_mm_maskz_shufflelo_epi16() { |
| 19364 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 19365 | let r = _mm_maskz_shufflelo_epi16::<0b00_01_01_11>(0, a); |
| 19366 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 19367 | let r = _mm_maskz_shufflelo_epi16::<0b00_01_01_11>(0b11111111, a); |
| 19368 | let e = _mm_set_epi16(0, 1, 2, 3, 7, 6, 6, 4); |
| 19369 | assert_eq_m128i(r, e); |
| 19370 | } |
| 19371 | |
| 19372 | #[simd_test(enable = "avx512bw" )] |
| 19373 | unsafe fn test_mm512_shufflehi_epi16() { |
| 19374 | #[rustfmt::skip] |
| 19375 | let a = _mm512_set_epi16( |
| 19376 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 19377 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
| 19378 | ); |
| 19379 | #[rustfmt::skip] |
| 19380 | let e = _mm512_set_epi16( |
| 19381 | 3, 2, 2, 0, 4, 5, 6, 7, 11, 10, 10, 8, 12, 13, 14, 15, |
| 19382 | 19, 18, 18, 16, 20, 21, 22, 23, 27, 26, 26, 24, 28, 29, 30, 31, |
| 19383 | ); |
| 19384 | let r = _mm512_shufflehi_epi16::<0b00_01_01_11>(a); |
| 19385 | assert_eq_m512i(r, e); |
| 19386 | } |
| 19387 | |
| 19388 | #[simd_test(enable = "avx512bw" )] |
| 19389 | unsafe fn test_mm512_mask_shufflehi_epi16() { |
| 19390 | #[rustfmt::skip] |
| 19391 | let a = _mm512_set_epi16( |
| 19392 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 19393 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
| 19394 | ); |
| 19395 | let r = _mm512_mask_shufflehi_epi16::<0b00_01_01_11>(a, 0, a); |
| 19396 | assert_eq_m512i(r, a); |
| 19397 | let r = _mm512_mask_shufflehi_epi16::<0b00_01_01_11>( |
| 19398 | a, |
| 19399 | 0b11111111_11111111_11111111_11111111, |
| 19400 | a, |
| 19401 | ); |
| 19402 | #[rustfmt::skip] |
| 19403 | let e = _mm512_set_epi16( |
| 19404 | 3, 2, 2, 0, 4, 5, 6, 7, 11, 10, 10, 8, 12, 13, 14, 15, |
| 19405 | 19, 18, 18, 16, 20, 21, 22, 23, 27, 26, 26, 24, 28, 29, 30, 31, |
| 19406 | ); |
| 19407 | assert_eq_m512i(r, e); |
| 19408 | } |
| 19409 | |
| 19410 | #[simd_test(enable = "avx512bw" )] |
| 19411 | unsafe fn test_mm512_maskz_shufflehi_epi16() { |
| 19412 | #[rustfmt::skip] |
| 19413 | let a = _mm512_set_epi16( |
| 19414 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 19415 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
| 19416 | ); |
| 19417 | let r = _mm512_maskz_shufflehi_epi16::<0b00_01_01_11>(0, a); |
| 19418 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 19419 | let r = |
| 19420 | _mm512_maskz_shufflehi_epi16::<0b00_01_01_11>(0b11111111_11111111_11111111_11111111, a); |
| 19421 | #[rustfmt::skip] |
| 19422 | let e = _mm512_set_epi16( |
| 19423 | 3, 2, 2, 0, 4, 5, 6, 7, 11, 10, 10, 8, 12, 13, 14, 15, |
| 19424 | 19, 18, 18, 16, 20, 21, 22, 23, 27, 26, 26, 24, 28, 29, 30, 31, |
| 19425 | ); |
| 19426 | assert_eq_m512i(r, e); |
| 19427 | } |
| 19428 | |
| 19429 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19430 | unsafe fn test_mm256_mask_shufflehi_epi16() { |
| 19431 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 19432 | let r = _mm256_mask_shufflehi_epi16::<0b00_01_01_11>(a, 0, a); |
| 19433 | assert_eq_m256i(r, a); |
| 19434 | let r = _mm256_mask_shufflehi_epi16::<0b00_01_01_11>(a, 0b11111111_11111111, a); |
| 19435 | let e = _mm256_set_epi16(3, 2, 2, 0, 4, 5, 6, 7, 11, 10, 10, 8, 12, 13, 14, 15); |
| 19436 | assert_eq_m256i(r, e); |
| 19437 | } |
| 19438 | |
| 19439 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19440 | unsafe fn test_mm256_maskz_shufflehi_epi16() { |
| 19441 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 19442 | let r = _mm256_maskz_shufflehi_epi16::<0b00_01_01_11>(0, a); |
| 19443 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 19444 | let r = _mm256_maskz_shufflehi_epi16::<0b00_01_01_11>(0b11111111_11111111, a); |
| 19445 | let e = _mm256_set_epi16(3, 2, 2, 0, 4, 5, 6, 7, 11, 10, 10, 8, 12, 13, 14, 15); |
| 19446 | assert_eq_m256i(r, e); |
| 19447 | } |
| 19448 | |
| 19449 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19450 | unsafe fn test_mm_mask_shufflehi_epi16() { |
| 19451 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 19452 | let r = _mm_mask_shufflehi_epi16::<0b00_01_01_11>(a, 0, a); |
| 19453 | assert_eq_m128i(r, a); |
| 19454 | let r = _mm_mask_shufflehi_epi16::<0b00_01_01_11>(a, 0b11111111, a); |
| 19455 | let e = _mm_set_epi16(3, 2, 2, 0, 4, 5, 6, 7); |
| 19456 | assert_eq_m128i(r, e); |
| 19457 | } |
| 19458 | |
| 19459 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19460 | unsafe fn test_mm_maskz_shufflehi_epi16() { |
| 19461 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
| 19462 | let r = _mm_maskz_shufflehi_epi16::<0b00_01_01_11>(0, a); |
| 19463 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 19464 | let r = _mm_maskz_shufflehi_epi16::<0b00_01_01_11>(0b11111111, a); |
| 19465 | let e = _mm_set_epi16(3, 2, 2, 0, 4, 5, 6, 7); |
| 19466 | assert_eq_m128i(r, e); |
| 19467 | } |
| 19468 | |
| 19469 | #[simd_test(enable = "avx512bw" )] |
| 19470 | unsafe fn test_mm512_shuffle_epi8() { |
| 19471 | #[rustfmt::skip] |
| 19472 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 19473 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
| 19474 | 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, |
| 19475 | 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63); |
| 19476 | let b = _mm512_set1_epi8(1); |
| 19477 | let r = _mm512_shuffle_epi8(a, b); |
| 19478 | #[rustfmt::skip] |
| 19479 | let e = _mm512_set_epi8(14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
| 19480 | 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, |
| 19481 | 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, |
| 19482 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62); |
| 19483 | assert_eq_m512i(r, e); |
| 19484 | } |
| 19485 | |
| 19486 | #[simd_test(enable = "avx512bw" )] |
| 19487 | unsafe fn test_mm512_mask_shuffle_epi8() { |
| 19488 | #[rustfmt::skip] |
| 19489 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 19490 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
| 19491 | 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, |
| 19492 | 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63); |
| 19493 | let b = _mm512_set1_epi8(1); |
| 19494 | let r = _mm512_mask_shuffle_epi8(a, 0, a, b); |
| 19495 | assert_eq_m512i(r, a); |
| 19496 | let r = _mm512_mask_shuffle_epi8( |
| 19497 | a, |
| 19498 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 19499 | a, |
| 19500 | b, |
| 19501 | ); |
| 19502 | #[rustfmt::skip] |
| 19503 | let e = _mm512_set_epi8(14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
| 19504 | 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, |
| 19505 | 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, |
| 19506 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62); |
| 19507 | assert_eq_m512i(r, e); |
| 19508 | } |
| 19509 | |
| 19510 | #[simd_test(enable = "avx512bw" )] |
| 19511 | unsafe fn test_mm512_maskz_shuffle_epi8() { |
| 19512 | #[rustfmt::skip] |
| 19513 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 19514 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
| 19515 | 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, |
| 19516 | 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63); |
| 19517 | let b = _mm512_set1_epi8(1); |
| 19518 | let r = _mm512_maskz_shuffle_epi8(0, a, b); |
| 19519 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 19520 | let r = _mm512_maskz_shuffle_epi8( |
| 19521 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 19522 | a, |
| 19523 | b, |
| 19524 | ); |
| 19525 | #[rustfmt::skip] |
| 19526 | let e = _mm512_set_epi8(14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
| 19527 | 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, |
| 19528 | 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, |
| 19529 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62); |
| 19530 | assert_eq_m512i(r, e); |
| 19531 | } |
| 19532 | |
| 19533 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19534 | unsafe fn test_mm256_mask_shuffle_epi8() { |
| 19535 | #[rustfmt::skip] |
| 19536 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 19537 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
| 19538 | let b = _mm256_set1_epi8(1); |
| 19539 | let r = _mm256_mask_shuffle_epi8(a, 0, a, b); |
| 19540 | assert_eq_m256i(r, a); |
| 19541 | let r = _mm256_mask_shuffle_epi8(a, 0b11111111_11111111_11111111_11111111, a, b); |
| 19542 | #[rustfmt::skip] |
| 19543 | let e = _mm256_set_epi8(14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
| 19544 | 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30); |
| 19545 | assert_eq_m256i(r, e); |
| 19546 | } |
| 19547 | |
| 19548 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19549 | unsafe fn test_mm256_maskz_shuffle_epi8() { |
| 19550 | #[rustfmt::skip] |
| 19551 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 19552 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
| 19553 | let b = _mm256_set1_epi8(1); |
| 19554 | let r = _mm256_maskz_shuffle_epi8(0, a, b); |
| 19555 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 19556 | let r = _mm256_maskz_shuffle_epi8(0b11111111_11111111_11111111_11111111, a, b); |
| 19557 | #[rustfmt::skip] |
| 19558 | let e = _mm256_set_epi8(14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
| 19559 | 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30); |
| 19560 | assert_eq_m256i(r, e); |
| 19561 | } |
| 19562 | |
| 19563 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19564 | unsafe fn test_mm_mask_shuffle_epi8() { |
| 19565 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 19566 | let b = _mm_set1_epi8(1); |
| 19567 | let r = _mm_mask_shuffle_epi8(a, 0, a, b); |
| 19568 | assert_eq_m128i(r, a); |
| 19569 | let r = _mm_mask_shuffle_epi8(a, 0b11111111_11111111, a, b); |
| 19570 | let e = _mm_set_epi8( |
| 19571 | 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
| 19572 | ); |
| 19573 | assert_eq_m128i(r, e); |
| 19574 | } |
| 19575 | |
| 19576 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19577 | unsafe fn test_mm_maskz_shuffle_epi8() { |
| 19578 | #[rustfmt::skip] |
| 19579 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
| 19580 | let b = _mm_set1_epi8(1); |
| 19581 | let r = _mm_maskz_shuffle_epi8(0, a, b); |
| 19582 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 19583 | let r = _mm_maskz_shuffle_epi8(0b11111111_11111111, a, b); |
| 19584 | let e = _mm_set_epi8( |
| 19585 | 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
| 19586 | ); |
| 19587 | assert_eq_m128i(r, e); |
| 19588 | } |
| 19589 | |
| 19590 | #[simd_test(enable = "avx512bw" )] |
| 19591 | unsafe fn test_mm512_test_epi16_mask() { |
| 19592 | let a = _mm512_set1_epi16(1 << 0); |
| 19593 | let b = _mm512_set1_epi16(1 << 0 | 1 << 1); |
| 19594 | let r = _mm512_test_epi16_mask(a, b); |
| 19595 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
| 19596 | assert_eq!(r, e); |
| 19597 | } |
| 19598 | |
| 19599 | #[simd_test(enable = "avx512bw" )] |
| 19600 | unsafe fn test_mm512_mask_test_epi16_mask() { |
| 19601 | let a = _mm512_set1_epi16(1 << 0); |
| 19602 | let b = _mm512_set1_epi16(1 << 0 | 1 << 1); |
| 19603 | let r = _mm512_mask_test_epi16_mask(0, a, b); |
| 19604 | assert_eq!(r, 0); |
| 19605 | let r = _mm512_mask_test_epi16_mask(0b11111111_11111111_11111111_11111111, a, b); |
| 19606 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
| 19607 | assert_eq!(r, e); |
| 19608 | } |
| 19609 | |
| 19610 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19611 | unsafe fn test_mm256_test_epi16_mask() { |
| 19612 | let a = _mm256_set1_epi16(1 << 0); |
| 19613 | let b = _mm256_set1_epi16(1 << 0 | 1 << 1); |
| 19614 | let r = _mm256_test_epi16_mask(a, b); |
| 19615 | let e: __mmask16 = 0b11111111_11111111; |
| 19616 | assert_eq!(r, e); |
| 19617 | } |
| 19618 | |
| 19619 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19620 | unsafe fn test_mm256_mask_test_epi16_mask() { |
| 19621 | let a = _mm256_set1_epi16(1 << 0); |
| 19622 | let b = _mm256_set1_epi16(1 << 0 | 1 << 1); |
| 19623 | let r = _mm256_mask_test_epi16_mask(0, a, b); |
| 19624 | assert_eq!(r, 0); |
| 19625 | let r = _mm256_mask_test_epi16_mask(0b11111111_11111111, a, b); |
| 19626 | let e: __mmask16 = 0b11111111_11111111; |
| 19627 | assert_eq!(r, e); |
| 19628 | } |
| 19629 | |
| 19630 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19631 | unsafe fn test_mm_test_epi16_mask() { |
| 19632 | let a = _mm_set1_epi16(1 << 0); |
| 19633 | let b = _mm_set1_epi16(1 << 0 | 1 << 1); |
| 19634 | let r = _mm_test_epi16_mask(a, b); |
| 19635 | let e: __mmask8 = 0b11111111; |
| 19636 | assert_eq!(r, e); |
| 19637 | } |
| 19638 | |
| 19639 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19640 | unsafe fn test_mm_mask_test_epi16_mask() { |
| 19641 | let a = _mm_set1_epi16(1 << 0); |
| 19642 | let b = _mm_set1_epi16(1 << 0 | 1 << 1); |
| 19643 | let r = _mm_mask_test_epi16_mask(0, a, b); |
| 19644 | assert_eq!(r, 0); |
| 19645 | let r = _mm_mask_test_epi16_mask(0b11111111, a, b); |
| 19646 | let e: __mmask8 = 0b11111111; |
| 19647 | assert_eq!(r, e); |
| 19648 | } |
| 19649 | |
| 19650 | #[simd_test(enable = "avx512bw" )] |
| 19651 | unsafe fn test_mm512_test_epi8_mask() { |
| 19652 | let a = _mm512_set1_epi8(1 << 0); |
| 19653 | let b = _mm512_set1_epi8(1 << 0 | 1 << 1); |
| 19654 | let r = _mm512_test_epi8_mask(a, b); |
| 19655 | let e: __mmask64 = |
| 19656 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111; |
| 19657 | assert_eq!(r, e); |
| 19658 | } |
| 19659 | |
| 19660 | #[simd_test(enable = "avx512bw" )] |
| 19661 | unsafe fn test_mm512_mask_test_epi8_mask() { |
| 19662 | let a = _mm512_set1_epi8(1 << 0); |
| 19663 | let b = _mm512_set1_epi8(1 << 0 | 1 << 1); |
| 19664 | let r = _mm512_mask_test_epi8_mask(0, a, b); |
| 19665 | assert_eq!(r, 0); |
| 19666 | let r = _mm512_mask_test_epi8_mask( |
| 19667 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 19668 | a, |
| 19669 | b, |
| 19670 | ); |
| 19671 | let e: __mmask64 = |
| 19672 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111; |
| 19673 | assert_eq!(r, e); |
| 19674 | } |
| 19675 | |
| 19676 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19677 | unsafe fn test_mm256_test_epi8_mask() { |
| 19678 | let a = _mm256_set1_epi8(1 << 0); |
| 19679 | let b = _mm256_set1_epi8(1 << 0 | 1 << 1); |
| 19680 | let r = _mm256_test_epi8_mask(a, b); |
| 19681 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
| 19682 | assert_eq!(r, e); |
| 19683 | } |
| 19684 | |
| 19685 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19686 | unsafe fn test_mm256_mask_test_epi8_mask() { |
| 19687 | let a = _mm256_set1_epi8(1 << 0); |
| 19688 | let b = _mm256_set1_epi8(1 << 0 | 1 << 1); |
| 19689 | let r = _mm256_mask_test_epi8_mask(0, a, b); |
| 19690 | assert_eq!(r, 0); |
| 19691 | let r = _mm256_mask_test_epi8_mask(0b11111111_11111111_11111111_11111111, a, b); |
| 19692 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
| 19693 | assert_eq!(r, e); |
| 19694 | } |
| 19695 | |
| 19696 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19697 | unsafe fn test_mm_test_epi8_mask() { |
| 19698 | let a = _mm_set1_epi8(1 << 0); |
| 19699 | let b = _mm_set1_epi8(1 << 0 | 1 << 1); |
| 19700 | let r = _mm_test_epi8_mask(a, b); |
| 19701 | let e: __mmask16 = 0b11111111_11111111; |
| 19702 | assert_eq!(r, e); |
| 19703 | } |
| 19704 | |
| 19705 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19706 | unsafe fn test_mm_mask_test_epi8_mask() { |
| 19707 | let a = _mm_set1_epi8(1 << 0); |
| 19708 | let b = _mm_set1_epi8(1 << 0 | 1 << 1); |
| 19709 | let r = _mm_mask_test_epi8_mask(0, a, b); |
| 19710 | assert_eq!(r, 0); |
| 19711 | let r = _mm_mask_test_epi8_mask(0b11111111_11111111, a, b); |
| 19712 | let e: __mmask16 = 0b11111111_11111111; |
| 19713 | assert_eq!(r, e); |
| 19714 | } |
| 19715 | |
| 19716 | #[simd_test(enable = "avx512bw" )] |
| 19717 | unsafe fn test_mm512_testn_epi16_mask() { |
| 19718 | let a = _mm512_set1_epi16(1 << 0); |
| 19719 | let b = _mm512_set1_epi16(1 << 0 | 1 << 1); |
| 19720 | let r = _mm512_testn_epi16_mask(a, b); |
| 19721 | let e: __mmask32 = 0b00000000_00000000_00000000_00000000; |
| 19722 | assert_eq!(r, e); |
| 19723 | } |
| 19724 | |
| 19725 | #[simd_test(enable = "avx512bw" )] |
| 19726 | unsafe fn test_mm512_mask_testn_epi16_mask() { |
| 19727 | let a = _mm512_set1_epi16(1 << 0); |
| 19728 | let b = _mm512_set1_epi16(1 << 0 | 1 << 1); |
| 19729 | let r = _mm512_mask_testn_epi16_mask(0, a, b); |
| 19730 | assert_eq!(r, 0); |
| 19731 | let r = _mm512_mask_testn_epi16_mask(0b11111111_11111111_11111111_11111111, a, b); |
| 19732 | let e: __mmask32 = 0b00000000_00000000_00000000_00000000; |
| 19733 | assert_eq!(r, e); |
| 19734 | } |
| 19735 | |
| 19736 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19737 | unsafe fn test_mm256_testn_epi16_mask() { |
| 19738 | let a = _mm256_set1_epi16(1 << 0); |
| 19739 | let b = _mm256_set1_epi16(1 << 0 | 1 << 1); |
| 19740 | let r = _mm256_testn_epi16_mask(a, b); |
| 19741 | let e: __mmask16 = 0b00000000_00000000; |
| 19742 | assert_eq!(r, e); |
| 19743 | } |
| 19744 | |
| 19745 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19746 | unsafe fn test_mm256_mask_testn_epi16_mask() { |
| 19747 | let a = _mm256_set1_epi16(1 << 0); |
| 19748 | let b = _mm256_set1_epi16(1 << 0 | 1 << 1); |
| 19749 | let r = _mm256_mask_testn_epi16_mask(0, a, b); |
| 19750 | assert_eq!(r, 0); |
| 19751 | let r = _mm256_mask_testn_epi16_mask(0b11111111_11111111, a, b); |
| 19752 | let e: __mmask16 = 0b00000000_00000000; |
| 19753 | assert_eq!(r, e); |
| 19754 | } |
| 19755 | |
| 19756 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19757 | unsafe fn test_mm_testn_epi16_mask() { |
| 19758 | let a = _mm_set1_epi16(1 << 0); |
| 19759 | let b = _mm_set1_epi16(1 << 0 | 1 << 1); |
| 19760 | let r = _mm_testn_epi16_mask(a, b); |
| 19761 | let e: __mmask8 = 0b00000000; |
| 19762 | assert_eq!(r, e); |
| 19763 | } |
| 19764 | |
| 19765 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19766 | unsafe fn test_mm_mask_testn_epi16_mask() { |
| 19767 | let a = _mm_set1_epi16(1 << 0); |
| 19768 | let b = _mm_set1_epi16(1 << 0 | 1 << 1); |
| 19769 | let r = _mm_mask_testn_epi16_mask(0, a, b); |
| 19770 | assert_eq!(r, 0); |
| 19771 | let r = _mm_mask_testn_epi16_mask(0b11111111, a, b); |
| 19772 | let e: __mmask8 = 0b00000000; |
| 19773 | assert_eq!(r, e); |
| 19774 | } |
| 19775 | |
| 19776 | #[simd_test(enable = "avx512bw" )] |
| 19777 | unsafe fn test_mm512_testn_epi8_mask() { |
| 19778 | let a = _mm512_set1_epi8(1 << 0); |
| 19779 | let b = _mm512_set1_epi8(1 << 0 | 1 << 1); |
| 19780 | let r = _mm512_testn_epi8_mask(a, b); |
| 19781 | let e: __mmask64 = |
| 19782 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00000000; |
| 19783 | assert_eq!(r, e); |
| 19784 | } |
| 19785 | |
| 19786 | #[simd_test(enable = "avx512bw" )] |
| 19787 | unsafe fn test_mm512_mask_testn_epi8_mask() { |
| 19788 | let a = _mm512_set1_epi8(1 << 0); |
| 19789 | let b = _mm512_set1_epi8(1 << 0 | 1 << 1); |
| 19790 | let r = _mm512_mask_testn_epi8_mask(0, a, b); |
| 19791 | assert_eq!(r, 0); |
| 19792 | let r = _mm512_mask_testn_epi8_mask( |
| 19793 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 19794 | a, |
| 19795 | b, |
| 19796 | ); |
| 19797 | let e: __mmask64 = |
| 19798 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00000000; |
| 19799 | assert_eq!(r, e); |
| 19800 | } |
| 19801 | |
| 19802 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19803 | unsafe fn test_mm256_testn_epi8_mask() { |
| 19804 | let a = _mm256_set1_epi8(1 << 0); |
| 19805 | let b = _mm256_set1_epi8(1 << 0 | 1 << 1); |
| 19806 | let r = _mm256_testn_epi8_mask(a, b); |
| 19807 | let e: __mmask32 = 0b00000000_00000000_00000000_00000000; |
| 19808 | assert_eq!(r, e); |
| 19809 | } |
| 19810 | |
| 19811 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19812 | unsafe fn test_mm256_mask_testn_epi8_mask() { |
| 19813 | let a = _mm256_set1_epi8(1 << 0); |
| 19814 | let b = _mm256_set1_epi8(1 << 0 | 1 << 1); |
| 19815 | let r = _mm256_mask_testn_epi8_mask(0, a, b); |
| 19816 | assert_eq!(r, 0); |
| 19817 | let r = _mm256_mask_testn_epi8_mask(0b11111111_11111111_11111111_11111111, a, b); |
| 19818 | let e: __mmask32 = 0b00000000_00000000_00000000_00000000; |
| 19819 | assert_eq!(r, e); |
| 19820 | } |
| 19821 | |
| 19822 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19823 | unsafe fn test_mm_testn_epi8_mask() { |
| 19824 | let a = _mm_set1_epi8(1 << 0); |
| 19825 | let b = _mm_set1_epi8(1 << 0 | 1 << 1); |
| 19826 | let r = _mm_testn_epi8_mask(a, b); |
| 19827 | let e: __mmask16 = 0b00000000_00000000; |
| 19828 | assert_eq!(r, e); |
| 19829 | } |
| 19830 | |
| 19831 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19832 | unsafe fn test_mm_mask_testn_epi8_mask() { |
| 19833 | let a = _mm_set1_epi8(1 << 0); |
| 19834 | let b = _mm_set1_epi8(1 << 0 | 1 << 1); |
| 19835 | let r = _mm_mask_testn_epi8_mask(0, a, b); |
| 19836 | assert_eq!(r, 0); |
| 19837 | let r = _mm_mask_testn_epi8_mask(0b11111111_11111111, a, b); |
| 19838 | let e: __mmask16 = 0b00000000_00000000; |
| 19839 | assert_eq!(r, e); |
| 19840 | } |
| 19841 | |
| 19842 | #[simd_test(enable = "avx512bw" )] |
| 19843 | unsafe fn test_store_mask64() { |
| 19844 | let a: __mmask64 = |
| 19845 | 0b11111111_00000000_11111111_00000000_11111111_00000000_11111111_00000000; |
| 19846 | let mut r = 0; |
| 19847 | _store_mask64(&mut r, a); |
| 19848 | assert_eq!(r, a); |
| 19849 | } |
| 19850 | |
| 19851 | #[simd_test(enable = "avx512bw" )] |
| 19852 | unsafe fn test_store_mask32() { |
| 19853 | let a: __mmask32 = 0b11111111_00000000_11111111_00000000; |
| 19854 | let mut r = 0; |
| 19855 | _store_mask32(&mut r, a); |
| 19856 | assert_eq!(r, a); |
| 19857 | } |
| 19858 | |
| 19859 | #[simd_test(enable = "avx512bw" )] |
| 19860 | unsafe fn test_load_mask64() { |
| 19861 | let p: __mmask64 = |
| 19862 | 0b11111111_00000000_11111111_00000000_11111111_00000000_11111111_00000000; |
| 19863 | let r = _load_mask64(&p); |
| 19864 | let e: __mmask64 = |
| 19865 | 0b11111111_00000000_11111111_00000000_11111111_00000000_11111111_00000000; |
| 19866 | assert_eq!(r, e); |
| 19867 | } |
| 19868 | |
| 19869 | #[simd_test(enable = "avx512bw" )] |
| 19870 | unsafe fn test_load_mask32() { |
| 19871 | let p: __mmask32 = 0b11111111_00000000_11111111_00000000; |
| 19872 | let r = _load_mask32(&p); |
| 19873 | let e: __mmask32 = 0b11111111_00000000_11111111_00000000; |
| 19874 | assert_eq!(r, e); |
| 19875 | } |
| 19876 | |
| 19877 | #[simd_test(enable = "avx512bw" )] |
| 19878 | unsafe fn test_mm512_sad_epu8() { |
| 19879 | let a = _mm512_set1_epi8(2); |
| 19880 | let b = _mm512_set1_epi8(4); |
| 19881 | let r = _mm512_sad_epu8(a, b); |
| 19882 | let e = _mm512_set1_epi64(16); |
| 19883 | assert_eq_m512i(r, e); |
| 19884 | } |
| 19885 | |
| 19886 | #[simd_test(enable = "avx512bw" )] |
| 19887 | unsafe fn test_mm512_dbsad_epu8() { |
| 19888 | let a = _mm512_set1_epi8(2); |
| 19889 | let b = _mm512_set1_epi8(4); |
| 19890 | let r = _mm512_dbsad_epu8::<0>(a, b); |
| 19891 | let e = _mm512_set1_epi16(8); |
| 19892 | assert_eq_m512i(r, e); |
| 19893 | } |
| 19894 | |
| 19895 | #[simd_test(enable = "avx512bw" )] |
| 19896 | unsafe fn test_mm512_mask_dbsad_epu8() { |
| 19897 | let src = _mm512_set1_epi16(1); |
| 19898 | let a = _mm512_set1_epi8(2); |
| 19899 | let b = _mm512_set1_epi8(4); |
| 19900 | let r = _mm512_mask_dbsad_epu8::<0>(src, 0, a, b); |
| 19901 | assert_eq_m512i(r, src); |
| 19902 | let r = _mm512_mask_dbsad_epu8::<0>(src, 0b11111111_11111111_11111111_11111111, a, b); |
| 19903 | let e = _mm512_set1_epi16(8); |
| 19904 | assert_eq_m512i(r, e); |
| 19905 | } |
| 19906 | |
| 19907 | #[simd_test(enable = "avx512bw" )] |
| 19908 | unsafe fn test_mm512_maskz_dbsad_epu8() { |
| 19909 | let a = _mm512_set1_epi8(2); |
| 19910 | let b = _mm512_set1_epi8(4); |
| 19911 | let r = _mm512_maskz_dbsad_epu8::<0>(0, a, b); |
| 19912 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 19913 | let r = _mm512_maskz_dbsad_epu8::<0>(0b11111111_11111111_11111111_11111111, a, b); |
| 19914 | let e = _mm512_set1_epi16(8); |
| 19915 | assert_eq_m512i(r, e); |
| 19916 | } |
| 19917 | |
| 19918 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19919 | unsafe fn test_mm256_dbsad_epu8() { |
| 19920 | let a = _mm256_set1_epi8(2); |
| 19921 | let b = _mm256_set1_epi8(4); |
| 19922 | let r = _mm256_dbsad_epu8::<0>(a, b); |
| 19923 | let e = _mm256_set1_epi16(8); |
| 19924 | assert_eq_m256i(r, e); |
| 19925 | } |
| 19926 | |
| 19927 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19928 | unsafe fn test_mm256_mask_dbsad_epu8() { |
| 19929 | let src = _mm256_set1_epi16(1); |
| 19930 | let a = _mm256_set1_epi8(2); |
| 19931 | let b = _mm256_set1_epi8(4); |
| 19932 | let r = _mm256_mask_dbsad_epu8::<0>(src, 0, a, b); |
| 19933 | assert_eq_m256i(r, src); |
| 19934 | let r = _mm256_mask_dbsad_epu8::<0>(src, 0b11111111_11111111, a, b); |
| 19935 | let e = _mm256_set1_epi16(8); |
| 19936 | assert_eq_m256i(r, e); |
| 19937 | } |
| 19938 | |
| 19939 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19940 | unsafe fn test_mm256_maskz_dbsad_epu8() { |
| 19941 | let a = _mm256_set1_epi8(2); |
| 19942 | let b = _mm256_set1_epi8(4); |
| 19943 | let r = _mm256_maskz_dbsad_epu8::<0>(0, a, b); |
| 19944 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 19945 | let r = _mm256_maskz_dbsad_epu8::<0>(0b11111111_11111111, a, b); |
| 19946 | let e = _mm256_set1_epi16(8); |
| 19947 | assert_eq_m256i(r, e); |
| 19948 | } |
| 19949 | |
| 19950 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19951 | unsafe fn test_mm_dbsad_epu8() { |
| 19952 | let a = _mm_set1_epi8(2); |
| 19953 | let b = _mm_set1_epi8(4); |
| 19954 | let r = _mm_dbsad_epu8::<0>(a, b); |
| 19955 | let e = _mm_set1_epi16(8); |
| 19956 | assert_eq_m128i(r, e); |
| 19957 | } |
| 19958 | |
| 19959 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19960 | unsafe fn test_mm_mask_dbsad_epu8() { |
| 19961 | let src = _mm_set1_epi16(1); |
| 19962 | let a = _mm_set1_epi8(2); |
| 19963 | let b = _mm_set1_epi8(4); |
| 19964 | let r = _mm_mask_dbsad_epu8::<0>(src, 0, a, b); |
| 19965 | assert_eq_m128i(r, src); |
| 19966 | let r = _mm_mask_dbsad_epu8::<0>(src, 0b11111111, a, b); |
| 19967 | let e = _mm_set1_epi16(8); |
| 19968 | assert_eq_m128i(r, e); |
| 19969 | } |
| 19970 | |
| 19971 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19972 | unsafe fn test_mm_maskz_dbsad_epu8() { |
| 19973 | let a = _mm_set1_epi8(2); |
| 19974 | let b = _mm_set1_epi8(4); |
| 19975 | let r = _mm_maskz_dbsad_epu8::<0>(0, a, b); |
| 19976 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 19977 | let r = _mm_maskz_dbsad_epu8::<0>(0b11111111, a, b); |
| 19978 | let e = _mm_set1_epi16(8); |
| 19979 | assert_eq_m128i(r, e); |
| 19980 | } |
| 19981 | |
| 19982 | #[simd_test(enable = "avx512bw" )] |
| 19983 | unsafe fn test_mm512_movepi16_mask() { |
| 19984 | let a = _mm512_set1_epi16(1 << 15); |
| 19985 | let r = _mm512_movepi16_mask(a); |
| 19986 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
| 19987 | assert_eq!(r, e); |
| 19988 | } |
| 19989 | |
| 19990 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19991 | unsafe fn test_mm256_movepi16_mask() { |
| 19992 | let a = _mm256_set1_epi16(1 << 15); |
| 19993 | let r = _mm256_movepi16_mask(a); |
| 19994 | let e: __mmask16 = 0b11111111_11111111; |
| 19995 | assert_eq!(r, e); |
| 19996 | } |
| 19997 | |
| 19998 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 19999 | unsafe fn test_mm_movepi16_mask() { |
| 20000 | let a = _mm_set1_epi16(1 << 15); |
| 20001 | let r = _mm_movepi16_mask(a); |
| 20002 | let e: __mmask8 = 0b11111111; |
| 20003 | assert_eq!(r, e); |
| 20004 | } |
| 20005 | |
| 20006 | #[simd_test(enable = "avx512bw" )] |
| 20007 | unsafe fn test_mm512_movepi8_mask() { |
| 20008 | let a = _mm512_set1_epi8(1 << 7); |
| 20009 | let r = _mm512_movepi8_mask(a); |
| 20010 | let e: __mmask64 = |
| 20011 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111; |
| 20012 | assert_eq!(r, e); |
| 20013 | } |
| 20014 | |
| 20015 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20016 | unsafe fn test_mm256_movepi8_mask() { |
| 20017 | let a = _mm256_set1_epi8(1 << 7); |
| 20018 | let r = _mm256_movepi8_mask(a); |
| 20019 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
| 20020 | assert_eq!(r, e); |
| 20021 | } |
| 20022 | |
| 20023 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20024 | unsafe fn test_mm_movepi8_mask() { |
| 20025 | let a = _mm_set1_epi8(1 << 7); |
| 20026 | let r = _mm_movepi8_mask(a); |
| 20027 | let e: __mmask16 = 0b11111111_11111111; |
| 20028 | assert_eq!(r, e); |
| 20029 | } |
| 20030 | |
| 20031 | #[simd_test(enable = "avx512bw" )] |
| 20032 | unsafe fn test_mm512_movm_epi16() { |
| 20033 | let a: __mmask32 = 0b11111111_11111111_11111111_11111111; |
| 20034 | let r = _mm512_movm_epi16(a); |
| 20035 | let e = _mm512_set1_epi16( |
| 20036 | 1 << 15 |
| 20037 | | 1 << 14 |
| 20038 | | 1 << 13 |
| 20039 | | 1 << 12 |
| 20040 | | 1 << 11 |
| 20041 | | 1 << 10 |
| 20042 | | 1 << 9 |
| 20043 | | 1 << 8 |
| 20044 | | 1 << 7 |
| 20045 | | 1 << 6 |
| 20046 | | 1 << 5 |
| 20047 | | 1 << 4 |
| 20048 | | 1 << 3 |
| 20049 | | 1 << 2 |
| 20050 | | 1 << 1 |
| 20051 | | 1 << 0, |
| 20052 | ); |
| 20053 | assert_eq_m512i(r, e); |
| 20054 | } |
| 20055 | |
| 20056 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20057 | unsafe fn test_mm256_movm_epi16() { |
| 20058 | let a: __mmask16 = 0b11111111_11111111; |
| 20059 | let r = _mm256_movm_epi16(a); |
| 20060 | let e = _mm256_set1_epi16( |
| 20061 | 1 << 15 |
| 20062 | | 1 << 14 |
| 20063 | | 1 << 13 |
| 20064 | | 1 << 12 |
| 20065 | | 1 << 11 |
| 20066 | | 1 << 10 |
| 20067 | | 1 << 9 |
| 20068 | | 1 << 8 |
| 20069 | | 1 << 7 |
| 20070 | | 1 << 6 |
| 20071 | | 1 << 5 |
| 20072 | | 1 << 4 |
| 20073 | | 1 << 3 |
| 20074 | | 1 << 2 |
| 20075 | | 1 << 1 |
| 20076 | | 1 << 0, |
| 20077 | ); |
| 20078 | assert_eq_m256i(r, e); |
| 20079 | } |
| 20080 | |
| 20081 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20082 | unsafe fn test_mm_movm_epi16() { |
| 20083 | let a: __mmask8 = 0b11111111; |
| 20084 | let r = _mm_movm_epi16(a); |
| 20085 | let e = _mm_set1_epi16( |
| 20086 | 1 << 15 |
| 20087 | | 1 << 14 |
| 20088 | | 1 << 13 |
| 20089 | | 1 << 12 |
| 20090 | | 1 << 11 |
| 20091 | | 1 << 10 |
| 20092 | | 1 << 9 |
| 20093 | | 1 << 8 |
| 20094 | | 1 << 7 |
| 20095 | | 1 << 6 |
| 20096 | | 1 << 5 |
| 20097 | | 1 << 4 |
| 20098 | | 1 << 3 |
| 20099 | | 1 << 2 |
| 20100 | | 1 << 1 |
| 20101 | | 1 << 0, |
| 20102 | ); |
| 20103 | assert_eq_m128i(r, e); |
| 20104 | } |
| 20105 | |
| 20106 | #[simd_test(enable = "avx512bw" )] |
| 20107 | unsafe fn test_mm512_movm_epi8() { |
| 20108 | let a: __mmask64 = |
| 20109 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111; |
| 20110 | let r = _mm512_movm_epi8(a); |
| 20111 | let e = |
| 20112 | _mm512_set1_epi8(1 << 7 | 1 << 6 | 1 << 5 | 1 << 4 | 1 << 3 | 1 << 2 | 1 << 1 | 1 << 0); |
| 20113 | assert_eq_m512i(r, e); |
| 20114 | } |
| 20115 | |
| 20116 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20117 | unsafe fn test_mm256_movm_epi8() { |
| 20118 | let a: __mmask32 = 0b11111111_11111111_11111111_11111111; |
| 20119 | let r = _mm256_movm_epi8(a); |
| 20120 | let e = |
| 20121 | _mm256_set1_epi8(1 << 7 | 1 << 6 | 1 << 5 | 1 << 4 | 1 << 3 | 1 << 2 | 1 << 1 | 1 << 0); |
| 20122 | assert_eq_m256i(r, e); |
| 20123 | } |
| 20124 | |
| 20125 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20126 | unsafe fn test_mm_movm_epi8() { |
| 20127 | let a: __mmask16 = 0b11111111_11111111; |
| 20128 | let r = _mm_movm_epi8(a); |
| 20129 | let e = |
| 20130 | _mm_set1_epi8(1 << 7 | 1 << 6 | 1 << 5 | 1 << 4 | 1 << 3 | 1 << 2 | 1 << 1 | 1 << 0); |
| 20131 | assert_eq_m128i(r, e); |
| 20132 | } |
| 20133 | |
| 20134 | #[simd_test(enable = "avx512bw" )] |
| 20135 | unsafe fn test_cvtmask32_u32() { |
| 20136 | let a: __mmask32 = 0b11001100_00110011_01100110_10011001; |
| 20137 | let r = _cvtmask32_u32(a); |
| 20138 | let e: u32 = 0b11001100_00110011_01100110_10011001; |
| 20139 | assert_eq!(r, e); |
| 20140 | } |
| 20141 | |
| 20142 | #[simd_test(enable = "avx512bw" )] |
| 20143 | unsafe fn test_cvtu32_mask32() { |
| 20144 | let a: u32 = 0b11001100_00110011_01100110_10011001; |
| 20145 | let r = _cvtu32_mask32(a); |
| 20146 | let e: __mmask32 = 0b11001100_00110011_01100110_10011001; |
| 20147 | assert_eq!(r, e); |
| 20148 | } |
| 20149 | |
| 20150 | #[simd_test(enable = "avx512bw" )] |
| 20151 | unsafe fn test_kadd_mask32() { |
| 20152 | let a: __mmask32 = 11; |
| 20153 | let b: __mmask32 = 22; |
| 20154 | let r = _kadd_mask32(a, b); |
| 20155 | let e: __mmask32 = 33; |
| 20156 | assert_eq!(r, e); |
| 20157 | } |
| 20158 | |
| 20159 | #[simd_test(enable = "avx512bw" )] |
| 20160 | unsafe fn test_kadd_mask64() { |
| 20161 | let a: __mmask64 = 11; |
| 20162 | let b: __mmask64 = 22; |
| 20163 | let r = _kadd_mask64(a, b); |
| 20164 | let e: __mmask64 = 33; |
| 20165 | assert_eq!(r, e); |
| 20166 | } |
| 20167 | |
| 20168 | #[simd_test(enable = "avx512bw" )] |
| 20169 | unsafe fn test_kand_mask32() { |
| 20170 | let a: __mmask32 = 0b11001100_00110011_11001100_00110011; |
| 20171 | let b: __mmask32 = 0b11001100_00110011_11001100_00110011; |
| 20172 | let r = _kand_mask32(a, b); |
| 20173 | let e: __mmask32 = 0b11001100_00110011_11001100_00110011; |
| 20174 | assert_eq!(r, e); |
| 20175 | } |
| 20176 | |
| 20177 | #[simd_test(enable = "avx512bw" )] |
| 20178 | unsafe fn test_kand_mask64() { |
| 20179 | let a: __mmask64 = |
| 20180 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
| 20181 | let b: __mmask64 = |
| 20182 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
| 20183 | let r = _kand_mask64(a, b); |
| 20184 | let e: __mmask64 = |
| 20185 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
| 20186 | assert_eq!(r, e); |
| 20187 | } |
| 20188 | |
| 20189 | #[simd_test(enable = "avx512bw" )] |
| 20190 | unsafe fn test_knot_mask32() { |
| 20191 | let a: __mmask32 = 0b11001100_00110011_11001100_00110011; |
| 20192 | let r = _knot_mask32(a); |
| 20193 | let e: __mmask32 = 0b00110011_11001100_00110011_11001100; |
| 20194 | assert_eq!(r, e); |
| 20195 | } |
| 20196 | |
| 20197 | #[simd_test(enable = "avx512bw" )] |
| 20198 | unsafe fn test_knot_mask64() { |
| 20199 | let a: __mmask64 = |
| 20200 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
| 20201 | let r = _knot_mask64(a); |
| 20202 | let e: __mmask64 = |
| 20203 | 0b00110011_11001100_00110011_11001100_00110011_11001100_00110011_11001100; |
| 20204 | assert_eq!(r, e); |
| 20205 | } |
| 20206 | |
| 20207 | #[simd_test(enable = "avx512bw" )] |
| 20208 | unsafe fn test_kandn_mask32() { |
| 20209 | let a: __mmask32 = 0b11001100_00110011_11001100_00110011; |
| 20210 | let b: __mmask32 = 0b11001100_00110011_11001100_00110011; |
| 20211 | let r = _kandn_mask32(a, b); |
| 20212 | let e: __mmask32 = 0b00000000_00000000_00000000_00000000; |
| 20213 | assert_eq!(r, e); |
| 20214 | } |
| 20215 | |
| 20216 | #[simd_test(enable = "avx512bw" )] |
| 20217 | unsafe fn test_kandn_mask64() { |
| 20218 | let a: __mmask64 = |
| 20219 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
| 20220 | let b: __mmask64 = |
| 20221 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
| 20222 | let r = _kandn_mask64(a, b); |
| 20223 | let e: __mmask64 = |
| 20224 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00000000; |
| 20225 | assert_eq!(r, e); |
| 20226 | } |
| 20227 | |
| 20228 | #[simd_test(enable = "avx512bw" )] |
| 20229 | unsafe fn test_kor_mask32() { |
| 20230 | let a: __mmask32 = 0b00110011_11001100_00110011_11001100; |
| 20231 | let b: __mmask32 = 0b11001100_00110011_11001100_00110011; |
| 20232 | let r = _kor_mask32(a, b); |
| 20233 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
| 20234 | assert_eq!(r, e); |
| 20235 | } |
| 20236 | |
| 20237 | #[simd_test(enable = "avx512bw" )] |
| 20238 | unsafe fn test_kor_mask64() { |
| 20239 | let a: __mmask64 = |
| 20240 | 0b00110011_11001100_00110011_11001100_00110011_11001100_00110011_11001100; |
| 20241 | let b: __mmask64 = |
| 20242 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
| 20243 | let r = _kor_mask64(a, b); |
| 20244 | let e: __mmask64 = |
| 20245 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111; |
| 20246 | assert_eq!(r, e); |
| 20247 | } |
| 20248 | |
| 20249 | #[simd_test(enable = "avx512bw" )] |
| 20250 | unsafe fn test_kxor_mask32() { |
| 20251 | let a: __mmask32 = 0b00110011_11001100_00110011_11001100; |
| 20252 | let b: __mmask32 = 0b11001100_00110011_11001100_00110011; |
| 20253 | let r = _kxor_mask32(a, b); |
| 20254 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
| 20255 | assert_eq!(r, e); |
| 20256 | } |
| 20257 | |
| 20258 | #[simd_test(enable = "avx512bw" )] |
| 20259 | unsafe fn test_kxor_mask64() { |
| 20260 | let a: __mmask64 = |
| 20261 | 0b00110011_11001100_00110011_11001100_00110011_11001100_00110011_11001100; |
| 20262 | let b: __mmask64 = |
| 20263 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
| 20264 | let r = _kxor_mask64(a, b); |
| 20265 | let e: __mmask64 = |
| 20266 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111; |
| 20267 | assert_eq!(r, e); |
| 20268 | } |
| 20269 | |
| 20270 | #[simd_test(enable = "avx512bw" )] |
| 20271 | unsafe fn test_kxnor_mask32() { |
| 20272 | let a: __mmask32 = 0b00110011_11001100_00110011_11001100; |
| 20273 | let b: __mmask32 = 0b11001100_00110011_11001100_00110011; |
| 20274 | let r = _kxnor_mask32(a, b); |
| 20275 | let e: __mmask32 = 0b00000000_00000000_00000000_00000000; |
| 20276 | assert_eq!(r, e); |
| 20277 | } |
| 20278 | |
| 20279 | #[simd_test(enable = "avx512bw" )] |
| 20280 | unsafe fn test_kxnor_mask64() { |
| 20281 | let a: __mmask64 = |
| 20282 | 0b00110011_11001100_00110011_11001100_00110011_11001100_00110011_11001100; |
| 20283 | let b: __mmask64 = |
| 20284 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
| 20285 | let r = _kxnor_mask64(a, b); |
| 20286 | let e: __mmask64 = |
| 20287 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00000000; |
| 20288 | assert_eq!(r, e); |
| 20289 | } |
| 20290 | |
| 20291 | #[simd_test(enable = "avx512bw" )] |
| 20292 | unsafe fn test_kortest_mask32_u8() { |
| 20293 | let a: __mmask32 = 0b0110100101101001_0110100101101001; |
| 20294 | let b: __mmask32 = 0b1011011010110110_1011011010110110; |
| 20295 | let mut all_ones: u8 = 0; |
| 20296 | let r = _kortest_mask32_u8(a, b, &mut all_ones); |
| 20297 | assert_eq!(r, 0); |
| 20298 | assert_eq!(all_ones, 1); |
| 20299 | } |
| 20300 | |
| 20301 | #[simd_test(enable = "avx512bw" )] |
| 20302 | unsafe fn test_kortest_mask64_u8() { |
| 20303 | let a: __mmask64 = 0b0110100101101001_0110100101101001; |
| 20304 | let b: __mmask64 = 0b1011011010110110_1011011010110110; |
| 20305 | let mut all_ones: u8 = 0; |
| 20306 | let r = _kortest_mask64_u8(a, b, &mut all_ones); |
| 20307 | assert_eq!(r, 0); |
| 20308 | assert_eq!(all_ones, 0); |
| 20309 | } |
| 20310 | |
| 20311 | #[simd_test(enable = "avx512bw" )] |
| 20312 | unsafe fn test_kortestc_mask32_u8() { |
| 20313 | let a: __mmask32 = 0b0110100101101001_0110100101101001; |
| 20314 | let b: __mmask32 = 0b1011011010110110_1011011010110110; |
| 20315 | let r = _kortestc_mask32_u8(a, b); |
| 20316 | assert_eq!(r, 1); |
| 20317 | } |
| 20318 | |
| 20319 | #[simd_test(enable = "avx512bw" )] |
| 20320 | unsafe fn test_kortestc_mask64_u8() { |
| 20321 | let a: __mmask64 = 0b0110100101101001_0110100101101001; |
| 20322 | let b: __mmask64 = 0b1011011010110110_1011011010110110; |
| 20323 | let r = _kortestc_mask64_u8(a, b); |
| 20324 | assert_eq!(r, 0); |
| 20325 | } |
| 20326 | |
| 20327 | #[simd_test(enable = "avx512bw" )] |
| 20328 | unsafe fn test_kortestz_mask32_u8() { |
| 20329 | let a: __mmask32 = 0b0110100101101001_0110100101101001; |
| 20330 | let b: __mmask32 = 0b1011011010110110_1011011010110110; |
| 20331 | let r = _kortestz_mask32_u8(a, b); |
| 20332 | assert_eq!(r, 0); |
| 20333 | } |
| 20334 | |
| 20335 | #[simd_test(enable = "avx512bw" )] |
| 20336 | unsafe fn test_kortestz_mask64_u8() { |
| 20337 | let a: __mmask64 = 0b0110100101101001_0110100101101001; |
| 20338 | let b: __mmask64 = 0b1011011010110110_1011011010110110; |
| 20339 | let r = _kortestz_mask64_u8(a, b); |
| 20340 | assert_eq!(r, 0); |
| 20341 | } |
| 20342 | |
| 20343 | #[simd_test(enable = "avx512bw" )] |
| 20344 | unsafe fn test_kshiftli_mask32() { |
| 20345 | let a: __mmask32 = 0b0110100101101001_0110100101101001; |
| 20346 | let r = _kshiftli_mask32::<3>(a); |
| 20347 | let e: __mmask32 = 0b0100101101001011_0100101101001000; |
| 20348 | assert_eq!(r, e); |
| 20349 | |
| 20350 | let r = _kshiftli_mask32::<31>(a); |
| 20351 | let e: __mmask32 = 0b1000000000000000_0000000000000000; |
| 20352 | assert_eq!(r, e); |
| 20353 | |
| 20354 | let r = _kshiftli_mask32::<32>(a); |
| 20355 | let e: __mmask32 = 0b0000000000000000_0000000000000000; |
| 20356 | assert_eq!(r, e); |
| 20357 | |
| 20358 | let r = _kshiftli_mask32::<33>(a); |
| 20359 | let e: __mmask32 = 0b0000000000000000_0000000000000000; |
| 20360 | assert_eq!(r, e); |
| 20361 | } |
| 20362 | |
| 20363 | #[simd_test(enable = "avx512bw" )] |
| 20364 | unsafe fn test_kshiftli_mask64() { |
| 20365 | let a: __mmask64 = 0b0110100101101001_0110100101101001; |
| 20366 | let r = _kshiftli_mask64::<3>(a); |
| 20367 | let e: __mmask64 = 0b0110100101101001011_0100101101001000; |
| 20368 | assert_eq!(r, e); |
| 20369 | |
| 20370 | let r = _kshiftli_mask64::<63>(a); |
| 20371 | let e: __mmask64 = 0b1000000000000000_0000000000000000_0000000000000000_0000000000000000; |
| 20372 | assert_eq!(r, e); |
| 20373 | |
| 20374 | let r = _kshiftli_mask64::<64>(a); |
| 20375 | let e: __mmask64 = 0b0000000000000000_0000000000000000_0000000000000000_0000000000000000; |
| 20376 | assert_eq!(r, e); |
| 20377 | |
| 20378 | let r = _kshiftli_mask64::<65>(a); |
| 20379 | let e: __mmask64 = 0b0000000000000000_0000000000000000_0000000000000000_0000000000000000; |
| 20380 | assert_eq!(r, e); |
| 20381 | } |
| 20382 | |
| 20383 | #[simd_test(enable = "avx512bw" )] |
| 20384 | unsafe fn test_kshiftri_mask32() { |
| 20385 | let a: __mmask32 = 0b1010100101101001_0110100101101001; |
| 20386 | let r = _kshiftri_mask32::<3>(a); |
| 20387 | let e: __mmask32 = 0b0001010100101101_0010110100101101; |
| 20388 | assert_eq!(r, e); |
| 20389 | |
| 20390 | let r = _kshiftri_mask32::<31>(a); |
| 20391 | let e: __mmask32 = 0b0000000000000000_0000000000000001; |
| 20392 | assert_eq!(r, e); |
| 20393 | |
| 20394 | let r = _kshiftri_mask32::<32>(a); |
| 20395 | let e: __mmask32 = 0b0000000000000000_0000000000000000; |
| 20396 | assert_eq!(r, e); |
| 20397 | |
| 20398 | let r = _kshiftri_mask32::<33>(a); |
| 20399 | let e: __mmask32 = 0b0000000000000000_0000000000000000; |
| 20400 | assert_eq!(r, e); |
| 20401 | } |
| 20402 | |
| 20403 | #[simd_test(enable = "avx512bw" )] |
| 20404 | unsafe fn test_kshiftri_mask64() { |
| 20405 | let a: __mmask64 = 0b1010100101101001011_0100101101001000; |
| 20406 | let r = _kshiftri_mask64::<3>(a); |
| 20407 | let e: __mmask64 = 0b1010100101101001_0110100101101001; |
| 20408 | assert_eq!(r, e); |
| 20409 | |
| 20410 | let r = _kshiftri_mask64::<34>(a); |
| 20411 | let e: __mmask64 = 0b0000000000000000_0000000000000000_0000000000000000_0000000000000001; |
| 20412 | assert_eq!(r, e); |
| 20413 | |
| 20414 | let r = _kshiftri_mask64::<35>(a); |
| 20415 | let e: __mmask64 = 0b0000000000000000_0000000000000000_0000000000000000_0000000000000000; |
| 20416 | assert_eq!(r, e); |
| 20417 | |
| 20418 | let r = _kshiftri_mask64::<64>(a); |
| 20419 | let e: __mmask64 = 0b0000000000000000_0000000000000000_0000000000000000_0000000000000000; |
| 20420 | assert_eq!(r, e); |
| 20421 | |
| 20422 | let r = _kshiftri_mask64::<65>(a); |
| 20423 | let e: __mmask64 = 0b0000000000000000_0000000000000000_0000000000000000_0000000000000000; |
| 20424 | assert_eq!(r, e); |
| 20425 | } |
| 20426 | |
| 20427 | #[simd_test(enable = "avx512bw" )] |
| 20428 | unsafe fn test_ktest_mask32_u8() { |
| 20429 | let a: __mmask32 = 0b0110100100111100_0110100100111100; |
| 20430 | let b: __mmask32 = 0b1001011011000011_1001011011000011; |
| 20431 | let mut and_not: u8 = 0; |
| 20432 | let r = _ktest_mask32_u8(a, b, &mut and_not); |
| 20433 | assert_eq!(r, 1); |
| 20434 | assert_eq!(and_not, 0); |
| 20435 | } |
| 20436 | |
| 20437 | #[simd_test(enable = "avx512bw" )] |
| 20438 | unsafe fn test_ktestc_mask32_u8() { |
| 20439 | let a: __mmask32 = 0b0110100100111100_0110100100111100; |
| 20440 | let b: __mmask32 = 0b1001011011000011_1001011011000011; |
| 20441 | let r = _ktestc_mask32_u8(a, b); |
| 20442 | assert_eq!(r, 0); |
| 20443 | } |
| 20444 | |
| 20445 | #[simd_test(enable = "avx512bw" )] |
| 20446 | unsafe fn test_ktestz_mask32_u8() { |
| 20447 | let a: __mmask32 = 0b0110100100111100_0110100100111100; |
| 20448 | let b: __mmask32 = 0b1001011011000011_1001011011000011; |
| 20449 | let r = _ktestz_mask32_u8(a, b); |
| 20450 | assert_eq!(r, 1); |
| 20451 | } |
| 20452 | |
| 20453 | #[simd_test(enable = "avx512bw" )] |
| 20454 | unsafe fn test_ktest_mask64_u8() { |
| 20455 | let a: __mmask64 = 0b0110100100111100_0110100100111100; |
| 20456 | let b: __mmask64 = 0b1001011011000011_1001011011000011; |
| 20457 | let mut and_not: u8 = 0; |
| 20458 | let r = _ktest_mask64_u8(a, b, &mut and_not); |
| 20459 | assert_eq!(r, 1); |
| 20460 | assert_eq!(and_not, 0); |
| 20461 | } |
| 20462 | |
| 20463 | #[simd_test(enable = "avx512bw" )] |
| 20464 | unsafe fn test_ktestc_mask64_u8() { |
| 20465 | let a: __mmask64 = 0b0110100100111100_0110100100111100; |
| 20466 | let b: __mmask64 = 0b1001011011000011_1001011011000011; |
| 20467 | let r = _ktestc_mask64_u8(a, b); |
| 20468 | assert_eq!(r, 0); |
| 20469 | } |
| 20470 | |
| 20471 | #[simd_test(enable = "avx512bw" )] |
| 20472 | unsafe fn test_ktestz_mask64_u8() { |
| 20473 | let a: __mmask64 = 0b0110100100111100_0110100100111100; |
| 20474 | let b: __mmask64 = 0b1001011011000011_1001011011000011; |
| 20475 | let r = _ktestz_mask64_u8(a, b); |
| 20476 | assert_eq!(r, 1); |
| 20477 | } |
| 20478 | |
| 20479 | #[simd_test(enable = "avx512bw" )] |
| 20480 | unsafe fn test_mm512_kunpackw() { |
| 20481 | let a: u32 = 0x00110011; |
| 20482 | let b: u32 = 0x00001011; |
| 20483 | let r = _mm512_kunpackw(a, b); |
| 20484 | let e: u32 = 0x00111011; |
| 20485 | assert_eq!(r, e); |
| 20486 | } |
| 20487 | |
| 20488 | #[simd_test(enable = "avx512bw" )] |
| 20489 | unsafe fn test_mm512_kunpackd() { |
| 20490 | let a: u64 = 0x11001100_00110011; |
| 20491 | let b: u64 = 0x00101110_00001011; |
| 20492 | let r = _mm512_kunpackd(a, b); |
| 20493 | let e: u64 = 0x00110011_00001011; |
| 20494 | assert_eq!(r, e); |
| 20495 | } |
| 20496 | |
| 20497 | #[simd_test(enable = "avx512bw" )] |
| 20498 | unsafe fn test_mm512_cvtepi16_epi8() { |
| 20499 | let a = _mm512_set1_epi16(2); |
| 20500 | let r = _mm512_cvtepi16_epi8(a); |
| 20501 | let e = _mm256_set1_epi8(2); |
| 20502 | assert_eq_m256i(r, e); |
| 20503 | } |
| 20504 | |
| 20505 | #[simd_test(enable = "avx512bw" )] |
| 20506 | unsafe fn test_mm512_mask_cvtepi16_epi8() { |
| 20507 | let src = _mm256_set1_epi8(1); |
| 20508 | let a = _mm512_set1_epi16(2); |
| 20509 | let r = _mm512_mask_cvtepi16_epi8(src, 0, a); |
| 20510 | assert_eq_m256i(r, src); |
| 20511 | let r = _mm512_mask_cvtepi16_epi8(src, 0b11111111_11111111_11111111_11111111, a); |
| 20512 | let e = _mm256_set1_epi8(2); |
| 20513 | assert_eq_m256i(r, e); |
| 20514 | } |
| 20515 | |
| 20516 | #[simd_test(enable = "avx512bw" )] |
| 20517 | unsafe fn test_mm512_maskz_cvtepi16_epi8() { |
| 20518 | let a = _mm512_set1_epi16(2); |
| 20519 | let r = _mm512_maskz_cvtepi16_epi8(0, a); |
| 20520 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 20521 | let r = _mm512_maskz_cvtepi16_epi8(0b11111111_11111111_11111111_11111111, a); |
| 20522 | let e = _mm256_set1_epi8(2); |
| 20523 | assert_eq_m256i(r, e); |
| 20524 | } |
| 20525 | |
| 20526 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20527 | unsafe fn test_mm256_cvtepi16_epi8() { |
| 20528 | let a = _mm256_set1_epi16(2); |
| 20529 | let r = _mm256_cvtepi16_epi8(a); |
| 20530 | let e = _mm_set1_epi8(2); |
| 20531 | assert_eq_m128i(r, e); |
| 20532 | } |
| 20533 | |
| 20534 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20535 | unsafe fn test_mm256_mask_cvtepi16_epi8() { |
| 20536 | let src = _mm_set1_epi8(1); |
| 20537 | let a = _mm256_set1_epi16(2); |
| 20538 | let r = _mm256_mask_cvtepi16_epi8(src, 0, a); |
| 20539 | assert_eq_m128i(r, src); |
| 20540 | let r = _mm256_mask_cvtepi16_epi8(src, 0b11111111_11111111, a); |
| 20541 | let e = _mm_set1_epi8(2); |
| 20542 | assert_eq_m128i(r, e); |
| 20543 | } |
| 20544 | |
| 20545 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20546 | unsafe fn test_mm256_maskz_cvtepi16_epi8() { |
| 20547 | let a = _mm256_set1_epi16(2); |
| 20548 | let r = _mm256_maskz_cvtepi16_epi8(0, a); |
| 20549 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 20550 | let r = _mm256_maskz_cvtepi16_epi8(0b11111111_11111111, a); |
| 20551 | let e = _mm_set1_epi8(2); |
| 20552 | assert_eq_m128i(r, e); |
| 20553 | } |
| 20554 | |
| 20555 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20556 | unsafe fn test_mm_cvtepi16_epi8() { |
| 20557 | let a = _mm_set1_epi16(2); |
| 20558 | let r = _mm_cvtepi16_epi8(a); |
| 20559 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2); |
| 20560 | assert_eq_m128i(r, e); |
| 20561 | } |
| 20562 | |
| 20563 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20564 | unsafe fn test_mm_mask_cvtepi16_epi8() { |
| 20565 | let src = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
| 20566 | let a = _mm_set1_epi16(2); |
| 20567 | let r = _mm_mask_cvtepi16_epi8(src, 0, a); |
| 20568 | assert_eq_m128i(r, src); |
| 20569 | let r = _mm_mask_cvtepi16_epi8(src, 0b11111111, a); |
| 20570 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2); |
| 20571 | assert_eq_m128i(r, e); |
| 20572 | } |
| 20573 | |
| 20574 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20575 | unsafe fn test_mm_maskz_cvtepi16_epi8() { |
| 20576 | let a = _mm_set1_epi16(2); |
| 20577 | let r = _mm_maskz_cvtepi16_epi8(0, a); |
| 20578 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 20579 | let r = _mm_maskz_cvtepi16_epi8(0b11111111, a); |
| 20580 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2); |
| 20581 | assert_eq_m128i(r, e); |
| 20582 | } |
| 20583 | |
| 20584 | #[simd_test(enable = "avx512bw" )] |
| 20585 | unsafe fn test_mm512_cvtsepi16_epi8() { |
| 20586 | let a = _mm512_set1_epi16(i16::MAX); |
| 20587 | let r = _mm512_cvtsepi16_epi8(a); |
| 20588 | let e = _mm256_set1_epi8(i8::MAX); |
| 20589 | assert_eq_m256i(r, e); |
| 20590 | } |
| 20591 | |
| 20592 | #[simd_test(enable = "avx512bw" )] |
| 20593 | unsafe fn test_mm512_mask_cvtsepi16_epi8() { |
| 20594 | let src = _mm256_set1_epi8(1); |
| 20595 | let a = _mm512_set1_epi16(i16::MAX); |
| 20596 | let r = _mm512_mask_cvtsepi16_epi8(src, 0, a); |
| 20597 | assert_eq_m256i(r, src); |
| 20598 | let r = _mm512_mask_cvtsepi16_epi8(src, 0b11111111_11111111_11111111_11111111, a); |
| 20599 | let e = _mm256_set1_epi8(i8::MAX); |
| 20600 | assert_eq_m256i(r, e); |
| 20601 | } |
| 20602 | |
| 20603 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20604 | unsafe fn test_mm256_cvtsepi16_epi8() { |
| 20605 | let a = _mm256_set1_epi16(i16::MAX); |
| 20606 | let r = _mm256_cvtsepi16_epi8(a); |
| 20607 | let e = _mm_set1_epi8(i8::MAX); |
| 20608 | assert_eq_m128i(r, e); |
| 20609 | } |
| 20610 | |
| 20611 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20612 | unsafe fn test_mm256_mask_cvtsepi16_epi8() { |
| 20613 | let src = _mm_set1_epi8(1); |
| 20614 | let a = _mm256_set1_epi16(i16::MAX); |
| 20615 | let r = _mm256_mask_cvtsepi16_epi8(src, 0, a); |
| 20616 | assert_eq_m128i(r, src); |
| 20617 | let r = _mm256_mask_cvtsepi16_epi8(src, 0b11111111_11111111, a); |
| 20618 | let e = _mm_set1_epi8(i8::MAX); |
| 20619 | assert_eq_m128i(r, e); |
| 20620 | } |
| 20621 | |
| 20622 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20623 | unsafe fn test_mm256_maskz_cvtsepi16_epi8() { |
| 20624 | let a = _mm256_set1_epi16(i16::MAX); |
| 20625 | let r = _mm256_maskz_cvtsepi16_epi8(0, a); |
| 20626 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 20627 | let r = _mm256_maskz_cvtsepi16_epi8(0b11111111_11111111, a); |
| 20628 | let e = _mm_set1_epi8(i8::MAX); |
| 20629 | assert_eq_m128i(r, e); |
| 20630 | } |
| 20631 | |
| 20632 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20633 | unsafe fn test_mm_cvtsepi16_epi8() { |
| 20634 | let a = _mm_set1_epi16(i16::MAX); |
| 20635 | let r = _mm_cvtsepi16_epi8(a); |
| 20636 | #[rustfmt::skip] |
| 20637 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 20638 | assert_eq_m128i(r, e); |
| 20639 | } |
| 20640 | |
| 20641 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20642 | unsafe fn test_mm_mask_cvtsepi16_epi8() { |
| 20643 | let src = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
| 20644 | let a = _mm_set1_epi16(i16::MAX); |
| 20645 | let r = _mm_mask_cvtsepi16_epi8(src, 0, a); |
| 20646 | assert_eq_m128i(r, src); |
| 20647 | let r = _mm_mask_cvtsepi16_epi8(src, 0b11111111, a); |
| 20648 | #[rustfmt::skip] |
| 20649 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 20650 | assert_eq_m128i(r, e); |
| 20651 | } |
| 20652 | |
| 20653 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20654 | unsafe fn test_mm_maskz_cvtsepi16_epi8() { |
| 20655 | let a = _mm_set1_epi16(i16::MAX); |
| 20656 | let r = _mm_maskz_cvtsepi16_epi8(0, a); |
| 20657 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 20658 | let r = _mm_maskz_cvtsepi16_epi8(0b11111111, a); |
| 20659 | #[rustfmt::skip] |
| 20660 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
| 20661 | assert_eq_m128i(r, e); |
| 20662 | } |
| 20663 | |
| 20664 | #[simd_test(enable = "avx512bw" )] |
| 20665 | unsafe fn test_mm512_maskz_cvtsepi16_epi8() { |
| 20666 | let a = _mm512_set1_epi16(i16::MAX); |
| 20667 | let r = _mm512_maskz_cvtsepi16_epi8(0, a); |
| 20668 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 20669 | let r = _mm512_maskz_cvtsepi16_epi8(0b11111111_11111111_11111111_11111111, a); |
| 20670 | let e = _mm256_set1_epi8(i8::MAX); |
| 20671 | assert_eq_m256i(r, e); |
| 20672 | } |
| 20673 | |
| 20674 | #[simd_test(enable = "avx512bw" )] |
| 20675 | unsafe fn test_mm512_cvtusepi16_epi8() { |
| 20676 | let a = _mm512_set1_epi16(i16::MIN); |
| 20677 | let r = _mm512_cvtusepi16_epi8(a); |
| 20678 | let e = _mm256_set1_epi8(-1); |
| 20679 | assert_eq_m256i(r, e); |
| 20680 | } |
| 20681 | |
| 20682 | #[simd_test(enable = "avx512bw" )] |
| 20683 | unsafe fn test_mm512_mask_cvtusepi16_epi8() { |
| 20684 | let src = _mm256_set1_epi8(1); |
| 20685 | let a = _mm512_set1_epi16(i16::MIN); |
| 20686 | let r = _mm512_mask_cvtusepi16_epi8(src, 0, a); |
| 20687 | assert_eq_m256i(r, src); |
| 20688 | let r = _mm512_mask_cvtusepi16_epi8(src, 0b11111111_11111111_11111111_11111111, a); |
| 20689 | let e = _mm256_set1_epi8(-1); |
| 20690 | assert_eq_m256i(r, e); |
| 20691 | } |
| 20692 | |
| 20693 | #[simd_test(enable = "avx512bw" )] |
| 20694 | unsafe fn test_mm512_maskz_cvtusepi16_epi8() { |
| 20695 | let a = _mm512_set1_epi16(i16::MIN); |
| 20696 | let r = _mm512_maskz_cvtusepi16_epi8(0, a); |
| 20697 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 20698 | let r = _mm512_maskz_cvtusepi16_epi8(0b11111111_11111111_11111111_11111111, a); |
| 20699 | let e = _mm256_set1_epi8(-1); |
| 20700 | assert_eq_m256i(r, e); |
| 20701 | } |
| 20702 | |
| 20703 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20704 | unsafe fn test_mm256_cvtusepi16_epi8() { |
| 20705 | let a = _mm256_set1_epi16(i16::MIN); |
| 20706 | let r = _mm256_cvtusepi16_epi8(a); |
| 20707 | let e = _mm_set1_epi8(-1); |
| 20708 | assert_eq_m128i(r, e); |
| 20709 | } |
| 20710 | |
| 20711 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20712 | unsafe fn test_mm256_mask_cvtusepi16_epi8() { |
| 20713 | let src = _mm_set1_epi8(1); |
| 20714 | let a = _mm256_set1_epi16(i16::MIN); |
| 20715 | let r = _mm256_mask_cvtusepi16_epi8(src, 0, a); |
| 20716 | assert_eq_m128i(r, src); |
| 20717 | let r = _mm256_mask_cvtusepi16_epi8(src, 0b11111111_11111111, a); |
| 20718 | let e = _mm_set1_epi8(-1); |
| 20719 | assert_eq_m128i(r, e); |
| 20720 | } |
| 20721 | |
| 20722 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20723 | unsafe fn test_mm256_maskz_cvtusepi16_epi8() { |
| 20724 | let a = _mm256_set1_epi16(i16::MIN); |
| 20725 | let r = _mm256_maskz_cvtusepi16_epi8(0, a); |
| 20726 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 20727 | let r = _mm256_maskz_cvtusepi16_epi8(0b11111111_11111111, a); |
| 20728 | let e = _mm_set1_epi8(-1); |
| 20729 | assert_eq_m128i(r, e); |
| 20730 | } |
| 20731 | |
| 20732 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20733 | unsafe fn test_mm_cvtusepi16_epi8() { |
| 20734 | let a = _mm_set1_epi16(i16::MIN); |
| 20735 | let r = _mm_cvtusepi16_epi8(a); |
| 20736 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
| 20737 | assert_eq_m128i(r, e); |
| 20738 | } |
| 20739 | |
| 20740 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20741 | unsafe fn test_mm_mask_cvtusepi16_epi8() { |
| 20742 | let src = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
| 20743 | let a = _mm_set1_epi16(i16::MIN); |
| 20744 | let r = _mm_mask_cvtusepi16_epi8(src, 0, a); |
| 20745 | assert_eq_m128i(r, src); |
| 20746 | let r = _mm_mask_cvtusepi16_epi8(src, 0b11111111, a); |
| 20747 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
| 20748 | assert_eq_m128i(r, e); |
| 20749 | } |
| 20750 | |
| 20751 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20752 | unsafe fn test_mm_maskz_cvtusepi16_epi8() { |
| 20753 | let a = _mm_set1_epi16(i16::MIN); |
| 20754 | let r = _mm_maskz_cvtusepi16_epi8(0, a); |
| 20755 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 20756 | let r = _mm_maskz_cvtusepi16_epi8(0b11111111, a); |
| 20757 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
| 20758 | assert_eq_m128i(r, e); |
| 20759 | } |
| 20760 | |
| 20761 | #[simd_test(enable = "avx512bw" )] |
| 20762 | unsafe fn test_mm512_cvtepi8_epi16() { |
| 20763 | let a = _mm256_set1_epi8(2); |
| 20764 | let r = _mm512_cvtepi8_epi16(a); |
| 20765 | let e = _mm512_set1_epi16(2); |
| 20766 | assert_eq_m512i(r, e); |
| 20767 | } |
| 20768 | |
| 20769 | #[simd_test(enable = "avx512bw" )] |
| 20770 | unsafe fn test_mm512_mask_cvtepi8_epi16() { |
| 20771 | let src = _mm512_set1_epi16(1); |
| 20772 | let a = _mm256_set1_epi8(2); |
| 20773 | let r = _mm512_mask_cvtepi8_epi16(src, 0, a); |
| 20774 | assert_eq_m512i(r, src); |
| 20775 | let r = _mm512_mask_cvtepi8_epi16(src, 0b11111111_11111111_11111111_11111111, a); |
| 20776 | let e = _mm512_set1_epi16(2); |
| 20777 | assert_eq_m512i(r, e); |
| 20778 | } |
| 20779 | |
| 20780 | #[simd_test(enable = "avx512bw" )] |
| 20781 | unsafe fn test_mm512_maskz_cvtepi8_epi16() { |
| 20782 | let a = _mm256_set1_epi8(2); |
| 20783 | let r = _mm512_maskz_cvtepi8_epi16(0, a); |
| 20784 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 20785 | let r = _mm512_maskz_cvtepi8_epi16(0b11111111_11111111_11111111_11111111, a); |
| 20786 | let e = _mm512_set1_epi16(2); |
| 20787 | assert_eq_m512i(r, e); |
| 20788 | } |
| 20789 | |
| 20790 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20791 | unsafe fn test_mm256_mask_cvtepi8_epi16() { |
| 20792 | let src = _mm256_set1_epi16(1); |
| 20793 | let a = _mm_set1_epi8(2); |
| 20794 | let r = _mm256_mask_cvtepi8_epi16(src, 0, a); |
| 20795 | assert_eq_m256i(r, src); |
| 20796 | let r = _mm256_mask_cvtepi8_epi16(src, 0b11111111_11111111, a); |
| 20797 | let e = _mm256_set1_epi16(2); |
| 20798 | assert_eq_m256i(r, e); |
| 20799 | } |
| 20800 | |
| 20801 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20802 | unsafe fn test_mm256_maskz_cvtepi8_epi16() { |
| 20803 | let a = _mm_set1_epi8(2); |
| 20804 | let r = _mm256_maskz_cvtepi8_epi16(0, a); |
| 20805 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 20806 | let r = _mm256_maskz_cvtepi8_epi16(0b11111111_11111111, a); |
| 20807 | let e = _mm256_set1_epi16(2); |
| 20808 | assert_eq_m256i(r, e); |
| 20809 | } |
| 20810 | |
| 20811 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20812 | unsafe fn test_mm_mask_cvtepi8_epi16() { |
| 20813 | let src = _mm_set1_epi16(1); |
| 20814 | let a = _mm_set1_epi8(2); |
| 20815 | let r = _mm_mask_cvtepi8_epi16(src, 0, a); |
| 20816 | assert_eq_m128i(r, src); |
| 20817 | let r = _mm_mask_cvtepi8_epi16(src, 0b11111111, a); |
| 20818 | let e = _mm_set1_epi16(2); |
| 20819 | assert_eq_m128i(r, e); |
| 20820 | } |
| 20821 | |
| 20822 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20823 | unsafe fn test_mm_maskz_cvtepi8_epi16() { |
| 20824 | let a = _mm_set1_epi8(2); |
| 20825 | let r = _mm_maskz_cvtepi8_epi16(0, a); |
| 20826 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 20827 | let r = _mm_maskz_cvtepi8_epi16(0b11111111, a); |
| 20828 | let e = _mm_set1_epi16(2); |
| 20829 | assert_eq_m128i(r, e); |
| 20830 | } |
| 20831 | |
| 20832 | #[simd_test(enable = "avx512bw" )] |
| 20833 | unsafe fn test_mm512_cvtepu8_epi16() { |
| 20834 | let a = _mm256_set1_epi8(2); |
| 20835 | let r = _mm512_cvtepu8_epi16(a); |
| 20836 | let e = _mm512_set1_epi16(2); |
| 20837 | assert_eq_m512i(r, e); |
| 20838 | } |
| 20839 | |
| 20840 | #[simd_test(enable = "avx512bw" )] |
| 20841 | unsafe fn test_mm512_mask_cvtepu8_epi16() { |
| 20842 | let src = _mm512_set1_epi16(1); |
| 20843 | let a = _mm256_set1_epi8(2); |
| 20844 | let r = _mm512_mask_cvtepu8_epi16(src, 0, a); |
| 20845 | assert_eq_m512i(r, src); |
| 20846 | let r = _mm512_mask_cvtepu8_epi16(src, 0b11111111_11111111_11111111_11111111, a); |
| 20847 | let e = _mm512_set1_epi16(2); |
| 20848 | assert_eq_m512i(r, e); |
| 20849 | } |
| 20850 | |
| 20851 | #[simd_test(enable = "avx512bw" )] |
| 20852 | unsafe fn test_mm512_maskz_cvtepu8_epi16() { |
| 20853 | let a = _mm256_set1_epi8(2); |
| 20854 | let r = _mm512_maskz_cvtepu8_epi16(0, a); |
| 20855 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 20856 | let r = _mm512_maskz_cvtepu8_epi16(0b11111111_11111111_11111111_11111111, a); |
| 20857 | let e = _mm512_set1_epi16(2); |
| 20858 | assert_eq_m512i(r, e); |
| 20859 | } |
| 20860 | |
| 20861 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20862 | unsafe fn test_mm256_mask_cvtepu8_epi16() { |
| 20863 | let src = _mm256_set1_epi16(1); |
| 20864 | let a = _mm_set1_epi8(2); |
| 20865 | let r = _mm256_mask_cvtepu8_epi16(src, 0, a); |
| 20866 | assert_eq_m256i(r, src); |
| 20867 | let r = _mm256_mask_cvtepu8_epi16(src, 0b11111111_11111111, a); |
| 20868 | let e = _mm256_set1_epi16(2); |
| 20869 | assert_eq_m256i(r, e); |
| 20870 | } |
| 20871 | |
| 20872 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20873 | unsafe fn test_mm256_maskz_cvtepu8_epi16() { |
| 20874 | let a = _mm_set1_epi8(2); |
| 20875 | let r = _mm256_maskz_cvtepu8_epi16(0, a); |
| 20876 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 20877 | let r = _mm256_maskz_cvtepu8_epi16(0b11111111_11111111, a); |
| 20878 | let e = _mm256_set1_epi16(2); |
| 20879 | assert_eq_m256i(r, e); |
| 20880 | } |
| 20881 | |
| 20882 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20883 | unsafe fn test_mm_mask_cvtepu8_epi16() { |
| 20884 | let src = _mm_set1_epi16(1); |
| 20885 | let a = _mm_set1_epi8(2); |
| 20886 | let r = _mm_mask_cvtepu8_epi16(src, 0, a); |
| 20887 | assert_eq_m128i(r, src); |
| 20888 | let r = _mm_mask_cvtepu8_epi16(src, 0b11111111, a); |
| 20889 | let e = _mm_set1_epi16(2); |
| 20890 | assert_eq_m128i(r, e); |
| 20891 | } |
| 20892 | |
| 20893 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 20894 | unsafe fn test_mm_maskz_cvtepu8_epi16() { |
| 20895 | let a = _mm_set1_epi8(2); |
| 20896 | let r = _mm_maskz_cvtepu8_epi16(0, a); |
| 20897 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 20898 | let r = _mm_maskz_cvtepu8_epi16(0b11111111, a); |
| 20899 | let e = _mm_set1_epi16(2); |
| 20900 | assert_eq_m128i(r, e); |
| 20901 | } |
| 20902 | |
| 20903 | #[simd_test(enable = "avx512bw" )] |
| 20904 | unsafe fn test_mm512_bslli_epi128() { |
| 20905 | #[rustfmt::skip] |
| 20906 | let a = _mm512_set_epi8( |
| 20907 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20908 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20909 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20910 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20911 | ); |
| 20912 | let r = _mm512_bslli_epi128::<9>(a); |
| 20913 | #[rustfmt::skip] |
| 20914 | let e = _mm512_set_epi8( |
| 20915 | 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 20916 | 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 20917 | 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 20918 | 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
| 20919 | ); |
| 20920 | assert_eq_m512i(r, e); |
| 20921 | } |
| 20922 | |
| 20923 | #[simd_test(enable = "avx512bw" )] |
| 20924 | unsafe fn test_mm512_bsrli_epi128() { |
| 20925 | #[rustfmt::skip] |
| 20926 | let a = _mm512_set_epi8( |
| 20927 | 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
| 20928 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
| 20929 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
| 20930 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, |
| 20931 | ); |
| 20932 | let r = _mm512_bsrli_epi128::<3>(a); |
| 20933 | #[rustfmt::skip] |
| 20934 | let e = _mm512_set_epi8( |
| 20935 | 0, 0, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, |
| 20936 | 0, 0, 0, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, |
| 20937 | 0, 0, 0, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, |
| 20938 | 0, 0, 0, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, |
| 20939 | ); |
| 20940 | assert_eq_m512i(r, e); |
| 20941 | } |
| 20942 | |
| 20943 | #[simd_test(enable = "avx512bw" )] |
| 20944 | unsafe fn test_mm512_alignr_epi8() { |
| 20945 | #[rustfmt::skip] |
| 20946 | let a = _mm512_set_epi8( |
| 20947 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20948 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20949 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20950 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20951 | ); |
| 20952 | let b = _mm512_set1_epi8(1); |
| 20953 | let r = _mm512_alignr_epi8::<14>(a, b); |
| 20954 | #[rustfmt::skip] |
| 20955 | let e = _mm512_set_epi8( |
| 20956 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 20957 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 20958 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 20959 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 20960 | ); |
| 20961 | assert_eq_m512i(r, e); |
| 20962 | } |
| 20963 | |
| 20964 | #[simd_test(enable = "avx512bw" )] |
| 20965 | unsafe fn test_mm512_mask_alignr_epi8() { |
| 20966 | #[rustfmt::skip] |
| 20967 | let a = _mm512_set_epi8( |
| 20968 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20969 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20970 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20971 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20972 | ); |
| 20973 | let b = _mm512_set1_epi8(1); |
| 20974 | let r = _mm512_mask_alignr_epi8::<14>(a, 0, a, b); |
| 20975 | assert_eq_m512i(r, a); |
| 20976 | let r = _mm512_mask_alignr_epi8::<14>( |
| 20977 | a, |
| 20978 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 20979 | a, |
| 20980 | b, |
| 20981 | ); |
| 20982 | #[rustfmt::skip] |
| 20983 | let e = _mm512_set_epi8( |
| 20984 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 20985 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 20986 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 20987 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 20988 | ); |
| 20989 | assert_eq_m512i(r, e); |
| 20990 | } |
| 20991 | |
| 20992 | #[simd_test(enable = "avx512bw" )] |
| 20993 | unsafe fn test_mm512_maskz_alignr_epi8() { |
| 20994 | #[rustfmt::skip] |
| 20995 | let a = _mm512_set_epi8( |
| 20996 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20997 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20998 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 20999 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 21000 | ); |
| 21001 | let b = _mm512_set1_epi8(1); |
| 21002 | let r = _mm512_maskz_alignr_epi8::<14>(0, a, b); |
| 21003 | assert_eq_m512i(r, _mm512_setzero_si512()); |
| 21004 | let r = _mm512_maskz_alignr_epi8::<14>( |
| 21005 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
| 21006 | a, |
| 21007 | b, |
| 21008 | ); |
| 21009 | #[rustfmt::skip] |
| 21010 | let e = _mm512_set_epi8( |
| 21011 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 21012 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 21013 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 21014 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 21015 | ); |
| 21016 | assert_eq_m512i(r, e); |
| 21017 | } |
| 21018 | |
| 21019 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 21020 | unsafe fn test_mm256_mask_alignr_epi8() { |
| 21021 | #[rustfmt::skip] |
| 21022 | let a = _mm256_set_epi8( |
| 21023 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 21024 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 21025 | ); |
| 21026 | let b = _mm256_set1_epi8(1); |
| 21027 | let r = _mm256_mask_alignr_epi8::<14>(a, 0, a, b); |
| 21028 | assert_eq_m256i(r, a); |
| 21029 | let r = _mm256_mask_alignr_epi8::<14>(a, 0b11111111_11111111_11111111_11111111, a, b); |
| 21030 | #[rustfmt::skip] |
| 21031 | let e = _mm256_set_epi8( |
| 21032 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 21033 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 21034 | ); |
| 21035 | assert_eq_m256i(r, e); |
| 21036 | } |
| 21037 | |
| 21038 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 21039 | unsafe fn test_mm256_maskz_alignr_epi8() { |
| 21040 | #[rustfmt::skip] |
| 21041 | let a = _mm256_set_epi8( |
| 21042 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 21043 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
| 21044 | ); |
| 21045 | let b = _mm256_set1_epi8(1); |
| 21046 | let r = _mm256_maskz_alignr_epi8::<14>(0, a, b); |
| 21047 | assert_eq_m256i(r, _mm256_setzero_si256()); |
| 21048 | let r = _mm256_maskz_alignr_epi8::<14>(0b11111111_11111111_11111111_11111111, a, b); |
| 21049 | #[rustfmt::skip] |
| 21050 | let e = _mm256_set_epi8( |
| 21051 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 21052 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
| 21053 | ); |
| 21054 | assert_eq_m256i(r, e); |
| 21055 | } |
| 21056 | |
| 21057 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 21058 | unsafe fn test_mm_mask_alignr_epi8() { |
| 21059 | let a = _mm_set_epi8(1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0); |
| 21060 | let b = _mm_set1_epi8(1); |
| 21061 | let r = _mm_mask_alignr_epi8::<14>(a, 0, a, b); |
| 21062 | assert_eq_m128i(r, a); |
| 21063 | let r = _mm_mask_alignr_epi8::<14>(a, 0b11111111_11111111, a, b); |
| 21064 | let e = _mm_set_epi8(0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1); |
| 21065 | assert_eq_m128i(r, e); |
| 21066 | } |
| 21067 | |
| 21068 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 21069 | unsafe fn test_mm_maskz_alignr_epi8() { |
| 21070 | let a = _mm_set_epi8(1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0); |
| 21071 | let b = _mm_set1_epi8(1); |
| 21072 | let r = _mm_maskz_alignr_epi8::<14>(0, a, b); |
| 21073 | assert_eq_m128i(r, _mm_setzero_si128()); |
| 21074 | let r = _mm_maskz_alignr_epi8::<14>(0b11111111_11111111, a, b); |
| 21075 | let e = _mm_set_epi8(0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1); |
| 21076 | assert_eq_m128i(r, e); |
| 21077 | } |
| 21078 | |
| 21079 | #[simd_test(enable = "avx512bw" )] |
| 21080 | unsafe fn test_mm512_mask_cvtsepi16_storeu_epi8() { |
| 21081 | let a = _mm512_set1_epi16(i16::MAX); |
| 21082 | let mut r = _mm256_undefined_si256(); |
| 21083 | _mm512_mask_cvtsepi16_storeu_epi8( |
| 21084 | &mut r as *mut _ as *mut i8, |
| 21085 | 0b11111111_11111111_11111111_11111111, |
| 21086 | a, |
| 21087 | ); |
| 21088 | let e = _mm256_set1_epi8(i8::MAX); |
| 21089 | assert_eq_m256i(r, e); |
| 21090 | } |
| 21091 | |
| 21092 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 21093 | unsafe fn test_mm256_mask_cvtsepi16_storeu_epi8() { |
| 21094 | let a = _mm256_set1_epi16(i16::MAX); |
| 21095 | let mut r = _mm_undefined_si128(); |
| 21096 | _mm256_mask_cvtsepi16_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111_11111111, a); |
| 21097 | let e = _mm_set1_epi8(i8::MAX); |
| 21098 | assert_eq_m128i(r, e); |
| 21099 | } |
| 21100 | |
| 21101 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 21102 | unsafe fn test_mm_mask_cvtsepi16_storeu_epi8() { |
| 21103 | let a = _mm_set1_epi16(i16::MAX); |
| 21104 | let mut r = _mm_set1_epi8(0); |
| 21105 | _mm_mask_cvtsepi16_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
| 21106 | #[rustfmt::skip] |
| 21107 | let e = _mm_set_epi8( |
| 21108 | 0, 0, 0, 0, 0, 0, 0, 0, |
| 21109 | i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, |
| 21110 | ); |
| 21111 | assert_eq_m128i(r, e); |
| 21112 | } |
| 21113 | |
| 21114 | #[simd_test(enable = "avx512bw" )] |
| 21115 | unsafe fn test_mm512_mask_cvtepi16_storeu_epi8() { |
| 21116 | let a = _mm512_set1_epi16(8); |
| 21117 | let mut r = _mm256_undefined_si256(); |
| 21118 | _mm512_mask_cvtepi16_storeu_epi8( |
| 21119 | &mut r as *mut _ as *mut i8, |
| 21120 | 0b11111111_11111111_11111111_11111111, |
| 21121 | a, |
| 21122 | ); |
| 21123 | let e = _mm256_set1_epi8(8); |
| 21124 | assert_eq_m256i(r, e); |
| 21125 | } |
| 21126 | |
| 21127 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 21128 | unsafe fn test_mm256_mask_cvtepi16_storeu_epi8() { |
| 21129 | let a = _mm256_set1_epi16(8); |
| 21130 | let mut r = _mm_undefined_si128(); |
| 21131 | _mm256_mask_cvtepi16_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111_11111111, a); |
| 21132 | let e = _mm_set1_epi8(8); |
| 21133 | assert_eq_m128i(r, e); |
| 21134 | } |
| 21135 | |
| 21136 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 21137 | unsafe fn test_mm_mask_cvtepi16_storeu_epi8() { |
| 21138 | let a = _mm_set1_epi16(8); |
| 21139 | let mut r = _mm_set1_epi8(0); |
| 21140 | _mm_mask_cvtepi16_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
| 21141 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 8, 8, 8, 8, 8, 8, 8); |
| 21142 | assert_eq_m128i(r, e); |
| 21143 | } |
| 21144 | |
| 21145 | #[simd_test(enable = "avx512bw" )] |
| 21146 | unsafe fn test_mm512_mask_cvtusepi16_storeu_epi8() { |
| 21147 | let a = _mm512_set1_epi16(i16::MAX); |
| 21148 | let mut r = _mm256_undefined_si256(); |
| 21149 | _mm512_mask_cvtusepi16_storeu_epi8( |
| 21150 | &mut r as *mut _ as *mut i8, |
| 21151 | 0b11111111_11111111_11111111_11111111, |
| 21152 | a, |
| 21153 | ); |
| 21154 | let e = _mm256_set1_epi8(u8::MAX as i8); |
| 21155 | assert_eq_m256i(r, e); |
| 21156 | } |
| 21157 | |
| 21158 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 21159 | unsafe fn test_mm256_mask_cvtusepi16_storeu_epi8() { |
| 21160 | let a = _mm256_set1_epi16(i16::MAX); |
| 21161 | let mut r = _mm_undefined_si128(); |
| 21162 | _mm256_mask_cvtusepi16_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111_11111111, a); |
| 21163 | let e = _mm_set1_epi8(u8::MAX as i8); |
| 21164 | assert_eq_m128i(r, e); |
| 21165 | } |
| 21166 | |
| 21167 | #[simd_test(enable = "avx512bw,avx512vl" )] |
| 21168 | unsafe fn test_mm_mask_cvtusepi16_storeu_epi8() { |
| 21169 | let a = _mm_set1_epi16(i16::MAX); |
| 21170 | let mut r = _mm_set1_epi8(0); |
| 21171 | _mm_mask_cvtusepi16_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
| 21172 | #[rustfmt::skip] |
| 21173 | let e = _mm_set_epi8( |
| 21174 | 0, 0, 0, 0, |
| 21175 | 0, 0, 0, 0, |
| 21176 | u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, |
| 21177 | u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, |
| 21178 | ); |
| 21179 | assert_eq_m128i(r, e); |
| 21180 | } |
| 21181 | } |
| 21182 | |