1 | use crate::{ |
2 | arch::asm, |
3 | core_arch::{simd::*, x86::*}, |
4 | intrinsics::simd::*, |
5 | mem, ptr, |
6 | }; |
7 | |
8 | #[cfg (test)] |
9 | use stdarch_test::assert_instr; |
10 | |
11 | use super::avx512f::{vpl, vps}; |
12 | |
13 | /// Compute the absolute value of packed signed 16-bit integers in a, and store the unsigned results in dst. |
14 | /// |
15 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_abs_epi16&expand=30) |
16 | #[inline ] |
17 | #[target_feature (enable = "avx512bw" )] |
18 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
19 | #[cfg_attr (test, assert_instr(vpabsw))] |
20 | pub unsafe fn _mm512_abs_epi16(a: __m512i) -> __m512i { |
21 | let a: i16x32 = a.as_i16x32(); |
22 | // all-0 is a properly initialized i16x32 |
23 | let zero: i16x32 = mem::zeroed(); |
24 | let sub: i16x32 = simd_sub(lhs:zero, rhs:a); |
25 | let cmp: i16x32 = simd_gt(x:a, y:zero); |
26 | transmute(src:simd_select(mask:cmp, if_true:a, if_false:sub)) |
27 | } |
28 | |
29 | /// 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). |
30 | /// |
31 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_abs_epi16&expand=31) |
32 | #[inline ] |
33 | #[target_feature (enable = "avx512bw" )] |
34 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
35 | #[cfg_attr (test, assert_instr(vpabsw))] |
36 | pub unsafe fn _mm512_mask_abs_epi16(src: __m512i, k: __mmask32, a: __m512i) -> __m512i { |
37 | let abs: i16x32 = _mm512_abs_epi16(a).as_i16x32(); |
38 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:src.as_i16x32())) |
39 | } |
40 | |
41 | /// 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). |
42 | /// |
43 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_abs_epi16&expand=32) |
44 | #[inline ] |
45 | #[target_feature (enable = "avx512bw" )] |
46 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
47 | #[cfg_attr (test, assert_instr(vpabsw))] |
48 | pub unsafe fn _mm512_maskz_abs_epi16(k: __mmask32, a: __m512i) -> __m512i { |
49 | let abs: i16x32 = _mm512_abs_epi16(a).as_i16x32(); |
50 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
51 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:zero)) |
52 | } |
53 | |
54 | /// 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). |
55 | /// |
56 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_abs_epi16&expand=28) |
57 | #[inline ] |
58 | #[target_feature (enable = "avx512bw,avx512vl" )] |
59 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
60 | #[cfg_attr (test, assert_instr(vpabsw))] |
61 | pub unsafe fn _mm256_mask_abs_epi16(src: __m256i, k: __mmask16, a: __m256i) -> __m256i { |
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 | /// 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). |
67 | /// |
68 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_abs_epi16&expand=29) |
69 | #[inline ] |
70 | #[target_feature (enable = "avx512bw,avx512vl" )] |
71 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
72 | #[cfg_attr (test, assert_instr(vpabsw))] |
73 | pub unsafe fn _mm256_maskz_abs_epi16(k: __mmask16, a: __m256i) -> __m256i { |
74 | let abs: i16x16 = _mm256_abs_epi16(a).as_i16x16(); |
75 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
76 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:zero)) |
77 | } |
78 | |
79 | /// 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). |
80 | /// |
81 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_abs_epi16&expand=25) |
82 | #[inline ] |
83 | #[target_feature (enable = "avx512bw,avx512vl" )] |
84 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
85 | #[cfg_attr (test, assert_instr(vpabsw))] |
86 | pub unsafe fn _mm_mask_abs_epi16(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
87 | let abs: i16x8 = _mm_abs_epi16(a).as_i16x8(); |
88 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:src.as_i16x8())) |
89 | } |
90 | |
91 | /// 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). |
92 | /// |
93 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_abs_epi16&expand=26) |
94 | #[inline ] |
95 | #[target_feature (enable = "avx512bw,avx512vl" )] |
96 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
97 | #[cfg_attr (test, assert_instr(vpabsw))] |
98 | pub unsafe fn _mm_maskz_abs_epi16(k: __mmask8, a: __m128i) -> __m128i { |
99 | let abs: i16x8 = _mm_abs_epi16(a).as_i16x8(); |
100 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
101 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:zero)) |
102 | } |
103 | |
104 | /// Compute the absolute value of packed signed 8-bit integers in a, and store the unsigned results in dst. |
105 | /// |
106 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_abs_epi8&expand=57) |
107 | #[inline ] |
108 | #[target_feature (enable = "avx512bw" )] |
109 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
110 | #[cfg_attr (test, assert_instr(vpabsb))] |
111 | pub unsafe fn _mm512_abs_epi8(a: __m512i) -> __m512i { |
112 | let a: i8x64 = a.as_i8x64(); |
113 | // all-0 is a properly initialized i8x64 |
114 | let zero: i8x64 = mem::zeroed(); |
115 | let sub: i8x64 = simd_sub(lhs:zero, rhs:a); |
116 | let cmp: i8x64 = simd_gt(x:a, y:zero); |
117 | transmute(src:simd_select(mask:cmp, if_true:a, if_false:sub)) |
118 | } |
119 | |
120 | /// 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). |
121 | /// |
122 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_abs_epi8&expand=58) |
123 | #[inline ] |
124 | #[target_feature (enable = "avx512bw" )] |
125 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
126 | #[cfg_attr (test, assert_instr(vpabsb))] |
127 | pub unsafe fn _mm512_mask_abs_epi8(src: __m512i, k: __mmask64, a: __m512i) -> __m512i { |
128 | let abs: i8x64 = _mm512_abs_epi8(a).as_i8x64(); |
129 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:src.as_i8x64())) |
130 | } |
131 | |
132 | /// 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). |
133 | /// |
134 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_abs_epi8&expand=59) |
135 | #[inline ] |
136 | #[target_feature (enable = "avx512bw" )] |
137 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
138 | #[cfg_attr (test, assert_instr(vpabsb))] |
139 | pub unsafe fn _mm512_maskz_abs_epi8(k: __mmask64, a: __m512i) -> __m512i { |
140 | let abs: i8x64 = _mm512_abs_epi8(a).as_i8x64(); |
141 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
142 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:zero)) |
143 | } |
144 | |
145 | /// 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). |
146 | /// |
147 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_abs_epi8&expand=55) |
148 | #[inline ] |
149 | #[target_feature (enable = "avx512bw,avx512vl" )] |
150 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
151 | #[cfg_attr (test, assert_instr(vpabsb))] |
152 | pub unsafe fn _mm256_mask_abs_epi8(src: __m256i, k: __mmask32, a: __m256i) -> __m256i { |
153 | let abs: i8x32 = _mm256_abs_epi8(a).as_i8x32(); |
154 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:src.as_i8x32())) |
155 | } |
156 | |
157 | /// 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). |
158 | /// |
159 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_abs_epi8&expand=56) |
160 | #[inline ] |
161 | #[target_feature (enable = "avx512bw,avx512vl" )] |
162 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
163 | #[cfg_attr (test, assert_instr(vpabsb))] |
164 | pub unsafe fn _mm256_maskz_abs_epi8(k: __mmask32, a: __m256i) -> __m256i { |
165 | let abs: i8x32 = _mm256_abs_epi8(a).as_i8x32(); |
166 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
167 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:zero)) |
168 | } |
169 | |
170 | /// 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) |
171 | /// |
172 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_abs_epi8&expand=52) |
173 | #[inline ] |
174 | #[target_feature (enable = "avx512bw,avx512vl" )] |
175 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
176 | #[cfg_attr (test, assert_instr(vpabsb))] |
177 | pub unsafe fn _mm_mask_abs_epi8(src: __m128i, k: __mmask16, a: __m128i) -> __m128i { |
178 | let abs: i8x16 = _mm_abs_epi8(a).as_i8x16(); |
179 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:src.as_i8x16())) |
180 | } |
181 | |
182 | /// 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). |
183 | /// |
184 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_abs_epi8&expand=53) |
185 | #[inline ] |
186 | #[target_feature (enable = "avx512bw,avx512vl" )] |
187 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
188 | #[cfg_attr (test, assert_instr(vpabsb))] |
189 | pub unsafe fn _mm_maskz_abs_epi8(k: __mmask16, a: __m128i) -> __m128i { |
190 | let abs: i8x16 = _mm_abs_epi8(a).as_i8x16(); |
191 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
192 | transmute(src:simd_select_bitmask(m:k, yes:abs, no:zero)) |
193 | } |
194 | |
195 | /// Add packed 16-bit integers in a and b, and store the results in dst. |
196 | /// |
197 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_add_epi16&expand=91) |
198 | #[inline ] |
199 | #[target_feature (enable = "avx512bw" )] |
200 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
201 | #[cfg_attr (test, assert_instr(vpaddw))] |
202 | pub unsafe fn _mm512_add_epi16(a: __m512i, b: __m512i) -> __m512i { |
203 | transmute(src:simd_add(x:a.as_i16x32(), y:b.as_i16x32())) |
204 | } |
205 | |
206 | /// 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). |
207 | /// |
208 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_add_epi16&expand=92) |
209 | #[inline ] |
210 | #[target_feature (enable = "avx512bw" )] |
211 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
212 | #[cfg_attr (test, assert_instr(vpaddw))] |
213 | pub unsafe fn _mm512_mask_add_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
214 | let add: i16x32 = _mm512_add_epi16(a, b).as_i16x32(); |
215 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i16x32())) |
216 | } |
217 | |
218 | /// 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). |
219 | /// |
220 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_add_epi16&expand=93) |
221 | #[inline ] |
222 | #[target_feature (enable = "avx512bw" )] |
223 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
224 | #[cfg_attr (test, assert_instr(vpaddw))] |
225 | pub unsafe fn _mm512_maskz_add_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
226 | let add: i16x32 = _mm512_add_epi16(a, b).as_i16x32(); |
227 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
228 | transmute(src:simd_select_bitmask(m:k, yes:add, no:zero)) |
229 | } |
230 | |
231 | /// 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). |
232 | /// |
233 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_add_epi&expand=89) |
234 | #[inline ] |
235 | #[target_feature (enable = "avx512bw,avx512vl" )] |
236 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
237 | #[cfg_attr (test, assert_instr(vpaddw))] |
238 | pub unsafe fn _mm256_mask_add_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
239 | let add: i16x16 = _mm256_add_epi16(a, b).as_i16x16(); |
240 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i16x16())) |
241 | } |
242 | |
243 | /// 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). |
244 | /// |
245 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_add_epi16&expand=90) |
246 | #[inline ] |
247 | #[target_feature (enable = "avx512bw,avx512vl" )] |
248 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
249 | #[cfg_attr (test, assert_instr(vpaddw))] |
250 | pub unsafe fn _mm256_maskz_add_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
251 | let add: i16x16 = _mm256_add_epi16(a, b).as_i16x16(); |
252 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
253 | transmute(src:simd_select_bitmask(m:k, yes:add, no:zero)) |
254 | } |
255 | |
256 | /// 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). |
257 | /// |
258 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_add_epi16&expand=86) |
259 | #[inline ] |
260 | #[target_feature (enable = "avx512bw,avx512vl" )] |
261 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
262 | #[cfg_attr (test, assert_instr(vpaddw))] |
263 | pub unsafe fn _mm_mask_add_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
264 | let add: i16x8 = _mm_add_epi16(a, b).as_i16x8(); |
265 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i16x8())) |
266 | } |
267 | |
268 | /// 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). |
269 | /// |
270 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_add_epi16&expand=87) |
271 | #[inline ] |
272 | #[target_feature (enable = "avx512bw,avx512vl" )] |
273 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
274 | #[cfg_attr (test, assert_instr(vpaddw))] |
275 | pub unsafe fn _mm_maskz_add_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
276 | let add: i16x8 = _mm_add_epi16(a, b).as_i16x8(); |
277 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
278 | transmute(src:simd_select_bitmask(m:k, yes:add, no:zero)) |
279 | } |
280 | |
281 | /// Add packed 8-bit integers in a and b, and store the results in dst. |
282 | /// |
283 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_add_epi8&expand=118) |
284 | #[inline ] |
285 | #[target_feature (enable = "avx512bw" )] |
286 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
287 | #[cfg_attr (test, assert_instr(vpaddb))] |
288 | pub unsafe fn _mm512_add_epi8(a: __m512i, b: __m512i) -> __m512i { |
289 | transmute(src:simd_add(x:a.as_i8x64(), y:b.as_i8x64())) |
290 | } |
291 | |
292 | /// 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). |
293 | /// |
294 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_add_epi8&expand=119) |
295 | #[inline ] |
296 | #[target_feature (enable = "avx512bw" )] |
297 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
298 | #[cfg_attr (test, assert_instr(vpaddb))] |
299 | pub unsafe fn _mm512_mask_add_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
300 | let add: i8x64 = _mm512_add_epi8(a, b).as_i8x64(); |
301 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i8x64())) |
302 | } |
303 | |
304 | /// 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). |
305 | /// |
306 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_add_epi8&expand=120) |
307 | #[inline ] |
308 | #[target_feature (enable = "avx512bw" )] |
309 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
310 | #[cfg_attr (test, assert_instr(vpaddb))] |
311 | pub unsafe fn _mm512_maskz_add_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
312 | let add: i8x64 = _mm512_add_epi8(a, b).as_i8x64(); |
313 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
314 | transmute(src:simd_select_bitmask(m:k, yes:add, no:zero)) |
315 | } |
316 | |
317 | /// 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). |
318 | /// |
319 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_add_epi8&expand=116) |
320 | #[inline ] |
321 | #[target_feature (enable = "avx512bw,avx512vl" )] |
322 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
323 | #[cfg_attr (test, assert_instr(vpaddb))] |
324 | pub unsafe fn _mm256_mask_add_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
325 | let add: i8x32 = _mm256_add_epi8(a, b).as_i8x32(); |
326 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i8x32())) |
327 | } |
328 | |
329 | /// 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). |
330 | /// |
331 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_add_epi8&expand=117) |
332 | #[inline ] |
333 | #[target_feature (enable = "avx512bw,avx512vl" )] |
334 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
335 | #[cfg_attr (test, assert_instr(vpaddb))] |
336 | pub unsafe fn _mm256_maskz_add_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
337 | let add: i8x32 = _mm256_add_epi8(a, b).as_i8x32(); |
338 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
339 | transmute(src:simd_select_bitmask(m:k, yes:add, no:zero)) |
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=_mm_mask_add_epi8&expand=113) |
345 | #[inline ] |
346 | #[target_feature (enable = "avx512bw,avx512vl" )] |
347 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
348 | #[cfg_attr (test, assert_instr(vpaddb))] |
349 | pub unsafe fn _mm_mask_add_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
350 | let add: i8x16 = _mm_add_epi8(a, b).as_i8x16(); |
351 | transmute(src:simd_select_bitmask(m:k, yes:add, no:src.as_i8x16())) |
352 | } |
353 | |
354 | /// 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). |
355 | /// |
356 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_add_epi8&expand=114) |
357 | #[inline ] |
358 | #[target_feature (enable = "avx512bw,avx512vl" )] |
359 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
360 | #[cfg_attr (test, assert_instr(vpaddb))] |
361 | pub unsafe fn _mm_maskz_add_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
362 | let add: i8x16 = _mm_add_epi8(a, b).as_i8x16(); |
363 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
364 | transmute(src:simd_select_bitmask(m:k, yes:add, no:zero)) |
365 | } |
366 | |
367 | /// Add packed unsigned 16-bit integers in a and b using saturation, and store the results in dst. |
368 | /// |
369 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_adds_epu16&expand=197) |
370 | #[inline ] |
371 | #[target_feature (enable = "avx512bw" )] |
372 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
373 | #[cfg_attr (test, assert_instr(vpaddusw))] |
374 | pub unsafe fn _mm512_adds_epu16(a: __m512i, b: __m512i) -> __m512i { |
375 | transmute(src:vpaddusw( |
376 | a:a.as_u16x32(), |
377 | b:b.as_u16x32(), |
378 | src:_mm512_setzero_si512().as_u16x32(), |
379 | mask:0b11111111_11111111_11111111_11111111, |
380 | )) |
381 | } |
382 | |
383 | /// 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). |
384 | /// |
385 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_adds_epu16&expand=198) |
386 | #[inline ] |
387 | #[target_feature (enable = "avx512bw" )] |
388 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
389 | #[cfg_attr (test, assert_instr(vpaddusw))] |
390 | pub unsafe fn _mm512_mask_adds_epu16( |
391 | src: __m512i, |
392 | k: __mmask32, |
393 | a: __m512i, |
394 | b: __m512i, |
395 | ) -> __m512i { |
396 | transmute(src:vpaddusw(a:a.as_u16x32(), b:b.as_u16x32(), src:src.as_u16x32(), mask:k)) |
397 | } |
398 | |
399 | /// 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). |
400 | /// |
401 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_adds_epu16&expand=199) |
402 | #[inline ] |
403 | #[target_feature (enable = "avx512bw" )] |
404 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
405 | #[cfg_attr (test, assert_instr(vpaddusw))] |
406 | pub unsafe fn _mm512_maskz_adds_epu16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
407 | transmute(src:vpaddusw( |
408 | a:a.as_u16x32(), |
409 | b:b.as_u16x32(), |
410 | src:_mm512_setzero_si512().as_u16x32(), |
411 | mask:k, |
412 | )) |
413 | } |
414 | |
415 | /// 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). |
416 | /// |
417 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_adds_epu16&expand=195) |
418 | #[inline ] |
419 | #[target_feature (enable = "avx512bw,avx512vl" )] |
420 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
421 | #[cfg_attr (test, assert_instr(vpaddusw))] |
422 | pub unsafe fn _mm256_mask_adds_epu16( |
423 | src: __m256i, |
424 | k: __mmask16, |
425 | a: __m256i, |
426 | b: __m256i, |
427 | ) -> __m256i { |
428 | transmute(src:vpaddusw256( |
429 | a:a.as_u16x16(), |
430 | b:b.as_u16x16(), |
431 | src:src.as_u16x16(), |
432 | mask:k, |
433 | )) |
434 | } |
435 | |
436 | /// 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). |
437 | /// |
438 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_adds_epu16&expand=196) |
439 | #[inline ] |
440 | #[target_feature (enable = "avx512bw,avx512vl" )] |
441 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
442 | #[cfg_attr (test, assert_instr(vpaddusw))] |
443 | pub unsafe fn _mm256_maskz_adds_epu16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
444 | transmute(src:vpaddusw256( |
445 | a:a.as_u16x16(), |
446 | b:b.as_u16x16(), |
447 | src:_mm256_setzero_si256().as_u16x16(), |
448 | mask:k, |
449 | )) |
450 | } |
451 | |
452 | /// 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). |
453 | /// |
454 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_adds_epu16&expand=192) |
455 | #[inline ] |
456 | #[target_feature (enable = "avx512bw,avx512vl" )] |
457 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
458 | #[cfg_attr (test, assert_instr(vpaddusw))] |
459 | pub unsafe fn _mm_mask_adds_epu16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
460 | transmute(src:vpaddusw128(a:a.as_u16x8(), b:b.as_u16x8(), src:src.as_u16x8(), mask:k)) |
461 | } |
462 | |
463 | /// 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). |
464 | /// |
465 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_adds_epu16&expand=193) |
466 | #[inline ] |
467 | #[target_feature (enable = "avx512bw,avx512vl" )] |
468 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
469 | #[cfg_attr (test, assert_instr(vpaddusw))] |
470 | pub unsafe fn _mm_maskz_adds_epu16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
471 | transmute(src:vpaddusw128( |
472 | a:a.as_u16x8(), |
473 | b:b.as_u16x8(), |
474 | src:_mm_setzero_si128().as_u16x8(), |
475 | mask:k, |
476 | )) |
477 | } |
478 | |
479 | /// Add packed unsigned 8-bit integers in a and b using saturation, and store the results in dst. |
480 | /// |
481 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_adds_epu8&expand=206) |
482 | #[inline ] |
483 | #[target_feature (enable = "avx512bw" )] |
484 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
485 | #[cfg_attr (test, assert_instr(vpaddusb))] |
486 | pub unsafe fn _mm512_adds_epu8(a: __m512i, b: __m512i) -> __m512i { |
487 | transmute(src:vpaddusb( |
488 | a:a.as_u8x64(), |
489 | b:b.as_u8x64(), |
490 | src:_mm512_setzero_si512().as_u8x64(), |
491 | mask:0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
492 | )) |
493 | } |
494 | |
495 | /// 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). |
496 | /// |
497 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_adds_epu8&expand=207) |
498 | #[inline ] |
499 | #[target_feature (enable = "avx512bw" )] |
500 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
501 | #[cfg_attr (test, assert_instr(vpaddusb))] |
502 | pub unsafe fn _mm512_mask_adds_epu8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
503 | transmute(src:vpaddusb(a:a.as_u8x64(), b:b.as_u8x64(), src:src.as_u8x64(), mask:k)) |
504 | } |
505 | |
506 | /// 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). |
507 | /// |
508 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_adds_epu8&expand=208) |
509 | #[inline ] |
510 | #[target_feature (enable = "avx512bw" )] |
511 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
512 | #[cfg_attr (test, assert_instr(vpaddusb))] |
513 | pub unsafe fn _mm512_maskz_adds_epu8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
514 | transmute(src:vpaddusb( |
515 | a:a.as_u8x64(), |
516 | b:b.as_u8x64(), |
517 | src:_mm512_setzero_si512().as_u8x64(), |
518 | mask:k, |
519 | )) |
520 | } |
521 | |
522 | /// 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). |
523 | /// |
524 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_adds_epu8&expand=204) |
525 | #[inline ] |
526 | #[target_feature (enable = "avx512bw,avx512vl" )] |
527 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
528 | #[cfg_attr (test, assert_instr(vpaddusb))] |
529 | pub unsafe fn _mm256_mask_adds_epu8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
530 | transmute(src:vpaddusb256(a:a.as_u8x32(), b:b.as_u8x32(), src:src.as_u8x32(), mask:k)) |
531 | } |
532 | |
533 | /// 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). |
534 | /// |
535 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_adds_epu8&expand=205) |
536 | #[inline ] |
537 | #[target_feature (enable = "avx512bw,avx512vl" )] |
538 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
539 | #[cfg_attr (test, assert_instr(vpaddusb))] |
540 | pub unsafe fn _mm256_maskz_adds_epu8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
541 | transmute(src:vpaddusb256( |
542 | a:a.as_u8x32(), |
543 | b:b.as_u8x32(), |
544 | src:_mm256_setzero_si256().as_u8x32(), |
545 | mask:k, |
546 | )) |
547 | } |
548 | |
549 | /// 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). |
550 | /// |
551 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_adds_epu8&expand=201) |
552 | #[inline ] |
553 | #[target_feature (enable = "avx512bw,avx512vl" )] |
554 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
555 | #[cfg_attr (test, assert_instr(vpaddusb))] |
556 | pub unsafe fn _mm_mask_adds_epu8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
557 | transmute(src:vpaddusb128(a:a.as_u8x16(), b:b.as_u8x16(), src:src.as_u8x16(), mask:k)) |
558 | } |
559 | |
560 | /// 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). |
561 | /// |
562 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_adds_epu8&expand=202) |
563 | #[inline ] |
564 | #[target_feature (enable = "avx512bw,avx512vl" )] |
565 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
566 | #[cfg_attr (test, assert_instr(vpaddusb))] |
567 | pub unsafe fn _mm_maskz_adds_epu8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
568 | transmute(src:vpaddusb128( |
569 | a:a.as_u8x16(), |
570 | b:b.as_u8x16(), |
571 | src:_mm_setzero_si128().as_u8x16(), |
572 | mask:k, |
573 | )) |
574 | } |
575 | |
576 | /// Add packed signed 16-bit integers in a and b using saturation, and store the results in dst. |
577 | /// |
578 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_adds_epi16&expand=179) |
579 | #[inline ] |
580 | #[target_feature (enable = "avx512bw" )] |
581 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
582 | #[cfg_attr (test, assert_instr(vpaddsw))] |
583 | pub unsafe fn _mm512_adds_epi16(a: __m512i, b: __m512i) -> __m512i { |
584 | transmute(src:vpaddsw( |
585 | a:a.as_i16x32(), |
586 | b:b.as_i16x32(), |
587 | src:_mm512_setzero_si512().as_i16x32(), |
588 | mask:0b11111111_11111111_11111111_11111111, |
589 | )) |
590 | } |
591 | |
592 | /// 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). |
593 | /// |
594 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_adds_epi16&expand=180) |
595 | #[inline ] |
596 | #[target_feature (enable = "avx512bw" )] |
597 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
598 | #[cfg_attr (test, assert_instr(vpaddsw))] |
599 | pub unsafe fn _mm512_mask_adds_epi16( |
600 | src: __m512i, |
601 | k: __mmask32, |
602 | a: __m512i, |
603 | b: __m512i, |
604 | ) -> __m512i { |
605 | transmute(src:vpaddsw(a:a.as_i16x32(), b:b.as_i16x32(), src:src.as_i16x32(), mask:k)) |
606 | } |
607 | |
608 | /// 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). |
609 | /// |
610 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_adds_epi16&expand=181) |
611 | #[inline ] |
612 | #[target_feature (enable = "avx512bw" )] |
613 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
614 | #[cfg_attr (test, assert_instr(vpaddsw))] |
615 | pub unsafe fn _mm512_maskz_adds_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
616 | transmute(src:vpaddsw( |
617 | a:a.as_i16x32(), |
618 | b:b.as_i16x32(), |
619 | src:_mm512_setzero_si512().as_i16x32(), |
620 | mask:k, |
621 | )) |
622 | } |
623 | |
624 | /// 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). |
625 | /// |
626 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_adds_epi16&expand=177) |
627 | #[inline ] |
628 | #[target_feature (enable = "avx512bw,avx512vl" )] |
629 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
630 | #[cfg_attr (test, assert_instr(vpaddsw))] |
631 | pub unsafe fn _mm256_mask_adds_epi16( |
632 | src: __m256i, |
633 | k: __mmask16, |
634 | a: __m256i, |
635 | b: __m256i, |
636 | ) -> __m256i { |
637 | transmute(src:vpaddsw256(a:a.as_i16x16(), b:b.as_i16x16(), src:src.as_i16x16(), mask:k)) |
638 | } |
639 | |
640 | /// 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). |
641 | /// |
642 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_adds_epi16&expand=178) |
643 | #[inline ] |
644 | #[target_feature (enable = "avx512bw,avx512vl" )] |
645 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
646 | #[cfg_attr (test, assert_instr(vpaddsw))] |
647 | pub unsafe fn _mm256_maskz_adds_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
648 | transmute(src:vpaddsw256( |
649 | a:a.as_i16x16(), |
650 | b:b.as_i16x16(), |
651 | src:_mm256_setzero_si256().as_i16x16(), |
652 | mask:k, |
653 | )) |
654 | } |
655 | |
656 | /// 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). |
657 | /// |
658 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_adds_epi16&expand=174) |
659 | #[inline ] |
660 | #[target_feature (enable = "avx512bw,avx512vl" )] |
661 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
662 | #[cfg_attr (test, assert_instr(vpaddsw))] |
663 | pub unsafe fn _mm_mask_adds_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
664 | transmute(src:vpaddsw128(a:a.as_i16x8(), b:b.as_i16x8(), src:src.as_i16x8(), mask:k)) |
665 | } |
666 | |
667 | /// 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). |
668 | /// |
669 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_adds_epi16&expand=175) |
670 | #[inline ] |
671 | #[target_feature (enable = "avx512bw,avx512vl" )] |
672 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
673 | #[cfg_attr (test, assert_instr(vpaddsw))] |
674 | pub unsafe fn _mm_maskz_adds_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
675 | transmute(src:vpaddsw128( |
676 | a:a.as_i16x8(), |
677 | b:b.as_i16x8(), |
678 | src:_mm_setzero_si128().as_i16x8(), |
679 | mask:k, |
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 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
689 | #[cfg_attr (test, assert_instr(vpaddsb))] |
690 | pub unsafe fn _mm512_adds_epi8(a: __m512i, b: __m512i) -> __m512i { |
691 | transmute(src:vpaddsb( |
692 | a:a.as_i8x64(), |
693 | b:b.as_i8x64(), |
694 | src:_mm512_setzero_si512().as_i8x64(), |
695 | mask:0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
696 | )) |
697 | } |
698 | |
699 | /// 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). |
700 | /// |
701 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_adds_epi8&expand=189) |
702 | #[inline ] |
703 | #[target_feature (enable = "avx512bw" )] |
704 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
705 | #[cfg_attr (test, assert_instr(vpaddsb))] |
706 | pub unsafe fn _mm512_mask_adds_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
707 | transmute(src:vpaddsb(a:a.as_i8x64(), b:b.as_i8x64(), src:src.as_i8x64(), mask:k)) |
708 | } |
709 | |
710 | /// 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). |
711 | /// |
712 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_adds_epi8&expand=190) |
713 | #[inline ] |
714 | #[target_feature (enable = "avx512bw" )] |
715 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
716 | #[cfg_attr (test, assert_instr(vpaddsb))] |
717 | pub unsafe fn _mm512_maskz_adds_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
718 | transmute(src:vpaddsb( |
719 | a:a.as_i8x64(), |
720 | b:b.as_i8x64(), |
721 | src:_mm512_setzero_si512().as_i8x64(), |
722 | mask:k, |
723 | )) |
724 | } |
725 | |
726 | /// 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). |
727 | /// |
728 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_adds_epi8&expand=186) |
729 | #[inline ] |
730 | #[target_feature (enable = "avx512bw,avx512vl" )] |
731 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
732 | #[cfg_attr (test, assert_instr(vpaddsb))] |
733 | pub unsafe fn _mm256_mask_adds_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
734 | transmute(src:vpaddsb256(a:a.as_i8x32(), b:b.as_i8x32(), src:src.as_i8x32(), mask:k)) |
735 | } |
736 | |
737 | /// 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). |
738 | /// |
739 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_adds_epi8&expand=187) |
740 | #[inline ] |
741 | #[target_feature (enable = "avx512bw,avx512vl" )] |
742 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
743 | #[cfg_attr (test, assert_instr(vpaddsb))] |
744 | pub unsafe fn _mm256_maskz_adds_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
745 | transmute(src:vpaddsb256( |
746 | a:a.as_i8x32(), |
747 | b:b.as_i8x32(), |
748 | src:_mm256_setzero_si256().as_i8x32(), |
749 | mask:k, |
750 | )) |
751 | } |
752 | |
753 | /// 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). |
754 | /// |
755 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_adds_epi8&expand=183) |
756 | #[inline ] |
757 | #[target_feature (enable = "avx512bw,avx512vl" )] |
758 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
759 | #[cfg_attr (test, assert_instr(vpaddsb))] |
760 | pub unsafe fn _mm_mask_adds_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
761 | transmute(src:vpaddsb128(a:a.as_i8x16(), b:b.as_i8x16(), src:src.as_i8x16(), mask:k)) |
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 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
770 | #[cfg_attr (test, assert_instr(vpaddsb))] |
771 | pub unsafe fn _mm_maskz_adds_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
772 | transmute(src:vpaddsb128( |
773 | a:a.as_i8x16(), |
774 | b:b.as_i8x16(), |
775 | src:_mm_setzero_si128().as_i8x16(), |
776 | mask:k, |
777 | )) |
778 | } |
779 | |
780 | /// Subtract packed 16-bit integers in b from packed 16-bit integers in a, and store the results in dst. |
781 | /// |
782 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_sub_epi16&expand=5685) |
783 | #[inline ] |
784 | #[target_feature (enable = "avx512bw" )] |
785 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
786 | #[cfg_attr (test, assert_instr(vpsubw))] |
787 | pub unsafe fn _mm512_sub_epi16(a: __m512i, b: __m512i) -> __m512i { |
788 | transmute(src:simd_sub(lhs:a.as_i16x32(), rhs:b.as_i16x32())) |
789 | } |
790 | |
791 | /// 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). |
792 | /// |
793 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_sub_epi16&expand=5683) |
794 | #[inline ] |
795 | #[target_feature (enable = "avx512bw" )] |
796 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
797 | #[cfg_attr (test, assert_instr(vpsubw))] |
798 | pub unsafe fn _mm512_mask_sub_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
799 | let sub: i16x32 = _mm512_sub_epi16(a, b).as_i16x32(); |
800 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i16x32())) |
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 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
809 | #[cfg_attr (test, assert_instr(vpsubw))] |
810 | pub unsafe fn _mm512_maskz_sub_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
811 | let sub: i16x32 = _mm512_sub_epi16(a, b).as_i16x32(); |
812 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
813 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:zero)) |
814 | } |
815 | |
816 | /// 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). |
817 | /// |
818 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_sub_epi16&expand=5680) |
819 | #[inline ] |
820 | #[target_feature (enable = "avx512bw,avx512vl" )] |
821 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
822 | #[cfg_attr (test, assert_instr(vpsubw))] |
823 | pub unsafe fn _mm256_mask_sub_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
824 | let sub: i16x16 = _mm256_sub_epi16(a, b).as_i16x16(); |
825 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i16x16())) |
826 | } |
827 | |
828 | /// 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). |
829 | /// |
830 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_sub_epi16&expand=5681) |
831 | #[inline ] |
832 | #[target_feature (enable = "avx512bw,avx512vl" )] |
833 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
834 | #[cfg_attr (test, assert_instr(vpsubw))] |
835 | pub unsafe fn _mm256_maskz_sub_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
836 | let sub: i16x16 = _mm256_sub_epi16(a, b).as_i16x16(); |
837 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
838 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:zero)) |
839 | } |
840 | |
841 | /// 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). |
842 | /// |
843 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_sub_epi16&expand=5677) |
844 | #[inline ] |
845 | #[target_feature (enable = "avx512bw,avx512vl" )] |
846 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
847 | #[cfg_attr (test, assert_instr(vpsubw))] |
848 | pub unsafe fn _mm_mask_sub_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
849 | let sub: i16x8 = _mm_sub_epi16(a, b).as_i16x8(); |
850 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i16x8())) |
851 | } |
852 | |
853 | /// 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). |
854 | /// |
855 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_sub_epi16&expand=5678) |
856 | #[inline ] |
857 | #[target_feature (enable = "avx512bw,avx512vl" )] |
858 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
859 | #[cfg_attr (test, assert_instr(vpsubw))] |
860 | pub unsafe fn _mm_maskz_sub_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
861 | let sub: i16x8 = _mm_sub_epi16(a, b).as_i16x8(); |
862 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
863 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:zero)) |
864 | } |
865 | |
866 | /// Subtract packed 8-bit integers in b from packed 8-bit integers in a, and store the results in dst. |
867 | /// |
868 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_sub_epi8&expand=5712) |
869 | #[inline ] |
870 | #[target_feature (enable = "avx512bw" )] |
871 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
872 | #[cfg_attr (test, assert_instr(vpsubb))] |
873 | pub unsafe fn _mm512_sub_epi8(a: __m512i, b: __m512i) -> __m512i { |
874 | transmute(src:simd_sub(lhs:a.as_i8x64(), rhs:b.as_i8x64())) |
875 | } |
876 | |
877 | /// 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). |
878 | /// |
879 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_sub_epi8&expand=5710) |
880 | #[inline ] |
881 | #[target_feature (enable = "avx512bw" )] |
882 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
883 | #[cfg_attr (test, assert_instr(vpsubb))] |
884 | pub unsafe fn _mm512_mask_sub_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
885 | let sub: i8x64 = _mm512_sub_epi8(a, b).as_i8x64(); |
886 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i8x64())) |
887 | } |
888 | |
889 | /// 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). |
890 | /// |
891 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_sub_epi8&expand=5711) |
892 | #[inline ] |
893 | #[target_feature (enable = "avx512bw" )] |
894 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
895 | #[cfg_attr (test, assert_instr(vpsubb))] |
896 | pub unsafe fn _mm512_maskz_sub_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
897 | let sub: i8x64 = _mm512_sub_epi8(a, b).as_i8x64(); |
898 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
899 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:zero)) |
900 | } |
901 | |
902 | /// 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). |
903 | /// |
904 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_sub_epi8&expand=5707) |
905 | #[inline ] |
906 | #[target_feature (enable = "avx512bw,avx512vl" )] |
907 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
908 | #[cfg_attr (test, assert_instr(vpsubb))] |
909 | pub unsafe fn _mm256_mask_sub_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
910 | let sub: i8x32 = _mm256_sub_epi8(a, b).as_i8x32(); |
911 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i8x32())) |
912 | } |
913 | |
914 | /// 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). |
915 | /// |
916 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_sub_epi8&expand=5708) |
917 | #[inline ] |
918 | #[target_feature (enable = "avx512bw,avx512vl" )] |
919 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
920 | #[cfg_attr (test, assert_instr(vpsubb))] |
921 | pub unsafe fn _mm256_maskz_sub_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
922 | let sub: i8x32 = _mm256_sub_epi8(a, b).as_i8x32(); |
923 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
924 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:zero)) |
925 | } |
926 | |
927 | /// 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). |
928 | /// |
929 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_sub_epi8&expand=5704) |
930 | #[inline ] |
931 | #[target_feature (enable = "avx512bw,avx512vl" )] |
932 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
933 | #[cfg_attr (test, assert_instr(vpsubb))] |
934 | pub unsafe fn _mm_mask_sub_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
935 | let sub: i8x16 = _mm_sub_epi8(a, b).as_i8x16(); |
936 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:src.as_i8x16())) |
937 | } |
938 | |
939 | /// 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). |
940 | /// |
941 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_sub_epi8&expand=5705) |
942 | #[inline ] |
943 | #[target_feature (enable = "avx512bw,avx512vl" )] |
944 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
945 | #[cfg_attr (test, assert_instr(vpsubb))] |
946 | pub unsafe fn _mm_maskz_sub_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
947 | let sub: i8x16 = _mm_sub_epi8(a, b).as_i8x16(); |
948 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
949 | transmute(src:simd_select_bitmask(m:k, yes:sub, no:zero)) |
950 | } |
951 | |
952 | /// Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation, and store the results in dst. |
953 | /// |
954 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_subs_epu16&expand=5793) |
955 | #[inline ] |
956 | #[target_feature (enable = "avx512bw" )] |
957 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
958 | #[cfg_attr (test, assert_instr(vpsubusw))] |
959 | pub unsafe fn _mm512_subs_epu16(a: __m512i, b: __m512i) -> __m512i { |
960 | transmute(src:vpsubusw( |
961 | a:a.as_u16x32(), |
962 | b:b.as_u16x32(), |
963 | src:_mm512_setzero_si512().as_u16x32(), |
964 | mask:0b11111111_11111111_11111111_11111111, |
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 using writemask k (elements are copied from src when the corresponding mask bit is not set). |
969 | /// |
970 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_subs_epu16&expand=5791) |
971 | #[inline ] |
972 | #[target_feature (enable = "avx512bw" )] |
973 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
974 | #[cfg_attr (test, assert_instr(vpsubusw))] |
975 | pub unsafe fn _mm512_mask_subs_epu16( |
976 | src: __m512i, |
977 | k: __mmask32, |
978 | a: __m512i, |
979 | b: __m512i, |
980 | ) -> __m512i { |
981 | transmute(src:vpsubusw(a:a.as_u16x32(), b:b.as_u16x32(), src:src.as_u16x32(), mask:k)) |
982 | } |
983 | |
984 | /// 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). |
985 | /// |
986 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_subs_epu16&expand=5792) |
987 | #[inline ] |
988 | #[target_feature (enable = "avx512bw" )] |
989 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
990 | #[cfg_attr (test, assert_instr(vpsubusw))] |
991 | pub unsafe fn _mm512_maskz_subs_epu16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
992 | transmute(src:vpsubusw( |
993 | a:a.as_u16x32(), |
994 | b:b.as_u16x32(), |
995 | src:_mm512_setzero_si512().as_u16x32(), |
996 | mask:k, |
997 | )) |
998 | } |
999 | |
1000 | /// 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). |
1001 | /// |
1002 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_subs_epu16&expand=5788) |
1003 | #[inline ] |
1004 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1005 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1006 | #[cfg_attr (test, assert_instr(vpsubusw))] |
1007 | pub unsafe fn _mm256_mask_subs_epu16( |
1008 | src: __m256i, |
1009 | k: __mmask16, |
1010 | a: __m256i, |
1011 | b: __m256i, |
1012 | ) -> __m256i { |
1013 | transmute(src:vpsubusw256( |
1014 | a:a.as_u16x16(), |
1015 | b:b.as_u16x16(), |
1016 | src:src.as_u16x16(), |
1017 | mask:k, |
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 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1027 | #[cfg_attr (test, assert_instr(vpsubusw))] |
1028 | pub unsafe fn _mm256_maskz_subs_epu16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
1029 | transmute(src:vpsubusw256( |
1030 | a:a.as_u16x16(), |
1031 | b:b.as_u16x16(), |
1032 | src:_mm256_setzero_si256().as_u16x16(), |
1033 | mask:k, |
1034 | )) |
1035 | } |
1036 | |
1037 | /// 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). |
1038 | /// |
1039 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_subs_epu16&expand=5785) |
1040 | #[inline ] |
1041 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1042 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1043 | #[cfg_attr (test, assert_instr(vpsubusw))] |
1044 | pub unsafe fn _mm_mask_subs_epu16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1045 | transmute(src:vpsubusw128(a:a.as_u16x8(), b:b.as_u16x8(), src:src.as_u16x8(), mask:k)) |
1046 | } |
1047 | |
1048 | /// 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). |
1049 | /// |
1050 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_subs_epu16&expand=5786) |
1051 | #[inline ] |
1052 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1053 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1054 | #[cfg_attr (test, assert_instr(vpsubusw))] |
1055 | pub unsafe fn _mm_maskz_subs_epu16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1056 | transmute(src:vpsubusw128( |
1057 | a:a.as_u16x8(), |
1058 | b:b.as_u16x8(), |
1059 | src:_mm_setzero_si128().as_u16x8(), |
1060 | mask:k, |
1061 | )) |
1062 | } |
1063 | |
1064 | /// Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation, and store the results in dst. |
1065 | /// |
1066 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_subs_epu8&expand=5802) |
1067 | #[inline ] |
1068 | #[target_feature (enable = "avx512bw" )] |
1069 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1070 | #[cfg_attr (test, assert_instr(vpsubusb))] |
1071 | pub unsafe fn _mm512_subs_epu8(a: __m512i, b: __m512i) -> __m512i { |
1072 | transmute(src:vpsubusb( |
1073 | a:a.as_u8x64(), |
1074 | b:b.as_u8x64(), |
1075 | src:_mm512_setzero_si512().as_u8x64(), |
1076 | mask:0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
1077 | )) |
1078 | } |
1079 | |
1080 | /// 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). |
1081 | /// |
1082 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_subs_epu8&expand=5800) |
1083 | #[inline ] |
1084 | #[target_feature (enable = "avx512bw" )] |
1085 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1086 | #[cfg_attr (test, assert_instr(vpsubusb))] |
1087 | pub unsafe fn _mm512_mask_subs_epu8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
1088 | transmute(src:vpsubusb(a:a.as_u8x64(), b:b.as_u8x64(), src:src.as_u8x64(), mask:k)) |
1089 | } |
1090 | |
1091 | /// 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). |
1092 | /// |
1093 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_subs_epu8&expand=5801) |
1094 | #[inline ] |
1095 | #[target_feature (enable = "avx512bw" )] |
1096 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1097 | #[cfg_attr (test, assert_instr(vpsubusb))] |
1098 | pub unsafe fn _mm512_maskz_subs_epu8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
1099 | transmute(src:vpsubusb( |
1100 | a:a.as_u8x64(), |
1101 | b:b.as_u8x64(), |
1102 | src:_mm512_setzero_si512().as_u8x64(), |
1103 | mask:k, |
1104 | )) |
1105 | } |
1106 | |
1107 | /// 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). |
1108 | /// |
1109 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_subs_epu8&expand=5797) |
1110 | #[inline ] |
1111 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1112 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1113 | #[cfg_attr (test, assert_instr(vpsubusb))] |
1114 | pub unsafe fn _mm256_mask_subs_epu8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
1115 | transmute(src:vpsubusb256(a:a.as_u8x32(), b:b.as_u8x32(), src:src.as_u8x32(), mask:k)) |
1116 | } |
1117 | |
1118 | /// 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). |
1119 | /// |
1120 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_subs_epu8&expand=5798) |
1121 | #[inline ] |
1122 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1123 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1124 | #[cfg_attr (test, assert_instr(vpsubusb))] |
1125 | pub unsafe fn _mm256_maskz_subs_epu8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
1126 | transmute(src:vpsubusb256( |
1127 | a:a.as_u8x32(), |
1128 | b:b.as_u8x32(), |
1129 | src:_mm256_setzero_si256().as_u8x32(), |
1130 | mask:k, |
1131 | )) |
1132 | } |
1133 | |
1134 | /// 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). |
1135 | /// |
1136 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_subs_epu8&expand=5794) |
1137 | #[inline ] |
1138 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1139 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1140 | #[cfg_attr (test, assert_instr(vpsubusb))] |
1141 | pub unsafe fn _mm_mask_subs_epu8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
1142 | transmute(src:vpsubusb128(a:a.as_u8x16(), b:b.as_u8x16(), src:src.as_u8x16(), mask:k)) |
1143 | } |
1144 | |
1145 | /// 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). |
1146 | /// |
1147 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_subs_epu8&expand=5795) |
1148 | #[inline ] |
1149 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1150 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1151 | #[cfg_attr (test, assert_instr(vpsubusb))] |
1152 | pub unsafe fn _mm_maskz_subs_epu8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
1153 | transmute(src:vpsubusb128( |
1154 | a:a.as_u8x16(), |
1155 | b:b.as_u8x16(), |
1156 | src:_mm_setzero_si128().as_u8x16(), |
1157 | mask:k, |
1158 | )) |
1159 | } |
1160 | |
1161 | /// Subtract packed signed 16-bit integers in b from packed 16-bit integers in a using saturation, and store the results in dst. |
1162 | /// |
1163 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_subs_epi16&expand=5775) |
1164 | #[inline ] |
1165 | #[target_feature (enable = "avx512bw" )] |
1166 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1167 | #[cfg_attr (test, assert_instr(vpsubsw))] |
1168 | pub unsafe fn _mm512_subs_epi16(a: __m512i, b: __m512i) -> __m512i { |
1169 | transmute(src:vpsubsw( |
1170 | a:a.as_i16x32(), |
1171 | b:b.as_i16x32(), |
1172 | src:_mm512_setzero_si512().as_i16x32(), |
1173 | mask:0b11111111_11111111_11111111_11111111, |
1174 | )) |
1175 | } |
1176 | |
1177 | /// 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). |
1178 | /// |
1179 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_subs_epi16&expand=5773) |
1180 | #[inline ] |
1181 | #[target_feature (enable = "avx512bw" )] |
1182 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1183 | #[cfg_attr (test, assert_instr(vpsubsw))] |
1184 | pub unsafe fn _mm512_mask_subs_epi16( |
1185 | src: __m512i, |
1186 | k: __mmask32, |
1187 | a: __m512i, |
1188 | b: __m512i, |
1189 | ) -> __m512i { |
1190 | transmute(src:vpsubsw(a:a.as_i16x32(), b:b.as_i16x32(), src:src.as_i16x32(), mask:k)) |
1191 | } |
1192 | |
1193 | /// 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). |
1194 | /// |
1195 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_subs_epi16&expand=5774) |
1196 | #[inline ] |
1197 | #[target_feature (enable = "avx512bw" )] |
1198 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1199 | #[cfg_attr (test, assert_instr(vpsubsw))] |
1200 | pub unsafe fn _mm512_maskz_subs_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
1201 | transmute(src:vpsubsw( |
1202 | a:a.as_i16x32(), |
1203 | b:b.as_i16x32(), |
1204 | src:_mm512_setzero_si512().as_i16x32(), |
1205 | mask:k, |
1206 | )) |
1207 | } |
1208 | |
1209 | /// 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). |
1210 | /// |
1211 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_subs_epi16&expand=5770) |
1212 | #[inline ] |
1213 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1214 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1215 | #[cfg_attr (test, assert_instr(vpsubsw))] |
1216 | pub unsafe fn _mm256_mask_subs_epi16( |
1217 | src: __m256i, |
1218 | k: __mmask16, |
1219 | a: __m256i, |
1220 | b: __m256i, |
1221 | ) -> __m256i { |
1222 | transmute(src:vpsubsw256(a:a.as_i16x16(), b:b.as_i16x16(), src:src.as_i16x16(), mask:k)) |
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 zeromask k (elements are zeroed out 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=_mm256_maskz_subs_epi16&expand=5771) |
1228 | #[inline ] |
1229 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1230 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1231 | #[cfg_attr (test, assert_instr(vpsubsw))] |
1232 | pub unsafe fn _mm256_maskz_subs_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
1233 | transmute(src:vpsubsw256( |
1234 | a:a.as_i16x16(), |
1235 | b:b.as_i16x16(), |
1236 | src:_mm256_setzero_si256().as_i16x16(), |
1237 | mask:k, |
1238 | )) |
1239 | } |
1240 | |
1241 | /// 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). |
1242 | /// |
1243 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_subs_epi16&expand=5767) |
1244 | #[inline ] |
1245 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1246 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1247 | #[cfg_attr (test, assert_instr(vpsubsw))] |
1248 | pub unsafe fn _mm_mask_subs_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1249 | transmute(src:vpsubsw128(a:a.as_i16x8(), b:b.as_i16x8(), src:src.as_i16x8(), mask:k)) |
1250 | } |
1251 | |
1252 | /// 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). |
1253 | /// |
1254 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_subs_epi16&expand=5768) |
1255 | #[inline ] |
1256 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1257 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1258 | #[cfg_attr (test, assert_instr(vpsubsw))] |
1259 | pub unsafe fn _mm_maskz_subs_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1260 | transmute(src:vpsubsw128( |
1261 | a:a.as_i16x8(), |
1262 | b:b.as_i16x8(), |
1263 | src:_mm_setzero_si128().as_i16x8(), |
1264 | mask:k, |
1265 | )) |
1266 | } |
1267 | |
1268 | /// Subtract packed signed 8-bit integers in b from packed 8-bit integers in a using saturation, and store the results in dst. |
1269 | /// |
1270 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_subs_epi8&expand=5784) |
1271 | #[inline ] |
1272 | #[target_feature (enable = "avx512bw" )] |
1273 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1274 | #[cfg_attr (test, assert_instr(vpsubsb))] |
1275 | pub unsafe fn _mm512_subs_epi8(a: __m512i, b: __m512i) -> __m512i { |
1276 | transmute(src:vpsubsb( |
1277 | a:a.as_i8x64(), |
1278 | b:b.as_i8x64(), |
1279 | src:_mm512_setzero_si512().as_i8x64(), |
1280 | mask:0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
1281 | )) |
1282 | } |
1283 | |
1284 | /// 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). |
1285 | /// |
1286 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_subs_epi8&expand=5782) |
1287 | #[inline ] |
1288 | #[target_feature (enable = "avx512bw" )] |
1289 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1290 | #[cfg_attr (test, assert_instr(vpsubsb))] |
1291 | pub unsafe fn _mm512_mask_subs_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
1292 | transmute(src:vpsubsb(a:a.as_i8x64(), b:b.as_i8x64(), src:src.as_i8x64(), mask:k)) |
1293 | } |
1294 | |
1295 | /// 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). |
1296 | /// |
1297 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_subs_epi8&expand=5783) |
1298 | #[inline ] |
1299 | #[target_feature (enable = "avx512bw" )] |
1300 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1301 | #[cfg_attr (test, assert_instr(vpsubsb))] |
1302 | pub unsafe fn _mm512_maskz_subs_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
1303 | transmute(src:vpsubsb( |
1304 | a:a.as_i8x64(), |
1305 | b:b.as_i8x64(), |
1306 | src:_mm512_setzero_si512().as_i8x64(), |
1307 | mask:k, |
1308 | )) |
1309 | } |
1310 | |
1311 | /// 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). |
1312 | /// |
1313 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_subs_epi8&expand=5779) |
1314 | #[inline ] |
1315 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1316 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1317 | #[cfg_attr (test, assert_instr(vpsubsb))] |
1318 | pub unsafe fn _mm256_mask_subs_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
1319 | transmute(src:vpsubsb256(a:a.as_i8x32(), b:b.as_i8x32(), src:src.as_i8x32(), mask:k)) |
1320 | } |
1321 | |
1322 | /// 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). |
1323 | /// |
1324 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_subs_epi8&expand=5780) |
1325 | #[inline ] |
1326 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1327 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1328 | #[cfg_attr (test, assert_instr(vpsubsb))] |
1329 | pub unsafe fn _mm256_maskz_subs_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
1330 | transmute(src:vpsubsb256( |
1331 | a:a.as_i8x32(), |
1332 | b:b.as_i8x32(), |
1333 | src:_mm256_setzero_si256().as_i8x32(), |
1334 | mask:k, |
1335 | )) |
1336 | } |
1337 | |
1338 | /// 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). |
1339 | /// |
1340 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_subs_epi8&expand=5776) |
1341 | #[inline ] |
1342 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1343 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1344 | #[cfg_attr (test, assert_instr(vpsubsb))] |
1345 | pub unsafe fn _mm_mask_subs_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
1346 | transmute(src:vpsubsb128(a:a.as_i8x16(), b:b.as_i8x16(), src:src.as_i8x16(), mask:k)) |
1347 | } |
1348 | |
1349 | /// 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). |
1350 | /// |
1351 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_subs_epi8&expand=5777) |
1352 | #[inline ] |
1353 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1354 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1355 | #[cfg_attr (test, assert_instr(vpsubsb))] |
1356 | pub unsafe fn _mm_maskz_subs_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
1357 | transmute(src:vpsubsb128( |
1358 | a:a.as_i8x16(), |
1359 | b:b.as_i8x16(), |
1360 | src:_mm_setzero_si128().as_i8x16(), |
1361 | mask:k, |
1362 | )) |
1363 | } |
1364 | |
1365 | /// 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. |
1366 | /// |
1367 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mulhi_epu16&expand=3973) |
1368 | #[inline ] |
1369 | #[target_feature (enable = "avx512bw" )] |
1370 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1371 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
1372 | pub unsafe fn _mm512_mulhi_epu16(a: __m512i, b: __m512i) -> __m512i { |
1373 | transmute(src:vpmulhuw(a:a.as_u16x32(), b:b.as_u16x32())) |
1374 | } |
1375 | |
1376 | /// 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). |
1377 | /// |
1378 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_mulhi_epu16&expand=3971) |
1379 | #[inline ] |
1380 | #[target_feature (enable = "avx512bw" )] |
1381 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1382 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
1383 | pub unsafe fn _mm512_mask_mulhi_epu16( |
1384 | src: __m512i, |
1385 | k: __mmask32, |
1386 | a: __m512i, |
1387 | b: __m512i, |
1388 | ) -> __m512i { |
1389 | let mul: u16x32 = _mm512_mulhi_epu16(a, b).as_u16x32(); |
1390 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_u16x32())) |
1391 | } |
1392 | |
1393 | /// 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). |
1394 | /// |
1395 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_mulhi_epu16&expand=3972) |
1396 | #[inline ] |
1397 | #[target_feature (enable = "avx512bw" )] |
1398 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1399 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
1400 | pub unsafe fn _mm512_maskz_mulhi_epu16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
1401 | let mul: u16x32 = _mm512_mulhi_epu16(a, b).as_u16x32(); |
1402 | let zero: u16x32 = _mm512_setzero_si512().as_u16x32(); |
1403 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:zero)) |
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 writemask k (elements are copied from src 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_mask_mulhi_epu16&expand=3968) |
1409 | #[inline ] |
1410 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1411 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1412 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
1413 | pub unsafe fn _mm256_mask_mulhi_epu16( |
1414 | src: __m256i, |
1415 | k: __mmask16, |
1416 | a: __m256i, |
1417 | b: __m256i, |
1418 | ) -> __m256i { |
1419 | let mul: u16x16 = _mm256_mulhi_epu16(a, b).as_u16x16(); |
1420 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_u16x16())) |
1421 | } |
1422 | |
1423 | /// 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). |
1424 | /// |
1425 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_mulhi_epu16&expand=3969) |
1426 | #[inline ] |
1427 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1428 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1429 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
1430 | pub unsafe fn _mm256_maskz_mulhi_epu16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
1431 | let mul: u16x16 = _mm256_mulhi_epu16(a, b).as_u16x16(); |
1432 | let zero: u16x16 = _mm256_setzero_si256().as_u16x16(); |
1433 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:zero)) |
1434 | } |
1435 | |
1436 | /// 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). |
1437 | /// |
1438 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_mulhi_epu16&expand=3965) |
1439 | #[inline ] |
1440 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1441 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1442 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
1443 | pub unsafe fn _mm_mask_mulhi_epu16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1444 | let mul: u16x8 = _mm_mulhi_epu16(a, b).as_u16x8(); |
1445 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_u16x8())) |
1446 | } |
1447 | |
1448 | /// 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). |
1449 | /// |
1450 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_mulhi_epu16&expand=3966) |
1451 | #[inline ] |
1452 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1453 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1454 | #[cfg_attr (test, assert_instr(vpmulhuw))] |
1455 | pub unsafe fn _mm_maskz_mulhi_epu16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1456 | let mul: u16x8 = _mm_mulhi_epu16(a, b).as_u16x8(); |
1457 | let zero: u16x8 = _mm_setzero_si128().as_u16x8(); |
1458 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:zero)) |
1459 | } |
1460 | |
1461 | /// 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. |
1462 | /// |
1463 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mulhi_epi16&expand=3962) |
1464 | #[inline ] |
1465 | #[target_feature (enable = "avx512bw" )] |
1466 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1467 | #[cfg_attr (test, assert_instr(vpmulhw))] |
1468 | pub unsafe fn _mm512_mulhi_epi16(a: __m512i, b: __m512i) -> __m512i { |
1469 | transmute(src:vpmulhw(a:a.as_i16x32(), b:b.as_i16x32())) |
1470 | } |
1471 | |
1472 | /// 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). |
1473 | /// |
1474 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_mulhi_epi16&expand=3960) |
1475 | #[inline ] |
1476 | #[target_feature (enable = "avx512bw" )] |
1477 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1478 | #[cfg_attr (test, assert_instr(vpmulhw))] |
1479 | pub unsafe fn _mm512_mask_mulhi_epi16( |
1480 | src: __m512i, |
1481 | k: __mmask32, |
1482 | a: __m512i, |
1483 | b: __m512i, |
1484 | ) -> __m512i { |
1485 | let mul: i16x32 = _mm512_mulhi_epi16(a, b).as_i16x32(); |
1486 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x32())) |
1487 | } |
1488 | |
1489 | /// 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). |
1490 | /// |
1491 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_mulhi_epi16&expand=3961) |
1492 | #[inline ] |
1493 | #[target_feature (enable = "avx512bw" )] |
1494 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1495 | #[cfg_attr (test, assert_instr(vpmulhw))] |
1496 | pub unsafe fn _mm512_maskz_mulhi_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
1497 | let mul: i16x32 = _mm512_mulhi_epi16(a, b).as_i16x32(); |
1498 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
1499 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:zero)) |
1500 | } |
1501 | |
1502 | /// 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). |
1503 | /// |
1504 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_mulhi_epi16&expand=3957) |
1505 | #[inline ] |
1506 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1507 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1508 | #[cfg_attr (test, assert_instr(vpmulhw))] |
1509 | pub unsafe fn _mm256_mask_mulhi_epi16( |
1510 | src: __m256i, |
1511 | k: __mmask16, |
1512 | a: __m256i, |
1513 | b: __m256i, |
1514 | ) -> __m256i { |
1515 | let mul: i16x16 = _mm256_mulhi_epi16(a, b).as_i16x16(); |
1516 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x16())) |
1517 | } |
1518 | |
1519 | /// 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). |
1520 | /// |
1521 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_mulhi_epi16&expand=3958) |
1522 | #[inline ] |
1523 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1524 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1525 | #[cfg_attr (test, assert_instr(vpmulhw))] |
1526 | pub unsafe fn _mm256_maskz_mulhi_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
1527 | let mul: i16x16 = _mm256_mulhi_epi16(a, b).as_i16x16(); |
1528 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
1529 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:zero)) |
1530 | } |
1531 | |
1532 | /// 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). |
1533 | /// |
1534 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_mulhi_epi16&expand=3954) |
1535 | #[inline ] |
1536 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1537 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1538 | #[cfg_attr (test, assert_instr(vpmulhw))] |
1539 | pub unsafe fn _mm_mask_mulhi_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1540 | let mul: i16x8 = _mm_mulhi_epi16(a, b).as_i16x8(); |
1541 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x8())) |
1542 | } |
1543 | |
1544 | /// 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). |
1545 | /// |
1546 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_mulhi_epi16&expand=3955) |
1547 | #[inline ] |
1548 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1549 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1550 | #[cfg_attr (test, assert_instr(vpmulhw))] |
1551 | pub unsafe fn _mm_maskz_mulhi_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1552 | let mul: i16x8 = _mm_mulhi_epi16(a, b).as_i16x8(); |
1553 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
1554 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:zero)) |
1555 | } |
1556 | |
1557 | /// 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. |
1558 | /// |
1559 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mulhrs_epi16&expand=3986) |
1560 | #[inline ] |
1561 | #[target_feature (enable = "avx512bw" )] |
1562 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1563 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
1564 | pub unsafe fn _mm512_mulhrs_epi16(a: __m512i, b: __m512i) -> __m512i { |
1565 | transmute(src:vpmulhrsw(a:a.as_i16x32(), b:b.as_i16x32())) |
1566 | } |
1567 | |
1568 | /// 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). |
1569 | /// |
1570 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_mulhrs_epi16&expand=3984) |
1571 | #[inline ] |
1572 | #[target_feature (enable = "avx512bw" )] |
1573 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1574 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
1575 | pub unsafe fn _mm512_mask_mulhrs_epi16( |
1576 | src: __m512i, |
1577 | k: __mmask32, |
1578 | a: __m512i, |
1579 | b: __m512i, |
1580 | ) -> __m512i { |
1581 | let mul: i16x32 = _mm512_mulhrs_epi16(a, b).as_i16x32(); |
1582 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x32())) |
1583 | } |
1584 | |
1585 | /// 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). |
1586 | /// |
1587 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_mulhrs_epi16&expand=3985) |
1588 | #[inline ] |
1589 | #[target_feature (enable = "avx512bw" )] |
1590 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1591 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
1592 | pub unsafe fn _mm512_maskz_mulhrs_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
1593 | let mul: i16x32 = _mm512_mulhrs_epi16(a, b).as_i16x32(); |
1594 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
1595 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:zero)) |
1596 | } |
1597 | |
1598 | /// 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). |
1599 | /// |
1600 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_mulhrs_epi16&expand=3981) |
1601 | #[inline ] |
1602 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1603 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1604 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
1605 | pub unsafe fn _mm256_mask_mulhrs_epi16( |
1606 | src: __m256i, |
1607 | k: __mmask16, |
1608 | a: __m256i, |
1609 | b: __m256i, |
1610 | ) -> __m256i { |
1611 | let mul: i16x16 = _mm256_mulhrs_epi16(a, b).as_i16x16(); |
1612 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x16())) |
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 zeromask k (elements are zeroed out 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=_mm256_maskz_mulhrs_epi16&expand=3982) |
1618 | #[inline ] |
1619 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1620 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1621 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
1622 | pub unsafe fn _mm256_maskz_mulhrs_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
1623 | let mul: i16x16 = _mm256_mulhrs_epi16(a, b).as_i16x16(); |
1624 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
1625 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:zero)) |
1626 | } |
1627 | |
1628 | /// 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). |
1629 | /// |
1630 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_mulhrs_epi16&expand=3978) |
1631 | #[inline ] |
1632 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1633 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1634 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
1635 | pub unsafe fn _mm_mask_mulhrs_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1636 | let mul: i16x8 = _mm_mulhrs_epi16(a, b).as_i16x8(); |
1637 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x8())) |
1638 | } |
1639 | |
1640 | /// 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). |
1641 | /// |
1642 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_mulhrs_epi16&expand=3979) |
1643 | #[inline ] |
1644 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1645 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1646 | #[cfg_attr (test, assert_instr(vpmulhrsw))] |
1647 | pub unsafe fn _mm_maskz_mulhrs_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1648 | let mul: i16x8 = _mm_mulhrs_epi16(a, b).as_i16x8(); |
1649 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
1650 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:zero)) |
1651 | } |
1652 | |
1653 | /// 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. |
1654 | /// |
1655 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mullo_epi16&expand=3996) |
1656 | #[inline ] |
1657 | #[target_feature (enable = "avx512bw" )] |
1658 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1659 | #[cfg_attr (test, assert_instr(vpmullw))] |
1660 | pub unsafe fn _mm512_mullo_epi16(a: __m512i, b: __m512i) -> __m512i { |
1661 | transmute(src:simd_mul(x:a.as_i16x32(), y:b.as_i16x32())) |
1662 | } |
1663 | |
1664 | /// 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). |
1665 | /// |
1666 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_mullo_epi16&expand=3994) |
1667 | #[inline ] |
1668 | #[target_feature (enable = "avx512bw" )] |
1669 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1670 | #[cfg_attr (test, assert_instr(vpmullw))] |
1671 | pub unsafe fn _mm512_mask_mullo_epi16( |
1672 | src: __m512i, |
1673 | k: __mmask32, |
1674 | a: __m512i, |
1675 | b: __m512i, |
1676 | ) -> __m512i { |
1677 | let mul: i16x32 = _mm512_mullo_epi16(a, b).as_i16x32(); |
1678 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x32())) |
1679 | } |
1680 | |
1681 | /// 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). |
1682 | /// |
1683 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_mullo_epi16&expand=3995) |
1684 | #[inline ] |
1685 | #[target_feature (enable = "avx512bw" )] |
1686 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1687 | #[cfg_attr (test, assert_instr(vpmullw))] |
1688 | pub unsafe fn _mm512_maskz_mullo_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
1689 | let mul: i16x32 = _mm512_mullo_epi16(a, b).as_i16x32(); |
1690 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
1691 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:zero)) |
1692 | } |
1693 | |
1694 | /// 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). |
1695 | /// |
1696 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_mullo_epi16&expand=3991) |
1697 | #[inline ] |
1698 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1699 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1700 | #[cfg_attr (test, assert_instr(vpmullw))] |
1701 | pub unsafe fn _mm256_mask_mullo_epi16( |
1702 | src: __m256i, |
1703 | k: __mmask16, |
1704 | a: __m256i, |
1705 | b: __m256i, |
1706 | ) -> __m256i { |
1707 | let mul: i16x16 = _mm256_mullo_epi16(a, b).as_i16x16(); |
1708 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x16())) |
1709 | } |
1710 | |
1711 | /// 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). |
1712 | /// |
1713 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_mullo_epi16&expand=3992) |
1714 | #[inline ] |
1715 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1716 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1717 | #[cfg_attr (test, assert_instr(vpmullw))] |
1718 | pub unsafe fn _mm256_maskz_mullo_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
1719 | let mul: i16x16 = _mm256_mullo_epi16(a, b).as_i16x16(); |
1720 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
1721 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:zero)) |
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 writemask k (elements are copied from src 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_mask_mullo_epi16&expand=3988) |
1727 | #[inline ] |
1728 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1729 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1730 | #[cfg_attr (test, assert_instr(vpmullw))] |
1731 | pub unsafe fn _mm_mask_mullo_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1732 | let mul: i16x8 = _mm_mullo_epi16(a, b).as_i16x8(); |
1733 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:src.as_i16x8())) |
1734 | } |
1735 | |
1736 | /// 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). |
1737 | /// |
1738 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_mullo_epi16&expand=3989) |
1739 | #[inline ] |
1740 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1741 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1742 | #[cfg_attr (test, assert_instr(vpmullw))] |
1743 | pub unsafe fn _mm_maskz_mullo_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1744 | let mul: i16x8 = _mm_mullo_epi16(a, b).as_i16x8(); |
1745 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
1746 | transmute(src:simd_select_bitmask(m:k, yes:mul, no:zero)) |
1747 | } |
1748 | |
1749 | /// Compare packed unsigned 16-bit integers in a and b, and store packed maximum values in dst. |
1750 | /// |
1751 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_max_epu16&expand=3609) |
1752 | #[inline ] |
1753 | #[target_feature (enable = "avx512bw" )] |
1754 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1755 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
1756 | pub unsafe fn _mm512_max_epu16(a: __m512i, b: __m512i) -> __m512i { |
1757 | transmute(src:vpmaxuw(a:a.as_u16x32(), b:b.as_u16x32())) |
1758 | } |
1759 | |
1760 | /// 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). |
1761 | /// |
1762 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_max_epu16&expand=3607) |
1763 | #[inline ] |
1764 | #[target_feature (enable = "avx512bw" )] |
1765 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1766 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
1767 | pub unsafe fn _mm512_mask_max_epu16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
1768 | let max: u16x32 = _mm512_max_epu16(a, b).as_u16x32(); |
1769 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_u16x32())) |
1770 | } |
1771 | |
1772 | /// 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). |
1773 | /// |
1774 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_max_epu16&expand=3608) |
1775 | #[inline ] |
1776 | #[target_feature (enable = "avx512bw" )] |
1777 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1778 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
1779 | pub unsafe fn _mm512_maskz_max_epu16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
1780 | let max: u16x32 = _mm512_max_epu16(a, b).as_u16x32(); |
1781 | let zero: u16x32 = _mm512_setzero_si512().as_u16x32(); |
1782 | transmute(src:simd_select_bitmask(m:k, yes:max, no:zero)) |
1783 | } |
1784 | |
1785 | /// 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). |
1786 | /// |
1787 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_max_epu16&expand=3604) |
1788 | #[inline ] |
1789 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1790 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1791 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
1792 | pub unsafe fn _mm256_mask_max_epu16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
1793 | let max: u16x16 = _mm256_max_epu16(a, b).as_u16x16(); |
1794 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_u16x16())) |
1795 | } |
1796 | |
1797 | /// 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). |
1798 | /// |
1799 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_max_epu16&expand=3605) |
1800 | #[inline ] |
1801 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1802 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1803 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
1804 | pub unsafe fn _mm256_maskz_max_epu16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
1805 | let max: u16x16 = _mm256_max_epu16(a, b).as_u16x16(); |
1806 | let zero: u16x16 = _mm256_setzero_si256().as_u16x16(); |
1807 | transmute(src:simd_select_bitmask(m:k, yes:max, no:zero)) |
1808 | } |
1809 | |
1810 | /// 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). |
1811 | /// |
1812 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_max_epu16&expand=3601) |
1813 | #[inline ] |
1814 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1815 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1816 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
1817 | pub unsafe fn _mm_mask_max_epu16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
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 | /// 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). |
1823 | /// |
1824 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_max_epu16&expand=3602) |
1825 | #[inline ] |
1826 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1827 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1828 | #[cfg_attr (test, assert_instr(vpmaxuw))] |
1829 | pub unsafe fn _mm_maskz_max_epu16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1830 | let max: u16x8 = _mm_max_epu16(a, b).as_u16x8(); |
1831 | let zero: u16x8 = _mm_setzero_si128().as_u16x8(); |
1832 | transmute(src:simd_select_bitmask(m:k, yes:max, no:zero)) |
1833 | } |
1834 | |
1835 | /// Compare packed unsigned 8-bit integers in a and b, and store packed maximum values in dst. |
1836 | /// |
1837 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_max_epu8&expand=3636) |
1838 | #[inline ] |
1839 | #[target_feature (enable = "avx512bw" )] |
1840 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1841 | #[cfg_attr (test, assert_instr(vpmaxub))] |
1842 | pub unsafe fn _mm512_max_epu8(a: __m512i, b: __m512i) -> __m512i { |
1843 | transmute(src:vpmaxub(a:a.as_u8x64(), b:b.as_u8x64())) |
1844 | } |
1845 | |
1846 | /// 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). |
1847 | /// |
1848 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_max_epu8&expand=3634) |
1849 | #[inline ] |
1850 | #[target_feature (enable = "avx512bw" )] |
1851 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1852 | #[cfg_attr (test, assert_instr(vpmaxub))] |
1853 | pub unsafe fn _mm512_mask_max_epu8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
1854 | let max: u8x64 = _mm512_max_epu8(a, b).as_u8x64(); |
1855 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_u8x64())) |
1856 | } |
1857 | |
1858 | /// 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). |
1859 | /// |
1860 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_max_epu8&expand=3635) |
1861 | #[inline ] |
1862 | #[target_feature (enable = "avx512bw" )] |
1863 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1864 | #[cfg_attr (test, assert_instr(vpmaxub))] |
1865 | pub unsafe fn _mm512_maskz_max_epu8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
1866 | let max: u8x64 = _mm512_max_epu8(a, b).as_u8x64(); |
1867 | let zero: u8x64 = _mm512_setzero_si512().as_u8x64(); |
1868 | transmute(src:simd_select_bitmask(m:k, yes:max, no:zero)) |
1869 | } |
1870 | |
1871 | /// 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). |
1872 | /// |
1873 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_max_epu8&expand=3631) |
1874 | #[inline ] |
1875 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1876 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1877 | #[cfg_attr (test, assert_instr(vpmaxub))] |
1878 | pub unsafe fn _mm256_mask_max_epu8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
1879 | let max: u8x32 = _mm256_max_epu8(a, b).as_u8x32(); |
1880 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_u8x32())) |
1881 | } |
1882 | |
1883 | /// 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). |
1884 | /// |
1885 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_max_epu8&expand=3632) |
1886 | #[inline ] |
1887 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1888 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1889 | #[cfg_attr (test, assert_instr(vpmaxub))] |
1890 | pub unsafe fn _mm256_maskz_max_epu8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
1891 | let max: u8x32 = _mm256_max_epu8(a, b).as_u8x32(); |
1892 | let zero: u8x32 = _mm256_setzero_si256().as_u8x32(); |
1893 | transmute(src:simd_select_bitmask(m:k, yes:max, no:zero)) |
1894 | } |
1895 | |
1896 | /// 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). |
1897 | /// |
1898 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_max_epu8&expand=3628) |
1899 | #[inline ] |
1900 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1901 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1902 | #[cfg_attr (test, assert_instr(vpmaxub))] |
1903 | pub unsafe fn _mm_mask_max_epu8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
1904 | let max: u8x16 = _mm_max_epu8(a, b).as_u8x16(); |
1905 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_u8x16())) |
1906 | } |
1907 | |
1908 | /// 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). |
1909 | /// |
1910 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_max_epu8&expand=3629) |
1911 | #[inline ] |
1912 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1913 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1914 | #[cfg_attr (test, assert_instr(vpmaxub))] |
1915 | pub unsafe fn _mm_maskz_max_epu8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
1916 | let max: u8x16 = _mm_max_epu8(a, b).as_u8x16(); |
1917 | let zero: u8x16 = _mm_setzero_si128().as_u8x16(); |
1918 | transmute(src:simd_select_bitmask(m:k, yes:max, no:zero)) |
1919 | } |
1920 | |
1921 | /// Compare packed signed 16-bit integers in a and b, and store packed maximum values in dst. |
1922 | /// |
1923 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_max_epi16&expand=3573) |
1924 | #[inline ] |
1925 | #[target_feature (enable = "avx512bw" )] |
1926 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1927 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
1928 | pub unsafe fn _mm512_max_epi16(a: __m512i, b: __m512i) -> __m512i { |
1929 | transmute(src:vpmaxsw(a:a.as_i16x32(), b:b.as_i16x32())) |
1930 | } |
1931 | |
1932 | /// 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). |
1933 | /// |
1934 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_max_epi16&expand=3571) |
1935 | #[inline ] |
1936 | #[target_feature (enable = "avx512bw" )] |
1937 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1938 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
1939 | pub unsafe fn _mm512_mask_max_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
1940 | let max: i16x32 = _mm512_max_epi16(a, b).as_i16x32(); |
1941 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_i16x32())) |
1942 | } |
1943 | |
1944 | /// 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). |
1945 | /// |
1946 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_max_epi16&expand=3572) |
1947 | #[inline ] |
1948 | #[target_feature (enable = "avx512bw" )] |
1949 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1950 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
1951 | pub unsafe fn _mm512_maskz_max_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
1952 | let max: i16x32 = _mm512_max_epi16(a, b).as_i16x32(); |
1953 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
1954 | transmute(src:simd_select_bitmask(m:k, yes:max, no:zero)) |
1955 | } |
1956 | |
1957 | /// 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). |
1958 | /// |
1959 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_max_epi16&expand=3568) |
1960 | #[inline ] |
1961 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1962 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1963 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
1964 | pub unsafe fn _mm256_mask_max_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
1965 | let max: i16x16 = _mm256_max_epi16(a, b).as_i16x16(); |
1966 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_i16x16())) |
1967 | } |
1968 | |
1969 | /// 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). |
1970 | /// |
1971 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_max_epi16&expand=3569) |
1972 | #[inline ] |
1973 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1974 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1975 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
1976 | pub unsafe fn _mm256_maskz_max_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
1977 | let max: i16x16 = _mm256_max_epi16(a, b).as_i16x16(); |
1978 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
1979 | transmute(src:simd_select_bitmask(m:k, yes:max, no:zero)) |
1980 | } |
1981 | |
1982 | /// 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). |
1983 | /// |
1984 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_max_epi16&expand=3565) |
1985 | #[inline ] |
1986 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1987 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
1988 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
1989 | pub unsafe fn _mm_mask_max_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
1990 | let max: i16x8 = _mm_max_epi16(a, b).as_i16x8(); |
1991 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_i16x8())) |
1992 | } |
1993 | |
1994 | /// 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). |
1995 | /// |
1996 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_max_epi16&expand=3566) |
1997 | #[inline ] |
1998 | #[target_feature (enable = "avx512bw,avx512vl" )] |
1999 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2000 | #[cfg_attr (test, assert_instr(vpmaxsw))] |
2001 | pub unsafe fn _mm_maskz_max_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
2002 | let max: i16x8 = _mm_max_epi16(a, b).as_i16x8(); |
2003 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
2004 | transmute(src:simd_select_bitmask(m:k, yes:max, no:zero)) |
2005 | } |
2006 | |
2007 | /// Compare packed signed 8-bit integers in a and b, and store packed maximum values in dst. |
2008 | /// |
2009 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_max_epi8&expand=3600) |
2010 | #[inline ] |
2011 | #[target_feature (enable = "avx512bw" )] |
2012 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2013 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
2014 | pub unsafe fn _mm512_max_epi8(a: __m512i, b: __m512i) -> __m512i { |
2015 | transmute(src:vpmaxsb(a:a.as_i8x64(), b:b.as_i8x64())) |
2016 | } |
2017 | |
2018 | /// 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). |
2019 | /// |
2020 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_max_epi8&expand=3598) |
2021 | #[inline ] |
2022 | #[target_feature (enable = "avx512bw" )] |
2023 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2024 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
2025 | pub unsafe fn _mm512_mask_max_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
2026 | let max: i8x64 = _mm512_max_epi8(a, b).as_i8x64(); |
2027 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_i8x64())) |
2028 | } |
2029 | |
2030 | /// 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). |
2031 | /// |
2032 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_max_epi8&expand=3599) |
2033 | #[inline ] |
2034 | #[target_feature (enable = "avx512bw" )] |
2035 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2036 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
2037 | pub unsafe fn _mm512_maskz_max_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
2038 | let max: i8x64 = _mm512_max_epi8(a, b).as_i8x64(); |
2039 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
2040 | transmute(src:simd_select_bitmask(m:k, yes:max, no:zero)) |
2041 | } |
2042 | |
2043 | /// 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). |
2044 | /// |
2045 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_max_epi8&expand=3595) |
2046 | #[inline ] |
2047 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2048 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2049 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
2050 | pub unsafe fn _mm256_mask_max_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
2051 | let max: i8x32 = _mm256_max_epi8(a, b).as_i8x32(); |
2052 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_i8x32())) |
2053 | } |
2054 | |
2055 | /// 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). |
2056 | /// |
2057 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_max_epi8&expand=3596) |
2058 | #[inline ] |
2059 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2060 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2061 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
2062 | pub unsafe fn _mm256_maskz_max_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
2063 | let max: i8x32 = _mm256_max_epi8(a, b).as_i8x32(); |
2064 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
2065 | transmute(src:simd_select_bitmask(m:k, yes:max, no:zero)) |
2066 | } |
2067 | |
2068 | /// 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). |
2069 | /// |
2070 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_max_epi8&expand=3592) |
2071 | #[inline ] |
2072 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2073 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2074 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
2075 | pub unsafe fn _mm_mask_max_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
2076 | let max: i8x16 = _mm_max_epi8(a, b).as_i8x16(); |
2077 | transmute(src:simd_select_bitmask(m:k, yes:max, no:src.as_i8x16())) |
2078 | } |
2079 | |
2080 | /// 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). |
2081 | /// |
2082 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_max_epi8&expand=3593) |
2083 | #[inline ] |
2084 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2085 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2086 | #[cfg_attr (test, assert_instr(vpmaxsb))] |
2087 | pub unsafe fn _mm_maskz_max_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
2088 | let max: i8x16 = _mm_max_epi8(a, b).as_i8x16(); |
2089 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
2090 | transmute(src:simd_select_bitmask(m:k, yes:max, no:zero)) |
2091 | } |
2092 | |
2093 | /// Compare packed unsigned 16-bit integers in a and b, and store packed minimum values in dst. |
2094 | /// |
2095 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_min_epu16&expand=3723) |
2096 | #[inline ] |
2097 | #[target_feature (enable = "avx512bw" )] |
2098 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2099 | #[cfg_attr (test, assert_instr(vpminuw))] |
2100 | pub unsafe fn _mm512_min_epu16(a: __m512i, b: __m512i) -> __m512i { |
2101 | transmute(src:vpminuw(a:a.as_u16x32(), b:b.as_u16x32())) |
2102 | } |
2103 | |
2104 | /// 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). |
2105 | /// |
2106 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_min_epu16&expand=3721) |
2107 | #[inline ] |
2108 | #[target_feature (enable = "avx512bw" )] |
2109 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2110 | #[cfg_attr (test, assert_instr(vpminuw))] |
2111 | pub unsafe fn _mm512_mask_min_epu16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
2112 | let min: u16x32 = _mm512_min_epu16(a, b).as_u16x32(); |
2113 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_u16x32())) |
2114 | } |
2115 | |
2116 | /// 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). |
2117 | /// |
2118 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_min_epu16&expand=3722) |
2119 | #[inline ] |
2120 | #[target_feature (enable = "avx512bw" )] |
2121 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2122 | #[cfg_attr (test, assert_instr(vpminuw))] |
2123 | pub unsafe fn _mm512_maskz_min_epu16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
2124 | let min: u16x32 = _mm512_min_epu16(a, b).as_u16x32(); |
2125 | let zero: u16x32 = _mm512_setzero_si512().as_u16x32(); |
2126 | transmute(src:simd_select_bitmask(m:k, yes:min, no:zero)) |
2127 | } |
2128 | |
2129 | /// 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). |
2130 | /// |
2131 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_min_epu16&expand=3718) |
2132 | #[inline ] |
2133 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2134 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2135 | #[cfg_attr (test, assert_instr(vpminuw))] |
2136 | pub unsafe fn _mm256_mask_min_epu16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
2137 | let min: u16x16 = _mm256_min_epu16(a, b).as_u16x16(); |
2138 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_u16x16())) |
2139 | } |
2140 | |
2141 | /// 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). |
2142 | /// |
2143 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_min_epu16&expand=3719) |
2144 | #[inline ] |
2145 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2146 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2147 | #[cfg_attr (test, assert_instr(vpminuw))] |
2148 | pub unsafe fn _mm256_maskz_min_epu16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
2149 | let min: u16x16 = _mm256_min_epu16(a, b).as_u16x16(); |
2150 | let zero: u16x16 = _mm256_setzero_si256().as_u16x16(); |
2151 | transmute(src:simd_select_bitmask(m:k, yes:min, no:zero)) |
2152 | } |
2153 | |
2154 | /// 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). |
2155 | /// |
2156 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_min_epu16&expand=3715) |
2157 | #[inline ] |
2158 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2159 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2160 | #[cfg_attr (test, assert_instr(vpminuw))] |
2161 | pub unsafe fn _mm_mask_min_epu16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
2162 | let min: u16x8 = _mm_min_epu16(a, b).as_u16x8(); |
2163 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_u16x8())) |
2164 | } |
2165 | |
2166 | /// 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). |
2167 | /// |
2168 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_min_epu16&expand=3716) |
2169 | #[inline ] |
2170 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2171 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2172 | #[cfg_attr (test, assert_instr(vpminuw))] |
2173 | pub unsafe fn _mm_maskz_min_epu16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
2174 | let min: u16x8 = _mm_min_epu16(a, b).as_u16x8(); |
2175 | let zero: u16x8 = _mm_setzero_si128().as_u16x8(); |
2176 | transmute(src:simd_select_bitmask(m:k, yes:min, no:zero)) |
2177 | } |
2178 | |
2179 | /// Compare packed unsigned 8-bit integers in a and b, and store packed minimum values in dst. |
2180 | /// |
2181 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_min_epu8&expand=3750) |
2182 | #[inline ] |
2183 | #[target_feature (enable = "avx512bw" )] |
2184 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2185 | #[cfg_attr (test, assert_instr(vpminub))] |
2186 | pub unsafe fn _mm512_min_epu8(a: __m512i, b: __m512i) -> __m512i { |
2187 | transmute(src:vpminub(a:a.as_u8x64(), b:b.as_u8x64())) |
2188 | } |
2189 | |
2190 | /// 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). |
2191 | /// |
2192 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_min_epu8&expand=3748) |
2193 | #[inline ] |
2194 | #[target_feature (enable = "avx512bw" )] |
2195 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2196 | #[cfg_attr (test, assert_instr(vpminub))] |
2197 | pub unsafe fn _mm512_mask_min_epu8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
2198 | let min: u8x64 = _mm512_min_epu8(a, b).as_u8x64(); |
2199 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_u8x64())) |
2200 | } |
2201 | |
2202 | /// 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). |
2203 | /// |
2204 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_min_epu8&expand=3749) |
2205 | #[inline ] |
2206 | #[target_feature (enable = "avx512bw" )] |
2207 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2208 | #[cfg_attr (test, assert_instr(vpminub))] |
2209 | pub unsafe fn _mm512_maskz_min_epu8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
2210 | let min: u8x64 = _mm512_min_epu8(a, b).as_u8x64(); |
2211 | let zero: u8x64 = _mm512_setzero_si512().as_u8x64(); |
2212 | transmute(src:simd_select_bitmask(m:k, yes:min, no:zero)) |
2213 | } |
2214 | |
2215 | /// 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). |
2216 | /// |
2217 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_min_epu8&expand=3745) |
2218 | #[inline ] |
2219 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2220 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2221 | #[cfg_attr (test, assert_instr(vpminub))] |
2222 | pub unsafe fn _mm256_mask_min_epu8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
2223 | let min: u8x32 = _mm256_min_epu8(a, b).as_u8x32(); |
2224 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_u8x32())) |
2225 | } |
2226 | |
2227 | /// 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). |
2228 | /// |
2229 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_min_epu8&expand=3746) |
2230 | #[inline ] |
2231 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2232 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2233 | #[cfg_attr (test, assert_instr(vpminub))] |
2234 | pub unsafe fn _mm256_maskz_min_epu8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
2235 | let min: u8x32 = _mm256_min_epu8(a, b).as_u8x32(); |
2236 | let zero: u8x32 = _mm256_setzero_si256().as_u8x32(); |
2237 | transmute(src:simd_select_bitmask(m:k, yes:min, no:zero)) |
2238 | } |
2239 | |
2240 | /// 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). |
2241 | /// |
2242 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_min_epu8&expand=3742) |
2243 | #[inline ] |
2244 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2245 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2246 | #[cfg_attr (test, assert_instr(vpminub))] |
2247 | pub unsafe fn _mm_mask_min_epu8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
2248 | let min: u8x16 = _mm_min_epu8(a, b).as_u8x16(); |
2249 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_u8x16())) |
2250 | } |
2251 | |
2252 | /// 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). |
2253 | /// |
2254 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_min_epu8&expand=3743) |
2255 | #[inline ] |
2256 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2257 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2258 | #[cfg_attr (test, assert_instr(vpminub))] |
2259 | pub unsafe fn _mm_maskz_min_epu8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
2260 | let min: u8x16 = _mm_min_epu8(a, b).as_u8x16(); |
2261 | let zero: u8x16 = _mm_setzero_si128().as_u8x16(); |
2262 | transmute(src:simd_select_bitmask(m:k, yes:min, no:zero)) |
2263 | } |
2264 | |
2265 | /// Compare packed signed 16-bit integers in a and b, and store packed minimum values in dst. |
2266 | /// |
2267 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_min_epi16&expand=3687) |
2268 | #[inline ] |
2269 | #[target_feature (enable = "avx512bw" )] |
2270 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2271 | #[cfg_attr (test, assert_instr(vpminsw))] |
2272 | pub unsafe fn _mm512_min_epi16(a: __m512i, b: __m512i) -> __m512i { |
2273 | transmute(src:vpminsw(a:a.as_i16x32(), b:b.as_i16x32())) |
2274 | } |
2275 | |
2276 | /// 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). |
2277 | /// |
2278 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_min_epi16&expand=3685) |
2279 | #[inline ] |
2280 | #[target_feature (enable = "avx512bw" )] |
2281 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2282 | #[cfg_attr (test, assert_instr(vpminsw))] |
2283 | pub unsafe fn _mm512_mask_min_epi16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
2284 | let min: i16x32 = _mm512_min_epi16(a, b).as_i16x32(); |
2285 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_i16x32())) |
2286 | } |
2287 | |
2288 | /// 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). |
2289 | /// |
2290 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_min_epi16&expand=3686) |
2291 | #[inline ] |
2292 | #[target_feature (enable = "avx512bw" )] |
2293 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2294 | #[cfg_attr (test, assert_instr(vpminsw))] |
2295 | pub unsafe fn _mm512_maskz_min_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
2296 | let min: i16x32 = _mm512_min_epi16(a, b).as_i16x32(); |
2297 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
2298 | transmute(src:simd_select_bitmask(m:k, yes:min, no:zero)) |
2299 | } |
2300 | |
2301 | /// 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). |
2302 | /// |
2303 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_min_epi16&expand=3682) |
2304 | #[inline ] |
2305 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2306 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2307 | #[cfg_attr (test, assert_instr(vpminsw))] |
2308 | pub unsafe fn _mm256_mask_min_epi16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
2309 | let min: i16x16 = _mm256_min_epi16(a, b).as_i16x16(); |
2310 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_i16x16())) |
2311 | } |
2312 | |
2313 | /// 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). |
2314 | /// |
2315 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_min_epi16&expand=3683) |
2316 | #[inline ] |
2317 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2318 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2319 | #[cfg_attr (test, assert_instr(vpminsw))] |
2320 | pub unsafe fn _mm256_maskz_min_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
2321 | let min: i16x16 = _mm256_min_epi16(a, b).as_i16x16(); |
2322 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
2323 | transmute(src:simd_select_bitmask(m:k, yes:min, no:zero)) |
2324 | } |
2325 | |
2326 | /// 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). |
2327 | /// |
2328 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_min_epi16&expand=3679) |
2329 | #[inline ] |
2330 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2331 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2332 | #[cfg_attr (test, assert_instr(vpminsw))] |
2333 | pub unsafe fn _mm_mask_min_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
2334 | let min: i16x8 = _mm_min_epi16(a, b).as_i16x8(); |
2335 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_i16x8())) |
2336 | } |
2337 | |
2338 | /// 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). |
2339 | /// |
2340 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_min_epi16&expand=3680) |
2341 | #[inline ] |
2342 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2343 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2344 | #[cfg_attr (test, assert_instr(vpminsw))] |
2345 | pub unsafe fn _mm_maskz_min_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
2346 | let min: i16x8 = _mm_min_epi16(a, b).as_i16x8(); |
2347 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
2348 | transmute(src:simd_select_bitmask(m:k, yes:min, no:zero)) |
2349 | } |
2350 | |
2351 | /// Compare packed signed 8-bit integers in a and b, and store packed minimum values in dst. |
2352 | /// |
2353 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_min_epi8&expand=3714) |
2354 | #[inline ] |
2355 | #[target_feature (enable = "avx512bw" )] |
2356 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2357 | #[cfg_attr (test, assert_instr(vpminsb))] |
2358 | pub unsafe fn _mm512_min_epi8(a: __m512i, b: __m512i) -> __m512i { |
2359 | transmute(src:vpminsb(a:a.as_i8x64(), b:b.as_i8x64())) |
2360 | } |
2361 | |
2362 | /// 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). |
2363 | /// |
2364 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_min_epi8&expand=3712) |
2365 | #[inline ] |
2366 | #[target_feature (enable = "avx512bw" )] |
2367 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2368 | #[cfg_attr (test, assert_instr(vpminsb))] |
2369 | pub unsafe fn _mm512_mask_min_epi8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
2370 | let min: i8x64 = _mm512_min_epi8(a, b).as_i8x64(); |
2371 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_i8x64())) |
2372 | } |
2373 | |
2374 | /// 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). |
2375 | /// |
2376 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_min_epi8&expand=3713) |
2377 | #[inline ] |
2378 | #[target_feature (enable = "avx512bw" )] |
2379 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2380 | #[cfg_attr (test, assert_instr(vpminsb))] |
2381 | pub unsafe fn _mm512_maskz_min_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
2382 | let min: i8x64 = _mm512_min_epi8(a, b).as_i8x64(); |
2383 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
2384 | transmute(src:simd_select_bitmask(m:k, yes:min, no:zero)) |
2385 | } |
2386 | |
2387 | /// 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). |
2388 | /// |
2389 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_min_epi8&expand=3709) |
2390 | #[inline ] |
2391 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2392 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2393 | #[cfg_attr (test, assert_instr(vpminsb))] |
2394 | pub unsafe fn _mm256_mask_min_epi8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
2395 | let min: i8x32 = _mm256_min_epi8(a, b).as_i8x32(); |
2396 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_i8x32())) |
2397 | } |
2398 | |
2399 | /// 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). |
2400 | /// |
2401 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_min_epi8&expand=3710) |
2402 | #[inline ] |
2403 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2404 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2405 | #[cfg_attr (test, assert_instr(vpminsb))] |
2406 | pub unsafe fn _mm256_maskz_min_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
2407 | let min: i8x32 = _mm256_min_epi8(a, b).as_i8x32(); |
2408 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
2409 | transmute(src:simd_select_bitmask(m:k, yes:min, no:zero)) |
2410 | } |
2411 | |
2412 | /// 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). |
2413 | /// |
2414 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_min_epi8&expand=3706) |
2415 | #[inline ] |
2416 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2417 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2418 | #[cfg_attr (test, assert_instr(vpminsb))] |
2419 | pub unsafe fn _mm_mask_min_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
2420 | let min: i8x16 = _mm_min_epi8(a, b).as_i8x16(); |
2421 | transmute(src:simd_select_bitmask(m:k, yes:min, no:src.as_i8x16())) |
2422 | } |
2423 | |
2424 | /// 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). |
2425 | /// |
2426 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_min_epi8&expand=3707) |
2427 | #[inline ] |
2428 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2429 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2430 | #[cfg_attr (test, assert_instr(vpminsb))] |
2431 | pub unsafe fn _mm_maskz_min_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
2432 | let min: i8x16 = _mm_min_epi8(a, b).as_i8x16(); |
2433 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
2434 | transmute(src:simd_select_bitmask(m:k, yes:min, no:zero)) |
2435 | } |
2436 | |
2437 | /// Compare packed unsigned 16-bit integers in a and b for less-than, and store the results in mask vector k. |
2438 | /// |
2439 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=512_cmplt_epu16_mask&expand=1050) |
2440 | #[inline ] |
2441 | #[target_feature (enable = "avx512bw" )] |
2442 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2443 | #[cfg_attr (test, assert_instr(vpcmp))] |
2444 | pub unsafe fn _mm512_cmplt_epu16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
2445 | simd_bitmask::<u16x32, _>(simd_lt(x:a.as_u16x32(), y:b.as_u16x32())) |
2446 | } |
2447 | |
2448 | /// 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). |
2449 | /// |
2450 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmplt_epu16_mask&expand=1051) |
2451 | #[inline ] |
2452 | #[target_feature (enable = "avx512bw" )] |
2453 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2454 | #[cfg_attr (test, assert_instr(vpcmp))] |
2455 | pub unsafe fn _mm512_mask_cmplt_epu16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
2456 | _mm512_cmplt_epu16_mask(a, b) & k1 |
2457 | } |
2458 | |
2459 | /// Compare packed unsigned 16-bit integers in a and b for less-than, and store the results in mask vector k. |
2460 | /// |
2461 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=512_cmplt_epu16_mask&expand=1050) |
2462 | #[inline ] |
2463 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2464 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2465 | #[cfg_attr (test, assert_instr(vpcmp))] |
2466 | pub unsafe fn _mm256_cmplt_epu16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
2467 | simd_bitmask::<u16x16, _>(simd_lt(x:a.as_u16x16(), y:b.as_u16x16())) |
2468 | } |
2469 | |
2470 | /// 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). |
2471 | /// |
2472 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmplt_epu16_mask&expand=1049) |
2473 | #[inline ] |
2474 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2475 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2476 | #[cfg_attr (test, assert_instr(vpcmp))] |
2477 | pub unsafe fn _mm256_mask_cmplt_epu16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
2478 | _mm256_cmplt_epu16_mask(a, b) & k1 |
2479 | } |
2480 | |
2481 | /// Compare packed unsigned 16-bit integers in a and b for less-than, and store the results in mask vector k. |
2482 | /// |
2483 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmplt_epi16_mask&expand=1018) |
2484 | #[inline ] |
2485 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2486 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2487 | #[cfg_attr (test, assert_instr(vpcmp))] |
2488 | pub unsafe fn _mm_cmplt_epu16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
2489 | simd_bitmask::<u16x8, _>(simd_lt(x:a.as_u16x8(), y:b.as_u16x8())) |
2490 | } |
2491 | |
2492 | /// 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). |
2493 | /// |
2494 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmplt_epi16_mask&expand=1019) |
2495 | #[inline ] |
2496 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2497 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2498 | #[cfg_attr (test, assert_instr(vpcmp))] |
2499 | pub unsafe fn _mm_mask_cmplt_epu16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
2500 | _mm_cmplt_epu16_mask(a, b) & k1 |
2501 | } |
2502 | |
2503 | /// Compare packed unsigned 8-bit integers in a and b for less-than, and store the results in mask vector k. |
2504 | /// |
2505 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=mm512_cmplt_epu8_mask&expand=1068) |
2506 | #[inline ] |
2507 | #[target_feature (enable = "avx512bw" )] |
2508 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2509 | #[cfg_attr (test, assert_instr(vpcmp))] |
2510 | pub unsafe fn _mm512_cmplt_epu8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
2511 | simd_bitmask::<u8x64, _>(simd_lt(x:a.as_u8x64(), y:b.as_u8x64())) |
2512 | } |
2513 | |
2514 | /// 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). |
2515 | /// |
2516 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmplt_epu8_mask&expand=1069) |
2517 | #[inline ] |
2518 | #[target_feature (enable = "avx512bw" )] |
2519 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2520 | #[cfg_attr (test, assert_instr(vpcmp))] |
2521 | pub unsafe fn _mm512_mask_cmplt_epu8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
2522 | _mm512_cmplt_epu8_mask(a, b) & k1 |
2523 | } |
2524 | |
2525 | /// Compare packed unsigned 8-bit integers in a and b for less-than, and store the results in mask vector k. |
2526 | /// |
2527 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmplt_epu8_mask&expand=1066) |
2528 | #[inline ] |
2529 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2530 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2531 | #[cfg_attr (test, assert_instr(vpcmp))] |
2532 | pub unsafe fn _mm256_cmplt_epu8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
2533 | simd_bitmask::<u8x32, _>(simd_lt(x:a.as_u8x32(), y:b.as_u8x32())) |
2534 | } |
2535 | |
2536 | /// 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). |
2537 | /// |
2538 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmplt_epu8_mask&expand=1067) |
2539 | #[inline ] |
2540 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2541 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2542 | #[cfg_attr (test, assert_instr(vpcmp))] |
2543 | pub unsafe fn _mm256_mask_cmplt_epu8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
2544 | _mm256_cmplt_epu8_mask(a, b) & k1 |
2545 | } |
2546 | |
2547 | /// Compare packed unsigned 8-bit integers in a and b for less-than, and store the results in mask vector k. |
2548 | /// |
2549 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmplt_epu8_mask&expand=1064) |
2550 | #[inline ] |
2551 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2552 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2553 | #[cfg_attr (test, assert_instr(vpcmp))] |
2554 | pub unsafe fn _mm_cmplt_epu8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
2555 | simd_bitmask::<u8x16, _>(simd_lt(x:a.as_u8x16(), y:b.as_u8x16())) |
2556 | } |
2557 | |
2558 | /// 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). |
2559 | /// |
2560 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmplt_epu8_mask&expand=1065) |
2561 | #[inline ] |
2562 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2563 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2564 | #[cfg_attr (test, assert_instr(vpcmp))] |
2565 | pub unsafe fn _mm_mask_cmplt_epu8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
2566 | _mm_cmplt_epu8_mask(a, b) & k1 |
2567 | } |
2568 | |
2569 | /// Compare packed signed 16-bit integers in a and b for less-than, and store the results in mask vector k. |
2570 | /// |
2571 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmplt_epi16_mask&expand=1022) |
2572 | #[inline ] |
2573 | #[target_feature (enable = "avx512bw" )] |
2574 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2575 | #[cfg_attr (test, assert_instr(vpcmp))] |
2576 | pub unsafe fn _mm512_cmplt_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
2577 | simd_bitmask::<i16x32, _>(simd_lt(x:a.as_i16x32(), y:b.as_i16x32())) |
2578 | } |
2579 | |
2580 | /// 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). |
2581 | /// |
2582 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmplt_epi16_mask&expand=1023) |
2583 | #[inline ] |
2584 | #[target_feature (enable = "avx512bw" )] |
2585 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2586 | #[cfg_attr (test, assert_instr(vpcmp))] |
2587 | pub unsafe fn _mm512_mask_cmplt_epi16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
2588 | _mm512_cmplt_epi16_mask(a, b) & k1 |
2589 | } |
2590 | |
2591 | /// Compare packed signed 16-bit integers in a and b for less-than, and store the results in mask vector k. |
2592 | /// |
2593 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmplt_epi16_mask&expand=1020) |
2594 | #[inline ] |
2595 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2596 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2597 | #[cfg_attr (test, assert_instr(vpcmp))] |
2598 | pub unsafe fn _mm256_cmplt_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
2599 | simd_bitmask::<i16x16, _>(simd_lt(x:a.as_i16x16(), y:b.as_i16x16())) |
2600 | } |
2601 | |
2602 | /// 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). |
2603 | /// |
2604 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmplt_epi16_mask&expand=1021) |
2605 | #[inline ] |
2606 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2607 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2608 | #[cfg_attr (test, assert_instr(vpcmp))] |
2609 | pub unsafe fn _mm256_mask_cmplt_epi16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
2610 | _mm256_cmplt_epi16_mask(a, b) & k1 |
2611 | } |
2612 | |
2613 | /// Compare packed signed 16-bit integers in a and b for less-than, and store the results in mask vector k. |
2614 | /// |
2615 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmplt_epi16_mask&expand=1018) |
2616 | #[inline ] |
2617 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2618 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2619 | #[cfg_attr (test, assert_instr(vpcmp))] |
2620 | pub unsafe fn _mm_cmplt_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
2621 | simd_bitmask::<i16x8, _>(simd_lt(x:a.as_i16x8(), y:b.as_i16x8())) |
2622 | } |
2623 | |
2624 | /// 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). |
2625 | /// |
2626 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmplt_epi16_mask&expand=1019) |
2627 | #[inline ] |
2628 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2629 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2630 | #[cfg_attr (test, assert_instr(vpcmp))] |
2631 | pub unsafe fn _mm_mask_cmplt_epi16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
2632 | _mm_cmplt_epi16_mask(a, b) & k1 |
2633 | } |
2634 | |
2635 | /// Compare packed signed 8-bit integers in a and b for less-than, and store the results in mask vector k. |
2636 | /// |
2637 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmplt_epi8_mask&expand=1044) |
2638 | #[inline ] |
2639 | #[target_feature (enable = "avx512bw" )] |
2640 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2641 | #[cfg_attr (test, assert_instr(vpcmp))] |
2642 | pub unsafe fn _mm512_cmplt_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
2643 | simd_bitmask::<i8x64, _>(simd_lt(x:a.as_i8x64(), y:b.as_i8x64())) |
2644 | } |
2645 | |
2646 | /// 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). |
2647 | /// |
2648 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmplt_epi8_mask&expand=1045) |
2649 | #[inline ] |
2650 | #[target_feature (enable = "avx512bw" )] |
2651 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2652 | #[cfg_attr (test, assert_instr(vpcmp))] |
2653 | pub unsafe fn _mm512_mask_cmplt_epi8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
2654 | _mm512_cmplt_epi8_mask(a, b) & k1 |
2655 | } |
2656 | |
2657 | /// Compare packed signed 8-bit integers in a and b for less-than, and store the results in mask vector k. |
2658 | /// |
2659 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmplt_epi8_mask&expand=1042) |
2660 | #[inline ] |
2661 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2662 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2663 | #[cfg_attr (test, assert_instr(vpcmp))] |
2664 | pub unsafe fn _mm256_cmplt_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
2665 | simd_bitmask::<i8x32, _>(simd_lt(x:a.as_i8x32(), y:b.as_i8x32())) |
2666 | } |
2667 | |
2668 | /// 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). |
2669 | /// |
2670 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmplt_epi8_mask&expand=1043) |
2671 | #[inline ] |
2672 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2673 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2674 | #[cfg_attr (test, assert_instr(vpcmp))] |
2675 | pub unsafe fn _mm256_mask_cmplt_epi8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
2676 | _mm256_cmplt_epi8_mask(a, b) & k1 |
2677 | } |
2678 | |
2679 | /// Compare packed signed 8-bit integers in a and b for less-than, and store the results in mask vector k. |
2680 | /// |
2681 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmplt_epi8_mask&expand=1040) |
2682 | #[inline ] |
2683 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2684 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2685 | #[cfg_attr (test, assert_instr(vpcmp))] |
2686 | pub unsafe fn _mm_cmplt_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
2687 | simd_bitmask::<i8x16, _>(simd_lt(x:a.as_i8x16(), y:b.as_i8x16())) |
2688 | } |
2689 | |
2690 | /// 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). |
2691 | /// |
2692 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmplt_epi8_mask&expand=1041) |
2693 | #[inline ] |
2694 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2695 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2696 | #[cfg_attr (test, assert_instr(vpcmp))] |
2697 | pub unsafe fn _mm_mask_cmplt_epi8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
2698 | _mm_cmplt_epi8_mask(a, b) & k1 |
2699 | } |
2700 | |
2701 | /// Compare packed unsigned 16-bit integers in a and b for greater-than, and store the results in mask vector k. |
2702 | /// |
2703 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpgt_epu16_mask&expand=927) |
2704 | #[inline ] |
2705 | #[target_feature (enable = "avx512bw" )] |
2706 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2707 | #[cfg_attr (test, assert_instr(vpcmp))] |
2708 | pub unsafe fn _mm512_cmpgt_epu16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
2709 | simd_bitmask::<u16x32, _>(simd_gt(x:a.as_u16x32(), y:b.as_u16x32())) |
2710 | } |
2711 | |
2712 | /// 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). |
2713 | /// |
2714 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpgt_epu16_mask&expand=928) |
2715 | #[inline ] |
2716 | #[target_feature (enable = "avx512bw" )] |
2717 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2718 | #[cfg_attr (test, assert_instr(vpcmp))] |
2719 | pub unsafe fn _mm512_mask_cmpgt_epu16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
2720 | _mm512_cmpgt_epu16_mask(a, b) & k1 |
2721 | } |
2722 | |
2723 | /// Compare packed unsigned 16-bit integers in a and b for greater-than, and store the results in mask vector k. |
2724 | /// |
2725 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpgt_epu16_mask&expand=925) |
2726 | #[inline ] |
2727 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2728 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2729 | #[cfg_attr (test, assert_instr(vpcmp))] |
2730 | pub unsafe fn _mm256_cmpgt_epu16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
2731 | simd_bitmask::<u16x16, _>(simd_gt(x:a.as_u16x16(), y:b.as_u16x16())) |
2732 | } |
2733 | |
2734 | /// 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). |
2735 | /// |
2736 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpgt_epu16_mask&expand=926) |
2737 | #[inline ] |
2738 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2739 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2740 | #[cfg_attr (test, assert_instr(vpcmp))] |
2741 | pub unsafe fn _mm256_mask_cmpgt_epu16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
2742 | _mm256_cmpgt_epu16_mask(a, b) & k1 |
2743 | } |
2744 | |
2745 | /// Compare packed unsigned 16-bit integers in a and b for greater-than, and store the results in mask vector k. |
2746 | /// |
2747 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpgt_epu16_mask&expand=923) |
2748 | #[inline ] |
2749 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2750 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2751 | #[cfg_attr (test, assert_instr(vpcmp))] |
2752 | pub unsafe fn _mm_cmpgt_epu16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
2753 | simd_bitmask::<u16x8, _>(simd_gt(x:a.as_u16x8(), y:b.as_u16x8())) |
2754 | } |
2755 | |
2756 | /// 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). |
2757 | /// |
2758 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpgt_epu16_mask&expand=924) |
2759 | #[inline ] |
2760 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2761 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2762 | #[cfg_attr (test, assert_instr(vpcmp))] |
2763 | pub unsafe fn _mm_mask_cmpgt_epu16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
2764 | _mm_cmpgt_epu16_mask(a, b) & k1 |
2765 | } |
2766 | |
2767 | /// Compare packed unsigned 8-bit integers in a and b for greater-than, and store the results in mask vector k. |
2768 | /// |
2769 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpgt_epu8_mask&expand=945) |
2770 | #[inline ] |
2771 | #[target_feature (enable = "avx512bw" )] |
2772 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2773 | #[cfg_attr (test, assert_instr(vpcmp))] |
2774 | pub unsafe fn _mm512_cmpgt_epu8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
2775 | simd_bitmask::<u8x64, _>(simd_gt(x:a.as_u8x64(), y:b.as_u8x64())) |
2776 | } |
2777 | |
2778 | /// 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). |
2779 | /// |
2780 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpgt_epu8_mask&expand=946) |
2781 | #[inline ] |
2782 | #[target_feature (enable = "avx512bw" )] |
2783 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2784 | #[cfg_attr (test, assert_instr(vpcmp))] |
2785 | pub unsafe fn _mm512_mask_cmpgt_epu8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
2786 | _mm512_cmpgt_epu8_mask(a, b) & k1 |
2787 | } |
2788 | |
2789 | /// Compare packed unsigned 8-bit integers in a and b for greater-than, and store the results in mask vector k. |
2790 | /// |
2791 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpgt_epu8_mask&expand=943) |
2792 | #[inline ] |
2793 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2794 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2795 | #[cfg_attr (test, assert_instr(vpcmp))] |
2796 | pub unsafe fn _mm256_cmpgt_epu8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
2797 | simd_bitmask::<u8x32, _>(simd_gt(x:a.as_u8x32(), y:b.as_u8x32())) |
2798 | } |
2799 | |
2800 | /// 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). |
2801 | /// |
2802 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpgt_epu8_mask&expand=944) |
2803 | #[inline ] |
2804 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2805 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2806 | #[cfg_attr (test, assert_instr(vpcmp))] |
2807 | pub unsafe fn _mm256_mask_cmpgt_epu8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
2808 | _mm256_cmpgt_epu8_mask(a, b) & k1 |
2809 | } |
2810 | |
2811 | /// Compare packed unsigned 8-bit integers in a and b for greater-than, and store the results in mask vector k. |
2812 | /// |
2813 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpgt_epu8_mask&expand=941) |
2814 | #[inline ] |
2815 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2816 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2817 | #[cfg_attr (test, assert_instr(vpcmp))] |
2818 | pub unsafe fn _mm_cmpgt_epu8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
2819 | simd_bitmask::<u8x16, _>(simd_gt(x:a.as_u8x16(), y:b.as_u8x16())) |
2820 | } |
2821 | |
2822 | /// 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). |
2823 | /// |
2824 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpgt_epu8_mask&expand=942) |
2825 | #[inline ] |
2826 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2827 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2828 | #[cfg_attr (test, assert_instr(vpcmp))] |
2829 | pub unsafe fn _mm_mask_cmpgt_epu8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
2830 | _mm_cmpgt_epu8_mask(a, b) & k1 |
2831 | } |
2832 | |
2833 | /// Compare packed signed 16-bit integers in a and b for greater-than, and store the results in mask vector k. |
2834 | /// |
2835 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpgt_epi16_mask&expand=897) |
2836 | #[inline ] |
2837 | #[target_feature (enable = "avx512bw" )] |
2838 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2839 | #[cfg_attr (test, assert_instr(vpcmp))] |
2840 | pub unsafe fn _mm512_cmpgt_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
2841 | simd_bitmask::<i16x32, _>(simd_gt(x:a.as_i16x32(), y:b.as_i16x32())) |
2842 | } |
2843 | |
2844 | /// 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). |
2845 | /// |
2846 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpgt_epi16_mask&expand=898) |
2847 | #[inline ] |
2848 | #[target_feature (enable = "avx512bw" )] |
2849 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2850 | #[cfg_attr (test, assert_instr(vpcmp))] |
2851 | pub unsafe fn _mm512_mask_cmpgt_epi16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
2852 | _mm512_cmpgt_epi16_mask(a, b) & k1 |
2853 | } |
2854 | |
2855 | /// Compare packed signed 16-bit integers in a and b for greater-than, and store the results in mask vector k. |
2856 | /// |
2857 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpgt_epi16_mask&expand=895) |
2858 | #[inline ] |
2859 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2860 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2861 | #[cfg_attr (test, assert_instr(vpcmp))] |
2862 | pub unsafe fn _mm256_cmpgt_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
2863 | simd_bitmask::<i16x16, _>(simd_gt(x:a.as_i16x16(), y:b.as_i16x16())) |
2864 | } |
2865 | |
2866 | /// 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). |
2867 | /// |
2868 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpgt_epi16_mask&expand=896) |
2869 | #[inline ] |
2870 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2871 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2872 | #[cfg_attr (test, assert_instr(vpcmp))] |
2873 | pub unsafe fn _mm256_mask_cmpgt_epi16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
2874 | _mm256_cmpgt_epi16_mask(a, b) & k1 |
2875 | } |
2876 | |
2877 | /// Compare packed signed 16-bit integers in a and b for greater-than, and store the results in mask vector k. |
2878 | /// |
2879 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpgt_epi16_mask&expand=893) |
2880 | #[inline ] |
2881 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2882 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2883 | #[cfg_attr (test, assert_instr(vpcmp))] |
2884 | pub unsafe fn _mm_cmpgt_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
2885 | simd_bitmask::<i16x8, _>(simd_gt(x:a.as_i16x8(), y:b.as_i16x8())) |
2886 | } |
2887 | |
2888 | /// 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). |
2889 | /// |
2890 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpgt_epi16_mask&expand=894) |
2891 | #[inline ] |
2892 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2893 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2894 | #[cfg_attr (test, assert_instr(vpcmp))] |
2895 | pub unsafe fn _mm_mask_cmpgt_epi16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
2896 | _mm_cmpgt_epi16_mask(a, b) & k1 |
2897 | } |
2898 | |
2899 | /// Compare packed signed 8-bit integers in a and b for greater-than, and store the results in mask vector k. |
2900 | /// |
2901 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpgt_epi8_mask&expand=921) |
2902 | #[inline ] |
2903 | #[target_feature (enable = "avx512bw" )] |
2904 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2905 | #[cfg_attr (test, assert_instr(vpcmp))] |
2906 | pub unsafe fn _mm512_cmpgt_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
2907 | simd_bitmask::<i8x64, _>(simd_gt(x:a.as_i8x64(), y:b.as_i8x64())) |
2908 | } |
2909 | |
2910 | /// 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). |
2911 | /// |
2912 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpgt_epi8_mask&expand=922) |
2913 | #[inline ] |
2914 | #[target_feature (enable = "avx512bw" )] |
2915 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2916 | #[cfg_attr (test, assert_instr(vpcmp))] |
2917 | pub unsafe fn _mm512_mask_cmpgt_epi8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
2918 | _mm512_cmpgt_epi8_mask(a, b) & k1 |
2919 | } |
2920 | |
2921 | /// Compare packed signed 8-bit integers in a and b for greater-than, and store the results in mask vector k. |
2922 | /// |
2923 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpgt_epi8_mask&expand=919) |
2924 | #[inline ] |
2925 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2926 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2927 | #[cfg_attr (test, assert_instr(vpcmp))] |
2928 | pub unsafe fn _mm256_cmpgt_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
2929 | simd_bitmask::<i8x32, _>(simd_gt(x:a.as_i8x32(), y:b.as_i8x32())) |
2930 | } |
2931 | |
2932 | /// 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). |
2933 | /// |
2934 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpgt_epi8_mask&expand=920) |
2935 | #[inline ] |
2936 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2937 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2938 | #[cfg_attr (test, assert_instr(vpcmp))] |
2939 | pub unsafe fn _mm256_mask_cmpgt_epi8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
2940 | _mm256_cmpgt_epi8_mask(a, b) & k1 |
2941 | } |
2942 | |
2943 | /// Compare packed signed 8-bit integers in a and b for greater-than, and store the results in mask vector k. |
2944 | /// |
2945 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpgt_epi8_mask&expand=917) |
2946 | #[inline ] |
2947 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2948 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2949 | #[cfg_attr (test, assert_instr(vpcmp))] |
2950 | pub unsafe fn _mm_cmpgt_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
2951 | simd_bitmask::<i8x16, _>(simd_gt(x:a.as_i8x16(), y:b.as_i8x16())) |
2952 | } |
2953 | |
2954 | /// 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). |
2955 | /// |
2956 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpgt_epi8_mask&expand=918) |
2957 | #[inline ] |
2958 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2959 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2960 | #[cfg_attr (test, assert_instr(vpcmp))] |
2961 | pub unsafe fn _mm_mask_cmpgt_epi8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
2962 | _mm_cmpgt_epi8_mask(a, b) & k1 |
2963 | } |
2964 | |
2965 | /// Compare packed unsigned 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
2966 | /// |
2967 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmple_epu16_mask&expand=989) |
2968 | #[inline ] |
2969 | #[target_feature (enable = "avx512bw" )] |
2970 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2971 | #[cfg_attr (test, assert_instr(vpcmp))] |
2972 | pub unsafe fn _mm512_cmple_epu16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
2973 | simd_bitmask::<u16x32, _>(simd_le(x:a.as_u16x32(), y:b.as_u16x32())) |
2974 | } |
2975 | |
2976 | /// 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). |
2977 | /// |
2978 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmple_epu16_mask&expand=990) |
2979 | #[inline ] |
2980 | #[target_feature (enable = "avx512bw" )] |
2981 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2982 | #[cfg_attr (test, assert_instr(vpcmp))] |
2983 | pub unsafe fn _mm512_mask_cmple_epu16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
2984 | _mm512_cmple_epu16_mask(a, b) & k1 |
2985 | } |
2986 | |
2987 | /// Compare packed unsigned 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
2988 | /// |
2989 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmple_epu16_mask&expand=987) |
2990 | #[inline ] |
2991 | #[target_feature (enable = "avx512bw,avx512vl" )] |
2992 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
2993 | #[cfg_attr (test, assert_instr(vpcmp))] |
2994 | pub unsafe fn _mm256_cmple_epu16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
2995 | simd_bitmask::<u16x16, _>(simd_le(x:a.as_u16x16(), y:b.as_u16x16())) |
2996 | } |
2997 | |
2998 | /// 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). |
2999 | /// |
3000 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmple_epu16_mask&expand=988) |
3001 | #[inline ] |
3002 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3003 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3004 | #[cfg_attr (test, assert_instr(vpcmp))] |
3005 | pub unsafe fn _mm256_mask_cmple_epu16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
3006 | _mm256_cmple_epu16_mask(a, b) & k1 |
3007 | } |
3008 | |
3009 | /// Compare packed unsigned 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
3010 | /// |
3011 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmple_epu16_mask&expand=985) |
3012 | #[inline ] |
3013 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3014 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3015 | #[cfg_attr (test, assert_instr(vpcmp))] |
3016 | pub unsafe fn _mm_cmple_epu16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
3017 | simd_bitmask::<u16x8, _>(simd_le(x:a.as_u16x8(), y:b.as_u16x8())) |
3018 | } |
3019 | |
3020 | /// 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). |
3021 | /// |
3022 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmple_epu16_mask&expand=986) |
3023 | #[inline ] |
3024 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3025 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3026 | #[cfg_attr (test, assert_instr(vpcmp))] |
3027 | pub unsafe fn _mm_mask_cmple_epu16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
3028 | _mm_cmple_epu16_mask(a, b) & k1 |
3029 | } |
3030 | |
3031 | /// Compare packed unsigned 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
3032 | /// |
3033 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmple_epu8_mask&expand=1007) |
3034 | #[inline ] |
3035 | #[target_feature (enable = "avx512bw" )] |
3036 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3037 | #[cfg_attr (test, assert_instr(vpcmp))] |
3038 | pub unsafe fn _mm512_cmple_epu8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
3039 | simd_bitmask::<u8x64, _>(simd_le(x:a.as_u8x64(), y:b.as_u8x64())) |
3040 | } |
3041 | |
3042 | /// 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). |
3043 | /// |
3044 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmple_epu8_mask&expand=1008) |
3045 | #[inline ] |
3046 | #[target_feature (enable = "avx512bw" )] |
3047 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3048 | #[cfg_attr (test, assert_instr(vpcmp))] |
3049 | pub unsafe fn _mm512_mask_cmple_epu8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
3050 | _mm512_cmple_epu8_mask(a, b) & k1 |
3051 | } |
3052 | |
3053 | /// Compare packed unsigned 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
3054 | /// |
3055 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmple_epu8_mask&expand=1005) |
3056 | #[inline ] |
3057 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3058 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3059 | #[cfg_attr (test, assert_instr(vpcmp))] |
3060 | pub unsafe fn _mm256_cmple_epu8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
3061 | simd_bitmask::<u8x32, _>(simd_le(x:a.as_u8x32(), y:b.as_u8x32())) |
3062 | } |
3063 | |
3064 | /// 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). |
3065 | /// |
3066 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmple_epu8_mask&expand=1006) |
3067 | #[inline ] |
3068 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3069 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3070 | #[cfg_attr (test, assert_instr(vpcmp))] |
3071 | pub unsafe fn _mm256_mask_cmple_epu8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
3072 | _mm256_cmple_epu8_mask(a, b) & k1 |
3073 | } |
3074 | |
3075 | /// Compare packed unsigned 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
3076 | /// |
3077 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmple_epu8_mask&expand=1003) |
3078 | #[inline ] |
3079 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3080 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3081 | #[cfg_attr (test, assert_instr(vpcmp))] |
3082 | pub unsafe fn _mm_cmple_epu8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
3083 | simd_bitmask::<u8x16, _>(simd_le(x:a.as_u8x16(), y:b.as_u8x16())) |
3084 | } |
3085 | |
3086 | /// 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). |
3087 | /// |
3088 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmple_epu8_mask&expand=1004) |
3089 | #[inline ] |
3090 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3091 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3092 | #[cfg_attr (test, assert_instr(vpcmp))] |
3093 | pub unsafe fn _mm_mask_cmple_epu8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
3094 | _mm_cmple_epu8_mask(a, b) & k1 |
3095 | } |
3096 | |
3097 | /// Compare packed signed 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
3098 | /// |
3099 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmple_epi16_mask&expand=965) |
3100 | #[inline ] |
3101 | #[target_feature (enable = "avx512bw" )] |
3102 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3103 | #[cfg_attr (test, assert_instr(vpcmp))] |
3104 | pub unsafe fn _mm512_cmple_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
3105 | simd_bitmask::<i16x32, _>(simd_le(x:a.as_i16x32(), y:b.as_i16x32())) |
3106 | } |
3107 | |
3108 | /// 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). |
3109 | /// |
3110 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmple_epi16_mask&expand=966) |
3111 | #[inline ] |
3112 | #[target_feature (enable = "avx512bw" )] |
3113 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3114 | #[cfg_attr (test, assert_instr(vpcmp))] |
3115 | pub unsafe fn _mm512_mask_cmple_epi16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
3116 | _mm512_cmple_epi16_mask(a, b) & k1 |
3117 | } |
3118 | |
3119 | /// Compare packed signed 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
3120 | /// |
3121 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmple_epi16_mask&expand=963) |
3122 | #[inline ] |
3123 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3124 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3125 | #[cfg_attr (test, assert_instr(vpcmp))] |
3126 | pub unsafe fn _mm256_cmple_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
3127 | simd_bitmask::<i16x16, _>(simd_le(x:a.as_i16x16(), y:b.as_i16x16())) |
3128 | } |
3129 | |
3130 | /// 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). |
3131 | /// |
3132 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmple_epi16_mask&expand=964) |
3133 | #[inline ] |
3134 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3135 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3136 | #[cfg_attr (test, assert_instr(vpcmp))] |
3137 | pub unsafe fn _mm256_mask_cmple_epi16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
3138 | _mm256_cmple_epi16_mask(a, b) & k1 |
3139 | } |
3140 | |
3141 | /// Compare packed signed 16-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
3142 | /// |
3143 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmple_epi16_mask&expand=961) |
3144 | #[inline ] |
3145 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3146 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3147 | #[cfg_attr (test, assert_instr(vpcmp))] |
3148 | pub unsafe fn _mm_cmple_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
3149 | simd_bitmask::<i16x8, _>(simd_le(x:a.as_i16x8(), y:b.as_i16x8())) |
3150 | } |
3151 | |
3152 | /// 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). |
3153 | /// |
3154 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmple_epi16_mask&expand=962) |
3155 | #[inline ] |
3156 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3157 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3158 | #[cfg_attr (test, assert_instr(vpcmp))] |
3159 | pub unsafe fn _mm_mask_cmple_epi16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
3160 | _mm_cmple_epi16_mask(a, b) & k1 |
3161 | } |
3162 | |
3163 | /// Compare packed signed 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
3164 | /// |
3165 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmple_epi8_mask&expand=983) |
3166 | #[inline ] |
3167 | #[target_feature (enable = "avx512bw" )] |
3168 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3169 | #[cfg_attr (test, assert_instr(vpcmp))] |
3170 | pub unsafe fn _mm512_cmple_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
3171 | simd_bitmask::<i8x64, _>(simd_le(x:a.as_i8x64(), y:b.as_i8x64())) |
3172 | } |
3173 | |
3174 | /// 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). |
3175 | /// |
3176 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmple_epi8_mask&expand=984) |
3177 | #[inline ] |
3178 | #[target_feature (enable = "avx512bw" )] |
3179 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3180 | #[cfg_attr (test, assert_instr(vpcmp))] |
3181 | pub unsafe fn _mm512_mask_cmple_epi8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
3182 | _mm512_cmple_epi8_mask(a, b) & k1 |
3183 | } |
3184 | |
3185 | /// Compare packed signed 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
3186 | /// |
3187 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmple_epi8_mask&expand=981) |
3188 | #[inline ] |
3189 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3190 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3191 | #[cfg_attr (test, assert_instr(vpcmp))] |
3192 | pub unsafe fn _mm256_cmple_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
3193 | simd_bitmask::<i8x32, _>(simd_le(x:a.as_i8x32(), y:b.as_i8x32())) |
3194 | } |
3195 | |
3196 | /// 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). |
3197 | /// |
3198 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmple_epi8_mask&expand=982) |
3199 | #[inline ] |
3200 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3201 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3202 | #[cfg_attr (test, assert_instr(vpcmp))] |
3203 | pub unsafe fn _mm256_mask_cmple_epi8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
3204 | _mm256_cmple_epi8_mask(a, b) & k1 |
3205 | } |
3206 | |
3207 | /// Compare packed signed 8-bit integers in a and b for less-than-or-equal, and store the results in mask vector k. |
3208 | /// |
3209 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmple_epi8_mask&expand=979) |
3210 | #[inline ] |
3211 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3212 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3213 | #[cfg_attr (test, assert_instr(vpcmp))] |
3214 | pub unsafe fn _mm_cmple_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
3215 | simd_bitmask::<i8x16, _>(simd_le(x:a.as_i8x16(), y:b.as_i8x16())) |
3216 | } |
3217 | |
3218 | /// 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). |
3219 | /// |
3220 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmple_epi8_mask&expand=980) |
3221 | #[inline ] |
3222 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3223 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3224 | #[cfg_attr (test, assert_instr(vpcmp))] |
3225 | pub unsafe fn _mm_mask_cmple_epi8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
3226 | _mm_cmple_epi8_mask(a, b) & k1 |
3227 | } |
3228 | |
3229 | /// Compare packed unsigned 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
3230 | /// |
3231 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpge_epu16_mask&expand=867) |
3232 | #[inline ] |
3233 | #[target_feature (enable = "avx512bw" )] |
3234 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3235 | #[cfg_attr (test, assert_instr(vpcmp))] |
3236 | pub unsafe fn _mm512_cmpge_epu16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
3237 | simd_bitmask::<u16x32, _>(simd_ge(x:a.as_u16x32(), y:b.as_u16x32())) |
3238 | } |
3239 | |
3240 | /// 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). |
3241 | /// |
3242 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpge_epu16_mask&expand=868) |
3243 | #[inline ] |
3244 | #[target_feature (enable = "avx512bw" )] |
3245 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3246 | #[cfg_attr (test, assert_instr(vpcmp))] |
3247 | pub unsafe fn _mm512_mask_cmpge_epu16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
3248 | _mm512_cmpge_epu16_mask(a, b) & k1 |
3249 | } |
3250 | |
3251 | /// Compare packed unsigned 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
3252 | /// |
3253 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpge_epu16_mask&expand=865) |
3254 | #[inline ] |
3255 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3256 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3257 | #[cfg_attr (test, assert_instr(vpcmp))] |
3258 | pub unsafe fn _mm256_cmpge_epu16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
3259 | simd_bitmask::<u16x16, _>(simd_ge(x:a.as_u16x16(), y:b.as_u16x16())) |
3260 | } |
3261 | |
3262 | /// 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). |
3263 | /// |
3264 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpge_epu16_mask&expand=866) |
3265 | #[inline ] |
3266 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3267 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3268 | #[cfg_attr (test, assert_instr(vpcmp))] |
3269 | pub unsafe fn _mm256_mask_cmpge_epu16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
3270 | _mm256_cmpge_epu16_mask(a, b) & k1 |
3271 | } |
3272 | |
3273 | /// Compare packed unsigned 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
3274 | /// |
3275 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpge_epu16_mask&expand=863) |
3276 | #[inline ] |
3277 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3278 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3279 | #[cfg_attr (test, assert_instr(vpcmp))] |
3280 | pub unsafe fn _mm_cmpge_epu16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
3281 | simd_bitmask::<u16x8, _>(simd_ge(x:a.as_u16x8(), y:b.as_u16x8())) |
3282 | } |
3283 | |
3284 | /// 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). |
3285 | /// |
3286 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpge_epu16_mask&expand=864) |
3287 | #[inline ] |
3288 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3289 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3290 | #[cfg_attr (test, assert_instr(vpcmp))] |
3291 | pub unsafe fn _mm_mask_cmpge_epu16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
3292 | _mm_cmpge_epu16_mask(a, b) & k1 |
3293 | } |
3294 | |
3295 | /// Compare packed unsigned 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
3296 | /// |
3297 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpge_epu8_mask&expand=885) |
3298 | #[inline ] |
3299 | #[target_feature (enable = "avx512bw" )] |
3300 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3301 | #[cfg_attr (test, assert_instr(vpcmp))] |
3302 | pub unsafe fn _mm512_cmpge_epu8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
3303 | simd_bitmask::<u8x64, _>(simd_ge(x:a.as_u8x64(), y:b.as_u8x64())) |
3304 | } |
3305 | |
3306 | /// 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). |
3307 | /// |
3308 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpge_epu8_mask&expand=886) |
3309 | #[inline ] |
3310 | #[target_feature (enable = "avx512bw" )] |
3311 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3312 | #[cfg_attr (test, assert_instr(vpcmp))] |
3313 | pub unsafe fn _mm512_mask_cmpge_epu8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
3314 | _mm512_cmpge_epu8_mask(a, b) & k1 |
3315 | } |
3316 | |
3317 | /// Compare packed unsigned 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
3318 | /// |
3319 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpge_epu8_mask&expand=883) |
3320 | #[inline ] |
3321 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3322 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3323 | #[cfg_attr (test, assert_instr(vpcmp))] |
3324 | pub unsafe fn _mm256_cmpge_epu8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
3325 | simd_bitmask::<u8x32, _>(simd_ge(x:a.as_u8x32(), y:b.as_u8x32())) |
3326 | } |
3327 | |
3328 | /// 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). |
3329 | /// |
3330 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpge_epu8_mask&expand=884) |
3331 | #[inline ] |
3332 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3333 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3334 | #[cfg_attr (test, assert_instr(vpcmp))] |
3335 | pub unsafe fn _mm256_mask_cmpge_epu8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
3336 | _mm256_cmpge_epu8_mask(a, b) & k1 |
3337 | } |
3338 | |
3339 | /// Compare packed unsigned 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
3340 | /// |
3341 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpge_epu8_mask&expand=881) |
3342 | #[inline ] |
3343 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3344 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3345 | #[cfg_attr (test, assert_instr(vpcmp))] |
3346 | pub unsafe fn _mm_cmpge_epu8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
3347 | simd_bitmask::<u8x16, _>(simd_ge(x:a.as_u8x16(), y:b.as_u8x16())) |
3348 | } |
3349 | |
3350 | /// 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). |
3351 | /// |
3352 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpge_epu8_mask&expand=882) |
3353 | #[inline ] |
3354 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3355 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3356 | #[cfg_attr (test, assert_instr(vpcmp))] |
3357 | pub unsafe fn _mm_mask_cmpge_epu8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
3358 | _mm_cmpge_epu8_mask(a, b) & k1 |
3359 | } |
3360 | |
3361 | /// Compare packed signed 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
3362 | /// |
3363 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpge_epi16_mask&expand=843) |
3364 | #[inline ] |
3365 | #[target_feature (enable = "avx512bw" )] |
3366 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3367 | #[cfg_attr (test, assert_instr(vpcmp))] |
3368 | pub unsafe fn _mm512_cmpge_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
3369 | simd_bitmask::<i16x32, _>(simd_ge(x:a.as_i16x32(), y:b.as_i16x32())) |
3370 | } |
3371 | |
3372 | /// 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). |
3373 | /// |
3374 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpge_epi16_mask&expand=844) |
3375 | #[inline ] |
3376 | #[target_feature (enable = "avx512bw" )] |
3377 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3378 | #[cfg_attr (test, assert_instr(vpcmp))] |
3379 | pub unsafe fn _mm512_mask_cmpge_epi16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
3380 | _mm512_cmpge_epi16_mask(a, b) & k1 |
3381 | } |
3382 | |
3383 | /// Compare packed signed 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
3384 | /// |
3385 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpge_epi16_mask&expand=841) |
3386 | #[inline ] |
3387 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3388 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3389 | #[cfg_attr (test, assert_instr(vpcmp))] |
3390 | pub unsafe fn _mm256_cmpge_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
3391 | simd_bitmask::<i16x16, _>(simd_ge(x:a.as_i16x16(), y:b.as_i16x16())) |
3392 | } |
3393 | |
3394 | /// 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). |
3395 | /// |
3396 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpge_epi16_mask&expand=842) |
3397 | #[inline ] |
3398 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3399 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3400 | #[cfg_attr (test, assert_instr(vpcmp))] |
3401 | pub unsafe fn _mm256_mask_cmpge_epi16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
3402 | _mm256_cmpge_epi16_mask(a, b) & k1 |
3403 | } |
3404 | |
3405 | /// Compare packed signed 16-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
3406 | /// |
3407 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpge_epi16_mask&expand=839) |
3408 | #[inline ] |
3409 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3410 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3411 | #[cfg_attr (test, assert_instr(vpcmp))] |
3412 | pub unsafe fn _mm_cmpge_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
3413 | simd_bitmask::<i16x8, _>(simd_ge(x:a.as_i16x8(), y:b.as_i16x8())) |
3414 | } |
3415 | |
3416 | /// 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). |
3417 | /// |
3418 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpge_epi16_mask&expand=840) |
3419 | #[inline ] |
3420 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3421 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3422 | #[cfg_attr (test, assert_instr(vpcmp))] |
3423 | pub unsafe fn _mm_mask_cmpge_epi16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
3424 | _mm_cmpge_epi16_mask(a, b) & k1 |
3425 | } |
3426 | |
3427 | /// Compare packed signed 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
3428 | /// |
3429 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpge_epi8_mask&expand=861) |
3430 | #[inline ] |
3431 | #[target_feature (enable = "avx512bw" )] |
3432 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3433 | #[cfg_attr (test, assert_instr(vpcmp))] |
3434 | pub unsafe fn _mm512_cmpge_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
3435 | simd_bitmask::<i8x64, _>(simd_ge(x:a.as_i8x64(), y:b.as_i8x64())) |
3436 | } |
3437 | |
3438 | /// 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). |
3439 | /// |
3440 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpge_epi8_mask&expand=862) |
3441 | #[inline ] |
3442 | #[target_feature (enable = "avx512bw" )] |
3443 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3444 | #[cfg_attr (test, assert_instr(vpcmp))] |
3445 | pub unsafe fn _mm512_mask_cmpge_epi8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
3446 | _mm512_cmpge_epi8_mask(a, b) & k1 |
3447 | } |
3448 | |
3449 | /// Compare packed signed 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
3450 | /// |
3451 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpge_epi8_mask&expand=859) |
3452 | #[inline ] |
3453 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3454 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3455 | #[cfg_attr (test, assert_instr(vpcmp))] |
3456 | pub unsafe fn _mm256_cmpge_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
3457 | simd_bitmask::<i8x32, _>(simd_ge(x:a.as_i8x32(), y:b.as_i8x32())) |
3458 | } |
3459 | |
3460 | /// 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). |
3461 | /// |
3462 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpge_epi8_mask&expand=860) |
3463 | #[inline ] |
3464 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3465 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3466 | #[cfg_attr (test, assert_instr(vpcmp))] |
3467 | pub unsafe fn _mm256_mask_cmpge_epi8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
3468 | _mm256_cmpge_epi8_mask(a, b) & k1 |
3469 | } |
3470 | |
3471 | /// Compare packed signed 8-bit integers in a and b for greater-than-or-equal, and store the results in mask vector k. |
3472 | /// |
3473 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpge_epi8_mask&expand=857) |
3474 | #[inline ] |
3475 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3476 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3477 | #[cfg_attr (test, assert_instr(vpcmp))] |
3478 | pub unsafe fn _mm_cmpge_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
3479 | simd_bitmask::<i8x16, _>(simd_ge(x:a.as_i8x16(), y:b.as_i8x16())) |
3480 | } |
3481 | |
3482 | /// 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). |
3483 | /// |
3484 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpge_epi8_mask&expand=858) |
3485 | #[inline ] |
3486 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3487 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3488 | #[cfg_attr (test, assert_instr(vpcmp))] |
3489 | pub unsafe fn _mm_mask_cmpge_epi8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
3490 | _mm_cmpge_epi8_mask(a, b) & k1 |
3491 | } |
3492 | |
3493 | /// Compare packed unsigned 16-bit integers in a and b for equality, and store the results in mask vector k. |
3494 | /// |
3495 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpeq_epu16_mask&expand=801) |
3496 | #[inline ] |
3497 | #[target_feature (enable = "avx512bw" )] |
3498 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3499 | #[cfg_attr (test, assert_instr(vpcmp))] |
3500 | pub unsafe fn _mm512_cmpeq_epu16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
3501 | simd_bitmask::<u16x32, _>(simd_eq(x:a.as_u16x32(), y:b.as_u16x32())) |
3502 | } |
3503 | |
3504 | /// 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). |
3505 | /// |
3506 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpeq_epu16_mask&expand=802) |
3507 | #[inline ] |
3508 | #[target_feature (enable = "avx512bw" )] |
3509 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3510 | #[cfg_attr (test, assert_instr(vpcmp))] |
3511 | pub unsafe fn _mm512_mask_cmpeq_epu16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
3512 | _mm512_cmpeq_epu16_mask(a, b) & k1 |
3513 | } |
3514 | |
3515 | /// Compare packed unsigned 16-bit integers in a and b for equality, and store the results in mask vector k. |
3516 | /// |
3517 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpeq_epu16_mask&expand=799) |
3518 | #[inline ] |
3519 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3520 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3521 | #[cfg_attr (test, assert_instr(vpcmp))] |
3522 | pub unsafe fn _mm256_cmpeq_epu16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
3523 | simd_bitmask::<u16x16, _>(simd_eq(x:a.as_u16x16(), y:b.as_u16x16())) |
3524 | } |
3525 | |
3526 | /// 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). |
3527 | /// |
3528 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpeq_epu16_mask&expand=800) |
3529 | #[inline ] |
3530 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3531 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3532 | #[cfg_attr (test, assert_instr(vpcmp))] |
3533 | pub unsafe fn _mm256_mask_cmpeq_epu16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
3534 | _mm256_cmpeq_epu16_mask(a, b) & k1 |
3535 | } |
3536 | |
3537 | /// Compare packed unsigned 16-bit integers in a and b for equality, and store the results in mask vector k. |
3538 | /// |
3539 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpeq_epu16_mask&expand=797) |
3540 | #[inline ] |
3541 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3542 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3543 | #[cfg_attr (test, assert_instr(vpcmp))] |
3544 | pub unsafe fn _mm_cmpeq_epu16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
3545 | simd_bitmask::<u16x8, _>(simd_eq(x:a.as_u16x8(), y:b.as_u16x8())) |
3546 | } |
3547 | |
3548 | /// 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). |
3549 | /// |
3550 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpeq_epu16_mask&expand=798) |
3551 | #[inline ] |
3552 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3553 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3554 | #[cfg_attr (test, assert_instr(vpcmp))] |
3555 | pub unsafe fn _mm_mask_cmpeq_epu16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
3556 | _mm_cmpeq_epu16_mask(a, b) & k1 |
3557 | } |
3558 | |
3559 | /// Compare packed unsigned 8-bit integers in a and b for equality, and store the results in mask vector k. |
3560 | /// |
3561 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpeq_epu8_mask&expand=819) |
3562 | #[inline ] |
3563 | #[target_feature (enable = "avx512bw" )] |
3564 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3565 | #[cfg_attr (test, assert_instr(vpcmp))] |
3566 | pub unsafe fn _mm512_cmpeq_epu8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
3567 | simd_bitmask::<u8x64, _>(simd_eq(x:a.as_u8x64(), y:b.as_u8x64())) |
3568 | } |
3569 | |
3570 | /// 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). |
3571 | /// |
3572 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpeq_epu8_mask&expand=820) |
3573 | #[inline ] |
3574 | #[target_feature (enable = "avx512bw" )] |
3575 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3576 | #[cfg_attr (test, assert_instr(vpcmp))] |
3577 | pub unsafe fn _mm512_mask_cmpeq_epu8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
3578 | _mm512_cmpeq_epu8_mask(a, b) & k1 |
3579 | } |
3580 | |
3581 | /// Compare packed unsigned 8-bit integers in a and b for equality, and store the results in mask vector k. |
3582 | /// |
3583 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpeq_epu8_mask&expand=817) |
3584 | #[inline ] |
3585 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3586 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3587 | #[cfg_attr (test, assert_instr(vpcmp))] |
3588 | pub unsafe fn _mm256_cmpeq_epu8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
3589 | simd_bitmask::<u8x32, _>(simd_eq(x:a.as_u8x32(), y:b.as_u8x32())) |
3590 | } |
3591 | |
3592 | /// 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). |
3593 | /// |
3594 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpeq_epu8_mask&expand=818) |
3595 | #[inline ] |
3596 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3597 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3598 | #[cfg_attr (test, assert_instr(vpcmp))] |
3599 | pub unsafe fn _mm256_mask_cmpeq_epu8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
3600 | _mm256_cmpeq_epu8_mask(a, b) & k1 |
3601 | } |
3602 | |
3603 | /// Compare packed unsigned 8-bit integers in a and b for equality, and store the results in mask vector k. |
3604 | /// |
3605 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpeq_epu8_mask&expand=815) |
3606 | #[inline ] |
3607 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3608 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3609 | #[cfg_attr (test, assert_instr(vpcmp))] |
3610 | pub unsafe fn _mm_cmpeq_epu8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
3611 | simd_bitmask::<u8x16, _>(simd_eq(x:a.as_u8x16(), y:b.as_u8x16())) |
3612 | } |
3613 | |
3614 | /// 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). |
3615 | /// |
3616 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpeq_epu8_mask&expand=816) |
3617 | #[inline ] |
3618 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3619 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3620 | #[cfg_attr (test, assert_instr(vpcmp))] |
3621 | pub unsafe fn _mm_mask_cmpeq_epu8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
3622 | _mm_cmpeq_epu8_mask(a, b) & k1 |
3623 | } |
3624 | |
3625 | /// Compare packed signed 16-bit integers in a and b for equality, and store the results in mask vector k. |
3626 | /// |
3627 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpeq_epi16_mask&expand=771) |
3628 | #[inline ] |
3629 | #[target_feature (enable = "avx512bw" )] |
3630 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3631 | #[cfg_attr (test, assert_instr(vpcmp))] |
3632 | pub unsafe fn _mm512_cmpeq_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
3633 | simd_bitmask::<i16x32, _>(simd_eq(x:a.as_i16x32(), y:b.as_i16x32())) |
3634 | } |
3635 | |
3636 | /// 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). |
3637 | /// |
3638 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpeq_epi16_mask&expand=772) |
3639 | #[inline ] |
3640 | #[target_feature (enable = "avx512bw" )] |
3641 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3642 | #[cfg_attr (test, assert_instr(vpcmp))] |
3643 | pub unsafe fn _mm512_mask_cmpeq_epi16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
3644 | _mm512_cmpeq_epi16_mask(a, b) & k1 |
3645 | } |
3646 | |
3647 | /// Compare packed signed 16-bit integers in a and b for equality, and store the results in mask vector k. |
3648 | /// |
3649 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpeq_epi16_mask&expand=769) |
3650 | #[inline ] |
3651 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3652 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3653 | #[cfg_attr (test, assert_instr(vpcmp))] |
3654 | pub unsafe fn _mm256_cmpeq_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
3655 | simd_bitmask::<i16x16, _>(simd_eq(x:a.as_i16x16(), y:b.as_i16x16())) |
3656 | } |
3657 | |
3658 | /// 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). |
3659 | /// |
3660 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpeq_epi16_mask&expand=770) |
3661 | #[inline ] |
3662 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3663 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3664 | #[cfg_attr (test, assert_instr(vpcmp))] |
3665 | pub unsafe fn _mm256_mask_cmpeq_epi16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
3666 | _mm256_cmpeq_epi16_mask(a, b) & k1 |
3667 | } |
3668 | |
3669 | /// Compare packed signed 16-bit integers in a and b for equality, and store the results in mask vector k. |
3670 | /// |
3671 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpeq_epi16_mask&expand=767) |
3672 | #[inline ] |
3673 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3674 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3675 | #[cfg_attr (test, assert_instr(vpcmp))] |
3676 | pub unsafe fn _mm_cmpeq_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
3677 | simd_bitmask::<i16x8, _>(simd_eq(x:a.as_i16x8(), y:b.as_i16x8())) |
3678 | } |
3679 | |
3680 | /// 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). |
3681 | /// |
3682 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpeq_epi16_mask&expand=768) |
3683 | #[inline ] |
3684 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3685 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3686 | #[cfg_attr (test, assert_instr(vpcmp))] |
3687 | pub unsafe fn _mm_mask_cmpeq_epi16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
3688 | _mm_cmpeq_epi16_mask(a, b) & k1 |
3689 | } |
3690 | |
3691 | /// Compare packed signed 8-bit integers in a and b for equality, and store the results in mask vector k. |
3692 | /// |
3693 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpeq_epi8_mask&expand=795) |
3694 | #[inline ] |
3695 | #[target_feature (enable = "avx512bw" )] |
3696 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3697 | #[cfg_attr (test, assert_instr(vpcmp))] |
3698 | pub unsafe fn _mm512_cmpeq_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
3699 | simd_bitmask::<i8x64, _>(simd_eq(x:a.as_i8x64(), y:b.as_i8x64())) |
3700 | } |
3701 | |
3702 | /// 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). |
3703 | /// |
3704 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpeq_epi8_mask&expand=796) |
3705 | #[inline ] |
3706 | #[target_feature (enable = "avx512bw" )] |
3707 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3708 | #[cfg_attr (test, assert_instr(vpcmp))] |
3709 | pub unsafe fn _mm512_mask_cmpeq_epi8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
3710 | _mm512_cmpeq_epi8_mask(a, b) & k1 |
3711 | } |
3712 | |
3713 | /// Compare packed signed 8-bit integers in a and b for equality, and store the results in mask vector k. |
3714 | /// |
3715 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpeq_epi8_mask&expand=793) |
3716 | #[inline ] |
3717 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3718 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3719 | #[cfg_attr (test, assert_instr(vpcmp))] |
3720 | pub unsafe fn _mm256_cmpeq_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
3721 | simd_bitmask::<i8x32, _>(simd_eq(x:a.as_i8x32(), y:b.as_i8x32())) |
3722 | } |
3723 | |
3724 | /// 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). |
3725 | /// |
3726 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpeq_epi8_mask&expand=794) |
3727 | #[inline ] |
3728 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3729 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3730 | #[cfg_attr (test, assert_instr(vpcmp))] |
3731 | pub unsafe fn _mm256_mask_cmpeq_epi8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
3732 | _mm256_cmpeq_epi8_mask(a, b) & k1 |
3733 | } |
3734 | |
3735 | /// Compare packed signed 8-bit integers in a and b for equality, and store the results in mask vector k. |
3736 | /// |
3737 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpeq_epi8_mask&expand=791) |
3738 | #[inline ] |
3739 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3740 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3741 | #[cfg_attr (test, assert_instr(vpcmp))] |
3742 | pub unsafe fn _mm_cmpeq_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
3743 | simd_bitmask::<i8x16, _>(simd_eq(x:a.as_i8x16(), y:b.as_i8x16())) |
3744 | } |
3745 | |
3746 | /// 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). |
3747 | /// |
3748 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpeq_epi8_mask&expand=792) |
3749 | #[inline ] |
3750 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3751 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3752 | #[cfg_attr (test, assert_instr(vpcmp))] |
3753 | pub unsafe fn _mm_mask_cmpeq_epi8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
3754 | _mm_cmpeq_epi8_mask(a, b) & k1 |
3755 | } |
3756 | |
3757 | /// Compare packed unsigned 16-bit integers in a and b for not-equal, and store the results in mask vector k. |
3758 | /// |
3759 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpneq_epu16_mask&expand=1106) |
3760 | #[inline ] |
3761 | #[target_feature (enable = "avx512bw" )] |
3762 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3763 | #[cfg_attr (test, assert_instr(vpcmp))] |
3764 | pub unsafe fn _mm512_cmpneq_epu16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
3765 | simd_bitmask::<u16x32, _>(simd_ne(x:a.as_u16x32(), y:b.as_u16x32())) |
3766 | } |
3767 | |
3768 | /// 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). |
3769 | /// |
3770 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpneq_epu16_mask&expand=1107) |
3771 | #[inline ] |
3772 | #[target_feature (enable = "avx512bw" )] |
3773 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3774 | #[cfg_attr (test, assert_instr(vpcmp))] |
3775 | pub unsafe fn _mm512_mask_cmpneq_epu16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
3776 | _mm512_cmpneq_epu16_mask(a, b) & k1 |
3777 | } |
3778 | |
3779 | /// Compare packed unsigned 16-bit integers in a and b for not-equal, and store the results in mask vector k. |
3780 | /// |
3781 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpneq_epu16_mask&expand=1104) |
3782 | #[inline ] |
3783 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3784 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3785 | #[cfg_attr (test, assert_instr(vpcmp))] |
3786 | pub unsafe fn _mm256_cmpneq_epu16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
3787 | simd_bitmask::<u16x16, _>(simd_ne(x:a.as_u16x16(), y:b.as_u16x16())) |
3788 | } |
3789 | |
3790 | /// 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). |
3791 | /// |
3792 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpneq_epu16_mask&expand=1105) |
3793 | #[inline ] |
3794 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3795 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3796 | #[cfg_attr (test, assert_instr(vpcmp))] |
3797 | pub unsafe fn _mm256_mask_cmpneq_epu16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
3798 | _mm256_cmpneq_epu16_mask(a, b) & k1 |
3799 | } |
3800 | |
3801 | /// Compare packed unsigned 16-bit integers in a and b for not-equal, and store the results in mask vector k. |
3802 | /// |
3803 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpneq_epu16_mask&expand=1102) |
3804 | #[inline ] |
3805 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3806 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3807 | #[cfg_attr (test, assert_instr(vpcmp))] |
3808 | pub unsafe fn _mm_cmpneq_epu16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
3809 | simd_bitmask::<u16x8, _>(simd_ne(x:a.as_u16x8(), y:b.as_u16x8())) |
3810 | } |
3811 | |
3812 | /// 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). |
3813 | /// |
3814 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpneq_epu16_mask&expand=1103) |
3815 | #[inline ] |
3816 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3817 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3818 | #[cfg_attr (test, assert_instr(vpcmp))] |
3819 | pub unsafe fn _mm_mask_cmpneq_epu16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
3820 | _mm_cmpneq_epu16_mask(a, b) & k1 |
3821 | } |
3822 | |
3823 | /// Compare packed unsigned 8-bit integers in a and b for not-equal, and store the results in mask vector k. |
3824 | /// |
3825 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpneq_epu8_mask&expand=1124) |
3826 | #[inline ] |
3827 | #[target_feature (enable = "avx512bw" )] |
3828 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3829 | #[cfg_attr (test, assert_instr(vpcmp))] |
3830 | pub unsafe fn _mm512_cmpneq_epu8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
3831 | simd_bitmask::<u8x64, _>(simd_ne(x:a.as_u8x64(), y:b.as_u8x64())) |
3832 | } |
3833 | |
3834 | /// 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). |
3835 | /// |
3836 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpneq_epu8_mask&expand=1125) |
3837 | #[inline ] |
3838 | #[target_feature (enable = "avx512bw" )] |
3839 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3840 | #[cfg_attr (test, assert_instr(vpcmp))] |
3841 | pub unsafe fn _mm512_mask_cmpneq_epu8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
3842 | _mm512_cmpneq_epu8_mask(a, b) & k1 |
3843 | } |
3844 | |
3845 | /// Compare packed unsigned 8-bit integers in a and b for not-equal, and store the results in mask vector k. |
3846 | /// |
3847 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpneq_epu8_mask&expand=1122) |
3848 | #[inline ] |
3849 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3850 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3851 | #[cfg_attr (test, assert_instr(vpcmp))] |
3852 | pub unsafe fn _mm256_cmpneq_epu8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
3853 | simd_bitmask::<u8x32, _>(simd_ne(x:a.as_u8x32(), y:b.as_u8x32())) |
3854 | } |
3855 | |
3856 | /// 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). |
3857 | /// |
3858 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpneq_epu8_mask&expand=1123) |
3859 | #[inline ] |
3860 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3861 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3862 | #[cfg_attr (test, assert_instr(vpcmp))] |
3863 | pub unsafe fn _mm256_mask_cmpneq_epu8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
3864 | _mm256_cmpneq_epu8_mask(a, b) & k1 |
3865 | } |
3866 | |
3867 | /// Compare packed unsigned 8-bit integers in a and b for not-equal, and store the results in mask vector k. |
3868 | /// |
3869 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpneq_epu8_mask&expand=1120) |
3870 | #[inline ] |
3871 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3872 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3873 | #[cfg_attr (test, assert_instr(vpcmp))] |
3874 | pub unsafe fn _mm_cmpneq_epu8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
3875 | simd_bitmask::<u8x16, _>(simd_ne(x:a.as_u8x16(), y:b.as_u8x16())) |
3876 | } |
3877 | |
3878 | /// 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). |
3879 | /// |
3880 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpneq_epu8_mask&expand=1121) |
3881 | #[inline ] |
3882 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3883 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3884 | #[cfg_attr (test, assert_instr(vpcmp))] |
3885 | pub unsafe fn _mm_mask_cmpneq_epu8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
3886 | _mm_cmpneq_epu8_mask(a, b) & k1 |
3887 | } |
3888 | |
3889 | /// Compare packed signed 16-bit integers in a and b for not-equal, and store the results in mask vector k. |
3890 | /// |
3891 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpneq_epi16_mask&expand=1082) |
3892 | #[inline ] |
3893 | #[target_feature (enable = "avx512bw" )] |
3894 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3895 | #[cfg_attr (test, assert_instr(vpcmp))] |
3896 | pub unsafe fn _mm512_cmpneq_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
3897 | simd_bitmask::<i16x32, _>(simd_ne(x:a.as_i16x32(), y:b.as_i16x32())) |
3898 | } |
3899 | |
3900 | /// 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). |
3901 | /// |
3902 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpneq_epi16_mask&expand=1083) |
3903 | #[inline ] |
3904 | #[target_feature (enable = "avx512bw" )] |
3905 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3906 | #[cfg_attr (test, assert_instr(vpcmp))] |
3907 | pub unsafe fn _mm512_mask_cmpneq_epi16_mask(k1: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
3908 | _mm512_cmpneq_epi16_mask(a, b) & k1 |
3909 | } |
3910 | |
3911 | /// Compare packed signed 16-bit integers in a and b for not-equal, and store the results in mask vector k. |
3912 | /// |
3913 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpneq_epi16_mask&expand=1080) |
3914 | #[inline ] |
3915 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3916 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3917 | #[cfg_attr (test, assert_instr(vpcmp))] |
3918 | pub unsafe fn _mm256_cmpneq_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
3919 | simd_bitmask::<i16x16, _>(simd_ne(x:a.as_i16x16(), y:b.as_i16x16())) |
3920 | } |
3921 | |
3922 | /// 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). |
3923 | /// |
3924 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpneq_epi16_mask&expand=1081) |
3925 | #[inline ] |
3926 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3927 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3928 | #[cfg_attr (test, assert_instr(vpcmp))] |
3929 | pub unsafe fn _mm256_mask_cmpneq_epi16_mask(k1: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
3930 | _mm256_cmpneq_epi16_mask(a, b) & k1 |
3931 | } |
3932 | |
3933 | /// Compare packed signed 16-bit integers in a and b for not-equal, and store the results in mask vector k. |
3934 | /// |
3935 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpneq_epi16_mask&expand=1078) |
3936 | #[inline ] |
3937 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3938 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3939 | #[cfg_attr (test, assert_instr(vpcmp))] |
3940 | pub unsafe fn _mm_cmpneq_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
3941 | simd_bitmask::<i16x8, _>(simd_ne(x:a.as_i16x8(), y:b.as_i16x8())) |
3942 | } |
3943 | |
3944 | /// 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). |
3945 | /// |
3946 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpneq_epi16_mask&expand=1079) |
3947 | #[inline ] |
3948 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3949 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3950 | #[cfg_attr (test, assert_instr(vpcmp))] |
3951 | pub unsafe fn _mm_mask_cmpneq_epi16_mask(k1: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
3952 | _mm_cmpneq_epi16_mask(a, b) & k1 |
3953 | } |
3954 | |
3955 | /// Compare packed signed 8-bit integers in a and b for not-equal, and store the results in mask vector k. |
3956 | /// |
3957 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmpneq_epi8_mask&expand=1100) |
3958 | #[inline ] |
3959 | #[target_feature (enable = "avx512bw" )] |
3960 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3961 | #[cfg_attr (test, assert_instr(vpcmp))] |
3962 | pub unsafe fn _mm512_cmpneq_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
3963 | simd_bitmask::<i8x64, _>(simd_ne(x:a.as_i8x64(), y:b.as_i8x64())) |
3964 | } |
3965 | |
3966 | /// 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). |
3967 | /// |
3968 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmpneq_epi8_mask&expand=1101) |
3969 | #[inline ] |
3970 | #[target_feature (enable = "avx512bw" )] |
3971 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3972 | #[cfg_attr (test, assert_instr(vpcmp))] |
3973 | pub unsafe fn _mm512_mask_cmpneq_epi8_mask(k1: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
3974 | _mm512_cmpneq_epi8_mask(a, b) & k1 |
3975 | } |
3976 | |
3977 | /// Compare packed signed 8-bit integers in a and b for not-equal, and store the results in mask vector k. |
3978 | /// |
3979 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmpneq_epi8_mask&expand=1098) |
3980 | #[inline ] |
3981 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3982 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3983 | #[cfg_attr (test, assert_instr(vpcmp))] |
3984 | pub unsafe fn _mm256_cmpneq_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
3985 | simd_bitmask::<i8x32, _>(simd_ne(x:a.as_i8x32(), y:b.as_i8x32())) |
3986 | } |
3987 | |
3988 | /// 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). |
3989 | /// |
3990 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmpneq_epi8_mask&expand=1099) |
3991 | #[inline ] |
3992 | #[target_feature (enable = "avx512bw,avx512vl" )] |
3993 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
3994 | #[cfg_attr (test, assert_instr(vpcmp))] |
3995 | pub unsafe fn _mm256_mask_cmpneq_epi8_mask(k1: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
3996 | _mm256_cmpneq_epi8_mask(a, b) & k1 |
3997 | } |
3998 | |
3999 | /// Compare packed signed 8-bit integers in a and b for not-equal, and store the results in mask vector k. |
4000 | /// |
4001 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpneq_epi8_mask&expand=1096) |
4002 | #[inline ] |
4003 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4004 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4005 | #[cfg_attr (test, assert_instr(vpcmp))] |
4006 | pub unsafe fn _mm_cmpneq_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
4007 | simd_bitmask::<i8x16, _>(simd_ne(x:a.as_i8x16(), y:b.as_i8x16())) |
4008 | } |
4009 | |
4010 | /// 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). |
4011 | /// |
4012 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmpneq_epi8_mask&expand=1097) |
4013 | #[inline ] |
4014 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4015 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4016 | #[cfg_attr (test, assert_instr(vpcmp))] |
4017 | pub unsafe fn _mm_mask_cmpneq_epi8_mask(k1: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
4018 | _mm_cmpneq_epi8_mask(a, b) & k1 |
4019 | } |
4020 | |
4021 | /// 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. |
4022 | /// |
4023 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmp_epu16_mask&expand=715) |
4024 | #[inline ] |
4025 | #[target_feature (enable = "avx512bw" )] |
4026 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4027 | #[rustc_legacy_const_generics (2)] |
4028 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4029 | pub unsafe fn _mm512_cmp_epu16_mask<const IMM8: i32>(a: __m512i, b: __m512i) -> __mmask32 { |
4030 | static_assert_uimm_bits!(IMM8, 3); |
4031 | let a: u16x32 = a.as_u16x32(); |
4032 | let b: u16x32 = b.as_u16x32(); |
4033 | vpcmpuw(a, b, IMM8, mask:0b11111111_11111111_11111111_11111111) |
4034 | } |
4035 | |
4036 | /// 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). |
4037 | /// |
4038 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmp_epu16_mask&expand=716) |
4039 | #[inline ] |
4040 | #[target_feature (enable = "avx512bw" )] |
4041 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4042 | #[rustc_legacy_const_generics (3)] |
4043 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4044 | pub unsafe fn _mm512_mask_cmp_epu16_mask<const IMM8: i32>( |
4045 | k1: __mmask32, |
4046 | a: __m512i, |
4047 | b: __m512i, |
4048 | ) -> __mmask32 { |
4049 | static_assert_uimm_bits!(IMM8, 3); |
4050 | let a: u16x32 = a.as_u16x32(); |
4051 | let b: u16x32 = b.as_u16x32(); |
4052 | vpcmpuw(a, b, IMM8, mask:k1) |
4053 | } |
4054 | |
4055 | /// 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. |
4056 | /// |
4057 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmp_epu16_mask&expand=713) |
4058 | #[inline ] |
4059 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4060 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4061 | #[rustc_legacy_const_generics (2)] |
4062 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4063 | pub unsafe fn _mm256_cmp_epu16_mask<const IMM8: i32>(a: __m256i, b: __m256i) -> __mmask16 { |
4064 | static_assert_uimm_bits!(IMM8, 3); |
4065 | let a: u16x16 = a.as_u16x16(); |
4066 | let b: u16x16 = b.as_u16x16(); |
4067 | vpcmpuw256(a, b, IMM8, mask:0b11111111_11111111) |
4068 | } |
4069 | |
4070 | /// 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). |
4071 | /// |
4072 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmp_epu16_mask&expand=714) |
4073 | #[inline ] |
4074 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4075 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4076 | #[rustc_legacy_const_generics (3)] |
4077 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4078 | pub unsafe fn _mm256_mask_cmp_epu16_mask<const IMM8: i32>( |
4079 | k1: __mmask16, |
4080 | a: __m256i, |
4081 | b: __m256i, |
4082 | ) -> __mmask16 { |
4083 | static_assert_uimm_bits!(IMM8, 3); |
4084 | let a: u16x16 = a.as_u16x16(); |
4085 | let b: u16x16 = b.as_u16x16(); |
4086 | vpcmpuw256(a, b, IMM8, mask:k1) |
4087 | } |
4088 | |
4089 | /// 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. |
4090 | /// |
4091 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmp_epu16_mask&expand=711) |
4092 | #[inline ] |
4093 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4094 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4095 | #[rustc_legacy_const_generics (2)] |
4096 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4097 | pub unsafe fn _mm_cmp_epu16_mask<const IMM8: i32>(a: __m128i, b: __m128i) -> __mmask8 { |
4098 | static_assert_uimm_bits!(IMM8, 3); |
4099 | let a: u16x8 = a.as_u16x8(); |
4100 | let b: u16x8 = b.as_u16x8(); |
4101 | vpcmpuw128(a, b, IMM8, mask:0b11111111) |
4102 | } |
4103 | |
4104 | /// 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). |
4105 | /// |
4106 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmp_epu16_mask&expand=712) |
4107 | #[inline ] |
4108 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4109 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4110 | #[rustc_legacy_const_generics (3)] |
4111 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4112 | pub unsafe fn _mm_mask_cmp_epu16_mask<const IMM8: i32>( |
4113 | k1: __mmask8, |
4114 | a: __m128i, |
4115 | b: __m128i, |
4116 | ) -> __mmask8 { |
4117 | static_assert_uimm_bits!(IMM8, 3); |
4118 | let a: u16x8 = a.as_u16x8(); |
4119 | let b: u16x8 = b.as_u16x8(); |
4120 | vpcmpuw128(a, b, IMM8, mask:k1) |
4121 | } |
4122 | |
4123 | /// 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. |
4124 | /// |
4125 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmp_epu8_mask&expand=733) |
4126 | #[inline ] |
4127 | #[target_feature (enable = "avx512bw" )] |
4128 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4129 | #[rustc_legacy_const_generics (2)] |
4130 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4131 | pub unsafe fn _mm512_cmp_epu8_mask<const IMM8: i32>(a: __m512i, b: __m512i) -> __mmask64 { |
4132 | static_assert_uimm_bits!(IMM8, 3); |
4133 | let a: u8x64 = a.as_u8x64(); |
4134 | let b: u8x64 = b.as_u8x64(); |
4135 | vpcmpub( |
4136 | a, |
4137 | b, |
4138 | IMM8, |
4139 | mask:0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
4140 | ) |
4141 | } |
4142 | |
4143 | /// 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). |
4144 | /// |
4145 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmp_epu8_mask&expand=734) |
4146 | #[inline ] |
4147 | #[target_feature (enable = "avx512bw" )] |
4148 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4149 | #[rustc_legacy_const_generics (3)] |
4150 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4151 | pub unsafe fn _mm512_mask_cmp_epu8_mask<const IMM8: i32>( |
4152 | k1: __mmask64, |
4153 | a: __m512i, |
4154 | b: __m512i, |
4155 | ) -> __mmask64 { |
4156 | static_assert_uimm_bits!(IMM8, 3); |
4157 | let a: u8x64 = a.as_u8x64(); |
4158 | let b: u8x64 = b.as_u8x64(); |
4159 | vpcmpub(a, b, IMM8, mask:k1) |
4160 | } |
4161 | |
4162 | /// 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. |
4163 | /// |
4164 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmp_epu8_mask&expand=731) |
4165 | #[inline ] |
4166 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4167 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4168 | #[rustc_legacy_const_generics (2)] |
4169 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4170 | pub unsafe fn _mm256_cmp_epu8_mask<const IMM8: i32>(a: __m256i, b: __m256i) -> __mmask32 { |
4171 | static_assert_uimm_bits!(IMM8, 3); |
4172 | let a: u8x32 = a.as_u8x32(); |
4173 | let b: u8x32 = b.as_u8x32(); |
4174 | vpcmpub256(a, b, IMM8, mask:0b11111111_11111111_11111111_11111111) |
4175 | } |
4176 | |
4177 | /// 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). |
4178 | /// |
4179 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmp_epu8_mask&expand=732) |
4180 | #[inline ] |
4181 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4182 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4183 | #[rustc_legacy_const_generics (3)] |
4184 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4185 | pub unsafe fn _mm256_mask_cmp_epu8_mask<const IMM8: i32>( |
4186 | k1: __mmask32, |
4187 | a: __m256i, |
4188 | b: __m256i, |
4189 | ) -> __mmask32 { |
4190 | static_assert_uimm_bits!(IMM8, 3); |
4191 | let a: u8x32 = a.as_u8x32(); |
4192 | let b: u8x32 = b.as_u8x32(); |
4193 | vpcmpub256(a, b, IMM8, mask:k1) |
4194 | } |
4195 | |
4196 | /// 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. |
4197 | /// |
4198 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmp_epu8_mask&expand=729) |
4199 | #[inline ] |
4200 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4201 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4202 | #[rustc_legacy_const_generics (2)] |
4203 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4204 | pub unsafe fn _mm_cmp_epu8_mask<const IMM8: i32>(a: __m128i, b: __m128i) -> __mmask16 { |
4205 | static_assert_uimm_bits!(IMM8, 3); |
4206 | let a: u8x16 = a.as_u8x16(); |
4207 | let b: u8x16 = b.as_u8x16(); |
4208 | vpcmpub128(a, b, IMM8, mask:0b11111111_11111111) |
4209 | } |
4210 | |
4211 | /// 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). |
4212 | /// |
4213 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmp_epu8_mask&expand=730) |
4214 | #[inline ] |
4215 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4216 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4217 | #[rustc_legacy_const_generics (3)] |
4218 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4219 | pub unsafe fn _mm_mask_cmp_epu8_mask<const IMM8: i32>( |
4220 | k1: __mmask16, |
4221 | a: __m128i, |
4222 | b: __m128i, |
4223 | ) -> __mmask16 { |
4224 | static_assert_uimm_bits!(IMM8, 3); |
4225 | let a: u8x16 = a.as_u8x16(); |
4226 | let b: u8x16 = b.as_u8x16(); |
4227 | vpcmpub128(a, b, IMM8, mask:k1) |
4228 | } |
4229 | |
4230 | /// 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. |
4231 | /// |
4232 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmp_epi16_mask&expand=691) |
4233 | #[inline ] |
4234 | #[target_feature (enable = "avx512bw" )] |
4235 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4236 | #[rustc_legacy_const_generics (2)] |
4237 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4238 | pub unsafe fn _mm512_cmp_epi16_mask<const IMM8: i32>(a: __m512i, b: __m512i) -> __mmask32 { |
4239 | static_assert_uimm_bits!(IMM8, 3); |
4240 | let a: i16x32 = a.as_i16x32(); |
4241 | let b: i16x32 = b.as_i16x32(); |
4242 | vpcmpw(a, b, IMM8, mask:0b11111111_11111111_11111111_11111111) |
4243 | } |
4244 | |
4245 | /// 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). |
4246 | /// |
4247 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmp_epi16_mask&expand=692) |
4248 | #[inline ] |
4249 | #[target_feature (enable = "avx512bw" )] |
4250 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4251 | #[rustc_legacy_const_generics (3)] |
4252 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4253 | pub unsafe fn _mm512_mask_cmp_epi16_mask<const IMM8: i32>( |
4254 | k1: __mmask32, |
4255 | a: __m512i, |
4256 | b: __m512i, |
4257 | ) -> __mmask32 { |
4258 | static_assert_uimm_bits!(IMM8, 3); |
4259 | let a: i16x32 = a.as_i16x32(); |
4260 | let b: i16x32 = b.as_i16x32(); |
4261 | vpcmpw(a, b, IMM8, mask:k1) |
4262 | } |
4263 | |
4264 | /// 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. |
4265 | /// |
4266 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmp_epi16_mask&expand=689) |
4267 | #[inline ] |
4268 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4269 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4270 | #[rustc_legacy_const_generics (2)] |
4271 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4272 | pub unsafe fn _mm256_cmp_epi16_mask<const IMM8: i32>(a: __m256i, b: __m256i) -> __mmask16 { |
4273 | static_assert_uimm_bits!(IMM8, 3); |
4274 | let a: i16x16 = a.as_i16x16(); |
4275 | let b: i16x16 = b.as_i16x16(); |
4276 | vpcmpw256(a, b, IMM8, mask:0b11111111_11111111) |
4277 | } |
4278 | |
4279 | /// 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). |
4280 | /// |
4281 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmp_epi16_mask&expand=690) |
4282 | #[inline ] |
4283 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4284 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4285 | #[rustc_legacy_const_generics (3)] |
4286 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4287 | pub unsafe fn _mm256_mask_cmp_epi16_mask<const IMM8: i32>( |
4288 | k1: __mmask16, |
4289 | a: __m256i, |
4290 | b: __m256i, |
4291 | ) -> __mmask16 { |
4292 | static_assert_uimm_bits!(IMM8, 3); |
4293 | let a: i16x16 = a.as_i16x16(); |
4294 | let b: i16x16 = b.as_i16x16(); |
4295 | vpcmpw256(a, b, IMM8, mask:k1) |
4296 | } |
4297 | |
4298 | /// 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. |
4299 | /// |
4300 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmp_epi16_mask&expand=687) |
4301 | #[inline ] |
4302 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4303 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4304 | #[rustc_legacy_const_generics (2)] |
4305 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4306 | pub unsafe fn _mm_cmp_epi16_mask<const IMM8: i32>(a: __m128i, b: __m128i) -> __mmask8 { |
4307 | static_assert_uimm_bits!(IMM8, 3); |
4308 | let a: i16x8 = a.as_i16x8(); |
4309 | let b: i16x8 = b.as_i16x8(); |
4310 | vpcmpw128(a, b, IMM8, mask:0b11111111) |
4311 | } |
4312 | |
4313 | /// 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). |
4314 | /// |
4315 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmp_epi16_mask&expand=688) |
4316 | #[inline ] |
4317 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4318 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4319 | #[rustc_legacy_const_generics (3)] |
4320 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4321 | pub unsafe fn _mm_mask_cmp_epi16_mask<const IMM8: i32>( |
4322 | k1: __mmask8, |
4323 | a: __m128i, |
4324 | b: __m128i, |
4325 | ) -> __mmask8 { |
4326 | static_assert_uimm_bits!(IMM8, 3); |
4327 | let a: i16x8 = a.as_i16x8(); |
4328 | let b: i16x8 = b.as_i16x8(); |
4329 | vpcmpw128(a, b, IMM8, mask:k1) |
4330 | } |
4331 | |
4332 | /// 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. |
4333 | /// |
4334 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cmp_epi8_mask&expand=709) |
4335 | #[inline ] |
4336 | #[target_feature (enable = "avx512bw" )] |
4337 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4338 | #[rustc_legacy_const_generics (2)] |
4339 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4340 | pub unsafe fn _mm512_cmp_epi8_mask<const IMM8: i32>(a: __m512i, b: __m512i) -> __mmask64 { |
4341 | static_assert_uimm_bits!(IMM8, 3); |
4342 | let a: i8x64 = a.as_i8x64(); |
4343 | let b: i8x64 = b.as_i8x64(); |
4344 | vpcmpb( |
4345 | a, |
4346 | b, |
4347 | IMM8, |
4348 | mask:0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
4349 | ) |
4350 | } |
4351 | |
4352 | /// 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). |
4353 | /// |
4354 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cmp_epi8_mask&expand=710) |
4355 | #[inline ] |
4356 | #[target_feature (enable = "avx512bw" )] |
4357 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4358 | #[rustc_legacy_const_generics (3)] |
4359 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4360 | pub unsafe fn _mm512_mask_cmp_epi8_mask<const IMM8: i32>( |
4361 | k1: __mmask64, |
4362 | a: __m512i, |
4363 | b: __m512i, |
4364 | ) -> __mmask64 { |
4365 | static_assert_uimm_bits!(IMM8, 3); |
4366 | let a: i8x64 = a.as_i8x64(); |
4367 | let b: i8x64 = b.as_i8x64(); |
4368 | vpcmpb(a, b, IMM8, mask:k1) |
4369 | } |
4370 | |
4371 | /// 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. |
4372 | /// |
4373 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cmp_epi8_mask&expand=707) |
4374 | #[inline ] |
4375 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4376 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4377 | #[rustc_legacy_const_generics (2)] |
4378 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4379 | pub unsafe fn _mm256_cmp_epi8_mask<const IMM8: i32>(a: __m256i, b: __m256i) -> __mmask32 { |
4380 | static_assert_uimm_bits!(IMM8, 3); |
4381 | let a: i8x32 = a.as_i8x32(); |
4382 | let b: i8x32 = b.as_i8x32(); |
4383 | vpcmpb256(a, b, IMM8, mask:0b11111111_11111111_11111111_11111111) |
4384 | } |
4385 | |
4386 | /// 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). |
4387 | /// |
4388 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cmp_epi8_mask&expand=708) |
4389 | #[inline ] |
4390 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4391 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4392 | #[rustc_legacy_const_generics (3)] |
4393 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4394 | pub unsafe fn _mm256_mask_cmp_epi8_mask<const IMM8: i32>( |
4395 | k1: __mmask32, |
4396 | a: __m256i, |
4397 | b: __m256i, |
4398 | ) -> __mmask32 { |
4399 | static_assert_uimm_bits!(IMM8, 3); |
4400 | let a: i8x32 = a.as_i8x32(); |
4401 | let b: i8x32 = b.as_i8x32(); |
4402 | vpcmpb256(a, b, IMM8, mask:k1) |
4403 | } |
4404 | |
4405 | /// 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. |
4406 | /// |
4407 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmp_epi8_mask&expand=705) |
4408 | #[inline ] |
4409 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4410 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4411 | #[rustc_legacy_const_generics (2)] |
4412 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4413 | pub unsafe fn _mm_cmp_epi8_mask<const IMM8: i32>(a: __m128i, b: __m128i) -> __mmask16 { |
4414 | static_assert_uimm_bits!(IMM8, 3); |
4415 | let a: i8x16 = a.as_i8x16(); |
4416 | let b: i8x16 = b.as_i8x16(); |
4417 | vpcmpb128(a, b, IMM8, mask:0b11111111_11111111) |
4418 | } |
4419 | |
4420 | /// 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). |
4421 | /// |
4422 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cmp_epi8_mask&expand=706) |
4423 | #[inline ] |
4424 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4425 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4426 | #[rustc_legacy_const_generics (3)] |
4427 | #[cfg_attr (test, assert_instr(vpcmp, IMM8 = 0))] |
4428 | pub unsafe fn _mm_mask_cmp_epi8_mask<const IMM8: i32>( |
4429 | k1: __mmask16, |
4430 | a: __m128i, |
4431 | b: __m128i, |
4432 | ) -> __mmask16 { |
4433 | static_assert_uimm_bits!(IMM8, 3); |
4434 | let a: i8x16 = a.as_i8x16(); |
4435 | let b: i8x16 = b.as_i8x16(); |
4436 | vpcmpb128(a, b, IMM8, mask:k1) |
4437 | } |
4438 | |
4439 | /// 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. |
4440 | /// |
4441 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_loadu_epi16&expand=3368) |
4442 | #[inline ] |
4443 | #[target_feature (enable = "avx512bw" )] |
4444 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4445 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu16 |
4446 | pub unsafe fn _mm512_loadu_epi16(mem_addr: *const i16) -> __m512i { |
4447 | ptr::read_unaligned(src:mem_addr as *const __m512i) |
4448 | } |
4449 | |
4450 | /// 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. |
4451 | /// |
4452 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_loadu_epi16&expand=3365) |
4453 | #[inline ] |
4454 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4455 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4456 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu16 |
4457 | pub unsafe fn _mm256_loadu_epi16(mem_addr: *const i16) -> __m256i { |
4458 | ptr::read_unaligned(src:mem_addr as *const __m256i) |
4459 | } |
4460 | |
4461 | /// 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. |
4462 | /// |
4463 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_loadu_epi16&expand=3362) |
4464 | #[inline ] |
4465 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4466 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4467 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu16 |
4468 | pub unsafe fn _mm_loadu_epi16(mem_addr: *const i16) -> __m128i { |
4469 | ptr::read_unaligned(src:mem_addr as *const __m128i) |
4470 | } |
4471 | |
4472 | /// 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. |
4473 | /// |
4474 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_loadu_epi8&expand=3395) |
4475 | #[inline ] |
4476 | #[target_feature (enable = "avx512bw" )] |
4477 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4478 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu8 |
4479 | pub unsafe fn _mm512_loadu_epi8(mem_addr: *const i8) -> __m512i { |
4480 | ptr::read_unaligned(src:mem_addr as *const __m512i) |
4481 | } |
4482 | |
4483 | /// 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. |
4484 | /// |
4485 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_loadu_epi8&expand=3392) |
4486 | #[inline ] |
4487 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4488 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4489 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu8 |
4490 | pub unsafe fn _mm256_loadu_epi8(mem_addr: *const i8) -> __m256i { |
4491 | ptr::read_unaligned(src:mem_addr as *const __m256i) |
4492 | } |
4493 | |
4494 | /// 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. |
4495 | /// |
4496 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_loadu_epi8&expand=3389) |
4497 | #[inline ] |
4498 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4499 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4500 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu8 |
4501 | pub unsafe fn _mm_loadu_epi8(mem_addr: *const i8) -> __m128i { |
4502 | ptr::read_unaligned(src:mem_addr as *const __m128i) |
4503 | } |
4504 | |
4505 | /// 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. |
4506 | /// |
4507 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_storeu_epi16&expand=5622) |
4508 | #[inline ] |
4509 | #[target_feature (enable = "avx512bw" )] |
4510 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4511 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu16 |
4512 | pub unsafe fn _mm512_storeu_epi16(mem_addr: *mut i16, a: __m512i) { |
4513 | ptr::write_unaligned(dst:mem_addr as *mut __m512i, src:a); |
4514 | } |
4515 | |
4516 | /// 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. |
4517 | /// |
4518 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_storeu_epi16&expand=5620) |
4519 | #[inline ] |
4520 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4521 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4522 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu16 |
4523 | pub unsafe fn _mm256_storeu_epi16(mem_addr: *mut i16, a: __m256i) { |
4524 | ptr::write_unaligned(dst:mem_addr as *mut __m256i, src:a); |
4525 | } |
4526 | |
4527 | /// 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. |
4528 | /// |
4529 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_storeu_epi16&expand=5618) |
4530 | #[inline ] |
4531 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4532 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4533 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu16 |
4534 | pub unsafe fn _mm_storeu_epi16(mem_addr: *mut i16, a: __m128i) { |
4535 | ptr::write_unaligned(dst:mem_addr as *mut __m128i, src:a); |
4536 | } |
4537 | |
4538 | /// 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. |
4539 | /// |
4540 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_storeu_epi8&expand=5640) |
4541 | #[inline ] |
4542 | #[target_feature (enable = "avx512bw" )] |
4543 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4544 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu8 |
4545 | pub unsafe fn _mm512_storeu_epi8(mem_addr: *mut i8, a: __m512i) { |
4546 | ptr::write_unaligned(dst:mem_addr as *mut __m512i, src:a); |
4547 | } |
4548 | |
4549 | /// 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. |
4550 | /// |
4551 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_storeu_epi8&expand=5638) |
4552 | #[inline ] |
4553 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4554 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4555 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu8 |
4556 | pub unsafe fn _mm256_storeu_epi8(mem_addr: *mut i8, a: __m256i) { |
4557 | ptr::write_unaligned(dst:mem_addr as *mut __m256i, src:a); |
4558 | } |
4559 | |
4560 | /// 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. |
4561 | /// |
4562 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_storeu_epi8&expand=5636) |
4563 | #[inline ] |
4564 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4565 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4566 | #[cfg_attr (test, assert_instr(vmovups))] //should be vmovdqu8 |
4567 | pub unsafe fn _mm_storeu_epi8(mem_addr: *mut i8, a: __m128i) { |
4568 | ptr::write_unaligned(dst:mem_addr as *mut __m128i, src:a); |
4569 | } |
4570 | |
4571 | /// Load packed 16-bit integers from memory into dst using writemask k |
4572 | /// (elements are copied from src when the corresponding mask bit is not set). |
4573 | /// mem_addr does not need to be aligned on any particular boundary. |
4574 | /// |
4575 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_loadu_epi16) |
4576 | #[inline ] |
4577 | #[target_feature (enable = "avx512f,avx512bw" )] |
4578 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4579 | pub unsafe fn _mm512_mask_loadu_epi16(src: __m512i, k: __mmask32, mem_addr: *const i16) -> __m512i { |
4580 | let mut dst: __m512i = src; |
4581 | asm!( |
4582 | vpl!("vmovdqu16 {dst}{{{k}}}" ), |
4583 | p = in(reg) mem_addr, |
4584 | k = in(kreg) k, |
4585 | dst = inout(zmm_reg) dst, |
4586 | options(pure, readonly, nostack) |
4587 | ); |
4588 | dst |
4589 | } |
4590 | |
4591 | /// Load packed 16-bit integers from memory into dst using zeromask k |
4592 | /// (elements are zeroed out when the corresponding mask bit is not set). |
4593 | /// mem_addr does not need to be aligned on any particular boundary. |
4594 | /// |
4595 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_loadu_epi16) |
4596 | #[inline ] |
4597 | #[target_feature (enable = "avx512f,avx512bw" )] |
4598 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4599 | pub unsafe fn _mm512_maskz_loadu_epi16(k: __mmask32, mem_addr: *const i16) -> __m512i { |
4600 | let mut dst: __m512i; |
4601 | asm!( |
4602 | vpl!("vmovdqu16 {dst}{{{k}}} {{z}}" ), |
4603 | p = in(reg) mem_addr, |
4604 | k = in(kreg) k, |
4605 | dst = out(zmm_reg) dst, |
4606 | options(pure, readonly, nostack) |
4607 | ); |
4608 | dst |
4609 | } |
4610 | |
4611 | /// Load packed 8-bit integers from memory into dst using writemask k |
4612 | /// (elements are copied from src when the corresponding mask bit is not set). |
4613 | /// mem_addr does not need to be aligned on any particular boundary. |
4614 | /// |
4615 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_loadu_epi8) |
4616 | #[inline ] |
4617 | #[target_feature (enable = "avx512f,avx512bw" )] |
4618 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4619 | pub unsafe fn _mm512_mask_loadu_epi8(src: __m512i, k: __mmask64, mem_addr: *const i8) -> __m512i { |
4620 | let mut dst: __m512i = src; |
4621 | asm!( |
4622 | vpl!("vmovdqu8 {dst}{{{k}}}" ), |
4623 | p = in(reg) mem_addr, |
4624 | k = in(kreg) k, |
4625 | dst = inout(zmm_reg) dst, |
4626 | options(pure, readonly, nostack) |
4627 | ); |
4628 | dst |
4629 | } |
4630 | |
4631 | /// Load packed 8-bit integers from memory into dst using zeromask k |
4632 | /// (elements are zeroed out when the corresponding mask bit is not set). |
4633 | /// mem_addr does not need to be aligned on any particular boundary. |
4634 | /// |
4635 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_loadu_epi8) |
4636 | #[inline ] |
4637 | #[target_feature (enable = "avx512f,avx512bw" )] |
4638 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4639 | pub unsafe fn _mm512_maskz_loadu_epi8(k: __mmask64, mem_addr: *const i8) -> __m512i { |
4640 | let mut dst: __m512i; |
4641 | asm!( |
4642 | vpl!("vmovdqu8 {dst}{{{k}}} {{z}}" ), |
4643 | p = in(reg) mem_addr, |
4644 | k = in(kreg) k, |
4645 | dst = out(zmm_reg) dst, |
4646 | options(pure, readonly, nostack) |
4647 | ); |
4648 | dst |
4649 | } |
4650 | |
4651 | /// Load packed 16-bit integers from memory into dst using writemask k |
4652 | /// (elements are copied from src when the corresponding mask bit is not set). |
4653 | /// mem_addr does not need to be aligned on any particular boundary. |
4654 | /// |
4655 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_loadu_epi16) |
4656 | #[inline ] |
4657 | #[target_feature (enable = "avx512f,avx512bw,avx512vl,avx" )] |
4658 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4659 | pub unsafe fn _mm256_mask_loadu_epi16(src: __m256i, k: __mmask16, mem_addr: *const i16) -> __m256i { |
4660 | let mut dst: __m256i = src; |
4661 | asm!( |
4662 | vpl!("vmovdqu16 {dst}{{{k}}}" ), |
4663 | p = in(reg) mem_addr, |
4664 | k = in(kreg) k, |
4665 | dst = inout(ymm_reg) dst, |
4666 | options(pure, readonly, nostack) |
4667 | ); |
4668 | dst |
4669 | } |
4670 | |
4671 | /// Load packed 16-bit integers from memory into dst using zeromask k |
4672 | /// (elements are zeroed out when the corresponding mask bit is not set). |
4673 | /// mem_addr does not need to be aligned on any particular boundary. |
4674 | /// |
4675 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_loadu_epi16) |
4676 | #[inline ] |
4677 | #[target_feature (enable = "avx512f,avx512bw,avx512vl,avx" )] |
4678 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4679 | pub unsafe fn _mm256_maskz_loadu_epi16(k: __mmask16, mem_addr: *const i16) -> __m256i { |
4680 | let mut dst: __m256i; |
4681 | asm!( |
4682 | vpl!("vmovdqu16 {dst}{{{k}}} {{z}}" ), |
4683 | p = in(reg) mem_addr, |
4684 | k = in(kreg) k, |
4685 | dst = out(ymm_reg) dst, |
4686 | options(pure, readonly, nostack) |
4687 | ); |
4688 | dst |
4689 | } |
4690 | |
4691 | /// Load packed 8-bit integers from memory into dst using writemask k |
4692 | /// (elements are copied from src when the corresponding mask bit is not set). |
4693 | /// mem_addr does not need to be aligned on any particular boundary. |
4694 | /// |
4695 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_loadu_epi8) |
4696 | #[inline ] |
4697 | #[target_feature (enable = "avx512f,avx512bw,avx512vl,avx" )] |
4698 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4699 | pub unsafe fn _mm256_mask_loadu_epi8(src: __m256i, k: __mmask32, mem_addr: *const i8) -> __m256i { |
4700 | let mut dst: __m256i = src; |
4701 | asm!( |
4702 | vpl!("vmovdqu8 {dst}{{{k}}}" ), |
4703 | p = in(reg) mem_addr, |
4704 | k = in(kreg) k, |
4705 | dst = inout(ymm_reg) dst, |
4706 | options(pure, readonly, nostack) |
4707 | ); |
4708 | dst |
4709 | } |
4710 | |
4711 | /// Load packed 8-bit integers from memory into dst using zeromask k |
4712 | /// (elements are zeroed out when the corresponding mask bit is not set). |
4713 | /// mem_addr does not need to be aligned on any particular boundary. |
4714 | /// |
4715 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_loadu_epi8) |
4716 | #[inline ] |
4717 | #[target_feature (enable = "avx512f,avx512bw,avx512vl,avx" )] |
4718 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4719 | pub unsafe fn _mm256_maskz_loadu_epi8(k: __mmask32, mem_addr: *const i8) -> __m256i { |
4720 | let mut dst: __m256i; |
4721 | asm!( |
4722 | vpl!("vmovdqu8 {dst}{{{k}}} {{z}}" ), |
4723 | p = in(reg) mem_addr, |
4724 | k = in(kreg) k, |
4725 | dst = out(ymm_reg) dst, |
4726 | options(pure, readonly, nostack) |
4727 | ); |
4728 | dst |
4729 | } |
4730 | |
4731 | /// Load packed 16-bit integers from memory into dst using writemask k |
4732 | /// (elements are copied from src when the corresponding mask bit is not set). |
4733 | /// mem_addr does not need to be aligned on any particular boundary. |
4734 | /// |
4735 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_loadu_epi16) |
4736 | #[inline ] |
4737 | #[target_feature (enable = "avx512f,avx512bw,avx512vl,avx,sse" )] |
4738 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4739 | pub unsafe fn _mm_mask_loadu_epi16(src: __m128i, k: __mmask8, mem_addr: *const i16) -> __m128i { |
4740 | let mut dst: __m128i = src; |
4741 | asm!( |
4742 | vpl!("vmovdqu16 {dst}{{{k}}}" ), |
4743 | p = in(reg) mem_addr, |
4744 | k = in(kreg) k, |
4745 | dst = inout(xmm_reg) dst, |
4746 | options(pure, readonly, nostack) |
4747 | ); |
4748 | dst |
4749 | } |
4750 | |
4751 | /// Load packed 16-bit integers from memory into dst using zeromask k |
4752 | /// (elements are zeroed out when the corresponding mask bit is not set). |
4753 | /// mem_addr does not need to be aligned on any particular boundary. |
4754 | /// |
4755 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_loadu_epi16) |
4756 | #[inline ] |
4757 | #[target_feature (enable = "avx512f,avx512bw,avx512vl,avx,sse" )] |
4758 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4759 | pub unsafe fn _mm_maskz_loadu_epi16(k: __mmask8, mem_addr: *const i16) -> __m128i { |
4760 | let mut dst: __m128i; |
4761 | asm!( |
4762 | vpl!("vmovdqu16 {dst}{{{k}}} {{z}}" ), |
4763 | p = in(reg) mem_addr, |
4764 | k = in(kreg) k, |
4765 | dst = out(xmm_reg) dst, |
4766 | options(pure, readonly, nostack) |
4767 | ); |
4768 | dst |
4769 | } |
4770 | |
4771 | /// Load packed 8-bit integers from memory into dst using writemask k |
4772 | /// (elements are copied from src when the corresponding mask bit is not set). |
4773 | /// mem_addr does not need to be aligned on any particular boundary. |
4774 | /// |
4775 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_loadu_epi8) |
4776 | #[inline ] |
4777 | #[target_feature (enable = "avx512f,avx512bw,avx512vl,avx,sse" )] |
4778 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4779 | pub unsafe fn _mm_mask_loadu_epi8(src: __m128i, k: __mmask16, mem_addr: *const i8) -> __m128i { |
4780 | let mut dst: __m128i = src; |
4781 | asm!( |
4782 | vpl!("vmovdqu8 {dst}{{{k}}}" ), |
4783 | p = in(reg) mem_addr, |
4784 | k = in(kreg) k, |
4785 | dst = inout(xmm_reg) dst, |
4786 | options(pure, readonly, nostack) |
4787 | ); |
4788 | dst |
4789 | } |
4790 | |
4791 | /// Load packed 8-bit integers from memory into dst using zeromask k |
4792 | /// (elements are zeroed out when the corresponding mask bit is not set). |
4793 | /// mem_addr does not need to be aligned on any particular boundary. |
4794 | /// |
4795 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_loadu_epi8) |
4796 | #[inline ] |
4797 | #[target_feature (enable = "avx512f,avx512bw,avx512vl,avx,sse" )] |
4798 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4799 | pub unsafe fn _mm_maskz_loadu_epi8(k: __mmask16, mem_addr: *const i8) -> __m128i { |
4800 | let mut dst: __m128i; |
4801 | asm!( |
4802 | vpl!("vmovdqu8 {dst}{{{k}}} {{z}}" ), |
4803 | p = in(reg) mem_addr, |
4804 | k = in(kreg) k, |
4805 | dst = out(xmm_reg) dst, |
4806 | options(pure, readonly, nostack) |
4807 | ); |
4808 | dst |
4809 | } |
4810 | |
4811 | /// Store packed 16-bit integers from a into memory using writemask k. |
4812 | /// mem_addr does not need to be aligned on any particular boundary. |
4813 | /// |
4814 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_storeu_epi16) |
4815 | #[inline ] |
4816 | #[target_feature (enable = "avx512f,avx512bw" )] |
4817 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4818 | pub unsafe fn _mm512_mask_storeu_epi16(mem_addr: *mut i16, mask: __mmask32, a: __m512i) { |
4819 | asm!( |
4820 | vps!("vmovdqu16" , "{{{mask}}}, {a}" ), |
4821 | p = in(reg) mem_addr, |
4822 | mask = in(kreg) mask, |
4823 | a = in(zmm_reg) a, |
4824 | options(nostack) |
4825 | ); |
4826 | } |
4827 | |
4828 | /// Store packed 8-bit integers from a into memory using writemask k. |
4829 | /// mem_addr does not need to be aligned on any particular boundary. |
4830 | /// |
4831 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_storeu_epi8) |
4832 | #[inline ] |
4833 | #[target_feature (enable = "avx512f,avx512bw" )] |
4834 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4835 | pub unsafe fn _mm512_mask_storeu_epi8(mem_addr: *mut i8, mask: __mmask64, a: __m512i) { |
4836 | asm!( |
4837 | vps!("vmovdqu8" , "{{{mask}}}, {a}" ), |
4838 | p = in(reg) mem_addr, |
4839 | mask = in(kreg) mask, |
4840 | a = in(zmm_reg) a, |
4841 | options(nostack) |
4842 | ); |
4843 | } |
4844 | |
4845 | /// Store packed 16-bit integers from a into memory using writemask k. |
4846 | /// mem_addr does not need to be aligned on any particular boundary. |
4847 | /// |
4848 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_storeu_epi16) |
4849 | #[inline ] |
4850 | #[target_feature (enable = "avx512f,avx512bw,avx512vl,avx" )] |
4851 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4852 | pub unsafe fn _mm256_mask_storeu_epi16(mem_addr: *mut i16, mask: __mmask16, a: __m256i) { |
4853 | asm!( |
4854 | vps!("vmovdqu16" , "{{{mask}}}, {a}" ), |
4855 | p = in(reg) mem_addr, |
4856 | mask = in(kreg) mask, |
4857 | a = in(ymm_reg) a, |
4858 | options(nostack) |
4859 | ); |
4860 | } |
4861 | |
4862 | /// Store packed 8-bit integers from a into memory using writemask k. |
4863 | /// mem_addr does not need to be aligned on any particular boundary. |
4864 | /// |
4865 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_storeu_epi8) |
4866 | #[inline ] |
4867 | #[target_feature (enable = "avx512f,avx512bw,avx512vl,avx" )] |
4868 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4869 | pub unsafe fn _mm256_mask_storeu_epi8(mem_addr: *mut i8, mask: __mmask32, a: __m256i) { |
4870 | asm!( |
4871 | vps!("vmovdqu8" , "{{{mask}}}, {a}" ), |
4872 | p = in(reg) mem_addr, |
4873 | mask = in(kreg) mask, |
4874 | a = in(ymm_reg) a, |
4875 | options(nostack) |
4876 | ); |
4877 | } |
4878 | |
4879 | /// Store packed 16-bit integers from a into memory using writemask k. |
4880 | /// mem_addr does not need to be aligned on any particular boundary. |
4881 | /// |
4882 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_storeu_epi16) |
4883 | #[inline ] |
4884 | #[target_feature (enable = "avx512f,avx512bw,avx512vl,avx,sse" )] |
4885 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4886 | pub unsafe fn _mm_mask_storeu_epi16(mem_addr: *mut i16, mask: __mmask8, a: __m128i) { |
4887 | asm!( |
4888 | vps!("vmovdqu16" , "{{{mask}}}, {a}" ), |
4889 | p = in(reg) mem_addr, |
4890 | mask = in(kreg) mask, |
4891 | a = in(xmm_reg) a, |
4892 | options(nostack) |
4893 | ); |
4894 | } |
4895 | |
4896 | /// Store packed 8-bit integers from a into memory using writemask k. |
4897 | /// mem_addr does not need to be aligned on any particular boundary. |
4898 | /// |
4899 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_storeu_epi8) |
4900 | #[inline ] |
4901 | #[target_feature (enable = "avx512f,avx512bw,avx512vl,avx,sse" )] |
4902 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4903 | pub unsafe fn _mm_mask_storeu_epi8(mem_addr: *mut i8, mask: __mmask16, a: __m128i) { |
4904 | asm!( |
4905 | vps!("vmovdqu8" , "{{{mask}}}, {a}" ), |
4906 | p = in(reg) mem_addr, |
4907 | mask = in(kreg) mask, |
4908 | a = in(xmm_reg) a, |
4909 | options(nostack) |
4910 | ); |
4911 | } |
4912 | |
4913 | /// 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. |
4914 | /// |
4915 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_madd_epi16&expand=3511) |
4916 | #[inline ] |
4917 | #[target_feature (enable = "avx512bw" )] |
4918 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4919 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
4920 | pub unsafe fn _mm512_madd_epi16(a: __m512i, b: __m512i) -> __m512i { |
4921 | transmute(src:vpmaddwd(a:a.as_i16x32(), b:b.as_i16x32())) |
4922 | } |
4923 | |
4924 | /// 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). |
4925 | /// |
4926 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_madd_epi16&expand=3512) |
4927 | #[inline ] |
4928 | #[target_feature (enable = "avx512bw" )] |
4929 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4930 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
4931 | pub unsafe fn _mm512_mask_madd_epi16( |
4932 | src: __m512i, |
4933 | k: __mmask16, |
4934 | a: __m512i, |
4935 | b: __m512i, |
4936 | ) -> __m512i { |
4937 | let madd: i32x16 = _mm512_madd_epi16(a, b).as_i32x16(); |
4938 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:src.as_i32x16())) |
4939 | } |
4940 | |
4941 | /// 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). |
4942 | /// |
4943 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_madd_epi16&expand=3513) |
4944 | #[inline ] |
4945 | #[target_feature (enable = "avx512bw" )] |
4946 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4947 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
4948 | pub unsafe fn _mm512_maskz_madd_epi16(k: __mmask16, a: __m512i, b: __m512i) -> __m512i { |
4949 | let madd: i32x16 = _mm512_madd_epi16(a, b).as_i32x16(); |
4950 | let zero: i32x16 = _mm512_setzero_si512().as_i32x16(); |
4951 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:zero)) |
4952 | } |
4953 | |
4954 | /// 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). |
4955 | /// |
4956 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_madd_epi16&expand=3509) |
4957 | #[inline ] |
4958 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4959 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4960 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
4961 | pub unsafe fn _mm256_mask_madd_epi16(src: __m256i, k: __mmask8, a: __m256i, b: __m256i) -> __m256i { |
4962 | let madd: i32x8 = _mm256_madd_epi16(a, b).as_i32x8(); |
4963 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:src.as_i32x8())) |
4964 | } |
4965 | |
4966 | /// 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). |
4967 | /// |
4968 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_madd_epi16&expand=3510) |
4969 | #[inline ] |
4970 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4971 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4972 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
4973 | pub unsafe fn _mm256_maskz_madd_epi16(k: __mmask8, a: __m256i, b: __m256i) -> __m256i { |
4974 | let madd: i32x8 = _mm256_madd_epi16(a, b).as_i32x8(); |
4975 | let zero: i32x8 = _mm256_setzero_si256().as_i32x8(); |
4976 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:zero)) |
4977 | } |
4978 | |
4979 | /// 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). |
4980 | /// |
4981 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_madd_epi16&expand=3506) |
4982 | #[inline ] |
4983 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4984 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4985 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
4986 | pub unsafe fn _mm_mask_madd_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
4987 | let madd: i32x4 = _mm_madd_epi16(a, b).as_i32x4(); |
4988 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:src.as_i32x4())) |
4989 | } |
4990 | |
4991 | /// 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). |
4992 | /// |
4993 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_madd_epi16&expand=3507) |
4994 | #[inline ] |
4995 | #[target_feature (enable = "avx512bw,avx512vl" )] |
4996 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
4997 | #[cfg_attr (test, assert_instr(vpmaddwd))] |
4998 | pub unsafe fn _mm_maskz_madd_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
4999 | let madd: i32x4 = _mm_madd_epi16(a, b).as_i32x4(); |
5000 | let zero: i32x4 = _mm_setzero_si128().as_i32x4(); |
5001 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:zero)) |
5002 | } |
5003 | |
5004 | /// 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. |
5005 | /// |
5006 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maddubs_epi16&expand=3539) |
5007 | #[inline ] |
5008 | #[target_feature (enable = "avx512bw" )] |
5009 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5010 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
5011 | pub unsafe fn _mm512_maddubs_epi16(a: __m512i, b: __m512i) -> __m512i { |
5012 | transmute(src:vpmaddubsw(a:a.as_i8x64(), b:b.as_i8x64())) |
5013 | } |
5014 | |
5015 | /// 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). |
5016 | /// |
5017 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_maddubs_epi16&expand=3540) |
5018 | #[inline ] |
5019 | #[target_feature (enable = "avx512bw" )] |
5020 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5021 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
5022 | pub unsafe fn _mm512_mask_maddubs_epi16( |
5023 | src: __m512i, |
5024 | k: __mmask32, |
5025 | a: __m512i, |
5026 | b: __m512i, |
5027 | ) -> __m512i { |
5028 | let madd: i16x32 = _mm512_maddubs_epi16(a, b).as_i16x32(); |
5029 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:src.as_i16x32())) |
5030 | } |
5031 | |
5032 | /// 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). |
5033 | /// |
5034 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_maddubs_epi16&expand=3541) |
5035 | #[inline ] |
5036 | #[target_feature (enable = "avx512bw" )] |
5037 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5038 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
5039 | pub unsafe fn _mm512_maskz_maddubs_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
5040 | let madd: i16x32 = _mm512_maddubs_epi16(a, b).as_i16x32(); |
5041 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
5042 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:zero)) |
5043 | } |
5044 | |
5045 | /// 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). |
5046 | /// |
5047 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_maddubs_epi16&expand=3537) |
5048 | #[inline ] |
5049 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5050 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5051 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
5052 | pub unsafe fn _mm256_mask_maddubs_epi16( |
5053 | src: __m256i, |
5054 | k: __mmask16, |
5055 | a: __m256i, |
5056 | b: __m256i, |
5057 | ) -> __m256i { |
5058 | let madd: i16x16 = _mm256_maddubs_epi16(a, b).as_i16x16(); |
5059 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:src.as_i16x16())) |
5060 | } |
5061 | |
5062 | /// 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). |
5063 | /// |
5064 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_maddubs_epi16&expand=3538) |
5065 | #[inline ] |
5066 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5067 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5068 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
5069 | pub unsafe fn _mm256_maskz_maddubs_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
5070 | let madd: i16x16 = _mm256_maddubs_epi16(a, b).as_i16x16(); |
5071 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
5072 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:zero)) |
5073 | } |
5074 | |
5075 | /// 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). |
5076 | /// |
5077 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_maddubs_epi16&expand=3534) |
5078 | #[inline ] |
5079 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5080 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5081 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
5082 | pub unsafe fn _mm_mask_maddubs_epi16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
5083 | let madd: i16x8 = _mm_maddubs_epi16(a, b).as_i16x8(); |
5084 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:src.as_i16x8())) |
5085 | } |
5086 | |
5087 | /// 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). |
5088 | /// |
5089 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_maddubs_epi16&expand=3535) |
5090 | #[inline ] |
5091 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5092 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5093 | #[cfg_attr (test, assert_instr(vpmaddubsw))] |
5094 | pub unsafe fn _mm_maskz_maddubs_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
5095 | let madd: i16x8 = _mm_maddubs_epi16(a, b).as_i16x8(); |
5096 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
5097 | transmute(src:simd_select_bitmask(m:k, yes:madd, no:zero)) |
5098 | } |
5099 | |
5100 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using signed saturation, and store the results in dst. |
5101 | /// |
5102 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_packs_epi32&expand=4091) |
5103 | #[inline ] |
5104 | #[target_feature (enable = "avx512bw" )] |
5105 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5106 | #[cfg_attr (test, assert_instr(vpackssdw))] |
5107 | pub unsafe fn _mm512_packs_epi32(a: __m512i, b: __m512i) -> __m512i { |
5108 | transmute(src:vpackssdw(a:a.as_i32x16(), b:b.as_i32x16())) |
5109 | } |
5110 | |
5111 | /// 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). |
5112 | /// |
5113 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_packs_epi32&expand=4089) |
5114 | #[inline ] |
5115 | #[target_feature (enable = "avx512bw" )] |
5116 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5117 | #[cfg_attr (test, assert_instr(vpackssdw))] |
5118 | pub unsafe fn _mm512_mask_packs_epi32( |
5119 | src: __m512i, |
5120 | k: __mmask32, |
5121 | a: __m512i, |
5122 | b: __m512i, |
5123 | ) -> __m512i { |
5124 | let pack: i16x32 = _mm512_packs_epi32(a, b).as_i16x32(); |
5125 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i16x32())) |
5126 | } |
5127 | |
5128 | /// 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). |
5129 | /// |
5130 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_packs_epi32&expand=4090) |
5131 | #[inline ] |
5132 | #[target_feature (enable = "avx512bw" )] |
5133 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5134 | #[cfg_attr (test, assert_instr(vpackssdw))] |
5135 | pub unsafe fn _mm512_maskz_packs_epi32(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
5136 | let pack: i16x32 = _mm512_packs_epi32(a, b).as_i16x32(); |
5137 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
5138 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:zero)) |
5139 | } |
5140 | |
5141 | /// 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). |
5142 | /// |
5143 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_packs_epi32&expand=4086) |
5144 | #[inline ] |
5145 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5146 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5147 | #[cfg_attr (test, assert_instr(vpackssdw))] |
5148 | pub unsafe fn _mm256_mask_packs_epi32( |
5149 | src: __m256i, |
5150 | k: __mmask16, |
5151 | a: __m256i, |
5152 | b: __m256i, |
5153 | ) -> __m256i { |
5154 | let pack: i16x16 = _mm256_packs_epi32(a, b).as_i16x16(); |
5155 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i16x16())) |
5156 | } |
5157 | |
5158 | /// 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). |
5159 | /// |
5160 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_packs_epi32&expand=4087) |
5161 | #[inline ] |
5162 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5163 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5164 | #[cfg_attr (test, assert_instr(vpackssdw))] |
5165 | pub unsafe fn _mm256_maskz_packs_epi32(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
5166 | let pack: i16x16 = _mm256_packs_epi32(a, b).as_i16x16(); |
5167 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
5168 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:zero)) |
5169 | } |
5170 | |
5171 | /// 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). |
5172 | /// |
5173 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_packs_epi32&expand=4083) |
5174 | #[inline ] |
5175 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5176 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5177 | #[cfg_attr (test, assert_instr(vpackssdw))] |
5178 | pub unsafe fn _mm_mask_packs_epi32(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
5179 | let pack: i16x8 = _mm_packs_epi32(a, b).as_i16x8(); |
5180 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i16x8())) |
5181 | } |
5182 | |
5183 | /// 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). |
5184 | /// |
5185 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_packs_epi32&expand=4084) |
5186 | #[inline ] |
5187 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5188 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5189 | #[cfg_attr (test, assert_instr(vpackssdw))] |
5190 | pub unsafe fn _mm_maskz_packs_epi32(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
5191 | let pack: i16x8 = _mm_packs_epi32(a, b).as_i16x8(); |
5192 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
5193 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:zero)) |
5194 | } |
5195 | |
5196 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using signed saturation, and store the results in dst. |
5197 | /// |
5198 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_packs_epi16&expand=4082) |
5199 | #[inline ] |
5200 | #[target_feature (enable = "avx512bw" )] |
5201 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5202 | #[cfg_attr (test, assert_instr(vpacksswb))] |
5203 | pub unsafe fn _mm512_packs_epi16(a: __m512i, b: __m512i) -> __m512i { |
5204 | transmute(src:vpacksswb(a:a.as_i16x32(), b:b.as_i16x32())) |
5205 | } |
5206 | |
5207 | /// 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). |
5208 | /// |
5209 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_packs_epi16&expand=4080) |
5210 | #[inline ] |
5211 | #[target_feature (enable = "avx512bw" )] |
5212 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5213 | #[cfg_attr (test, assert_instr(vpacksswb))] |
5214 | pub unsafe fn _mm512_mask_packs_epi16( |
5215 | src: __m512i, |
5216 | k: __mmask64, |
5217 | a: __m512i, |
5218 | b: __m512i, |
5219 | ) -> __m512i { |
5220 | let pack: i8x64 = _mm512_packs_epi16(a, b).as_i8x64(); |
5221 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i8x64())) |
5222 | } |
5223 | |
5224 | /// 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). |
5225 | /// |
5226 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_packs_epi16&expand=4081) |
5227 | #[inline ] |
5228 | #[target_feature (enable = "avx512bw" )] |
5229 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5230 | #[cfg_attr (test, assert_instr(vpacksswb))] |
5231 | pub unsafe fn _mm512_maskz_packs_epi16(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
5232 | let pack: i8x64 = _mm512_packs_epi16(a, b).as_i8x64(); |
5233 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
5234 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:zero)) |
5235 | } |
5236 | |
5237 | /// 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). |
5238 | /// |
5239 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_packs_epi16&expand=4077) |
5240 | #[inline ] |
5241 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5242 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5243 | #[cfg_attr (test, assert_instr(vpacksswb))] |
5244 | pub unsafe fn _mm256_mask_packs_epi16( |
5245 | src: __m256i, |
5246 | k: __mmask32, |
5247 | a: __m256i, |
5248 | b: __m256i, |
5249 | ) -> __m256i { |
5250 | let pack: i8x32 = _mm256_packs_epi16(a, b).as_i8x32(); |
5251 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i8x32())) |
5252 | } |
5253 | |
5254 | /// 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). |
5255 | /// |
5256 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=#text=_mm256_maskz_packs_epi16&expand=4078) |
5257 | #[inline ] |
5258 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5259 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5260 | #[cfg_attr (test, assert_instr(vpacksswb))] |
5261 | pub unsafe fn _mm256_maskz_packs_epi16(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
5262 | let pack: i8x32 = _mm256_packs_epi16(a, b).as_i8x32(); |
5263 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
5264 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:zero)) |
5265 | } |
5266 | |
5267 | /// 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). |
5268 | /// |
5269 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_packs_epi16&expand=4074) |
5270 | #[inline ] |
5271 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5272 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5273 | #[cfg_attr (test, assert_instr(vpacksswb))] |
5274 | pub unsafe fn _mm_mask_packs_epi16(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
5275 | let pack: i8x16 = _mm_packs_epi16(a, b).as_i8x16(); |
5276 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i8x16())) |
5277 | } |
5278 | |
5279 | /// 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). |
5280 | /// |
5281 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_packs_epi16&expand=4075) |
5282 | #[inline ] |
5283 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5284 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5285 | #[cfg_attr (test, assert_instr(vpacksswb))] |
5286 | pub unsafe fn _mm_maskz_packs_epi16(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
5287 | let pack: i8x16 = _mm_packs_epi16(a, b).as_i8x16(); |
5288 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
5289 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:zero)) |
5290 | } |
5291 | |
5292 | /// Convert packed signed 32-bit integers from a and b to packed 16-bit integers using unsigned saturation, and store the results in dst. |
5293 | /// |
5294 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_packus_epi32&expand=4130) |
5295 | #[inline ] |
5296 | #[target_feature (enable = "avx512bw" )] |
5297 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5298 | #[cfg_attr (test, assert_instr(vpackusdw))] |
5299 | pub unsafe fn _mm512_packus_epi32(a: __m512i, b: __m512i) -> __m512i { |
5300 | transmute(src:vpackusdw(a:a.as_i32x16(), b:b.as_i32x16())) |
5301 | } |
5302 | |
5303 | /// 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). |
5304 | /// |
5305 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_packus_epi32&expand=4128) |
5306 | #[inline ] |
5307 | #[target_feature (enable = "avx512bw" )] |
5308 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5309 | #[cfg_attr (test, assert_instr(vpackusdw))] |
5310 | pub unsafe fn _mm512_mask_packus_epi32( |
5311 | src: __m512i, |
5312 | k: __mmask32, |
5313 | a: __m512i, |
5314 | b: __m512i, |
5315 | ) -> __m512i { |
5316 | let pack: i16x32 = _mm512_packus_epi32(a, b).as_i16x32(); |
5317 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i16x32())) |
5318 | } |
5319 | |
5320 | /// 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). |
5321 | /// |
5322 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_packus_epi32&expand=4129) |
5323 | #[inline ] |
5324 | #[target_feature (enable = "avx512bw" )] |
5325 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5326 | #[cfg_attr (test, assert_instr(vpackusdw))] |
5327 | pub unsafe fn _mm512_maskz_packus_epi32(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
5328 | let pack: i16x32 = _mm512_packus_epi32(a, b).as_i16x32(); |
5329 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
5330 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:zero)) |
5331 | } |
5332 | |
5333 | /// 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). |
5334 | /// |
5335 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_packus_epi32&expand=4125) |
5336 | #[inline ] |
5337 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5338 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5339 | #[cfg_attr (test, assert_instr(vpackusdw))] |
5340 | pub unsafe fn _mm256_mask_packus_epi32( |
5341 | src: __m256i, |
5342 | k: __mmask16, |
5343 | a: __m256i, |
5344 | b: __m256i, |
5345 | ) -> __m256i { |
5346 | let pack: i16x16 = _mm256_packus_epi32(a, b).as_i16x16(); |
5347 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i16x16())) |
5348 | } |
5349 | |
5350 | /// 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). |
5351 | /// |
5352 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_packus_epi32&expand=4126) |
5353 | #[inline ] |
5354 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5355 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5356 | #[cfg_attr (test, assert_instr(vpackusdw))] |
5357 | pub unsafe fn _mm256_maskz_packus_epi32(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
5358 | let pack: i16x16 = _mm256_packus_epi32(a, b).as_i16x16(); |
5359 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
5360 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:zero)) |
5361 | } |
5362 | |
5363 | /// 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). |
5364 | /// |
5365 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_packus_epi32&expand=4122) |
5366 | #[inline ] |
5367 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5368 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5369 | #[cfg_attr (test, assert_instr(vpackusdw))] |
5370 | pub unsafe fn _mm_mask_packus_epi32(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
5371 | let pack: i16x8 = _mm_packus_epi32(a, b).as_i16x8(); |
5372 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i16x8())) |
5373 | } |
5374 | |
5375 | /// 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). |
5376 | /// |
5377 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_packus_epi32&expand=4123) |
5378 | #[inline ] |
5379 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5380 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5381 | #[cfg_attr (test, assert_instr(vpackusdw))] |
5382 | pub unsafe fn _mm_maskz_packus_epi32(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
5383 | let pack: i16x8 = _mm_packus_epi32(a, b).as_i16x8(); |
5384 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
5385 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:zero)) |
5386 | } |
5387 | |
5388 | /// Convert packed signed 16-bit integers from a and b to packed 8-bit integers using unsigned saturation, and store the results in dst. |
5389 | /// |
5390 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_packus_epi16&expand=4121) |
5391 | #[inline ] |
5392 | #[target_feature (enable = "avx512bw" )] |
5393 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5394 | #[cfg_attr (test, assert_instr(vpackuswb))] |
5395 | pub unsafe fn _mm512_packus_epi16(a: __m512i, b: __m512i) -> __m512i { |
5396 | transmute(src:vpackuswb(a:a.as_i16x32(), b:b.as_i16x32())) |
5397 | } |
5398 | |
5399 | /// 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). |
5400 | /// |
5401 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_packus_epi16&expand=4119) |
5402 | #[inline ] |
5403 | #[target_feature (enable = "avx512bw" )] |
5404 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5405 | #[cfg_attr (test, assert_instr(vpackuswb))] |
5406 | pub unsafe fn _mm512_mask_packus_epi16( |
5407 | src: __m512i, |
5408 | k: __mmask64, |
5409 | a: __m512i, |
5410 | b: __m512i, |
5411 | ) -> __m512i { |
5412 | let pack: i8x64 = _mm512_packus_epi16(a, b).as_i8x64(); |
5413 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i8x64())) |
5414 | } |
5415 | |
5416 | /// 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). |
5417 | /// |
5418 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_packus_epi16&expand=4120) |
5419 | #[inline ] |
5420 | #[target_feature (enable = "avx512bw" )] |
5421 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5422 | #[cfg_attr (test, assert_instr(vpackuswb))] |
5423 | pub unsafe fn _mm512_maskz_packus_epi16(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
5424 | let pack: i8x64 = _mm512_packus_epi16(a, b).as_i8x64(); |
5425 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
5426 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:zero)) |
5427 | } |
5428 | |
5429 | /// 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). |
5430 | /// |
5431 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_packus_epi16&expand=4116) |
5432 | #[inline ] |
5433 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5434 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5435 | #[cfg_attr (test, assert_instr(vpackuswb))] |
5436 | pub unsafe fn _mm256_mask_packus_epi16( |
5437 | src: __m256i, |
5438 | k: __mmask32, |
5439 | a: __m256i, |
5440 | b: __m256i, |
5441 | ) -> __m256i { |
5442 | let pack: i8x32 = _mm256_packus_epi16(a, b).as_i8x32(); |
5443 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i8x32())) |
5444 | } |
5445 | |
5446 | /// 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). |
5447 | /// |
5448 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_packus_epi16&expand=4117) |
5449 | #[inline ] |
5450 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5451 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5452 | #[cfg_attr (test, assert_instr(vpackuswb))] |
5453 | pub unsafe fn _mm256_maskz_packus_epi16(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
5454 | let pack: i8x32 = _mm256_packus_epi16(a, b).as_i8x32(); |
5455 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
5456 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:zero)) |
5457 | } |
5458 | |
5459 | /// 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). |
5460 | /// |
5461 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_packus_epi16&expand=4113) |
5462 | #[inline ] |
5463 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5464 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5465 | #[cfg_attr (test, assert_instr(vpackuswb))] |
5466 | pub unsafe fn _mm_mask_packus_epi16(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
5467 | let pack: i8x16 = _mm_packus_epi16(a, b).as_i8x16(); |
5468 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:src.as_i8x16())) |
5469 | } |
5470 | |
5471 | /// 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). |
5472 | /// |
5473 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_packus_epi16&expand=4114) |
5474 | #[inline ] |
5475 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5476 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5477 | #[cfg_attr (test, assert_instr(vpackuswb))] |
5478 | pub unsafe fn _mm_maskz_packus_epi16(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
5479 | let pack: i8x16 = _mm_packus_epi16(a, b).as_i8x16(); |
5480 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
5481 | transmute(src:simd_select_bitmask(m:k, yes:pack, no:zero)) |
5482 | } |
5483 | |
5484 | /// Average packed unsigned 16-bit integers in a and b, and store the results in dst. |
5485 | /// |
5486 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_avg_epu16&expand=388) |
5487 | #[inline ] |
5488 | #[target_feature (enable = "avx512bw" )] |
5489 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5490 | #[cfg_attr (test, assert_instr(vpavgw))] |
5491 | pub unsafe fn _mm512_avg_epu16(a: __m512i, b: __m512i) -> __m512i { |
5492 | transmute(src:vpavgw(a:a.as_u16x32(), b:b.as_u16x32())) |
5493 | } |
5494 | |
5495 | /// 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). |
5496 | /// |
5497 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_avg_epu16&expand=389) |
5498 | #[inline ] |
5499 | #[target_feature (enable = "avx512bw" )] |
5500 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5501 | #[cfg_attr (test, assert_instr(vpavgw))] |
5502 | pub unsafe fn _mm512_mask_avg_epu16(src: __m512i, k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
5503 | let avg: u16x32 = _mm512_avg_epu16(a, b).as_u16x32(); |
5504 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:src.as_u16x32())) |
5505 | } |
5506 | |
5507 | /// 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). |
5508 | /// |
5509 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_avg_epu16&expand=390) |
5510 | #[inline ] |
5511 | #[target_feature (enable = "avx512bw" )] |
5512 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5513 | #[cfg_attr (test, assert_instr(vpavgw))] |
5514 | pub unsafe fn _mm512_maskz_avg_epu16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
5515 | let avg: u16x32 = _mm512_avg_epu16(a, b).as_u16x32(); |
5516 | let zero: u16x32 = _mm512_setzero_si512().as_u16x32(); |
5517 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:zero)) |
5518 | } |
5519 | |
5520 | /// 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). |
5521 | /// |
5522 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_avg_epu16&expand=386) |
5523 | #[inline ] |
5524 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5525 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5526 | #[cfg_attr (test, assert_instr(vpavgw))] |
5527 | pub unsafe fn _mm256_mask_avg_epu16(src: __m256i, k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
5528 | let avg: u16x16 = _mm256_avg_epu16(a, b).as_u16x16(); |
5529 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:src.as_u16x16())) |
5530 | } |
5531 | |
5532 | /// 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). |
5533 | /// |
5534 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_avg_epu16&expand=387) |
5535 | #[inline ] |
5536 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5537 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5538 | #[cfg_attr (test, assert_instr(vpavgw))] |
5539 | pub unsafe fn _mm256_maskz_avg_epu16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
5540 | let avg: u16x16 = _mm256_avg_epu16(a, b).as_u16x16(); |
5541 | let zero: u16x16 = _mm256_setzero_si256().as_u16x16(); |
5542 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:zero)) |
5543 | } |
5544 | |
5545 | /// 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). |
5546 | /// |
5547 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_avg_epu16&expand=383) |
5548 | #[inline ] |
5549 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5550 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5551 | #[cfg_attr (test, assert_instr(vpavgw))] |
5552 | pub unsafe fn _mm_mask_avg_epu16(src: __m128i, k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
5553 | let avg: u16x8 = _mm_avg_epu16(a, b).as_u16x8(); |
5554 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:src.as_u16x8())) |
5555 | } |
5556 | |
5557 | /// 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). |
5558 | /// |
5559 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_avg_epu16&expand=384) |
5560 | #[inline ] |
5561 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5562 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5563 | #[cfg_attr (test, assert_instr(vpavgw))] |
5564 | pub unsafe fn _mm_maskz_avg_epu16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
5565 | let avg: u16x8 = _mm_avg_epu16(a, b).as_u16x8(); |
5566 | let zero: u16x8 = _mm_setzero_si128().as_u16x8(); |
5567 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:zero)) |
5568 | } |
5569 | |
5570 | /// Average packed unsigned 8-bit integers in a and b, and store the results in dst. |
5571 | /// |
5572 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_avg_epu8&expand=397) |
5573 | #[inline ] |
5574 | #[target_feature (enable = "avx512bw" )] |
5575 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5576 | #[cfg_attr (test, assert_instr(vpavgb))] |
5577 | pub unsafe fn _mm512_avg_epu8(a: __m512i, b: __m512i) -> __m512i { |
5578 | transmute(src:vpavgb(a:a.as_u8x64(), b:b.as_u8x64())) |
5579 | } |
5580 | |
5581 | /// 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). |
5582 | /// |
5583 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_avg_epu8&expand=398) |
5584 | #[inline ] |
5585 | #[target_feature (enable = "avx512bw" )] |
5586 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5587 | #[cfg_attr (test, assert_instr(vpavgb))] |
5588 | pub unsafe fn _mm512_mask_avg_epu8(src: __m512i, k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
5589 | let avg: u8x64 = _mm512_avg_epu8(a, b).as_u8x64(); |
5590 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:src.as_u8x64())) |
5591 | } |
5592 | |
5593 | /// 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). |
5594 | /// |
5595 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_avg_epu8&expand=399) |
5596 | #[inline ] |
5597 | #[target_feature (enable = "avx512bw" )] |
5598 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5599 | #[cfg_attr (test, assert_instr(vpavgb))] |
5600 | pub unsafe fn _mm512_maskz_avg_epu8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
5601 | let avg: u8x64 = _mm512_avg_epu8(a, b).as_u8x64(); |
5602 | let zero: u8x64 = _mm512_setzero_si512().as_u8x64(); |
5603 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:zero)) |
5604 | } |
5605 | |
5606 | /// 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). |
5607 | /// |
5608 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_avg_epu8&expand=395) |
5609 | #[inline ] |
5610 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5611 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5612 | #[cfg_attr (test, assert_instr(vpavgb))] |
5613 | pub unsafe fn _mm256_mask_avg_epu8(src: __m256i, k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
5614 | let avg: u8x32 = _mm256_avg_epu8(a, b).as_u8x32(); |
5615 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:src.as_u8x32())) |
5616 | } |
5617 | |
5618 | /// 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). |
5619 | /// |
5620 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_avg_epu8&expand=396) |
5621 | #[inline ] |
5622 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5623 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5624 | #[cfg_attr (test, assert_instr(vpavgb))] |
5625 | pub unsafe fn _mm256_maskz_avg_epu8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
5626 | let avg: u8x32 = _mm256_avg_epu8(a, b).as_u8x32(); |
5627 | let zero: u8x32 = _mm256_setzero_si256().as_u8x32(); |
5628 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:zero)) |
5629 | } |
5630 | |
5631 | /// 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). |
5632 | /// |
5633 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_avg_epu8&expand=392) |
5634 | #[inline ] |
5635 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5636 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5637 | #[cfg_attr (test, assert_instr(vpavgb))] |
5638 | pub unsafe fn _mm_mask_avg_epu8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
5639 | let avg: u8x16 = _mm_avg_epu8(a, b).as_u8x16(); |
5640 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:src.as_u8x16())) |
5641 | } |
5642 | |
5643 | /// 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). |
5644 | /// |
5645 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_avg_epu8&expand=393) |
5646 | #[inline ] |
5647 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5648 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5649 | #[cfg_attr (test, assert_instr(vpavgb))] |
5650 | pub unsafe fn _mm_maskz_avg_epu8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
5651 | let avg: u8x16 = _mm_avg_epu8(a, b).as_u8x16(); |
5652 | let zero: u8x16 = _mm_setzero_si128().as_u8x16(); |
5653 | transmute(src:simd_select_bitmask(m:k, yes:avg, no:zero)) |
5654 | } |
5655 | |
5656 | /// Shift packed 16-bit integers in a left by count while shifting in zeros, and store the results in dst. |
5657 | /// |
5658 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_sll_epi16&expand=5271) |
5659 | #[inline ] |
5660 | #[target_feature (enable = "avx512bw" )] |
5661 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5662 | #[cfg_attr (test, assert_instr(vpsllw))] |
5663 | pub unsafe fn _mm512_sll_epi16(a: __m512i, count: __m128i) -> __m512i { |
5664 | transmute(src:vpsllw(a:a.as_i16x32(), count:count.as_i16x8())) |
5665 | } |
5666 | |
5667 | /// 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). |
5668 | /// |
5669 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_sll_epi16&expand=5269) |
5670 | #[inline ] |
5671 | #[target_feature (enable = "avx512bw" )] |
5672 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5673 | #[cfg_attr (test, assert_instr(vpsllw))] |
5674 | pub unsafe fn _mm512_mask_sll_epi16( |
5675 | src: __m512i, |
5676 | k: __mmask32, |
5677 | a: __m512i, |
5678 | count: __m128i, |
5679 | ) -> __m512i { |
5680 | let shf: i16x32 = _mm512_sll_epi16(a, count).as_i16x32(); |
5681 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
5682 | } |
5683 | |
5684 | /// 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). |
5685 | /// |
5686 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_sll_epi16&expand=5270) |
5687 | #[inline ] |
5688 | #[target_feature (enable = "avx512bw" )] |
5689 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5690 | #[cfg_attr (test, assert_instr(vpsllw))] |
5691 | pub unsafe fn _mm512_maskz_sll_epi16(k: __mmask32, a: __m512i, count: __m128i) -> __m512i { |
5692 | let shf: i16x32 = _mm512_sll_epi16(a, count).as_i16x32(); |
5693 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
5694 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
5695 | } |
5696 | |
5697 | /// 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). |
5698 | /// |
5699 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_sll_epi16&expand=5266) |
5700 | #[inline ] |
5701 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5702 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5703 | #[cfg_attr (test, assert_instr(vpsllw))] |
5704 | pub unsafe fn _mm256_mask_sll_epi16( |
5705 | src: __m256i, |
5706 | k: __mmask16, |
5707 | a: __m256i, |
5708 | count: __m128i, |
5709 | ) -> __m256i { |
5710 | let shf: i16x16 = _mm256_sll_epi16(a, count).as_i16x16(); |
5711 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x16())) |
5712 | } |
5713 | |
5714 | /// 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). |
5715 | /// |
5716 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_sll_epi16&expand=5267) |
5717 | #[inline ] |
5718 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5719 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5720 | #[cfg_attr (test, assert_instr(vpsllw))] |
5721 | pub unsafe fn _mm256_maskz_sll_epi16(k: __mmask16, a: __m256i, count: __m128i) -> __m256i { |
5722 | let shf: i16x16 = _mm256_sll_epi16(a, count).as_i16x16(); |
5723 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
5724 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
5725 | } |
5726 | |
5727 | /// 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). |
5728 | /// |
5729 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_sll_epi16&expand=5263) |
5730 | #[inline ] |
5731 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5732 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5733 | #[cfg_attr (test, assert_instr(vpsllw))] |
5734 | pub unsafe fn _mm_mask_sll_epi16(src: __m128i, k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
5735 | let shf: i16x8 = _mm_sll_epi16(a, count).as_i16x8(); |
5736 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x8())) |
5737 | } |
5738 | |
5739 | /// 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). |
5740 | /// |
5741 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_sll_epi16&expand=5264) |
5742 | #[inline ] |
5743 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5744 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5745 | #[cfg_attr (test, assert_instr(vpsllw))] |
5746 | pub unsafe fn _mm_maskz_sll_epi16(k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
5747 | let shf: i16x8 = _mm_sll_epi16(a, count).as_i16x8(); |
5748 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
5749 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
5750 | } |
5751 | |
5752 | /// Shift packed 16-bit integers in a left by imm8 while shifting in zeros, and store the results in dst. |
5753 | /// |
5754 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_slli_epi16&expand=5301) |
5755 | #[inline ] |
5756 | #[target_feature (enable = "avx512bw" )] |
5757 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5758 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
5759 | #[rustc_legacy_const_generics (1)] |
5760 | pub unsafe fn _mm512_slli_epi16<const IMM8: u32>(a: __m512i) -> __m512i { |
5761 | static_assert_uimm_bits!(IMM8, 8); |
5762 | if IMM8 >= 16 { |
5763 | _mm512_setzero_si512() |
5764 | } else { |
5765 | transmute(src:simd_shl(lhs:a.as_u16x32(), rhs:u16x32::splat(IMM8 as u16))) |
5766 | } |
5767 | } |
5768 | |
5769 | /// 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). |
5770 | /// |
5771 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_slli_epi16&expand=5299) |
5772 | #[inline ] |
5773 | #[target_feature (enable = "avx512bw" )] |
5774 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5775 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
5776 | #[rustc_legacy_const_generics (3)] |
5777 | pub unsafe fn _mm512_mask_slli_epi16<const IMM8: u32>( |
5778 | src: __m512i, |
5779 | k: __mmask32, |
5780 | a: __m512i, |
5781 | ) -> __m512i { |
5782 | static_assert_uimm_bits!(IMM8, 8); |
5783 | let shf: u16x32 = if IMM8 >= 16 { |
5784 | u16x32::splat(0) |
5785 | } else { |
5786 | simd_shl(lhs:a.as_u16x32(), rhs:u16x32::splat(IMM8 as u16)) |
5787 | }; |
5788 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_u16x32())) |
5789 | } |
5790 | |
5791 | /// 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). |
5792 | /// |
5793 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_slli_epi16&expand=5300) |
5794 | #[inline ] |
5795 | #[target_feature (enable = "avx512bw" )] |
5796 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5797 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
5798 | #[rustc_legacy_const_generics (2)] |
5799 | pub unsafe fn _mm512_maskz_slli_epi16<const IMM8: u32>(k: __mmask32, a: __m512i) -> __m512i { |
5800 | static_assert_uimm_bits!(IMM8, 8); |
5801 | if IMM8 >= 16 { |
5802 | _mm512_setzero_si512() |
5803 | } else { |
5804 | let shf: u16x32 = simd_shl(lhs:a.as_u16x32(), rhs:u16x32::splat(IMM8 as u16)); |
5805 | let zero: u16x32 = u16x32::splat(0); |
5806 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
5807 | } |
5808 | } |
5809 | |
5810 | /// 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). |
5811 | /// |
5812 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_slli_epi16&expand=5296) |
5813 | #[inline ] |
5814 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5815 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5816 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
5817 | #[rustc_legacy_const_generics (3)] |
5818 | pub unsafe fn _mm256_mask_slli_epi16<const IMM8: u32>( |
5819 | src: __m256i, |
5820 | k: __mmask16, |
5821 | a: __m256i, |
5822 | ) -> __m256i { |
5823 | static_assert_uimm_bits!(IMM8, 8); |
5824 | let shf: u16x16 = if IMM8 >= 16 { |
5825 | u16x16::splat(0) |
5826 | } else { |
5827 | simd_shl(lhs:a.as_u16x16(), rhs:u16x16::splat(IMM8 as u16)) |
5828 | }; |
5829 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_u16x16())) |
5830 | } |
5831 | |
5832 | /// 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). |
5833 | /// |
5834 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_slli_epi16&expand=5297) |
5835 | #[inline ] |
5836 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5837 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5838 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
5839 | #[rustc_legacy_const_generics (2)] |
5840 | pub unsafe fn _mm256_maskz_slli_epi16<const IMM8: u32>(k: __mmask16, a: __m256i) -> __m256i { |
5841 | static_assert_uimm_bits!(IMM8, 8); |
5842 | if IMM8 >= 16 { |
5843 | _mm256_setzero_si256() |
5844 | } else { |
5845 | let shf: u16x16 = simd_shl(lhs:a.as_u16x16(), rhs:u16x16::splat(IMM8 as u16)); |
5846 | let zero: u16x16 = u16x16::splat(0); |
5847 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
5848 | } |
5849 | } |
5850 | |
5851 | /// 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). |
5852 | /// |
5853 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_slli_epi16&expand=5293) |
5854 | #[inline ] |
5855 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5856 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5857 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
5858 | #[rustc_legacy_const_generics (3)] |
5859 | pub unsafe fn _mm_mask_slli_epi16<const IMM8: u32>( |
5860 | src: __m128i, |
5861 | k: __mmask8, |
5862 | a: __m128i, |
5863 | ) -> __m128i { |
5864 | static_assert_uimm_bits!(IMM8, 8); |
5865 | let shf: u16x8 = if IMM8 >= 16 { |
5866 | u16x8::splat(0) |
5867 | } else { |
5868 | simd_shl(lhs:a.as_u16x8(), rhs:u16x8::splat(IMM8 as u16)) |
5869 | }; |
5870 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_u16x8())) |
5871 | } |
5872 | |
5873 | /// 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). |
5874 | /// |
5875 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_slli_epi16&expand=5294) |
5876 | #[inline ] |
5877 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5878 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5879 | #[cfg_attr (test, assert_instr(vpsllw, IMM8 = 5))] |
5880 | #[rustc_legacy_const_generics (2)] |
5881 | pub unsafe fn _mm_maskz_slli_epi16<const IMM8: u32>(k: __mmask8, a: __m128i) -> __m128i { |
5882 | static_assert_uimm_bits!(IMM8, 8); |
5883 | if IMM8 >= 16 { |
5884 | _mm_setzero_si128() |
5885 | } else { |
5886 | let shf: u16x8 = simd_shl(lhs:a.as_u16x8(), rhs:u16x8::splat(IMM8 as u16)); |
5887 | let zero: u16x8 = u16x8::splat(0); |
5888 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
5889 | } |
5890 | } |
5891 | |
5892 | /// 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. |
5893 | /// |
5894 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_sllv_epi16&expand=5333) |
5895 | #[inline ] |
5896 | #[target_feature (enable = "avx512bw" )] |
5897 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5898 | #[cfg_attr (test, assert_instr(vpsllvw))] |
5899 | pub unsafe fn _mm512_sllv_epi16(a: __m512i, count: __m512i) -> __m512i { |
5900 | transmute(src:vpsllvw(a:a.as_i16x32(), b:count.as_i16x32())) |
5901 | } |
5902 | |
5903 | /// 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). |
5904 | /// |
5905 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_sllv_epi16&expand=5331) |
5906 | #[inline ] |
5907 | #[target_feature (enable = "avx512bw" )] |
5908 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5909 | #[cfg_attr (test, assert_instr(vpsllvw))] |
5910 | pub unsafe fn _mm512_mask_sllv_epi16( |
5911 | src: __m512i, |
5912 | k: __mmask32, |
5913 | a: __m512i, |
5914 | count: __m512i, |
5915 | ) -> __m512i { |
5916 | let shf: i16x32 = _mm512_sllv_epi16(a, count).as_i16x32(); |
5917 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
5918 | } |
5919 | |
5920 | /// 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). |
5921 | /// |
5922 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_sllv_epi16&expand=5332) |
5923 | #[inline ] |
5924 | #[target_feature (enable = "avx512bw" )] |
5925 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5926 | #[cfg_attr (test, assert_instr(vpsllvw))] |
5927 | pub unsafe fn _mm512_maskz_sllv_epi16(k: __mmask32, a: __m512i, count: __m512i) -> __m512i { |
5928 | let shf: i16x32 = _mm512_sllv_epi16(a, count).as_i16x32(); |
5929 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
5930 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
5931 | } |
5932 | |
5933 | /// 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. |
5934 | /// |
5935 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_sllv_epi16&expand=5330) |
5936 | #[inline ] |
5937 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5938 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5939 | #[cfg_attr (test, assert_instr(vpsllvw))] |
5940 | pub unsafe fn _mm256_sllv_epi16(a: __m256i, count: __m256i) -> __m256i { |
5941 | transmute(src:vpsllvw256(a:a.as_i16x16(), b:count.as_i16x16())) |
5942 | } |
5943 | |
5944 | /// 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). |
5945 | /// |
5946 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_sllv_epi16&expand=5328) |
5947 | #[inline ] |
5948 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5949 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5950 | #[cfg_attr (test, assert_instr(vpsllvw))] |
5951 | pub unsafe fn _mm256_mask_sllv_epi16( |
5952 | src: __m256i, |
5953 | k: __mmask16, |
5954 | a: __m256i, |
5955 | count: __m256i, |
5956 | ) -> __m256i { |
5957 | let shf: i16x16 = _mm256_sllv_epi16(a, count).as_i16x16(); |
5958 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x16())) |
5959 | } |
5960 | |
5961 | /// 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). |
5962 | /// |
5963 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_sllv_epi16&expand=5329) |
5964 | #[inline ] |
5965 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5966 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5967 | #[cfg_attr (test, assert_instr(vpsllvw))] |
5968 | pub unsafe fn _mm256_maskz_sllv_epi16(k: __mmask16, a: __m256i, count: __m256i) -> __m256i { |
5969 | let shf: i16x16 = _mm256_sllv_epi16(a, count).as_i16x16(); |
5970 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
5971 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
5972 | } |
5973 | |
5974 | /// 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. |
5975 | /// |
5976 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_sllv_epi16&expand=5327) |
5977 | #[inline ] |
5978 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5979 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5980 | #[cfg_attr (test, assert_instr(vpsllvw))] |
5981 | pub unsafe fn _mm_sllv_epi16(a: __m128i, count: __m128i) -> __m128i { |
5982 | transmute(src:vpsllvw128(a:a.as_i16x8(), b:count.as_i16x8())) |
5983 | } |
5984 | |
5985 | /// 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). |
5986 | /// |
5987 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_sllv_epi16&expand=5325) |
5988 | #[inline ] |
5989 | #[target_feature (enable = "avx512bw,avx512vl" )] |
5990 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
5991 | #[cfg_attr (test, assert_instr(vpsllvw))] |
5992 | pub unsafe fn _mm_mask_sllv_epi16( |
5993 | src: __m128i, |
5994 | k: __mmask8, |
5995 | a: __m128i, |
5996 | count: __m128i, |
5997 | ) -> __m128i { |
5998 | let shf: i16x8 = _mm_sllv_epi16(a, count).as_i16x8(); |
5999 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x8())) |
6000 | } |
6001 | |
6002 | /// 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). |
6003 | /// |
6004 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_sllv_epi16&expand=5326) |
6005 | #[inline ] |
6006 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6007 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6008 | #[cfg_attr (test, assert_instr(vpsllvw))] |
6009 | pub unsafe fn _mm_maskz_sllv_epi16(k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
6010 | let shf: i16x8 = _mm_sllv_epi16(a, count).as_i16x8(); |
6011 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
6012 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6013 | } |
6014 | |
6015 | /// Shift packed 16-bit integers in a right by count while shifting in zeros, and store the results in dst. |
6016 | /// |
6017 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_srl_epi16&expand=5483) |
6018 | #[inline ] |
6019 | #[target_feature (enable = "avx512bw" )] |
6020 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6021 | #[cfg_attr (test, assert_instr(vpsrlw))] |
6022 | pub unsafe fn _mm512_srl_epi16(a: __m512i, count: __m128i) -> __m512i { |
6023 | transmute(src:vpsrlw(a:a.as_i16x32(), count:count.as_i16x8())) |
6024 | } |
6025 | |
6026 | /// 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). |
6027 | /// |
6028 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_srl_epi16&expand=5481) |
6029 | #[inline ] |
6030 | #[target_feature (enable = "avx512bw" )] |
6031 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6032 | #[cfg_attr (test, assert_instr(vpsrlw))] |
6033 | pub unsafe fn _mm512_mask_srl_epi16( |
6034 | src: __m512i, |
6035 | k: __mmask32, |
6036 | a: __m512i, |
6037 | count: __m128i, |
6038 | ) -> __m512i { |
6039 | let shf: i16x32 = _mm512_srl_epi16(a, count).as_i16x32(); |
6040 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
6041 | } |
6042 | |
6043 | /// 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). |
6044 | /// |
6045 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_srl_epi16&expand=5482) |
6046 | #[inline ] |
6047 | #[target_feature (enable = "avx512bw" )] |
6048 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6049 | #[cfg_attr (test, assert_instr(vpsrlw))] |
6050 | pub unsafe fn _mm512_maskz_srl_epi16(k: __mmask32, a: __m512i, count: __m128i) -> __m512i { |
6051 | let shf: i16x32 = _mm512_srl_epi16(a, count).as_i16x32(); |
6052 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
6053 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6054 | } |
6055 | |
6056 | /// 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). |
6057 | /// |
6058 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_srl_epi16&expand=5478) |
6059 | #[inline ] |
6060 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6061 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6062 | #[cfg_attr (test, assert_instr(vpsrlw))] |
6063 | pub unsafe fn _mm256_mask_srl_epi16( |
6064 | src: __m256i, |
6065 | k: __mmask16, |
6066 | a: __m256i, |
6067 | count: __m128i, |
6068 | ) -> __m256i { |
6069 | let shf: i16x16 = _mm256_srl_epi16(a, count).as_i16x16(); |
6070 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x16())) |
6071 | } |
6072 | |
6073 | /// 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). |
6074 | /// |
6075 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_srl_epi16&expand=5479) |
6076 | #[inline ] |
6077 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6078 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6079 | #[cfg_attr (test, assert_instr(vpsrlw))] |
6080 | pub unsafe fn _mm256_maskz_srl_epi16(k: __mmask16, a: __m256i, count: __m128i) -> __m256i { |
6081 | let shf: i16x16 = _mm256_srl_epi16(a, count).as_i16x16(); |
6082 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
6083 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6084 | } |
6085 | |
6086 | /// 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). |
6087 | /// |
6088 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_srl_epi16&expand=5475) |
6089 | #[inline ] |
6090 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6091 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6092 | #[cfg_attr (test, assert_instr(vpsrlw))] |
6093 | pub unsafe fn _mm_mask_srl_epi16(src: __m128i, k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
6094 | let shf: i16x8 = _mm_srl_epi16(a, count).as_i16x8(); |
6095 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x8())) |
6096 | } |
6097 | |
6098 | /// 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). |
6099 | /// |
6100 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_srl_epi16&expand=5476) |
6101 | #[inline ] |
6102 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6103 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6104 | #[cfg_attr (test, assert_instr(vpsrlw))] |
6105 | pub unsafe fn _mm_maskz_srl_epi16(k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
6106 | let shf: i16x8 = _mm_srl_epi16(a, count).as_i16x8(); |
6107 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
6108 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6109 | } |
6110 | |
6111 | /// Shift packed 16-bit integers in a right by imm8 while shifting in zeros, and store the results in dst. |
6112 | /// |
6113 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_srli_epi16&expand=5513) |
6114 | #[inline ] |
6115 | #[target_feature (enable = "avx512bw" )] |
6116 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6117 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
6118 | #[rustc_legacy_const_generics (1)] |
6119 | pub unsafe fn _mm512_srli_epi16<const IMM8: u32>(a: __m512i) -> __m512i { |
6120 | static_assert_uimm_bits!(IMM8, 8); |
6121 | if IMM8 >= 16 { |
6122 | _mm512_setzero_si512() |
6123 | } else { |
6124 | transmute(src:simd_shr(lhs:a.as_u16x32(), rhs:u16x32::splat(IMM8 as u16))) |
6125 | } |
6126 | } |
6127 | |
6128 | /// 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). |
6129 | /// |
6130 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_srli_epi16&expand=5511) |
6131 | #[inline ] |
6132 | #[target_feature (enable = "avx512bw" )] |
6133 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6134 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
6135 | #[rustc_legacy_const_generics (3)] |
6136 | pub unsafe fn _mm512_mask_srli_epi16<const IMM8: u32>( |
6137 | src: __m512i, |
6138 | k: __mmask32, |
6139 | a: __m512i, |
6140 | ) -> __m512i { |
6141 | static_assert_uimm_bits!(IMM8, 8); |
6142 | let shf: u16x32 = if IMM8 >= 16 { |
6143 | u16x32::splat(0) |
6144 | } else { |
6145 | simd_shr(lhs:a.as_u16x32(), rhs:u16x32::splat(IMM8 as u16)) |
6146 | }; |
6147 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_u16x32())) |
6148 | } |
6149 | |
6150 | /// 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). |
6151 | /// |
6152 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_srli_epi16&expand=5512) |
6153 | #[inline ] |
6154 | #[target_feature (enable = "avx512bw" )] |
6155 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6156 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
6157 | #[rustc_legacy_const_generics (2)] |
6158 | pub unsafe fn _mm512_maskz_srli_epi16<const IMM8: i32>(k: __mmask32, a: __m512i) -> __m512i { |
6159 | static_assert_uimm_bits!(IMM8, 8); |
6160 | //imm8 should be u32, it seems the document to verify is incorrect |
6161 | if IMM8 >= 16 { |
6162 | _mm512_setzero_si512() |
6163 | } else { |
6164 | let shf: u16x32 = simd_shr(lhs:a.as_u16x32(), rhs:u16x32::splat(IMM8 as u16)); |
6165 | let zero: u16x32 = u16x32::splat(0); |
6166 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6167 | } |
6168 | } |
6169 | |
6170 | /// 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). |
6171 | /// |
6172 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_srli_epi16&expand=5508) |
6173 | #[inline ] |
6174 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6175 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6176 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
6177 | #[rustc_legacy_const_generics (3)] |
6178 | pub unsafe fn _mm256_mask_srli_epi16<const IMM8: i32>( |
6179 | src: __m256i, |
6180 | k: __mmask16, |
6181 | a: __m256i, |
6182 | ) -> __m256i { |
6183 | static_assert_uimm_bits!(IMM8, 8); |
6184 | let shf: __m256i = _mm256_srli_epi16::<IMM8>(a); |
6185 | transmute(src:simd_select_bitmask(m:k, yes:shf.as_i16x16(), no:src.as_i16x16())) |
6186 | } |
6187 | |
6188 | /// 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). |
6189 | /// |
6190 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_srli_epi16&expand=5509) |
6191 | #[inline ] |
6192 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6193 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6194 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
6195 | #[rustc_legacy_const_generics (2)] |
6196 | pub unsafe fn _mm256_maskz_srli_epi16<const IMM8: i32>(k: __mmask16, a: __m256i) -> __m256i { |
6197 | static_assert_uimm_bits!(IMM8, 8); |
6198 | let shf: __m256i = _mm256_srli_epi16::<IMM8>(a); |
6199 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
6200 | transmute(src:simd_select_bitmask(m:k, yes:shf.as_i16x16(), no:zero)) |
6201 | } |
6202 | |
6203 | /// 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). |
6204 | /// |
6205 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_srli_epi16&expand=5505) |
6206 | #[inline ] |
6207 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6208 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6209 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
6210 | #[rustc_legacy_const_generics (3)] |
6211 | pub unsafe fn _mm_mask_srli_epi16<const IMM8: i32>( |
6212 | src: __m128i, |
6213 | k: __mmask8, |
6214 | a: __m128i, |
6215 | ) -> __m128i { |
6216 | static_assert_uimm_bits!(IMM8, 8); |
6217 | let shf: __m128i = _mm_srli_epi16::<IMM8>(a); |
6218 | transmute(src:simd_select_bitmask(m:k, yes:shf.as_i16x8(), no:src.as_i16x8())) |
6219 | } |
6220 | |
6221 | /// 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). |
6222 | /// |
6223 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_srli_epi16&expand=5506) |
6224 | #[inline ] |
6225 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6226 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6227 | #[cfg_attr (test, assert_instr(vpsrlw, IMM8 = 5))] |
6228 | #[rustc_legacy_const_generics (2)] |
6229 | pub unsafe fn _mm_maskz_srli_epi16<const IMM8: i32>(k: __mmask8, a: __m128i) -> __m128i { |
6230 | static_assert_uimm_bits!(IMM8, 8); |
6231 | let shf: __m128i = _mm_srli_epi16::<IMM8>(a); |
6232 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
6233 | transmute(src:simd_select_bitmask(m:k, yes:shf.as_i16x8(), no:zero)) |
6234 | } |
6235 | |
6236 | /// 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. |
6237 | /// |
6238 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_srlv_epi16&expand=5545) |
6239 | #[inline ] |
6240 | #[target_feature (enable = "avx512bw" )] |
6241 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6242 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
6243 | pub unsafe fn _mm512_srlv_epi16(a: __m512i, count: __m512i) -> __m512i { |
6244 | transmute(src:vpsrlvw(a:a.as_i16x32(), b:count.as_i16x32())) |
6245 | } |
6246 | |
6247 | /// 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). |
6248 | /// |
6249 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_srlv_epi16&expand=5543) |
6250 | #[inline ] |
6251 | #[target_feature (enable = "avx512bw" )] |
6252 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6253 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
6254 | pub unsafe fn _mm512_mask_srlv_epi16( |
6255 | src: __m512i, |
6256 | k: __mmask32, |
6257 | a: __m512i, |
6258 | count: __m512i, |
6259 | ) -> __m512i { |
6260 | let shf: i16x32 = _mm512_srlv_epi16(a, count).as_i16x32(); |
6261 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
6262 | } |
6263 | |
6264 | /// 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). |
6265 | /// |
6266 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_srlv_epi16&expand=5544) |
6267 | #[inline ] |
6268 | #[target_feature (enable = "avx512bw" )] |
6269 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6270 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
6271 | pub unsafe fn _mm512_maskz_srlv_epi16(k: __mmask32, a: __m512i, count: __m512i) -> __m512i { |
6272 | let shf: i16x32 = _mm512_srlv_epi16(a, count).as_i16x32(); |
6273 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
6274 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6275 | } |
6276 | |
6277 | /// 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. |
6278 | /// |
6279 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_srlv_epi16&expand=5542) |
6280 | #[inline ] |
6281 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6282 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6283 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
6284 | pub unsafe fn _mm256_srlv_epi16(a: __m256i, count: __m256i) -> __m256i { |
6285 | transmute(src:vpsrlvw256(a:a.as_i16x16(), b:count.as_i16x16())) |
6286 | } |
6287 | |
6288 | /// 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). |
6289 | /// |
6290 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_srlv_epi16&expand=5540) |
6291 | #[inline ] |
6292 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6293 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6294 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
6295 | pub unsafe fn _mm256_mask_srlv_epi16( |
6296 | src: __m256i, |
6297 | k: __mmask16, |
6298 | a: __m256i, |
6299 | count: __m256i, |
6300 | ) -> __m256i { |
6301 | let shf: i16x16 = _mm256_srlv_epi16(a, count).as_i16x16(); |
6302 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x16())) |
6303 | } |
6304 | |
6305 | /// 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). |
6306 | /// |
6307 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_srlv_epi16&expand=5541) |
6308 | #[inline ] |
6309 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6310 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6311 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
6312 | pub unsafe fn _mm256_maskz_srlv_epi16(k: __mmask16, a: __m256i, count: __m256i) -> __m256i { |
6313 | let shf: i16x16 = _mm256_srlv_epi16(a, count).as_i16x16(); |
6314 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
6315 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6316 | } |
6317 | |
6318 | /// 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. |
6319 | /// |
6320 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_srlv_epi16&expand=5539) |
6321 | #[inline ] |
6322 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6323 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6324 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
6325 | pub unsafe fn _mm_srlv_epi16(a: __m128i, count: __m128i) -> __m128i { |
6326 | transmute(src:vpsrlvw128(a:a.as_i16x8(), b:count.as_i16x8())) |
6327 | } |
6328 | |
6329 | /// 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). |
6330 | /// |
6331 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_srlv_epi16&expand=5537) |
6332 | #[inline ] |
6333 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6334 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6335 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
6336 | pub unsafe fn _mm_mask_srlv_epi16( |
6337 | src: __m128i, |
6338 | k: __mmask8, |
6339 | a: __m128i, |
6340 | count: __m128i, |
6341 | ) -> __m128i { |
6342 | let shf: i16x8 = _mm_srlv_epi16(a, count).as_i16x8(); |
6343 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x8())) |
6344 | } |
6345 | |
6346 | /// 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). |
6347 | /// |
6348 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_srlv_epi16&expand=5538) |
6349 | #[inline ] |
6350 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6351 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6352 | #[cfg_attr (test, assert_instr(vpsrlvw))] |
6353 | pub unsafe fn _mm_maskz_srlv_epi16(k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
6354 | let shf: i16x8 = _mm_srlv_epi16(a, count).as_i16x8(); |
6355 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
6356 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6357 | } |
6358 | |
6359 | /// Shift packed 16-bit integers in a right by count while shifting in sign bits, and store the results in dst. |
6360 | /// |
6361 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_sra_epi16&expand=5398) |
6362 | #[inline ] |
6363 | #[target_feature (enable = "avx512bw" )] |
6364 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6365 | #[cfg_attr (test, assert_instr(vpsraw))] |
6366 | pub unsafe fn _mm512_sra_epi16(a: __m512i, count: __m128i) -> __m512i { |
6367 | transmute(src:vpsraw(a:a.as_i16x32(), count:count.as_i16x8())) |
6368 | } |
6369 | |
6370 | /// 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). |
6371 | /// |
6372 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_sra_epi16&expand=5396) |
6373 | #[inline ] |
6374 | #[target_feature (enable = "avx512bw" )] |
6375 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6376 | #[cfg_attr (test, assert_instr(vpsraw))] |
6377 | pub unsafe fn _mm512_mask_sra_epi16( |
6378 | src: __m512i, |
6379 | k: __mmask32, |
6380 | a: __m512i, |
6381 | count: __m128i, |
6382 | ) -> __m512i { |
6383 | let shf: i16x32 = _mm512_sra_epi16(a, count).as_i16x32(); |
6384 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
6385 | } |
6386 | |
6387 | /// 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). |
6388 | /// |
6389 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_sra_epi16&expand=5397) |
6390 | #[inline ] |
6391 | #[target_feature (enable = "avx512bw" )] |
6392 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6393 | #[cfg_attr (test, assert_instr(vpsraw))] |
6394 | pub unsafe fn _mm512_maskz_sra_epi16(k: __mmask32, a: __m512i, count: __m128i) -> __m512i { |
6395 | let shf: i16x32 = _mm512_sra_epi16(a, count).as_i16x32(); |
6396 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
6397 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6398 | } |
6399 | |
6400 | /// 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). |
6401 | /// |
6402 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_sra_epi16&expand=5393) |
6403 | #[inline ] |
6404 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6405 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6406 | #[cfg_attr (test, assert_instr(vpsraw))] |
6407 | pub unsafe fn _mm256_mask_sra_epi16( |
6408 | src: __m256i, |
6409 | k: __mmask16, |
6410 | a: __m256i, |
6411 | count: __m128i, |
6412 | ) -> __m256i { |
6413 | let shf: i16x16 = _mm256_sra_epi16(a, count).as_i16x16(); |
6414 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x16())) |
6415 | } |
6416 | |
6417 | /// 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). |
6418 | /// |
6419 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_sra_epi16&expand=5394) |
6420 | #[inline ] |
6421 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6422 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6423 | #[cfg_attr (test, assert_instr(vpsraw))] |
6424 | pub unsafe fn _mm256_maskz_sra_epi16(k: __mmask16, a: __m256i, count: __m128i) -> __m256i { |
6425 | let shf: i16x16 = _mm256_sra_epi16(a, count).as_i16x16(); |
6426 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
6427 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6428 | } |
6429 | |
6430 | /// 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). |
6431 | /// |
6432 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_sra_epi16&expand=5390) |
6433 | #[inline ] |
6434 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6435 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6436 | #[cfg_attr (test, assert_instr(vpsraw))] |
6437 | pub unsafe fn _mm_mask_sra_epi16(src: __m128i, k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
6438 | let shf: i16x8 = _mm_sra_epi16(a, count).as_i16x8(); |
6439 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x8())) |
6440 | } |
6441 | |
6442 | /// 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). |
6443 | /// |
6444 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_sra_epi16&expand=5391) |
6445 | #[inline ] |
6446 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6447 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6448 | #[cfg_attr (test, assert_instr(vpsraw))] |
6449 | pub unsafe fn _mm_maskz_sra_epi16(k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
6450 | let shf: i16x8 = _mm_sra_epi16(a, count).as_i16x8(); |
6451 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
6452 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6453 | } |
6454 | |
6455 | /// Shift packed 16-bit integers in a right by imm8 while shifting in sign bits, and store the results in dst. |
6456 | /// |
6457 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_srai_epi16&expand=5427) |
6458 | #[inline ] |
6459 | #[target_feature (enable = "avx512bw" )] |
6460 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6461 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
6462 | #[rustc_legacy_const_generics (1)] |
6463 | pub unsafe fn _mm512_srai_epi16<const IMM8: u32>(a: __m512i) -> __m512i { |
6464 | static_assert_uimm_bits!(IMM8, 8); |
6465 | transmute(src:simd_shr(lhs:a.as_i16x32(), rhs:i16x32::splat(IMM8.min(15) as i16))) |
6466 | } |
6467 | |
6468 | /// 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). |
6469 | /// |
6470 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_srai_epi16&expand=5425) |
6471 | #[inline ] |
6472 | #[target_feature (enable = "avx512bw" )] |
6473 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6474 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
6475 | #[rustc_legacy_const_generics (3)] |
6476 | pub unsafe fn _mm512_mask_srai_epi16<const IMM8: u32>( |
6477 | src: __m512i, |
6478 | k: __mmask32, |
6479 | a: __m512i, |
6480 | ) -> __m512i { |
6481 | static_assert_uimm_bits!(IMM8, 8); |
6482 | let shf: i16x32 = simd_shr(lhs:a.as_i16x32(), rhs:i16x32::splat(IMM8.min(15) as i16)); |
6483 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
6484 | } |
6485 | |
6486 | /// 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). |
6487 | /// |
6488 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_srai_epi16&expand=5426) |
6489 | #[inline ] |
6490 | #[target_feature (enable = "avx512bw" )] |
6491 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6492 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
6493 | #[rustc_legacy_const_generics (2)] |
6494 | pub unsafe fn _mm512_maskz_srai_epi16<const IMM8: u32>(k: __mmask32, a: __m512i) -> __m512i { |
6495 | static_assert_uimm_bits!(IMM8, 8); |
6496 | let shf: i16x32 = simd_shr(lhs:a.as_i16x32(), rhs:i16x32::splat(IMM8.min(15) as i16)); |
6497 | let zero: i16x32 = i16x32::splat(0); |
6498 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6499 | } |
6500 | |
6501 | /// 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). |
6502 | /// |
6503 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_srai_epi16&expand=5422) |
6504 | #[inline ] |
6505 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6506 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6507 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
6508 | #[rustc_legacy_const_generics (3)] |
6509 | pub unsafe fn _mm256_mask_srai_epi16<const IMM8: u32>( |
6510 | src: __m256i, |
6511 | k: __mmask16, |
6512 | a: __m256i, |
6513 | ) -> __m256i { |
6514 | static_assert_uimm_bits!(IMM8, 8); |
6515 | let r: i16x16 = simd_shr(lhs:a.as_i16x16(), rhs:i16x16::splat(IMM8.min(15) as i16)); |
6516 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i16x16())) |
6517 | } |
6518 | |
6519 | /// 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). |
6520 | /// |
6521 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_srai_epi16&expand=5423) |
6522 | #[inline ] |
6523 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6524 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6525 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
6526 | #[rustc_legacy_const_generics (2)] |
6527 | pub unsafe fn _mm256_maskz_srai_epi16<const IMM8: u32>(k: __mmask16, a: __m256i) -> __m256i { |
6528 | static_assert_uimm_bits!(IMM8, 8); |
6529 | let r: i16x16 = simd_shr(lhs:a.as_i16x16(), rhs:i16x16::splat(IMM8.min(15) as i16)); |
6530 | let zero: i16x16 = i16x16::splat(0); |
6531 | transmute(src:simd_select_bitmask(m:k, yes:r, no:zero)) |
6532 | } |
6533 | |
6534 | /// 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). |
6535 | /// |
6536 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_srai_epi16&expand=5419) |
6537 | #[inline ] |
6538 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6539 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6540 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
6541 | #[rustc_legacy_const_generics (3)] |
6542 | pub unsafe fn _mm_mask_srai_epi16<const IMM8: u32>( |
6543 | src: __m128i, |
6544 | k: __mmask8, |
6545 | a: __m128i, |
6546 | ) -> __m128i { |
6547 | static_assert_uimm_bits!(IMM8, 8); |
6548 | let r: i16x8 = simd_shr(lhs:a.as_i16x8(), rhs:i16x8::splat(IMM8.min(15) as i16)); |
6549 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i16x8())) |
6550 | } |
6551 | |
6552 | /// 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). |
6553 | /// |
6554 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_srai_epi16&expand=5420) |
6555 | #[inline ] |
6556 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6557 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6558 | #[cfg_attr (test, assert_instr(vpsraw, IMM8 = 1))] |
6559 | #[rustc_legacy_const_generics (2)] |
6560 | pub unsafe fn _mm_maskz_srai_epi16<const IMM8: u32>(k: __mmask8, a: __m128i) -> __m128i { |
6561 | static_assert_uimm_bits!(IMM8, 8); |
6562 | let r: i16x8 = simd_shr(lhs:a.as_i16x8(), rhs:i16x8::splat(IMM8.min(15) as i16)); |
6563 | let zero: i16x8 = i16x8::splat(0); |
6564 | transmute(src:simd_select_bitmask(m:k, yes:r, no:zero)) |
6565 | } |
6566 | |
6567 | /// 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. |
6568 | /// |
6569 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_srav_epi16&expand=5456) |
6570 | #[inline ] |
6571 | #[target_feature (enable = "avx512bw" )] |
6572 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6573 | #[cfg_attr (test, assert_instr(vpsravw))] |
6574 | pub unsafe fn _mm512_srav_epi16(a: __m512i, count: __m512i) -> __m512i { |
6575 | transmute(src:vpsravw(a:a.as_i16x32(), count:count.as_i16x32())) |
6576 | } |
6577 | |
6578 | /// 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). |
6579 | /// |
6580 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_srav_epi16&expand=5454) |
6581 | #[inline ] |
6582 | #[target_feature (enable = "avx512bw" )] |
6583 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6584 | #[cfg_attr (test, assert_instr(vpsravw))] |
6585 | pub unsafe fn _mm512_mask_srav_epi16( |
6586 | src: __m512i, |
6587 | k: __mmask32, |
6588 | a: __m512i, |
6589 | count: __m512i, |
6590 | ) -> __m512i { |
6591 | let shf: i16x32 = _mm512_srav_epi16(a, count).as_i16x32(); |
6592 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x32())) |
6593 | } |
6594 | |
6595 | /// 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). |
6596 | /// |
6597 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_srav_epi16&expand=5455) |
6598 | #[inline ] |
6599 | #[target_feature (enable = "avx512bw" )] |
6600 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6601 | #[cfg_attr (test, assert_instr(vpsravw))] |
6602 | pub unsafe fn _mm512_maskz_srav_epi16(k: __mmask32, a: __m512i, count: __m512i) -> __m512i { |
6603 | let shf: i16x32 = _mm512_srav_epi16(a, count).as_i16x32(); |
6604 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
6605 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6606 | } |
6607 | |
6608 | /// 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. |
6609 | /// |
6610 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_srav_epi16&expand=5453) |
6611 | #[inline ] |
6612 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6613 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6614 | #[cfg_attr (test, assert_instr(vpsravw))] |
6615 | pub unsafe fn _mm256_srav_epi16(a: __m256i, count: __m256i) -> __m256i { |
6616 | transmute(src:vpsravw256(a:a.as_i16x16(), count:count.as_i16x16())) |
6617 | } |
6618 | |
6619 | /// 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). |
6620 | /// |
6621 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_srav_epi16&expand=5451) |
6622 | #[inline ] |
6623 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6624 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6625 | #[cfg_attr (test, assert_instr(vpsravw))] |
6626 | pub unsafe fn _mm256_mask_srav_epi16( |
6627 | src: __m256i, |
6628 | k: __mmask16, |
6629 | a: __m256i, |
6630 | count: __m256i, |
6631 | ) -> __m256i { |
6632 | let shf: i16x16 = _mm256_srav_epi16(a, count).as_i16x16(); |
6633 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x16())) |
6634 | } |
6635 | |
6636 | /// 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). |
6637 | /// |
6638 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_srav_epi16&expand=5452) |
6639 | #[inline ] |
6640 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6641 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6642 | #[cfg_attr (test, assert_instr(vpsravw))] |
6643 | pub unsafe fn _mm256_maskz_srav_epi16(k: __mmask16, a: __m256i, count: __m256i) -> __m256i { |
6644 | let shf: i16x16 = _mm256_srav_epi16(a, count).as_i16x16(); |
6645 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
6646 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6647 | } |
6648 | |
6649 | /// 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. |
6650 | /// |
6651 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_srav_epi16&expand=5450) |
6652 | #[inline ] |
6653 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6654 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6655 | #[cfg_attr (test, assert_instr(vpsravw))] |
6656 | pub unsafe fn _mm_srav_epi16(a: __m128i, count: __m128i) -> __m128i { |
6657 | transmute(src:vpsravw128(a:a.as_i16x8(), count:count.as_i16x8())) |
6658 | } |
6659 | |
6660 | /// 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). |
6661 | /// |
6662 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_srav_epi16&expand=5448) |
6663 | #[inline ] |
6664 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6665 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6666 | #[cfg_attr (test, assert_instr(vpsravw))] |
6667 | pub unsafe fn _mm_mask_srav_epi16( |
6668 | src: __m128i, |
6669 | k: __mmask8, |
6670 | a: __m128i, |
6671 | count: __m128i, |
6672 | ) -> __m128i { |
6673 | let shf: i16x8 = _mm_srav_epi16(a, count).as_i16x8(); |
6674 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:src.as_i16x8())) |
6675 | } |
6676 | |
6677 | /// 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). |
6678 | /// |
6679 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_srav_epi16&expand=5449) |
6680 | #[inline ] |
6681 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6682 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6683 | #[cfg_attr (test, assert_instr(vpsravw))] |
6684 | pub unsafe fn _mm_maskz_srav_epi16(k: __mmask8, a: __m128i, count: __m128i) -> __m128i { |
6685 | let shf: i16x8 = _mm_srav_epi16(a, count).as_i16x8(); |
6686 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
6687 | transmute(src:simd_select_bitmask(m:k, yes:shf, no:zero)) |
6688 | } |
6689 | |
6690 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst. |
6691 | /// |
6692 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_permutex2var_epi16&expand=4226) |
6693 | #[inline ] |
6694 | #[target_feature (enable = "avx512bw" )] |
6695 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6696 | #[cfg_attr (test, assert_instr(vperm))] //vpermi2w or vpermt2w |
6697 | pub unsafe fn _mm512_permutex2var_epi16(a: __m512i, idx: __m512i, b: __m512i) -> __m512i { |
6698 | transmute(src:vpermi2w(a:a.as_i16x32(), idx:idx.as_i16x32(), b:b.as_i16x32())) |
6699 | } |
6700 | |
6701 | /// 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). |
6702 | /// |
6703 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_permutex2var_epi16&expand=4223) |
6704 | #[inline ] |
6705 | #[target_feature (enable = "avx512bw" )] |
6706 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6707 | #[cfg_attr (test, assert_instr(vpermt2w))] |
6708 | pub unsafe fn _mm512_mask_permutex2var_epi16( |
6709 | a: __m512i, |
6710 | k: __mmask32, |
6711 | idx: __m512i, |
6712 | b: __m512i, |
6713 | ) -> __m512i { |
6714 | let permute: i16x32 = _mm512_permutex2var_epi16(a, idx, b).as_i16x32(); |
6715 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:a.as_i16x32())) |
6716 | } |
6717 | |
6718 | /// 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). |
6719 | /// |
6720 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_permutex2var_epi16&expand=4225) |
6721 | #[inline ] |
6722 | #[target_feature (enable = "avx512bw" )] |
6723 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6724 | #[cfg_attr (test, assert_instr(vperm))] //vpermi2w or vpermt2w |
6725 | pub unsafe fn _mm512_maskz_permutex2var_epi16( |
6726 | k: __mmask32, |
6727 | a: __m512i, |
6728 | idx: __m512i, |
6729 | b: __m512i, |
6730 | ) -> __m512i { |
6731 | let permute: i16x32 = _mm512_permutex2var_epi16(a, idx, b).as_i16x32(); |
6732 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
6733 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:zero)) |
6734 | } |
6735 | |
6736 | /// 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). |
6737 | /// |
6738 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask2_permutex2var_epi16&expand=4224) |
6739 | #[inline ] |
6740 | #[target_feature (enable = "avx512bw" )] |
6741 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6742 | #[cfg_attr (test, assert_instr(vpermi2w))] |
6743 | pub unsafe fn _mm512_mask2_permutex2var_epi16( |
6744 | a: __m512i, |
6745 | idx: __m512i, |
6746 | k: __mmask32, |
6747 | b: __m512i, |
6748 | ) -> __m512i { |
6749 | let permute: i16x32 = _mm512_permutex2var_epi16(a, idx, b).as_i16x32(); |
6750 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:idx.as_i16x32())) |
6751 | } |
6752 | |
6753 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst. |
6754 | /// |
6755 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_permutex2var_epi16&expand=4222) |
6756 | #[inline ] |
6757 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6758 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6759 | #[cfg_attr (test, assert_instr(vperm))] //vpermi2w or vpermt2w |
6760 | pub unsafe fn _mm256_permutex2var_epi16(a: __m256i, idx: __m256i, b: __m256i) -> __m256i { |
6761 | transmute(src:vpermi2w256(a:a.as_i16x16(), idx:idx.as_i16x16(), b:b.as_i16x16())) |
6762 | } |
6763 | |
6764 | /// 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). |
6765 | /// |
6766 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_permutex2var_epi16&expand=4219) |
6767 | #[inline ] |
6768 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6769 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6770 | #[cfg_attr (test, assert_instr(vpermt2w))] |
6771 | pub unsafe fn _mm256_mask_permutex2var_epi16( |
6772 | a: __m256i, |
6773 | k: __mmask16, |
6774 | idx: __m256i, |
6775 | b: __m256i, |
6776 | ) -> __m256i { |
6777 | let permute: i16x16 = _mm256_permutex2var_epi16(a, idx, b).as_i16x16(); |
6778 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:a.as_i16x16())) |
6779 | } |
6780 | |
6781 | /// 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). |
6782 | /// |
6783 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_permutex2var_epi16&expand=4221) |
6784 | #[inline ] |
6785 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6786 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6787 | #[cfg_attr (test, assert_instr(vperm))] //vpermi2w or vpermt2w |
6788 | pub unsafe fn _mm256_maskz_permutex2var_epi16( |
6789 | k: __mmask16, |
6790 | a: __m256i, |
6791 | idx: __m256i, |
6792 | b: __m256i, |
6793 | ) -> __m256i { |
6794 | let permute: i16x16 = _mm256_permutex2var_epi16(a, idx, b).as_i16x16(); |
6795 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
6796 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:zero)) |
6797 | } |
6798 | |
6799 | /// 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). |
6800 | /// |
6801 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask2_permutex2var_epi16&expand=4220) |
6802 | #[inline ] |
6803 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6804 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6805 | #[cfg_attr (test, assert_instr(vpermi2w))] |
6806 | pub unsafe fn _mm256_mask2_permutex2var_epi16( |
6807 | a: __m256i, |
6808 | idx: __m256i, |
6809 | k: __mmask16, |
6810 | b: __m256i, |
6811 | ) -> __m256i { |
6812 | let permute: i16x16 = _mm256_permutex2var_epi16(a, idx, b).as_i16x16(); |
6813 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:idx.as_i16x16())) |
6814 | } |
6815 | |
6816 | /// Shuffle 16-bit integers in a and b across lanes using the corresponding selector and index in idx, and store the results in dst. |
6817 | /// |
6818 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_permutex2var_epi16&expand=4218) |
6819 | #[inline ] |
6820 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6821 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6822 | #[cfg_attr (test, assert_instr(vperm))] //vpermi2w or vpermt2w |
6823 | pub unsafe fn _mm_permutex2var_epi16(a: __m128i, idx: __m128i, b: __m128i) -> __m128i { |
6824 | transmute(src:vpermi2w128(a:a.as_i16x8(), idx:idx.as_i16x8(), b:b.as_i16x8())) |
6825 | } |
6826 | |
6827 | /// 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). |
6828 | /// |
6829 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_permutex2var_epi16&expand=4215) |
6830 | #[inline ] |
6831 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6832 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6833 | #[cfg_attr (test, assert_instr(vpermt2w))] |
6834 | pub unsafe fn _mm_mask_permutex2var_epi16( |
6835 | a: __m128i, |
6836 | k: __mmask8, |
6837 | idx: __m128i, |
6838 | b: __m128i, |
6839 | ) -> __m128i { |
6840 | let permute: i16x8 = _mm_permutex2var_epi16(a, idx, b).as_i16x8(); |
6841 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:a.as_i16x8())) |
6842 | } |
6843 | |
6844 | /// 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). |
6845 | /// |
6846 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_permutex2var_epi16&expand=4217) |
6847 | #[inline ] |
6848 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6849 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6850 | #[cfg_attr (test, assert_instr(vperm))] //vpermi2w or vpermt2w |
6851 | pub unsafe fn _mm_maskz_permutex2var_epi16( |
6852 | k: __mmask8, |
6853 | a: __m128i, |
6854 | idx: __m128i, |
6855 | b: __m128i, |
6856 | ) -> __m128i { |
6857 | let permute: i16x8 = _mm_permutex2var_epi16(a, idx, b).as_i16x8(); |
6858 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
6859 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:zero)) |
6860 | } |
6861 | |
6862 | /// 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). |
6863 | /// |
6864 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask2_permutex2var_epi16&expand=4216) |
6865 | #[inline ] |
6866 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6867 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6868 | #[cfg_attr (test, assert_instr(vpermi2w))] |
6869 | pub unsafe fn _mm_mask2_permutex2var_epi16( |
6870 | a: __m128i, |
6871 | idx: __m128i, |
6872 | k: __mmask8, |
6873 | b: __m128i, |
6874 | ) -> __m128i { |
6875 | let permute: i16x8 = _mm_permutex2var_epi16(a, idx, b).as_i16x8(); |
6876 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:idx.as_i16x8())) |
6877 | } |
6878 | |
6879 | /// Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst. |
6880 | /// |
6881 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_permutexvar_epi16&expand=4295) |
6882 | #[inline ] |
6883 | #[target_feature (enable = "avx512bw" )] |
6884 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6885 | #[cfg_attr (test, assert_instr(vpermw))] |
6886 | pub unsafe fn _mm512_permutexvar_epi16(idx: __m512i, a: __m512i) -> __m512i { |
6887 | transmute(src:vpermw(a:a.as_i16x32(), idx:idx.as_i16x32())) |
6888 | } |
6889 | |
6890 | /// 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). |
6891 | /// |
6892 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_permutexvar_epi16&expand=4293) |
6893 | #[inline ] |
6894 | #[target_feature (enable = "avx512bw" )] |
6895 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6896 | #[cfg_attr (test, assert_instr(vpermw))] |
6897 | pub unsafe fn _mm512_mask_permutexvar_epi16( |
6898 | src: __m512i, |
6899 | k: __mmask32, |
6900 | idx: __m512i, |
6901 | a: __m512i, |
6902 | ) -> __m512i { |
6903 | let permute: i16x32 = _mm512_permutexvar_epi16(idx, a).as_i16x32(); |
6904 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:src.as_i16x32())) |
6905 | } |
6906 | |
6907 | /// 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). |
6908 | /// |
6909 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_permutexvar_epi16&expand=4294) |
6910 | #[inline ] |
6911 | #[target_feature (enable = "avx512bw" )] |
6912 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6913 | #[cfg_attr (test, assert_instr(vpermw))] |
6914 | pub unsafe fn _mm512_maskz_permutexvar_epi16(k: __mmask32, idx: __m512i, a: __m512i) -> __m512i { |
6915 | let permute: i16x32 = _mm512_permutexvar_epi16(idx, a).as_i16x32(); |
6916 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
6917 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:zero)) |
6918 | } |
6919 | |
6920 | /// Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst. |
6921 | /// |
6922 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_permutexvar_epi16&expand=4292) |
6923 | #[inline ] |
6924 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6925 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6926 | #[cfg_attr (test, assert_instr(vpermw))] |
6927 | pub unsafe fn _mm256_permutexvar_epi16(idx: __m256i, a: __m256i) -> __m256i { |
6928 | transmute(src:vpermw256(a:a.as_i16x16(), idx:idx.as_i16x16())) |
6929 | } |
6930 | |
6931 | /// 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). |
6932 | /// |
6933 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_permutexvar_epi16&expand=4290) |
6934 | #[inline ] |
6935 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6936 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6937 | #[cfg_attr (test, assert_instr(vpermw))] |
6938 | pub unsafe fn _mm256_mask_permutexvar_epi16( |
6939 | src: __m256i, |
6940 | k: __mmask16, |
6941 | idx: __m256i, |
6942 | a: __m256i, |
6943 | ) -> __m256i { |
6944 | let permute: i16x16 = _mm256_permutexvar_epi16(idx, a).as_i16x16(); |
6945 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:src.as_i16x16())) |
6946 | } |
6947 | |
6948 | /// 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). |
6949 | /// |
6950 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_permutexvar_epi16&expand=4291) |
6951 | #[inline ] |
6952 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6953 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6954 | #[cfg_attr (test, assert_instr(vpermw))] |
6955 | pub unsafe fn _mm256_maskz_permutexvar_epi16(k: __mmask16, idx: __m256i, a: __m256i) -> __m256i { |
6956 | let permute: i16x16 = _mm256_permutexvar_epi16(idx, a).as_i16x16(); |
6957 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
6958 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:zero)) |
6959 | } |
6960 | |
6961 | /// Shuffle 16-bit integers in a across lanes using the corresponding index in idx, and store the results in dst. |
6962 | /// |
6963 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_permutexvar_epi16&expand=4289) |
6964 | #[inline ] |
6965 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6966 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6967 | #[cfg_attr (test, assert_instr(vpermw))] |
6968 | pub unsafe fn _mm_permutexvar_epi16(idx: __m128i, a: __m128i) -> __m128i { |
6969 | transmute(src:vpermw128(a:a.as_i16x8(), idx:idx.as_i16x8())) |
6970 | } |
6971 | |
6972 | /// 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). |
6973 | /// |
6974 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_permutexvar_epi16&expand=4287) |
6975 | #[inline ] |
6976 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6977 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6978 | #[cfg_attr (test, assert_instr(vpermw))] |
6979 | pub unsafe fn _mm_mask_permutexvar_epi16( |
6980 | src: __m128i, |
6981 | k: __mmask8, |
6982 | idx: __m128i, |
6983 | a: __m128i, |
6984 | ) -> __m128i { |
6985 | let permute: i16x8 = _mm_permutexvar_epi16(idx, a).as_i16x8(); |
6986 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:src.as_i16x8())) |
6987 | } |
6988 | |
6989 | /// 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). |
6990 | /// |
6991 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_permutexvar_epi16&expand=4288) |
6992 | #[inline ] |
6993 | #[target_feature (enable = "avx512bw,avx512vl" )] |
6994 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
6995 | #[cfg_attr (test, assert_instr(vpermw))] |
6996 | pub unsafe fn _mm_maskz_permutexvar_epi16(k: __mmask8, idx: __m128i, a: __m128i) -> __m128i { |
6997 | let permute: i16x8 = _mm_permutexvar_epi16(idx, a).as_i16x8(); |
6998 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
6999 | transmute(src:simd_select_bitmask(m:k, yes:permute, no:zero)) |
7000 | } |
7001 | |
7002 | /// Blend packed 16-bit integers from a and b using control mask k, and store the results in dst. |
7003 | /// |
7004 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_blend_epi16&expand=430) |
7005 | #[inline ] |
7006 | #[target_feature (enable = "avx512bw" )] |
7007 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7008 | #[cfg_attr (test, assert_instr(vmovdqu16))] //should be vpblendmw |
7009 | pub unsafe fn _mm512_mask_blend_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
7010 | transmute(src:simd_select_bitmask(m:k, yes:b.as_i16x32(), no:a.as_i16x32())) |
7011 | } |
7012 | |
7013 | /// Blend packed 16-bit integers from a and b using control mask k, and store the results in dst. |
7014 | /// |
7015 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_blend_epi16&expand=429) |
7016 | #[inline ] |
7017 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7018 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7019 | #[cfg_attr (test, assert_instr(vmovdqu16))] //should be vpblendmw |
7020 | pub unsafe fn _mm256_mask_blend_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
7021 | transmute(src:simd_select_bitmask(m:k, yes:b.as_i16x16(), no:a.as_i16x16())) |
7022 | } |
7023 | |
7024 | /// Blend packed 16-bit integers from a and b using control mask k, and store the results in dst. |
7025 | /// |
7026 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_blend_epi16&expand=427) |
7027 | #[inline ] |
7028 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7029 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7030 | #[cfg_attr (test, assert_instr(vmovdqu16))] //should be vpblendmw |
7031 | pub unsafe fn _mm_mask_blend_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
7032 | transmute(src:simd_select_bitmask(m:k, yes:b.as_i16x8(), no:a.as_i16x8())) |
7033 | } |
7034 | |
7035 | /// Blend packed 8-bit integers from a and b using control mask k, and store the results in dst. |
7036 | /// |
7037 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_blend_epi8&expand=441) |
7038 | #[inline ] |
7039 | #[target_feature (enable = "avx512bw" )] |
7040 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7041 | #[cfg_attr (test, assert_instr(vmovdqu8))] //should be vpblendmb |
7042 | pub unsafe fn _mm512_mask_blend_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
7043 | transmute(src:simd_select_bitmask(m:k, yes:b.as_i8x64(), no:a.as_i8x64())) |
7044 | } |
7045 | |
7046 | /// Blend packed 8-bit integers from a and b using control mask k, and store the results in dst. |
7047 | /// |
7048 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_blend_epi8&expand=440) |
7049 | #[inline ] |
7050 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7051 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7052 | #[cfg_attr (test, assert_instr(vmovdqu8))] //should be vpblendmb |
7053 | pub unsafe fn _mm256_mask_blend_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
7054 | transmute(src:simd_select_bitmask(m:k, yes:b.as_i8x32(), no:a.as_i8x32())) |
7055 | } |
7056 | |
7057 | /// Blend packed 8-bit integers from a and b using control mask k, and store the results in dst. |
7058 | /// |
7059 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_blend_epi8&expand=439) |
7060 | #[inline ] |
7061 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7062 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7063 | #[cfg_attr (test, assert_instr(vmovdqu8))] //should be vpblendmb |
7064 | pub unsafe fn _mm_mask_blend_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
7065 | transmute(src:simd_select_bitmask(m:k, yes:b.as_i8x16(), no:a.as_i8x16())) |
7066 | } |
7067 | |
7068 | /// Broadcast the low packed 16-bit integer from a to all elements of dst. |
7069 | /// |
7070 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_broadcastw_epi16&expand=587) |
7071 | #[inline ] |
7072 | #[target_feature (enable = "avx512bw" )] |
7073 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7074 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7075 | pub unsafe fn _mm512_broadcastw_epi16(a: __m128i) -> __m512i { |
7076 | let a: i16x32 = _mm512_castsi128_si512(a).as_i16x32(); |
7077 | let ret: i16x32 = simd_shuffle!( |
7078 | a, |
7079 | a, |
7080 | [ |
7081 | 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, 0, |
7082 | 0, 0, 0, |
7083 | ], |
7084 | ); |
7085 | transmute(src:ret) |
7086 | } |
7087 | |
7088 | /// 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). |
7089 | /// |
7090 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_broadcastw_epi16&expand=588) |
7091 | #[inline ] |
7092 | #[target_feature (enable = "avx512bw" )] |
7093 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7094 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7095 | pub unsafe fn _mm512_mask_broadcastw_epi16(src: __m512i, k: __mmask32, a: __m128i) -> __m512i { |
7096 | let broadcast: i16x32 = _mm512_broadcastw_epi16(a).as_i16x32(); |
7097 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:src.as_i16x32())) |
7098 | } |
7099 | |
7100 | /// 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). |
7101 | /// |
7102 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_broadcastw_epi16&expand=589) |
7103 | #[inline ] |
7104 | #[target_feature (enable = "avx512bw" )] |
7105 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7106 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7107 | pub unsafe fn _mm512_maskz_broadcastw_epi16(k: __mmask32, a: __m128i) -> __m512i { |
7108 | let broadcast: i16x32 = _mm512_broadcastw_epi16(a).as_i16x32(); |
7109 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
7110 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:zero)) |
7111 | } |
7112 | |
7113 | /// 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). |
7114 | /// |
7115 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_broadcastw_epi16&expand=585) |
7116 | #[inline ] |
7117 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7118 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7119 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7120 | pub unsafe fn _mm256_mask_broadcastw_epi16(src: __m256i, k: __mmask16, a: __m128i) -> __m256i { |
7121 | let broadcast: i16x16 = _mm256_broadcastw_epi16(a).as_i16x16(); |
7122 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:src.as_i16x16())) |
7123 | } |
7124 | |
7125 | /// 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). |
7126 | /// |
7127 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_broadcastw_epi16&expand=586) |
7128 | #[inline ] |
7129 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7130 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7131 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7132 | pub unsafe fn _mm256_maskz_broadcastw_epi16(k: __mmask16, a: __m128i) -> __m256i { |
7133 | let broadcast: i16x16 = _mm256_broadcastw_epi16(a).as_i16x16(); |
7134 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
7135 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:zero)) |
7136 | } |
7137 | |
7138 | /// 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). |
7139 | /// |
7140 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_broadcastw_epi16&expand=582) |
7141 | #[inline ] |
7142 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7143 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7144 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7145 | pub unsafe fn _mm_mask_broadcastw_epi16(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
7146 | let broadcast: i16x8 = _mm_broadcastw_epi16(a).as_i16x8(); |
7147 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:src.as_i16x8())) |
7148 | } |
7149 | |
7150 | /// 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). |
7151 | /// |
7152 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_broadcastw_epi16&expand=583) |
7153 | #[inline ] |
7154 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7155 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7156 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7157 | pub unsafe fn _mm_maskz_broadcastw_epi16(k: __mmask8, a: __m128i) -> __m128i { |
7158 | let broadcast: i16x8 = _mm_broadcastw_epi16(a).as_i16x8(); |
7159 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
7160 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:zero)) |
7161 | } |
7162 | |
7163 | /// Broadcast the low packed 8-bit integer from a to all elements of dst. |
7164 | /// |
7165 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_broadcastb_epi8&expand=536) |
7166 | #[inline ] |
7167 | #[target_feature (enable = "avx512bw" )] |
7168 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7169 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
7170 | pub unsafe fn _mm512_broadcastb_epi8(a: __m128i) -> __m512i { |
7171 | let a: i8x64 = _mm512_castsi128_si512(a).as_i8x64(); |
7172 | let ret: i8x64 = simd_shuffle!( |
7173 | a, |
7174 | a, |
7175 | [ |
7176 | 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, 0, |
7177 | 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, 0, |
7178 | 0, 0, 0, 0, 0, 0, |
7179 | ], |
7180 | ); |
7181 | transmute(src:ret) |
7182 | } |
7183 | |
7184 | /// 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). |
7185 | /// |
7186 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_broadcastb_epi8&expand=537) |
7187 | #[inline ] |
7188 | #[target_feature (enable = "avx512bw" )] |
7189 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7190 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
7191 | pub unsafe fn _mm512_mask_broadcastb_epi8(src: __m512i, k: __mmask64, a: __m128i) -> __m512i { |
7192 | let broadcast: i8x64 = _mm512_broadcastb_epi8(a).as_i8x64(); |
7193 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:src.as_i8x64())) |
7194 | } |
7195 | |
7196 | /// 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). |
7197 | /// |
7198 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_broadcastb_epi8&expand=538) |
7199 | #[inline ] |
7200 | #[target_feature (enable = "avx512bw" )] |
7201 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7202 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
7203 | pub unsafe fn _mm512_maskz_broadcastb_epi8(k: __mmask64, a: __m128i) -> __m512i { |
7204 | let broadcast: i8x64 = _mm512_broadcastb_epi8(a).as_i8x64(); |
7205 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
7206 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:zero)) |
7207 | } |
7208 | |
7209 | /// 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). |
7210 | /// |
7211 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_broadcastb_epi8&expand=534) |
7212 | #[inline ] |
7213 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7214 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7215 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
7216 | pub unsafe fn _mm256_mask_broadcastb_epi8(src: __m256i, k: __mmask32, a: __m128i) -> __m256i { |
7217 | let broadcast: i8x32 = _mm256_broadcastb_epi8(a).as_i8x32(); |
7218 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:src.as_i8x32())) |
7219 | } |
7220 | |
7221 | /// 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). |
7222 | /// |
7223 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_broadcastb_epi8&expand=535) |
7224 | #[inline ] |
7225 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7226 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7227 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
7228 | pub unsafe fn _mm256_maskz_broadcastb_epi8(k: __mmask32, a: __m128i) -> __m256i { |
7229 | let broadcast: i8x32 = _mm256_broadcastb_epi8(a).as_i8x32(); |
7230 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
7231 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:zero)) |
7232 | } |
7233 | |
7234 | /// 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). |
7235 | /// |
7236 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_broadcastb_epi8&expand=531) |
7237 | #[inline ] |
7238 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7239 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7240 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
7241 | pub unsafe fn _mm_mask_broadcastb_epi8(src: __m128i, k: __mmask16, a: __m128i) -> __m128i { |
7242 | let broadcast: i8x16 = _mm_broadcastb_epi8(a).as_i8x16(); |
7243 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:src.as_i8x16())) |
7244 | } |
7245 | |
7246 | /// 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). |
7247 | /// |
7248 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_broadcastb_epi8&expand=532) |
7249 | #[inline ] |
7250 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7251 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7252 | #[cfg_attr (test, assert_instr(vpbroadcastb))] |
7253 | pub unsafe fn _mm_maskz_broadcastb_epi8(k: __mmask16, a: __m128i) -> __m128i { |
7254 | let broadcast: i8x16 = _mm_broadcastb_epi8(a).as_i8x16(); |
7255 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
7256 | transmute(src:simd_select_bitmask(m:k, yes:broadcast, no:zero)) |
7257 | } |
7258 | |
7259 | /// 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. |
7260 | /// |
7261 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_unpackhi_epi16&expand=6012) |
7262 | #[inline ] |
7263 | #[target_feature (enable = "avx512bw" )] |
7264 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7265 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
7266 | pub unsafe fn _mm512_unpackhi_epi16(a: __m512i, b: __m512i) -> __m512i { |
7267 | let a: i16x32 = a.as_i16x32(); |
7268 | let b: i16x32 = b.as_i16x32(); |
7269 | #[rustfmt::skip] |
7270 | let r: i16x32 = simd_shuffle!( |
7271 | a, |
7272 | b, |
7273 | [ |
7274 | 4, 32 + 4, 5, 32 + 5, |
7275 | 6, 32 + 6, 7, 32 + 7, |
7276 | 12, 32 + 12, 13, 32 + 13, |
7277 | 14, 32 + 14, 15, 32 + 15, |
7278 | 20, 32 + 20, 21, 32 + 21, |
7279 | 22, 32 + 22, 23, 32 + 23, |
7280 | 28, 32 + 28, 29, 32 + 29, |
7281 | 30, 32 + 30, 31, 32 + 31, |
7282 | ], |
7283 | ); |
7284 | transmute(src:r) |
7285 | } |
7286 | |
7287 | /// 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). |
7288 | /// |
7289 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_unpackhi_epi16&expand=6010) |
7290 | #[inline ] |
7291 | #[target_feature (enable = "avx512bw" )] |
7292 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7293 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
7294 | pub unsafe fn _mm512_mask_unpackhi_epi16( |
7295 | src: __m512i, |
7296 | k: __mmask32, |
7297 | a: __m512i, |
7298 | b: __m512i, |
7299 | ) -> __m512i { |
7300 | let unpackhi: i16x32 = _mm512_unpackhi_epi16(a, b).as_i16x32(); |
7301 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:src.as_i16x32())) |
7302 | } |
7303 | |
7304 | /// 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). |
7305 | /// |
7306 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_unpackhi_epi16&expand=6011) |
7307 | #[inline ] |
7308 | #[target_feature (enable = "avx512bw" )] |
7309 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7310 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
7311 | pub unsafe fn _mm512_maskz_unpackhi_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
7312 | let unpackhi: i16x32 = _mm512_unpackhi_epi16(a, b).as_i16x32(); |
7313 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
7314 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:zero)) |
7315 | } |
7316 | |
7317 | /// 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). |
7318 | /// |
7319 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_unpackhi_epi16&expand=6007) |
7320 | #[inline ] |
7321 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7322 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7323 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
7324 | pub unsafe fn _mm256_mask_unpackhi_epi16( |
7325 | src: __m256i, |
7326 | k: __mmask16, |
7327 | a: __m256i, |
7328 | b: __m256i, |
7329 | ) -> __m256i { |
7330 | let unpackhi: i16x16 = _mm256_unpackhi_epi16(a, b).as_i16x16(); |
7331 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:src.as_i16x16())) |
7332 | } |
7333 | |
7334 | /// 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). |
7335 | /// |
7336 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_unpackhi_epi16&expand=6008) |
7337 | #[inline ] |
7338 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7339 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7340 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
7341 | pub unsafe fn _mm256_maskz_unpackhi_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
7342 | let unpackhi: i16x16 = _mm256_unpackhi_epi16(a, b).as_i16x16(); |
7343 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
7344 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:zero)) |
7345 | } |
7346 | |
7347 | /// 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). |
7348 | /// |
7349 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_unpackhi_epi16&expand=6004) |
7350 | #[inline ] |
7351 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7352 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7353 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
7354 | pub unsafe fn _mm_mask_unpackhi_epi16( |
7355 | src: __m128i, |
7356 | k: __mmask8, |
7357 | a: __m128i, |
7358 | b: __m128i, |
7359 | ) -> __m128i { |
7360 | let unpackhi: i16x8 = _mm_unpackhi_epi16(a, b).as_i16x8(); |
7361 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:src.as_i16x8())) |
7362 | } |
7363 | |
7364 | /// 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). |
7365 | /// |
7366 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_unpackhi_epi16&expand=6005) |
7367 | #[inline ] |
7368 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7369 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7370 | #[cfg_attr (test, assert_instr(vpunpckhwd))] |
7371 | pub unsafe fn _mm_maskz_unpackhi_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
7372 | let unpackhi: i16x8 = _mm_unpackhi_epi16(a, b).as_i16x8(); |
7373 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
7374 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:zero)) |
7375 | } |
7376 | |
7377 | /// 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. |
7378 | /// |
7379 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_unpackhi_epi8&expand=6039) |
7380 | #[inline ] |
7381 | #[target_feature (enable = "avx512bw" )] |
7382 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7383 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
7384 | pub unsafe fn _mm512_unpackhi_epi8(a: __m512i, b: __m512i) -> __m512i { |
7385 | let a = a.as_i8x64(); |
7386 | let b = b.as_i8x64(); |
7387 | #[rustfmt::skip] |
7388 | let r: i8x64 = simd_shuffle!( |
7389 | a, |
7390 | b, |
7391 | [ |
7392 | 8, 64+8, 9, 64+9, |
7393 | 10, 64+10, 11, 64+11, |
7394 | 12, 64+12, 13, 64+13, |
7395 | 14, 64+14, 15, 64+15, |
7396 | 24, 64+24, 25, 64+25, |
7397 | 26, 64+26, 27, 64+27, |
7398 | 28, 64+28, 29, 64+29, |
7399 | 30, 64+30, 31, 64+31, |
7400 | 40, 64+40, 41, 64+41, |
7401 | 42, 64+42, 43, 64+43, |
7402 | 44, 64+44, 45, 64+45, |
7403 | 46, 64+46, 47, 64+47, |
7404 | 56, 64+56, 57, 64+57, |
7405 | 58, 64+58, 59, 64+59, |
7406 | 60, 64+60, 61, 64+61, |
7407 | 62, 64+62, 63, 64+63, |
7408 | ], |
7409 | ); |
7410 | transmute(r) |
7411 | } |
7412 | |
7413 | /// 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). |
7414 | /// |
7415 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_unpackhi_epi8&expand=6037) |
7416 | #[inline ] |
7417 | #[target_feature (enable = "avx512bw" )] |
7418 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7419 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
7420 | pub unsafe fn _mm512_mask_unpackhi_epi8( |
7421 | src: __m512i, |
7422 | k: __mmask64, |
7423 | a: __m512i, |
7424 | b: __m512i, |
7425 | ) -> __m512i { |
7426 | let unpackhi: i8x64 = _mm512_unpackhi_epi8(a, b).as_i8x64(); |
7427 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:src.as_i8x64())) |
7428 | } |
7429 | |
7430 | /// 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). |
7431 | /// |
7432 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_unpackhi_epi8&expand=6038) |
7433 | #[inline ] |
7434 | #[target_feature (enable = "avx512bw" )] |
7435 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7436 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
7437 | pub unsafe fn _mm512_maskz_unpackhi_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
7438 | let unpackhi: i8x64 = _mm512_unpackhi_epi8(a, b).as_i8x64(); |
7439 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
7440 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:zero)) |
7441 | } |
7442 | |
7443 | /// 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). |
7444 | /// |
7445 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_unpackhi_epi8&expand=6034) |
7446 | #[inline ] |
7447 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7448 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7449 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
7450 | pub unsafe fn _mm256_mask_unpackhi_epi8( |
7451 | src: __m256i, |
7452 | k: __mmask32, |
7453 | a: __m256i, |
7454 | b: __m256i, |
7455 | ) -> __m256i { |
7456 | let unpackhi: i8x32 = _mm256_unpackhi_epi8(a, b).as_i8x32(); |
7457 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:src.as_i8x32())) |
7458 | } |
7459 | |
7460 | /// 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). |
7461 | /// |
7462 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_unpackhi_epi8&expand=6035) |
7463 | #[inline ] |
7464 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7465 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7466 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
7467 | pub unsafe fn _mm256_maskz_unpackhi_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
7468 | let unpackhi: i8x32 = _mm256_unpackhi_epi8(a, b).as_i8x32(); |
7469 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
7470 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:zero)) |
7471 | } |
7472 | |
7473 | /// 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). |
7474 | /// |
7475 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_unpackhi_epi8&expand=6031) |
7476 | #[inline ] |
7477 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7478 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7479 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
7480 | pub unsafe fn _mm_mask_unpackhi_epi8( |
7481 | src: __m128i, |
7482 | k: __mmask16, |
7483 | a: __m128i, |
7484 | b: __m128i, |
7485 | ) -> __m128i { |
7486 | let unpackhi: i8x16 = _mm_unpackhi_epi8(a, b).as_i8x16(); |
7487 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:src.as_i8x16())) |
7488 | } |
7489 | |
7490 | /// 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). |
7491 | /// |
7492 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_unpackhi_epi8&expand=6032) |
7493 | #[inline ] |
7494 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7495 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7496 | #[cfg_attr (test, assert_instr(vpunpckhbw))] |
7497 | pub unsafe fn _mm_maskz_unpackhi_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
7498 | let unpackhi: i8x16 = _mm_unpackhi_epi8(a, b).as_i8x16(); |
7499 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
7500 | transmute(src:simd_select_bitmask(m:k, yes:unpackhi, no:zero)) |
7501 | } |
7502 | |
7503 | /// 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. |
7504 | /// |
7505 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_unpacklo_epi16&expand=6069) |
7506 | #[inline ] |
7507 | #[target_feature (enable = "avx512bw" )] |
7508 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7509 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
7510 | pub unsafe fn _mm512_unpacklo_epi16(a: __m512i, b: __m512i) -> __m512i { |
7511 | let a: i16x32 = a.as_i16x32(); |
7512 | let b: i16x32 = b.as_i16x32(); |
7513 | #[rustfmt::skip] |
7514 | let r: i16x32 = simd_shuffle!( |
7515 | a, |
7516 | b, |
7517 | [ |
7518 | 0, 32+0, 1, 32+1, |
7519 | 2, 32+2, 3, 32+3, |
7520 | 8, 32+8, 9, 32+9, |
7521 | 10, 32+10, 11, 32+11, |
7522 | 16, 32+16, 17, 32+17, |
7523 | 18, 32+18, 19, 32+19, |
7524 | 24, 32+24, 25, 32+25, |
7525 | 26, 32+26, 27, 32+27 |
7526 | ], |
7527 | ); |
7528 | transmute(src:r) |
7529 | } |
7530 | |
7531 | /// 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). |
7532 | /// |
7533 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_unpacklo_epi16&expand=6067) |
7534 | #[inline ] |
7535 | #[target_feature (enable = "avx512bw" )] |
7536 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7537 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
7538 | pub unsafe fn _mm512_mask_unpacklo_epi16( |
7539 | src: __m512i, |
7540 | k: __mmask32, |
7541 | a: __m512i, |
7542 | b: __m512i, |
7543 | ) -> __m512i { |
7544 | let unpacklo: i16x32 = _mm512_unpacklo_epi16(a, b).as_i16x32(); |
7545 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:src.as_i16x32())) |
7546 | } |
7547 | |
7548 | /// 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). |
7549 | /// |
7550 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_unpacklo_epi16&expand=6068) |
7551 | #[inline ] |
7552 | #[target_feature (enable = "avx512bw" )] |
7553 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7554 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
7555 | pub unsafe fn _mm512_maskz_unpacklo_epi16(k: __mmask32, a: __m512i, b: __m512i) -> __m512i { |
7556 | let unpacklo: i16x32 = _mm512_unpacklo_epi16(a, b).as_i16x32(); |
7557 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
7558 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:zero)) |
7559 | } |
7560 | |
7561 | /// 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). |
7562 | /// |
7563 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_unpacklo_epi16&expand=6064) |
7564 | #[inline ] |
7565 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7566 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7567 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
7568 | pub unsafe fn _mm256_mask_unpacklo_epi16( |
7569 | src: __m256i, |
7570 | k: __mmask16, |
7571 | a: __m256i, |
7572 | b: __m256i, |
7573 | ) -> __m256i { |
7574 | let unpacklo: i16x16 = _mm256_unpacklo_epi16(a, b).as_i16x16(); |
7575 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:src.as_i16x16())) |
7576 | } |
7577 | |
7578 | /// 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). |
7579 | /// |
7580 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_unpacklo_epi16&expand=6065) |
7581 | #[inline ] |
7582 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7583 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7584 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
7585 | pub unsafe fn _mm256_maskz_unpacklo_epi16(k: __mmask16, a: __m256i, b: __m256i) -> __m256i { |
7586 | let unpacklo: i16x16 = _mm256_unpacklo_epi16(a, b).as_i16x16(); |
7587 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
7588 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:zero)) |
7589 | } |
7590 | |
7591 | /// 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). |
7592 | /// |
7593 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_unpacklo_epi16&expand=6061) |
7594 | #[inline ] |
7595 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7596 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7597 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
7598 | pub unsafe fn _mm_mask_unpacklo_epi16( |
7599 | src: __m128i, |
7600 | k: __mmask8, |
7601 | a: __m128i, |
7602 | b: __m128i, |
7603 | ) -> __m128i { |
7604 | let unpacklo: i16x8 = _mm_unpacklo_epi16(a, b).as_i16x8(); |
7605 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:src.as_i16x8())) |
7606 | } |
7607 | |
7608 | /// 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). |
7609 | /// |
7610 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_unpacklo_epi16&expand=6062) |
7611 | #[inline ] |
7612 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7613 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7614 | #[cfg_attr (test, assert_instr(vpunpcklwd))] |
7615 | pub unsafe fn _mm_maskz_unpacklo_epi16(k: __mmask8, a: __m128i, b: __m128i) -> __m128i { |
7616 | let unpacklo: i16x8 = _mm_unpacklo_epi16(a, b).as_i16x8(); |
7617 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
7618 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:zero)) |
7619 | } |
7620 | |
7621 | /// 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. |
7622 | /// |
7623 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_unpacklo_epi8&expand=6096) |
7624 | #[inline ] |
7625 | #[target_feature (enable = "avx512bw" )] |
7626 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7627 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
7628 | pub unsafe fn _mm512_unpacklo_epi8(a: __m512i, b: __m512i) -> __m512i { |
7629 | let a = a.as_i8x64(); |
7630 | let b = b.as_i8x64(); |
7631 | #[rustfmt::skip] |
7632 | let r: i8x64 = simd_shuffle!( |
7633 | a, |
7634 | b, |
7635 | [ |
7636 | 0, 64+0, 1, 64+1, |
7637 | 2, 64+2, 3, 64+3, |
7638 | 4, 64+4, 5, 64+5, |
7639 | 6, 64+6, 7, 64+7, |
7640 | 16, 64+16, 17, 64+17, |
7641 | 18, 64+18, 19, 64+19, |
7642 | 20, 64+20, 21, 64+21, |
7643 | 22, 64+22, 23, 64+23, |
7644 | 32, 64+32, 33, 64+33, |
7645 | 34, 64+34, 35, 64+35, |
7646 | 36, 64+36, 37, 64+37, |
7647 | 38, 64+38, 39, 64+39, |
7648 | 48, 64+48, 49, 64+49, |
7649 | 50, 64+50, 51, 64+51, |
7650 | 52, 64+52, 53, 64+53, |
7651 | 54, 64+54, 55, 64+55, |
7652 | ], |
7653 | ); |
7654 | transmute(r) |
7655 | } |
7656 | |
7657 | /// 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). |
7658 | /// |
7659 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_unpacklo_epi8&expand=6094) |
7660 | #[inline ] |
7661 | #[target_feature (enable = "avx512bw" )] |
7662 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7663 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
7664 | pub unsafe fn _mm512_mask_unpacklo_epi8( |
7665 | src: __m512i, |
7666 | k: __mmask64, |
7667 | a: __m512i, |
7668 | b: __m512i, |
7669 | ) -> __m512i { |
7670 | let unpacklo: i8x64 = _mm512_unpacklo_epi8(a, b).as_i8x64(); |
7671 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:src.as_i8x64())) |
7672 | } |
7673 | |
7674 | /// 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). |
7675 | /// |
7676 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_unpacklo_epi8&expand=6095) |
7677 | #[inline ] |
7678 | #[target_feature (enable = "avx512bw" )] |
7679 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7680 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
7681 | pub unsafe fn _mm512_maskz_unpacklo_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
7682 | let unpacklo: i8x64 = _mm512_unpacklo_epi8(a, b).as_i8x64(); |
7683 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
7684 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:zero)) |
7685 | } |
7686 | |
7687 | /// 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). |
7688 | /// |
7689 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_unpacklo_epi8&expand=6091) |
7690 | #[inline ] |
7691 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7692 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7693 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
7694 | pub unsafe fn _mm256_mask_unpacklo_epi8( |
7695 | src: __m256i, |
7696 | k: __mmask32, |
7697 | a: __m256i, |
7698 | b: __m256i, |
7699 | ) -> __m256i { |
7700 | let unpacklo: i8x32 = _mm256_unpacklo_epi8(a, b).as_i8x32(); |
7701 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:src.as_i8x32())) |
7702 | } |
7703 | |
7704 | /// 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). |
7705 | /// |
7706 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_unpacklo_epi8&expand=6092) |
7707 | #[inline ] |
7708 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7709 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7710 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
7711 | pub unsafe fn _mm256_maskz_unpacklo_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
7712 | let unpacklo: i8x32 = _mm256_unpacklo_epi8(a, b).as_i8x32(); |
7713 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
7714 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:zero)) |
7715 | } |
7716 | |
7717 | /// 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). |
7718 | /// |
7719 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_unpacklo_epi8&expand=6088) |
7720 | #[inline ] |
7721 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7722 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7723 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
7724 | pub unsafe fn _mm_mask_unpacklo_epi8( |
7725 | src: __m128i, |
7726 | k: __mmask16, |
7727 | a: __m128i, |
7728 | b: __m128i, |
7729 | ) -> __m128i { |
7730 | let unpacklo: i8x16 = _mm_unpacklo_epi8(a, b).as_i8x16(); |
7731 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:src.as_i8x16())) |
7732 | } |
7733 | |
7734 | /// 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). |
7735 | /// |
7736 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_unpacklo_epi8&expand=6089) |
7737 | #[inline ] |
7738 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7739 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7740 | #[cfg_attr (test, assert_instr(vpunpcklbw))] |
7741 | pub unsafe fn _mm_maskz_unpacklo_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
7742 | let unpacklo: i8x16 = _mm_unpacklo_epi8(a, b).as_i8x16(); |
7743 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
7744 | transmute(src:simd_select_bitmask(m:k, yes:unpacklo, no:zero)) |
7745 | } |
7746 | |
7747 | /// 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). |
7748 | /// |
7749 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_mov_epi16&expand=3795) |
7750 | #[inline ] |
7751 | #[target_feature (enable = "avx512bw" )] |
7752 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7753 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
7754 | pub unsafe fn _mm512_mask_mov_epi16(src: __m512i, k: __mmask32, a: __m512i) -> __m512i { |
7755 | let mov: i16x32 = a.as_i16x32(); |
7756 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:src.as_i16x32())) |
7757 | } |
7758 | |
7759 | /// Move packed 16-bit integers from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
7760 | /// |
7761 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_mov_epi16&expand=3796) |
7762 | #[inline ] |
7763 | #[target_feature (enable = "avx512bw" )] |
7764 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7765 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
7766 | pub unsafe fn _mm512_maskz_mov_epi16(k: __mmask32, a: __m512i) -> __m512i { |
7767 | let mov: i16x32 = a.as_i16x32(); |
7768 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
7769 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:zero)) |
7770 | } |
7771 | |
7772 | /// 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). |
7773 | /// |
7774 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_mov_epi16&expand=3793) |
7775 | #[inline ] |
7776 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7777 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7778 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
7779 | pub unsafe fn _mm256_mask_mov_epi16(src: __m256i, k: __mmask16, a: __m256i) -> __m256i { |
7780 | let mov: i16x16 = a.as_i16x16(); |
7781 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:src.as_i16x16())) |
7782 | } |
7783 | |
7784 | /// Move packed 16-bit integers from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
7785 | /// |
7786 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_mov_epi16&expand=3794) |
7787 | #[inline ] |
7788 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7789 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7790 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
7791 | pub unsafe fn _mm256_maskz_mov_epi16(k: __mmask16, a: __m256i) -> __m256i { |
7792 | let mov: i16x16 = a.as_i16x16(); |
7793 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
7794 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:zero)) |
7795 | } |
7796 | |
7797 | /// 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). |
7798 | /// |
7799 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_mov_epi16&expand=3791) |
7800 | #[inline ] |
7801 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7802 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7803 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
7804 | pub unsafe fn _mm_mask_mov_epi16(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
7805 | let mov: i16x8 = a.as_i16x8(); |
7806 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:src.as_i16x8())) |
7807 | } |
7808 | |
7809 | /// Move packed 16-bit integers from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
7810 | /// |
7811 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_mov_epi16&expand=3792) |
7812 | #[inline ] |
7813 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7814 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7815 | #[cfg_attr (test, assert_instr(vmovdqu16))] |
7816 | pub unsafe fn _mm_maskz_mov_epi16(k: __mmask8, a: __m128i) -> __m128i { |
7817 | let mov: i16x8 = a.as_i16x8(); |
7818 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
7819 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:zero)) |
7820 | } |
7821 | |
7822 | /// 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). |
7823 | /// |
7824 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_mov_epi8&expand=3813) |
7825 | #[inline ] |
7826 | #[target_feature (enable = "avx512bw" )] |
7827 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7828 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
7829 | pub unsafe fn _mm512_mask_mov_epi8(src: __m512i, k: __mmask64, a: __m512i) -> __m512i { |
7830 | let mov: i8x64 = a.as_i8x64(); |
7831 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:src.as_i8x64())) |
7832 | } |
7833 | |
7834 | /// Move packed 8-bit integers from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
7835 | /// |
7836 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_mov_epi8&expand=3814) |
7837 | #[inline ] |
7838 | #[target_feature (enable = "avx512bw" )] |
7839 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7840 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
7841 | pub unsafe fn _mm512_maskz_mov_epi8(k: __mmask64, a: __m512i) -> __m512i { |
7842 | let mov: i8x64 = a.as_i8x64(); |
7843 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
7844 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:zero)) |
7845 | } |
7846 | |
7847 | /// 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). |
7848 | /// |
7849 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_mov_epi8&expand=3811) |
7850 | #[inline ] |
7851 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7852 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7853 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
7854 | pub unsafe fn _mm256_mask_mov_epi8(src: __m256i, k: __mmask32, a: __m256i) -> __m256i { |
7855 | let mov: i8x32 = a.as_i8x32(); |
7856 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:src.as_i8x32())) |
7857 | } |
7858 | |
7859 | /// Move packed 8-bit integers from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
7860 | /// |
7861 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_mov_epi8&expand=3812) |
7862 | #[inline ] |
7863 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7864 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7865 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
7866 | pub unsafe fn _mm256_maskz_mov_epi8(k: __mmask32, a: __m256i) -> __m256i { |
7867 | let mov: i8x32 = a.as_i8x32(); |
7868 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
7869 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:zero)) |
7870 | } |
7871 | |
7872 | /// 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). |
7873 | /// |
7874 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_mov_epi8&expand=3809) |
7875 | #[inline ] |
7876 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7877 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7878 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
7879 | pub unsafe fn _mm_mask_mov_epi8(src: __m128i, k: __mmask16, a: __m128i) -> __m128i { |
7880 | let mov: i8x16 = a.as_i8x16(); |
7881 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:src.as_i8x16())) |
7882 | } |
7883 | |
7884 | /// Move packed 8-bit integers from a into dst using zeromask k (elements are zeroed out when the corresponding mask bit is not set). |
7885 | /// |
7886 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_mov_epi8&expand=3810) |
7887 | #[inline ] |
7888 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7889 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7890 | #[cfg_attr (test, assert_instr(vmovdqu8))] |
7891 | pub unsafe fn _mm_maskz_mov_epi8(k: __mmask16, a: __m128i) -> __m128i { |
7892 | let mov: i8x16 = a.as_i8x16(); |
7893 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
7894 | transmute(src:simd_select_bitmask(m:k, yes:mov, no:zero)) |
7895 | } |
7896 | |
7897 | /// 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). |
7898 | /// |
7899 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_set1_epi16&expand=4942) |
7900 | #[inline ] |
7901 | #[target_feature (enable = "avx512bw" )] |
7902 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7903 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7904 | pub unsafe fn _mm512_mask_set1_epi16(src: __m512i, k: __mmask32, a: i16) -> __m512i { |
7905 | let r: i16x32 = _mm512_set1_epi16(a).as_i16x32(); |
7906 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i16x32())) |
7907 | } |
7908 | |
7909 | /// 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). |
7910 | /// |
7911 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_set1_epi16&expand=4943) |
7912 | #[inline ] |
7913 | #[target_feature (enable = "avx512bw" )] |
7914 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7915 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7916 | pub unsafe fn _mm512_maskz_set1_epi16(k: __mmask32, a: i16) -> __m512i { |
7917 | let r: i16x32 = _mm512_set1_epi16(a).as_i16x32(); |
7918 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
7919 | transmute(src:simd_select_bitmask(m:k, yes:r, no:zero)) |
7920 | } |
7921 | |
7922 | /// 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). |
7923 | /// |
7924 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_set1_epi16&expand=4939) |
7925 | #[inline ] |
7926 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7927 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7928 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7929 | pub unsafe fn _mm256_mask_set1_epi16(src: __m256i, k: __mmask16, a: i16) -> __m256i { |
7930 | let r: i16x16 = _mm256_set1_epi16(a).as_i16x16(); |
7931 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i16x16())) |
7932 | } |
7933 | |
7934 | /// 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). |
7935 | /// |
7936 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_set1_epi16&expand=4940) |
7937 | #[inline ] |
7938 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7939 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7940 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7941 | pub unsafe fn _mm256_maskz_set1_epi16(k: __mmask16, a: i16) -> __m256i { |
7942 | let r: i16x16 = _mm256_set1_epi16(a).as_i16x16(); |
7943 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
7944 | transmute(src:simd_select_bitmask(m:k, yes:r, no:zero)) |
7945 | } |
7946 | |
7947 | /// 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). |
7948 | /// |
7949 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_set1_epi16&expand=4936) |
7950 | #[inline ] |
7951 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7952 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7953 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7954 | pub unsafe fn _mm_mask_set1_epi16(src: __m128i, k: __mmask8, a: i16) -> __m128i { |
7955 | let r: i16x8 = _mm_set1_epi16(a).as_i16x8(); |
7956 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i16x8())) |
7957 | } |
7958 | |
7959 | /// 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). |
7960 | /// |
7961 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_set1_epi16&expand=4937) |
7962 | #[inline ] |
7963 | #[target_feature (enable = "avx512bw,avx512vl" )] |
7964 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7965 | #[cfg_attr (test, assert_instr(vpbroadcastw))] |
7966 | pub unsafe fn _mm_maskz_set1_epi16(k: __mmask8, a: i16) -> __m128i { |
7967 | let r: i16x8 = _mm_set1_epi16(a).as_i16x8(); |
7968 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
7969 | transmute(src:simd_select_bitmask(m:k, yes:r, no:zero)) |
7970 | } |
7971 | |
7972 | /// 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). |
7973 | /// |
7974 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_set1_epi8&expand=4970) |
7975 | #[inline ] |
7976 | #[target_feature (enable = "avx512bw" )] |
7977 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7978 | #[cfg_attr (test, assert_instr(vpbroadcast))] |
7979 | pub unsafe fn _mm512_mask_set1_epi8(src: __m512i, k: __mmask64, a: i8) -> __m512i { |
7980 | let r: i8x64 = _mm512_set1_epi8(a).as_i8x64(); |
7981 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i8x64())) |
7982 | } |
7983 | |
7984 | /// 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). |
7985 | /// |
7986 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_set1_epi8&expand=4971) |
7987 | #[inline ] |
7988 | #[target_feature (enable = "avx512bw" )] |
7989 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
7990 | #[cfg_attr (test, assert_instr(vpbroadcast))] |
7991 | pub unsafe fn _mm512_maskz_set1_epi8(k: __mmask64, a: i8) -> __m512i { |
7992 | let r: i8x64 = _mm512_set1_epi8(a).as_i8x64(); |
7993 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
7994 | transmute(src:simd_select_bitmask(m:k, yes:r, no:zero)) |
7995 | } |
7996 | |
7997 | /// 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). |
7998 | /// |
7999 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_set1_epi8&expand=4967) |
8000 | #[inline ] |
8001 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8002 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8003 | #[cfg_attr (test, assert_instr(vpbroadcast))] |
8004 | pub unsafe fn _mm256_mask_set1_epi8(src: __m256i, k: __mmask32, a: i8) -> __m256i { |
8005 | let r: i8x32 = _mm256_set1_epi8(a).as_i8x32(); |
8006 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i8x32())) |
8007 | } |
8008 | |
8009 | /// 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). |
8010 | /// |
8011 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_set1_epi8&expand=4968) |
8012 | #[inline ] |
8013 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8014 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8015 | #[cfg_attr (test, assert_instr(vpbroadcast))] |
8016 | pub unsafe fn _mm256_maskz_set1_epi8(k: __mmask32, a: i8) -> __m256i { |
8017 | let r: i8x32 = _mm256_set1_epi8(a).as_i8x32(); |
8018 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
8019 | transmute(src:simd_select_bitmask(m:k, yes:r, no:zero)) |
8020 | } |
8021 | |
8022 | /// 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). |
8023 | /// |
8024 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_set1_epi8&expand=4964) |
8025 | #[inline ] |
8026 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8027 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8028 | #[cfg_attr (test, assert_instr(vpbroadcast))] |
8029 | pub unsafe fn _mm_mask_set1_epi8(src: __m128i, k: __mmask16, a: i8) -> __m128i { |
8030 | let r: i8x16 = _mm_set1_epi8(a).as_i8x16(); |
8031 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_i8x16())) |
8032 | } |
8033 | |
8034 | /// 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). |
8035 | /// |
8036 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_set1_epi8&expand=4965) |
8037 | #[inline ] |
8038 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8039 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8040 | #[cfg_attr (test, assert_instr(vpbroadcast))] |
8041 | pub unsafe fn _mm_maskz_set1_epi8(k: __mmask16, a: i8) -> __m128i { |
8042 | let r: i8x16 = _mm_set1_epi8(a).as_i8x16(); |
8043 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
8044 | transmute(src:simd_select_bitmask(m:k, yes:r, no:zero)) |
8045 | } |
8046 | |
8047 | /// 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. |
8048 | /// |
8049 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_shufflelo_epi16&expand=5221) |
8050 | #[inline ] |
8051 | #[target_feature (enable = "avx512bw" )] |
8052 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8053 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 0))] |
8054 | #[rustc_legacy_const_generics (1)] |
8055 | pub unsafe fn _mm512_shufflelo_epi16<const IMM8: i32>(a: __m512i) -> __m512i { |
8056 | static_assert_uimm_bits!(IMM8, 8); |
8057 | let a = a.as_i16x32(); |
8058 | let r: i16x32 = simd_shuffle!( |
8059 | a, |
8060 | a, |
8061 | [ |
8062 | IMM8 as u32 & 0b11, |
8063 | (IMM8 as u32 >> 2) & 0b11, |
8064 | (IMM8 as u32 >> 4) & 0b11, |
8065 | (IMM8 as u32 >> 6) & 0b11, |
8066 | 4, |
8067 | 5, |
8068 | 6, |
8069 | 7, |
8070 | (IMM8 as u32 & 0b11) + 8, |
8071 | ((IMM8 as u32 >> 2) & 0b11) + 8, |
8072 | ((IMM8 as u32 >> 4) & 0b11) + 8, |
8073 | ((IMM8 as u32 >> 6) & 0b11) + 8, |
8074 | 12, |
8075 | 13, |
8076 | 14, |
8077 | 15, |
8078 | (IMM8 as u32 & 0b11) + 16, |
8079 | ((IMM8 as u32 >> 2) & 0b11) + 16, |
8080 | ((IMM8 as u32 >> 4) & 0b11) + 16, |
8081 | ((IMM8 as u32 >> 6) & 0b11) + 16, |
8082 | 20, |
8083 | 21, |
8084 | 22, |
8085 | 23, |
8086 | (IMM8 as u32 & 0b11) + 24, |
8087 | ((IMM8 as u32 >> 2) & 0b11) + 24, |
8088 | ((IMM8 as u32 >> 4) & 0b11) + 24, |
8089 | ((IMM8 as u32 >> 6) & 0b11) + 24, |
8090 | 28, |
8091 | 29, |
8092 | 30, |
8093 | 31, |
8094 | ], |
8095 | ); |
8096 | transmute(r) |
8097 | } |
8098 | |
8099 | /// 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). |
8100 | /// |
8101 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_shufflelo_epi16&expand=5219) |
8102 | #[inline ] |
8103 | #[target_feature (enable = "avx512bw" )] |
8104 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8105 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 0))] |
8106 | #[rustc_legacy_const_generics (3)] |
8107 | pub unsafe fn _mm512_mask_shufflelo_epi16<const IMM8: i32>( |
8108 | src: __m512i, |
8109 | k: __mmask32, |
8110 | a: __m512i, |
8111 | ) -> __m512i { |
8112 | static_assert_uimm_bits!(IMM8, 8); |
8113 | let r: __m512i = _mm512_shufflelo_epi16::<IMM8>(a); |
8114 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i16x32(), no:src.as_i16x32())) |
8115 | } |
8116 | |
8117 | /// 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). |
8118 | /// |
8119 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_shufflelo_epi16&expand=5220) |
8120 | #[inline ] |
8121 | #[target_feature (enable = "avx512bw" )] |
8122 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8123 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 0))] |
8124 | #[rustc_legacy_const_generics (2)] |
8125 | pub unsafe fn _mm512_maskz_shufflelo_epi16<const IMM8: i32>(k: __mmask32, a: __m512i) -> __m512i { |
8126 | static_assert_uimm_bits!(IMM8, 8); |
8127 | let r: __m512i = _mm512_shufflelo_epi16::<IMM8>(a); |
8128 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
8129 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i16x32(), no:zero)) |
8130 | } |
8131 | |
8132 | /// 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). |
8133 | /// |
8134 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_shufflelo_epi16&expand=5216) |
8135 | #[inline ] |
8136 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8137 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8138 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 5))] |
8139 | #[rustc_legacy_const_generics (3)] |
8140 | pub unsafe fn _mm256_mask_shufflelo_epi16<const IMM8: i32>( |
8141 | src: __m256i, |
8142 | k: __mmask16, |
8143 | a: __m256i, |
8144 | ) -> __m256i { |
8145 | static_assert_uimm_bits!(IMM8, 8); |
8146 | let shuffle: __m256i = _mm256_shufflelo_epi16::<IMM8>(a); |
8147 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x16(), no:src.as_i16x16())) |
8148 | } |
8149 | |
8150 | /// 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). |
8151 | /// |
8152 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_shufflelo_epi16&expand=5217) |
8153 | #[inline ] |
8154 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8155 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8156 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 5))] |
8157 | #[rustc_legacy_const_generics (2)] |
8158 | pub unsafe fn _mm256_maskz_shufflelo_epi16<const IMM8: i32>(k: __mmask16, a: __m256i) -> __m256i { |
8159 | static_assert_uimm_bits!(IMM8, 8); |
8160 | let shuffle: __m256i = _mm256_shufflelo_epi16::<IMM8>(a); |
8161 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
8162 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x16(), no:zero)) |
8163 | } |
8164 | |
8165 | /// 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). |
8166 | /// |
8167 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_shufflelo_epi16&expand=5213) |
8168 | #[inline ] |
8169 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8170 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8171 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 5))] |
8172 | #[rustc_legacy_const_generics (3)] |
8173 | pub unsafe fn _mm_mask_shufflelo_epi16<const IMM8: i32>( |
8174 | src: __m128i, |
8175 | k: __mmask8, |
8176 | a: __m128i, |
8177 | ) -> __m128i { |
8178 | static_assert_uimm_bits!(IMM8, 8); |
8179 | let shuffle: __m128i = _mm_shufflelo_epi16::<IMM8>(a); |
8180 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x8(), no:src.as_i16x8())) |
8181 | } |
8182 | |
8183 | /// 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). |
8184 | /// |
8185 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_shufflelo_epi16&expand=5214) |
8186 | #[inline ] |
8187 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8188 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8189 | #[cfg_attr (test, assert_instr(vpshuflw, IMM8 = 5))] |
8190 | #[rustc_legacy_const_generics (2)] |
8191 | pub unsafe fn _mm_maskz_shufflelo_epi16<const IMM8: i32>(k: __mmask8, a: __m128i) -> __m128i { |
8192 | static_assert_uimm_bits!(IMM8, 8); |
8193 | let shuffle: __m128i = _mm_shufflelo_epi16::<IMM8>(a); |
8194 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
8195 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x8(), no:zero)) |
8196 | } |
8197 | |
8198 | /// 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. |
8199 | /// |
8200 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_shufflehi_epi16&expand=5212) |
8201 | #[inline ] |
8202 | #[target_feature (enable = "avx512bw" )] |
8203 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8204 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 0))] |
8205 | #[rustc_legacy_const_generics (1)] |
8206 | pub unsafe fn _mm512_shufflehi_epi16<const IMM8: i32>(a: __m512i) -> __m512i { |
8207 | static_assert_uimm_bits!(IMM8, 8); |
8208 | let a = a.as_i16x32(); |
8209 | let r: i16x32 = simd_shuffle!( |
8210 | a, |
8211 | a, |
8212 | [ |
8213 | 0, |
8214 | 1, |
8215 | 2, |
8216 | 3, |
8217 | (IMM8 as u32 & 0b11) + 4, |
8218 | ((IMM8 as u32 >> 2) & 0b11) + 4, |
8219 | ((IMM8 as u32 >> 4) & 0b11) + 4, |
8220 | ((IMM8 as u32 >> 6) & 0b11) + 4, |
8221 | 8, |
8222 | 9, |
8223 | 10, |
8224 | 11, |
8225 | (IMM8 as u32 & 0b11) + 12, |
8226 | ((IMM8 as u32 >> 2) & 0b11) + 12, |
8227 | ((IMM8 as u32 >> 4) & 0b11) + 12, |
8228 | ((IMM8 as u32 >> 6) & 0b11) + 12, |
8229 | 16, |
8230 | 17, |
8231 | 18, |
8232 | 19, |
8233 | (IMM8 as u32 & 0b11) + 20, |
8234 | ((IMM8 as u32 >> 2) & 0b11) + 20, |
8235 | ((IMM8 as u32 >> 4) & 0b11) + 20, |
8236 | ((IMM8 as u32 >> 6) & 0b11) + 20, |
8237 | 24, |
8238 | 25, |
8239 | 26, |
8240 | 27, |
8241 | (IMM8 as u32 & 0b11) + 28, |
8242 | ((IMM8 as u32 >> 2) & 0b11) + 28, |
8243 | ((IMM8 as u32 >> 4) & 0b11) + 28, |
8244 | ((IMM8 as u32 >> 6) & 0b11) + 28, |
8245 | ], |
8246 | ); |
8247 | transmute(r) |
8248 | } |
8249 | |
8250 | /// 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). |
8251 | /// |
8252 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_shufflehi_epi16&expand=5210) |
8253 | #[inline ] |
8254 | #[target_feature (enable = "avx512bw" )] |
8255 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8256 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 0))] |
8257 | #[rustc_legacy_const_generics (3)] |
8258 | pub unsafe fn _mm512_mask_shufflehi_epi16<const IMM8: i32>( |
8259 | src: __m512i, |
8260 | k: __mmask32, |
8261 | a: __m512i, |
8262 | ) -> __m512i { |
8263 | static_assert_uimm_bits!(IMM8, 8); |
8264 | let r: __m512i = _mm512_shufflehi_epi16::<IMM8>(a); |
8265 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i16x32(), no:src.as_i16x32())) |
8266 | } |
8267 | |
8268 | /// 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). |
8269 | /// |
8270 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_shufflehi_epi16&expand=5211) |
8271 | #[inline ] |
8272 | #[target_feature (enable = "avx512bw" )] |
8273 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8274 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 0))] |
8275 | #[rustc_legacy_const_generics (2)] |
8276 | pub unsafe fn _mm512_maskz_shufflehi_epi16<const IMM8: i32>(k: __mmask32, a: __m512i) -> __m512i { |
8277 | static_assert_uimm_bits!(IMM8, 8); |
8278 | let r: __m512i = _mm512_shufflehi_epi16::<IMM8>(a); |
8279 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
8280 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i16x32(), no:zero)) |
8281 | } |
8282 | |
8283 | /// 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). |
8284 | /// |
8285 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_shufflehi_epi16&expand=5207) |
8286 | #[inline ] |
8287 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8288 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8289 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 5))] |
8290 | #[rustc_legacy_const_generics (3)] |
8291 | pub unsafe fn _mm256_mask_shufflehi_epi16<const IMM8: i32>( |
8292 | src: __m256i, |
8293 | k: __mmask16, |
8294 | a: __m256i, |
8295 | ) -> __m256i { |
8296 | static_assert_uimm_bits!(IMM8, 8); |
8297 | let shuffle: __m256i = _mm256_shufflehi_epi16::<IMM8>(a); |
8298 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x16(), no:src.as_i16x16())) |
8299 | } |
8300 | |
8301 | /// 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). |
8302 | /// |
8303 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_shufflehi_epi16&expand=5208) |
8304 | #[inline ] |
8305 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8306 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8307 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 5))] |
8308 | #[rustc_legacy_const_generics (2)] |
8309 | pub unsafe fn _mm256_maskz_shufflehi_epi16<const IMM8: i32>(k: __mmask16, a: __m256i) -> __m256i { |
8310 | static_assert_uimm_bits!(IMM8, 8); |
8311 | let shuffle: __m256i = _mm256_shufflehi_epi16::<IMM8>(a); |
8312 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
8313 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x16(), no:zero)) |
8314 | } |
8315 | |
8316 | /// 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). |
8317 | /// |
8318 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_shufflehi_epi16&expand=5204) |
8319 | #[inline ] |
8320 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8321 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8322 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 5))] |
8323 | #[rustc_legacy_const_generics (3)] |
8324 | pub unsafe fn _mm_mask_shufflehi_epi16<const IMM8: i32>( |
8325 | src: __m128i, |
8326 | k: __mmask8, |
8327 | a: __m128i, |
8328 | ) -> __m128i { |
8329 | static_assert_uimm_bits!(IMM8, 8); |
8330 | let shuffle: __m128i = _mm_shufflehi_epi16::<IMM8>(a); |
8331 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x8(), no:src.as_i16x8())) |
8332 | } |
8333 | |
8334 | /// 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). |
8335 | /// |
8336 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_shufflehi_epi16&expand=5205) |
8337 | #[inline ] |
8338 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8339 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8340 | #[cfg_attr (test, assert_instr(vpshufhw, IMM8 = 5))] |
8341 | #[rustc_legacy_const_generics (2)] |
8342 | pub unsafe fn _mm_maskz_shufflehi_epi16<const IMM8: i32>(k: __mmask8, a: __m128i) -> __m128i { |
8343 | static_assert_uimm_bits!(IMM8, 8); |
8344 | let shuffle: __m128i = _mm_shufflehi_epi16::<IMM8>(a); |
8345 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
8346 | transmute(src:simd_select_bitmask(m:k, yes:shuffle.as_i16x8(), no:zero)) |
8347 | } |
8348 | |
8349 | /// 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. |
8350 | /// |
8351 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_shuffle_epi8&expand=5159) |
8352 | #[inline ] |
8353 | #[target_feature (enable = "avx512bw" )] |
8354 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8355 | #[cfg_attr (test, assert_instr(vpshufb))] |
8356 | pub unsafe fn _mm512_shuffle_epi8(a: __m512i, b: __m512i) -> __m512i { |
8357 | transmute(src:vpshufb(a:a.as_i8x64(), b:b.as_i8x64())) |
8358 | } |
8359 | |
8360 | /// 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). |
8361 | /// |
8362 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_shuffle_epi8&expand=5157) |
8363 | #[inline ] |
8364 | #[target_feature (enable = "avx512bw" )] |
8365 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8366 | #[cfg_attr (test, assert_instr(vpshufb))] |
8367 | pub unsafe fn _mm512_mask_shuffle_epi8( |
8368 | src: __m512i, |
8369 | k: __mmask64, |
8370 | a: __m512i, |
8371 | b: __m512i, |
8372 | ) -> __m512i { |
8373 | let shuffle: i8x64 = _mm512_shuffle_epi8(a, b).as_i8x64(); |
8374 | transmute(src:simd_select_bitmask(m:k, yes:shuffle, no:src.as_i8x64())) |
8375 | } |
8376 | |
8377 | /// 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). |
8378 | /// |
8379 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_shuffle_epi8&expand=5158) |
8380 | #[inline ] |
8381 | #[target_feature (enable = "avx512bw" )] |
8382 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8383 | #[cfg_attr (test, assert_instr(vpshufb))] |
8384 | pub unsafe fn _mm512_maskz_shuffle_epi8(k: __mmask64, a: __m512i, b: __m512i) -> __m512i { |
8385 | let shuffle: i8x64 = _mm512_shuffle_epi8(a, b).as_i8x64(); |
8386 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
8387 | transmute(src:simd_select_bitmask(m:k, yes:shuffle, no:zero)) |
8388 | } |
8389 | |
8390 | /// 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). |
8391 | /// |
8392 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_shuffle_epi8&expand=5154) |
8393 | #[inline ] |
8394 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8395 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8396 | #[cfg_attr (test, assert_instr(vpshufb))] |
8397 | pub unsafe fn _mm256_mask_shuffle_epi8( |
8398 | src: __m256i, |
8399 | k: __mmask32, |
8400 | a: __m256i, |
8401 | b: __m256i, |
8402 | ) -> __m256i { |
8403 | let shuffle: i8x32 = _mm256_shuffle_epi8(a, b).as_i8x32(); |
8404 | transmute(src:simd_select_bitmask(m:k, yes:shuffle, no:src.as_i8x32())) |
8405 | } |
8406 | |
8407 | /// 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). |
8408 | /// |
8409 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_shuffle_epi8&expand=5155) |
8410 | #[inline ] |
8411 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8412 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8413 | #[cfg_attr (test, assert_instr(vpshufb))] |
8414 | pub unsafe fn _mm256_maskz_shuffle_epi8(k: __mmask32, a: __m256i, b: __m256i) -> __m256i { |
8415 | let shuffle: i8x32 = _mm256_shuffle_epi8(a, b).as_i8x32(); |
8416 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
8417 | transmute(src:simd_select_bitmask(m:k, yes:shuffle, no:zero)) |
8418 | } |
8419 | |
8420 | /// 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). |
8421 | /// |
8422 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_shuffle_epi8&expand=5151) |
8423 | #[inline ] |
8424 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8425 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8426 | #[cfg_attr (test, assert_instr(vpshufb))] |
8427 | pub unsafe fn _mm_mask_shuffle_epi8(src: __m128i, k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
8428 | let shuffle: i8x16 = _mm_shuffle_epi8(a, b).as_i8x16(); |
8429 | transmute(src:simd_select_bitmask(m:k, yes:shuffle, no:src.as_i8x16())) |
8430 | } |
8431 | |
8432 | /// 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). |
8433 | /// |
8434 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_shuffle_epi8&expand=5152) |
8435 | #[inline ] |
8436 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8437 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8438 | #[cfg_attr (test, assert_instr(vpshufb))] |
8439 | pub unsafe fn _mm_maskz_shuffle_epi8(k: __mmask16, a: __m128i, b: __m128i) -> __m128i { |
8440 | let shuffle: i8x16 = _mm_shuffle_epi8(a, b).as_i8x16(); |
8441 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
8442 | transmute(src:simd_select_bitmask(m:k, yes:shuffle, no:zero)) |
8443 | } |
8444 | |
8445 | /// 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. |
8446 | /// |
8447 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_test_epi16_mask&expand=5884) |
8448 | #[inline ] |
8449 | #[target_feature (enable = "avx512bw" )] |
8450 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8451 | #[cfg_attr (test, assert_instr(vptestmw))] |
8452 | pub unsafe fn _mm512_test_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
8453 | let and: __m512i = _mm512_and_si512(a, b); |
8454 | let zero: __m512i = _mm512_setzero_si512(); |
8455 | _mm512_cmpneq_epi16_mask(a:and, b:zero) |
8456 | } |
8457 | |
8458 | /// 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. |
8459 | /// |
8460 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_test_epi16_mask&expand=5883) |
8461 | #[inline ] |
8462 | #[target_feature (enable = "avx512bw" )] |
8463 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8464 | #[cfg_attr (test, assert_instr(vptestmw))] |
8465 | pub unsafe fn _mm512_mask_test_epi16_mask(k: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
8466 | let and: __m512i = _mm512_and_si512(a, b); |
8467 | let zero: __m512i = _mm512_setzero_si512(); |
8468 | _mm512_mask_cmpneq_epi16_mask(k1:k, a:and, b:zero) |
8469 | } |
8470 | |
8471 | /// 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. |
8472 | /// |
8473 | // [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_test_epi16_mask&expand=5882) |
8474 | #[inline ] |
8475 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8476 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8477 | #[cfg_attr (test, assert_instr(vptestmw))] |
8478 | pub unsafe fn _mm256_test_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
8479 | let and: __m256i = _mm256_and_si256(a, b); |
8480 | let zero: __m256i = _mm256_setzero_si256(); |
8481 | _mm256_cmpneq_epi16_mask(a:and, b:zero) |
8482 | } |
8483 | |
8484 | /// 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. |
8485 | /// |
8486 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_test_epi16_mask&expand=5881) |
8487 | #[inline ] |
8488 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8489 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8490 | #[cfg_attr (test, assert_instr(vptestmw))] |
8491 | pub unsafe fn _mm256_mask_test_epi16_mask(k: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
8492 | let and: __m256i = _mm256_and_si256(a, b); |
8493 | let zero: __m256i = _mm256_setzero_si256(); |
8494 | _mm256_mask_cmpneq_epi16_mask(k1:k, a:and, b:zero) |
8495 | } |
8496 | |
8497 | /// 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. |
8498 | /// |
8499 | // [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_test_epi16_mask&expand=5880) |
8500 | #[inline ] |
8501 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8502 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8503 | #[cfg_attr (test, assert_instr(vptestmw))] |
8504 | pub unsafe fn _mm_test_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
8505 | let and: __m128i = _mm_and_si128(a, b); |
8506 | let zero: __m128i = _mm_setzero_si128(); |
8507 | _mm_cmpneq_epi16_mask(a:and, b:zero) |
8508 | } |
8509 | |
8510 | /// 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. |
8511 | /// |
8512 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_test_epi16_mask&expand=5879) |
8513 | #[inline ] |
8514 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8515 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8516 | #[cfg_attr (test, assert_instr(vptestmw))] |
8517 | pub unsafe fn _mm_mask_test_epi16_mask(k: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
8518 | let and: __m128i = _mm_and_si128(a, b); |
8519 | let zero: __m128i = _mm_setzero_si128(); |
8520 | _mm_mask_cmpneq_epi16_mask(k1:k, a:and, b:zero) |
8521 | } |
8522 | |
8523 | /// 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. |
8524 | /// |
8525 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_test_epi8_mask&expand=5902) |
8526 | #[inline ] |
8527 | #[target_feature (enable = "avx512bw" )] |
8528 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8529 | #[cfg_attr (test, assert_instr(vptestmb))] |
8530 | pub unsafe fn _mm512_test_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
8531 | let and: __m512i = _mm512_and_si512(a, b); |
8532 | let zero: __m512i = _mm512_setzero_si512(); |
8533 | _mm512_cmpneq_epi8_mask(a:and, b:zero) |
8534 | } |
8535 | |
8536 | /// 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. |
8537 | /// |
8538 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_test_epi8_mask&expand=5901) |
8539 | #[inline ] |
8540 | #[target_feature (enable = "avx512bw" )] |
8541 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8542 | #[cfg_attr (test, assert_instr(vptestmb))] |
8543 | pub unsafe fn _mm512_mask_test_epi8_mask(k: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
8544 | let and: __m512i = _mm512_and_si512(a, b); |
8545 | let zero: __m512i = _mm512_setzero_si512(); |
8546 | _mm512_mask_cmpneq_epi8_mask(k1:k, a:and, b:zero) |
8547 | } |
8548 | |
8549 | /// 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. |
8550 | /// |
8551 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_test_epi8_mask&expand=5900) |
8552 | #[inline ] |
8553 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8554 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8555 | #[cfg_attr (test, assert_instr(vptestmb))] |
8556 | pub unsafe fn _mm256_test_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
8557 | let and: __m256i = _mm256_and_si256(a, b); |
8558 | let zero: __m256i = _mm256_setzero_si256(); |
8559 | _mm256_cmpneq_epi8_mask(a:and, b:zero) |
8560 | } |
8561 | |
8562 | /// 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. |
8563 | /// |
8564 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_test_epi8_mask&expand=5899) |
8565 | #[inline ] |
8566 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8567 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8568 | #[cfg_attr (test, assert_instr(vptestmb))] |
8569 | pub unsafe fn _mm256_mask_test_epi8_mask(k: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
8570 | let and: __m256i = _mm256_and_si256(a, b); |
8571 | let zero: __m256i = _mm256_setzero_si256(); |
8572 | _mm256_mask_cmpneq_epi8_mask(k1:k, a:and, b:zero) |
8573 | } |
8574 | |
8575 | /// 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. |
8576 | /// |
8577 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_test_epi8_mask&expand=5898) |
8578 | #[inline ] |
8579 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8580 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8581 | #[cfg_attr (test, assert_instr(vptestmb))] |
8582 | pub unsafe fn _mm_test_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
8583 | let and: __m128i = _mm_and_si128(a, b); |
8584 | let zero: __m128i = _mm_setzero_si128(); |
8585 | _mm_cmpneq_epi8_mask(a:and, b:zero) |
8586 | } |
8587 | |
8588 | /// 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. |
8589 | /// |
8590 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_test_epi8_mask&expand=5897) |
8591 | #[inline ] |
8592 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8593 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8594 | #[cfg_attr (test, assert_instr(vptestmb))] |
8595 | pub unsafe fn _mm_mask_test_epi8_mask(k: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
8596 | let and: __m128i = _mm_and_si128(a, b); |
8597 | let zero: __m128i = _mm_setzero_si128(); |
8598 | _mm_mask_cmpneq_epi8_mask(k1:k, a:and, b:zero) |
8599 | } |
8600 | |
8601 | /// 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. |
8602 | /// |
8603 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_testn_epi16_mask&expand=5915) |
8604 | #[inline ] |
8605 | #[target_feature (enable = "avx512bw" )] |
8606 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8607 | #[cfg_attr (test, assert_instr(vptestnmw))] |
8608 | pub unsafe fn _mm512_testn_epi16_mask(a: __m512i, b: __m512i) -> __mmask32 { |
8609 | let and: __m512i = _mm512_and_si512(a, b); |
8610 | let zero: __m512i = _mm512_setzero_si512(); |
8611 | _mm512_cmpeq_epi16_mask(a:and, b:zero) |
8612 | } |
8613 | |
8614 | /// 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. |
8615 | /// |
8616 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_testn_epi16&expand=5914) |
8617 | #[inline ] |
8618 | #[target_feature (enable = "avx512bw" )] |
8619 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8620 | #[cfg_attr (test, assert_instr(vptestnmw))] |
8621 | pub unsafe fn _mm512_mask_testn_epi16_mask(k: __mmask32, a: __m512i, b: __m512i) -> __mmask32 { |
8622 | let and: __m512i = _mm512_and_si512(a, b); |
8623 | let zero: __m512i = _mm512_setzero_si512(); |
8624 | _mm512_mask_cmpeq_epi16_mask(k1:k, a:and, b:zero) |
8625 | } |
8626 | |
8627 | /// 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. |
8628 | /// |
8629 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_testn_epi16_mask&expand=5913) |
8630 | #[inline ] |
8631 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8632 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8633 | #[cfg_attr (test, assert_instr(vptestnmw))] |
8634 | pub unsafe fn _mm256_testn_epi16_mask(a: __m256i, b: __m256i) -> __mmask16 { |
8635 | let and: __m256i = _mm256_and_si256(a, b); |
8636 | let zero: __m256i = _mm256_setzero_si256(); |
8637 | _mm256_cmpeq_epi16_mask(a:and, b:zero) |
8638 | } |
8639 | |
8640 | /// 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. |
8641 | /// |
8642 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_testn_epi16_mask&expand=5912) |
8643 | #[inline ] |
8644 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8645 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8646 | #[cfg_attr (test, assert_instr(vptestnmw))] |
8647 | pub unsafe fn _mm256_mask_testn_epi16_mask(k: __mmask16, a: __m256i, b: __m256i) -> __mmask16 { |
8648 | let and: __m256i = _mm256_and_si256(a, b); |
8649 | let zero: __m256i = _mm256_setzero_si256(); |
8650 | _mm256_mask_cmpeq_epi16_mask(k1:k, a:and, b:zero) |
8651 | } |
8652 | |
8653 | /// 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. |
8654 | /// |
8655 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_testn_epi16_mask&expand=5911) |
8656 | #[inline ] |
8657 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8658 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8659 | #[cfg_attr (test, assert_instr(vptestnmw))] |
8660 | pub unsafe fn _mm_testn_epi16_mask(a: __m128i, b: __m128i) -> __mmask8 { |
8661 | let and: __m128i = _mm_and_si128(a, b); |
8662 | let zero: __m128i = _mm_setzero_si128(); |
8663 | _mm_cmpeq_epi16_mask(a:and, b:zero) |
8664 | } |
8665 | |
8666 | /// 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. |
8667 | /// |
8668 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_testn_epi16_mask&expand=5910) |
8669 | #[inline ] |
8670 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8671 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8672 | #[cfg_attr (test, assert_instr(vptestnmw))] |
8673 | pub unsafe fn _mm_mask_testn_epi16_mask(k: __mmask8, a: __m128i, b: __m128i) -> __mmask8 { |
8674 | let and: __m128i = _mm_and_si128(a, b); |
8675 | let zero: __m128i = _mm_setzero_si128(); |
8676 | _mm_mask_cmpeq_epi16_mask(k1:k, a:and, b:zero) |
8677 | } |
8678 | |
8679 | /// 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. |
8680 | /// |
8681 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_testn_epi8_mask&expand=5933) |
8682 | #[inline ] |
8683 | #[target_feature (enable = "avx512bw" )] |
8684 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8685 | #[cfg_attr (test, assert_instr(vptestnmb))] |
8686 | pub unsafe fn _mm512_testn_epi8_mask(a: __m512i, b: __m512i) -> __mmask64 { |
8687 | let and: __m512i = _mm512_and_si512(a, b); |
8688 | let zero: __m512i = _mm512_setzero_si512(); |
8689 | _mm512_cmpeq_epi8_mask(a:and, b:zero) |
8690 | } |
8691 | |
8692 | /// 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. |
8693 | /// |
8694 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_testn_epi8_mask&expand=5932) |
8695 | #[inline ] |
8696 | #[target_feature (enable = "avx512bw" )] |
8697 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8698 | #[cfg_attr (test, assert_instr(vptestnmb))] |
8699 | pub unsafe fn _mm512_mask_testn_epi8_mask(k: __mmask64, a: __m512i, b: __m512i) -> __mmask64 { |
8700 | let and: __m512i = _mm512_and_si512(a, b); |
8701 | let zero: __m512i = _mm512_setzero_si512(); |
8702 | _mm512_mask_cmpeq_epi8_mask(k1:k, a:and, b:zero) |
8703 | } |
8704 | |
8705 | /// 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. |
8706 | /// |
8707 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_testn_epi8_mask&expand=5931) |
8708 | #[inline ] |
8709 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8710 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8711 | #[cfg_attr (test, assert_instr(vptestnmb))] |
8712 | pub unsafe fn _mm256_testn_epi8_mask(a: __m256i, b: __m256i) -> __mmask32 { |
8713 | let and: __m256i = _mm256_and_si256(a, b); |
8714 | let zero: __m256i = _mm256_setzero_si256(); |
8715 | _mm256_cmpeq_epi8_mask(a:and, b:zero) |
8716 | } |
8717 | |
8718 | /// 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. |
8719 | /// |
8720 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_testn_epi8_mask&expand=5930) |
8721 | #[inline ] |
8722 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8723 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8724 | #[cfg_attr (test, assert_instr(vptestnmb))] |
8725 | pub unsafe fn _mm256_mask_testn_epi8_mask(k: __mmask32, a: __m256i, b: __m256i) -> __mmask32 { |
8726 | let and: __m256i = _mm256_and_si256(a, b); |
8727 | let zero: __m256i = _mm256_setzero_si256(); |
8728 | _mm256_mask_cmpeq_epi8_mask(k1:k, a:and, b:zero) |
8729 | } |
8730 | |
8731 | /// 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. |
8732 | /// |
8733 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_testn_epi8_mask&expand=5929) |
8734 | #[inline ] |
8735 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8736 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8737 | #[cfg_attr (test, assert_instr(vptestnmb))] |
8738 | pub unsafe fn _mm_testn_epi8_mask(a: __m128i, b: __m128i) -> __mmask16 { |
8739 | let and: __m128i = _mm_and_si128(a, b); |
8740 | let zero: __m128i = _mm_setzero_si128(); |
8741 | _mm_cmpeq_epi8_mask(a:and, b:zero) |
8742 | } |
8743 | |
8744 | /// 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. |
8745 | /// |
8746 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_testn_epi8_mask&expand=5928) |
8747 | #[inline ] |
8748 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8749 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8750 | #[cfg_attr (test, assert_instr(vptestnmb))] |
8751 | pub unsafe fn _mm_mask_testn_epi8_mask(k: __mmask16, a: __m128i, b: __m128i) -> __mmask16 { |
8752 | let and: __m128i = _mm_and_si128(a, b); |
8753 | let zero: __m128i = _mm_setzero_si128(); |
8754 | _mm_mask_cmpeq_epi8_mask(k1:k, a:and, b:zero) |
8755 | } |
8756 | |
8757 | /// Store 64-bit mask from a into memory. |
8758 | /// |
8759 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_store_mask64&expand=5578) |
8760 | #[inline ] |
8761 | #[target_feature (enable = "avx512bw" )] |
8762 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8763 | #[cfg_attr (test, assert_instr(mov))] //should be kmovq |
8764 | pub unsafe fn _store_mask64(mem_addr: *mut u64, a: __mmask64) { |
8765 | ptr::write(dst:mem_addr as *mut __mmask64, src:a); |
8766 | } |
8767 | |
8768 | /// Store 32-bit mask from a into memory. |
8769 | /// |
8770 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_store_mask32&expand=5577) |
8771 | #[inline ] |
8772 | #[target_feature (enable = "avx512bw" )] |
8773 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8774 | #[cfg_attr (test, assert_instr(mov))] //should be kmovd |
8775 | pub unsafe fn _store_mask32(mem_addr: *mut u32, a: __mmask32) { |
8776 | ptr::write(dst:mem_addr as *mut __mmask32, src:a); |
8777 | } |
8778 | |
8779 | /// Load 64-bit mask from memory into k. |
8780 | /// |
8781 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_load_mask64&expand=3318) |
8782 | #[inline ] |
8783 | #[target_feature (enable = "avx512bw" )] |
8784 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8785 | #[cfg_attr (test, assert_instr(mov))] //should be kmovq |
8786 | pub unsafe fn _load_mask64(mem_addr: *const u64) -> __mmask64 { |
8787 | ptr::read(src:mem_addr as *const __mmask64) |
8788 | } |
8789 | |
8790 | /// Load 32-bit mask from memory into k. |
8791 | /// |
8792 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_load_mask32&expand=3317) |
8793 | #[inline ] |
8794 | #[target_feature (enable = "avx512bw" )] |
8795 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8796 | #[cfg_attr (test, assert_instr(mov))] //should be kmovd |
8797 | pub unsafe fn _load_mask32(mem_addr: *const u32) -> __mmask32 { |
8798 | ptr::read(src:mem_addr as *const __mmask32) |
8799 | } |
8800 | |
8801 | /// 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. |
8802 | /// |
8803 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_sad_epu8&expand=4855) |
8804 | #[inline ] |
8805 | #[target_feature (enable = "avx512bw" )] |
8806 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8807 | #[cfg_attr (test, assert_instr(vpsadbw))] |
8808 | pub unsafe fn _mm512_sad_epu8(a: __m512i, b: __m512i) -> __m512i { |
8809 | transmute(src:vpsadbw(a:a.as_u8x64(), b:b.as_u8x64())) |
8810 | } |
8811 | |
8812 | /// 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. |
8813 | /// |
8814 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_dbsad_epu8&expand=2114) |
8815 | #[inline ] |
8816 | #[target_feature (enable = "avx512bw" )] |
8817 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8818 | #[rustc_legacy_const_generics (2)] |
8819 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
8820 | pub unsafe fn _mm512_dbsad_epu8<const IMM8: i32>(a: __m512i, b: __m512i) -> __m512i { |
8821 | static_assert_uimm_bits!(IMM8, 8); |
8822 | let a: u8x64 = a.as_u8x64(); |
8823 | let b: u8x64 = b.as_u8x64(); |
8824 | let r: u16x32 = vdbpsadbw(a, b, IMM8); |
8825 | transmute(src:r) |
8826 | } |
8827 | |
8828 | /// 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. |
8829 | /// |
8830 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_dbsad_epu8&expand=2115) |
8831 | #[inline ] |
8832 | #[target_feature (enable = "avx512bw" )] |
8833 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8834 | #[rustc_legacy_const_generics (4)] |
8835 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
8836 | pub unsafe fn _mm512_mask_dbsad_epu8<const IMM8: i32>( |
8837 | src: __m512i, |
8838 | k: __mmask32, |
8839 | a: __m512i, |
8840 | b: __m512i, |
8841 | ) -> __m512i { |
8842 | static_assert_uimm_bits!(IMM8, 8); |
8843 | let a: u8x64 = a.as_u8x64(); |
8844 | let b: u8x64 = b.as_u8x64(); |
8845 | let r: u16x32 = vdbpsadbw(a, b, IMM8); |
8846 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_u16x32())) |
8847 | } |
8848 | |
8849 | /// 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. |
8850 | /// |
8851 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_dbsad_epu8&expand=2116) |
8852 | #[inline ] |
8853 | #[target_feature (enable = "avx512bw" )] |
8854 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8855 | #[rustc_legacy_const_generics (3)] |
8856 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
8857 | pub unsafe fn _mm512_maskz_dbsad_epu8<const IMM8: i32>( |
8858 | k: __mmask32, |
8859 | a: __m512i, |
8860 | b: __m512i, |
8861 | ) -> __m512i { |
8862 | static_assert_uimm_bits!(IMM8, 8); |
8863 | let a: u8x64 = a.as_u8x64(); |
8864 | let b: u8x64 = b.as_u8x64(); |
8865 | let r: u16x32 = vdbpsadbw(a, b, IMM8); |
8866 | transmute(src:simd_select_bitmask( |
8867 | m:k, |
8868 | yes:r, |
8869 | no:_mm512_setzero_si512().as_u16x32(), |
8870 | )) |
8871 | } |
8872 | |
8873 | /// 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. |
8874 | /// |
8875 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_dbsad_epu8&expand=2111) |
8876 | #[inline ] |
8877 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8878 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8879 | #[rustc_legacy_const_generics (2)] |
8880 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
8881 | pub unsafe fn _mm256_dbsad_epu8<const IMM8: i32>(a: __m256i, b: __m256i) -> __m256i { |
8882 | static_assert_uimm_bits!(IMM8, 8); |
8883 | let a: u8x32 = a.as_u8x32(); |
8884 | let b: u8x32 = b.as_u8x32(); |
8885 | let r: u16x16 = vdbpsadbw256(a, b, IMM8); |
8886 | transmute(src:r) |
8887 | } |
8888 | |
8889 | /// 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. |
8890 | /// |
8891 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_dbsad_epu8&expand=2112) |
8892 | #[inline ] |
8893 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8894 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8895 | #[rustc_legacy_const_generics (4)] |
8896 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
8897 | pub unsafe fn _mm256_mask_dbsad_epu8<const IMM8: i32>( |
8898 | src: __m256i, |
8899 | k: __mmask16, |
8900 | a: __m256i, |
8901 | b: __m256i, |
8902 | ) -> __m256i { |
8903 | static_assert_uimm_bits!(IMM8, 8); |
8904 | let a: u8x32 = a.as_u8x32(); |
8905 | let b: u8x32 = b.as_u8x32(); |
8906 | let r: u16x16 = vdbpsadbw256(a, b, IMM8); |
8907 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_u16x16())) |
8908 | } |
8909 | |
8910 | /// 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. |
8911 | /// |
8912 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_dbsad_epu8&expand=2113) |
8913 | #[inline ] |
8914 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8915 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8916 | #[rustc_legacy_const_generics (3)] |
8917 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
8918 | pub unsafe fn _mm256_maskz_dbsad_epu8<const IMM8: i32>( |
8919 | k: __mmask16, |
8920 | a: __m256i, |
8921 | b: __m256i, |
8922 | ) -> __m256i { |
8923 | static_assert_uimm_bits!(IMM8, 8); |
8924 | let a: u8x32 = a.as_u8x32(); |
8925 | let b: u8x32 = b.as_u8x32(); |
8926 | let r: u16x16 = vdbpsadbw256(a, b, IMM8); |
8927 | transmute(src:simd_select_bitmask( |
8928 | m:k, |
8929 | yes:r, |
8930 | no:_mm256_setzero_si256().as_u16x16(), |
8931 | )) |
8932 | } |
8933 | |
8934 | /// 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. |
8935 | /// |
8936 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_dbsad_epu8&expand=2108) |
8937 | #[inline ] |
8938 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8939 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8940 | #[rustc_legacy_const_generics (2)] |
8941 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
8942 | pub unsafe fn _mm_dbsad_epu8<const IMM8: i32>(a: __m128i, b: __m128i) -> __m128i { |
8943 | static_assert_uimm_bits!(IMM8, 8); |
8944 | let a: u8x16 = a.as_u8x16(); |
8945 | let b: u8x16 = b.as_u8x16(); |
8946 | let r: u16x8 = vdbpsadbw128(a, b, IMM8); |
8947 | transmute(src:r) |
8948 | } |
8949 | |
8950 | /// 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. |
8951 | /// |
8952 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_dbsad_epu8&expand=2109) |
8953 | #[inline ] |
8954 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8955 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8956 | #[rustc_legacy_const_generics (4)] |
8957 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
8958 | pub unsafe fn _mm_mask_dbsad_epu8<const IMM8: i32>( |
8959 | src: __m128i, |
8960 | k: __mmask8, |
8961 | a: __m128i, |
8962 | b: __m128i, |
8963 | ) -> __m128i { |
8964 | static_assert_uimm_bits!(IMM8, 8); |
8965 | let a: u8x16 = a.as_u8x16(); |
8966 | let b: u8x16 = b.as_u8x16(); |
8967 | let r: u16x8 = vdbpsadbw128(a, b, IMM8); |
8968 | transmute(src:simd_select_bitmask(m:k, yes:r, no:src.as_u16x8())) |
8969 | } |
8970 | |
8971 | /// 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. |
8972 | /// |
8973 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_dbsad_epu8&expand=2110) |
8974 | #[inline ] |
8975 | #[target_feature (enable = "avx512bw,avx512vl" )] |
8976 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8977 | #[rustc_legacy_const_generics (3)] |
8978 | #[cfg_attr (test, assert_instr(vdbpsadbw, IMM8 = 0))] |
8979 | pub unsafe fn _mm_maskz_dbsad_epu8<const IMM8: i32>( |
8980 | k: __mmask8, |
8981 | a: __m128i, |
8982 | b: __m128i, |
8983 | ) -> __m128i { |
8984 | static_assert_uimm_bits!(IMM8, 8); |
8985 | let a: u8x16 = a.as_u8x16(); |
8986 | let b: u8x16 = b.as_u8x16(); |
8987 | let r: u16x8 = vdbpsadbw128(a, b, IMM8); |
8988 | transmute(src:simd_select_bitmask(m:k, yes:r, no:_mm_setzero_si128().as_u16x8())) |
8989 | } |
8990 | |
8991 | /// Set each bit of mask register k based on the most significant bit of the corresponding packed 16-bit integer in a. |
8992 | /// |
8993 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_movepi16_mask&expand=3873) |
8994 | #[inline ] |
8995 | #[target_feature (enable = "avx512bw" )] |
8996 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
8997 | #[cfg_attr (test, assert_instr(vpmovw2m))] |
8998 | pub unsafe fn _mm512_movepi16_mask(a: __m512i) -> __mmask32 { |
8999 | let filter: __m512i = _mm512_set1_epi16(1 << 15); |
9000 | let a: __m512i = _mm512_and_si512(a, b:filter); |
9001 | _mm512_cmpeq_epi16_mask(a, b:filter) |
9002 | } |
9003 | |
9004 | /// Set each bit of mask register k based on the most significant bit of the corresponding packed 16-bit integer in a. |
9005 | /// |
9006 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_movepi16_mask&expand=3872) |
9007 | #[inline ] |
9008 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9009 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9010 | #[cfg_attr (test, assert_instr(vpmovw2m))] |
9011 | pub unsafe fn _mm256_movepi16_mask(a: __m256i) -> __mmask16 { |
9012 | let filter: __m256i = _mm256_set1_epi16(1 << 15); |
9013 | let a: __m256i = _mm256_and_si256(a, b:filter); |
9014 | _mm256_cmpeq_epi16_mask(a, b:filter) |
9015 | } |
9016 | |
9017 | /// Set each bit of mask register k based on the most significant bit of the corresponding packed 16-bit integer in a. |
9018 | /// |
9019 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_movepi16_mask&expand=3871) |
9020 | #[inline ] |
9021 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9022 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9023 | #[cfg_attr (test, assert_instr(vpmovw2m))] |
9024 | pub unsafe fn _mm_movepi16_mask(a: __m128i) -> __mmask8 { |
9025 | let filter: __m128i = _mm_set1_epi16(1 << 15); |
9026 | let a: __m128i = _mm_and_si128(a, b:filter); |
9027 | _mm_cmpeq_epi16_mask(a, b:filter) |
9028 | } |
9029 | |
9030 | /// Set each bit of mask register k based on the most significant bit of the corresponding packed 8-bit integer in a. |
9031 | /// |
9032 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_movepi8_mask&expand=3883) |
9033 | #[inline ] |
9034 | #[target_feature (enable = "avx512bw" )] |
9035 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9036 | #[cfg_attr (test, assert_instr(vpmovb2m))] |
9037 | pub unsafe fn _mm512_movepi8_mask(a: __m512i) -> __mmask64 { |
9038 | let filter: __m512i = _mm512_set1_epi8(1 << 7); |
9039 | let a: __m512i = _mm512_and_si512(a, b:filter); |
9040 | _mm512_cmpeq_epi8_mask(a, b:filter) |
9041 | } |
9042 | |
9043 | /// Set each bit of mask register k based on the most significant bit of the corresponding packed 8-bit integer in a. |
9044 | /// |
9045 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_movepi8_mask&expand=3882) |
9046 | #[inline ] |
9047 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9048 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9049 | #[cfg_attr (test, assert_instr(vpmovmskb))] // should be vpmovb2m but compiled to vpmovmskb in the test shim because that takes less cycles than |
9050 | // using vpmovb2m plus converting the mask register to a standard register. |
9051 | pub unsafe fn _mm256_movepi8_mask(a: __m256i) -> __mmask32 { |
9052 | let filter: __m256i = _mm256_set1_epi8(1 << 7); |
9053 | let a: __m256i = _mm256_and_si256(a, b:filter); |
9054 | _mm256_cmpeq_epi8_mask(a, b:filter) |
9055 | } |
9056 | |
9057 | /// Set each bit of mask register k based on the most significant bit of the corresponding packed 8-bit integer in a. |
9058 | /// |
9059 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_movepi8_mask&expand=3881) |
9060 | #[inline ] |
9061 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9062 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9063 | #[cfg_attr (test, assert_instr(vpmovmskb))] // should be vpmovb2m but compiled to vpmovmskb in the test shim because that takes less cycles than |
9064 | // using vpmovb2m plus converting the mask register to a standard register. |
9065 | pub unsafe fn _mm_movepi8_mask(a: __m128i) -> __mmask16 { |
9066 | let filter: __m128i = _mm_set1_epi8(1 << 7); |
9067 | let a: __m128i = _mm_and_si128(a, b:filter); |
9068 | _mm_cmpeq_epi8_mask(a, b:filter) |
9069 | } |
9070 | |
9071 | /// Set each packed 16-bit integer in dst to all ones or all zeros based on the value of the corresponding bit in k. |
9072 | /// |
9073 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_movm_epi16&expand=3886) |
9074 | #[inline ] |
9075 | #[target_feature (enable = "avx512bw" )] |
9076 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9077 | #[cfg_attr (test, assert_instr(vpmovm2w))] |
9078 | pub unsafe fn _mm512_movm_epi16(k: __mmask32) -> __m512i { |
9079 | let one: i16x32 = _mm512_set1_epi16__m512i( |
9080 | 1 << 15 |
9081 | | 1 << 14 |
9082 | | 1 << 13 |
9083 | | 1 << 12 |
9084 | | 1 << 11 |
9085 | | 1 << 10 |
9086 | | 1 << 9 |
9087 | | 1 << 8 |
9088 | | 1 << 7 |
9089 | | 1 << 6 |
9090 | | 1 << 5 |
9091 | | 1 << 4 |
9092 | | 1 << 3 |
9093 | | 1 << 2 |
9094 | | 1 << 1 |
9095 | | 1 << 0, |
9096 | ) |
9097 | .as_i16x32(); |
9098 | let zero: i16x32 = _mm512_setzero_si512().as_i16x32(); |
9099 | transmute(src:simd_select_bitmask(m:k, yes:one, no:zero)) |
9100 | } |
9101 | |
9102 | /// Set each packed 16-bit integer in dst to all ones or all zeros based on the value of the corresponding bit in k. |
9103 | /// |
9104 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_movm_epi16&expand=3885) |
9105 | #[inline ] |
9106 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9107 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9108 | #[cfg_attr (test, assert_instr(vpmovm2w))] |
9109 | pub unsafe fn _mm256_movm_epi16(k: __mmask16) -> __m256i { |
9110 | let one: i16x16 = _mm256_set1_epi16__m256i( |
9111 | 1 << 15 |
9112 | | 1 << 14 |
9113 | | 1 << 13 |
9114 | | 1 << 12 |
9115 | | 1 << 11 |
9116 | | 1 << 10 |
9117 | | 1 << 9 |
9118 | | 1 << 8 |
9119 | | 1 << 7 |
9120 | | 1 << 6 |
9121 | | 1 << 5 |
9122 | | 1 << 4 |
9123 | | 1 << 3 |
9124 | | 1 << 2 |
9125 | | 1 << 1 |
9126 | | 1 << 0, |
9127 | ) |
9128 | .as_i16x16(); |
9129 | let zero: i16x16 = _mm256_setzero_si256().as_i16x16(); |
9130 | transmute(src:simd_select_bitmask(m:k, yes:one, no:zero)) |
9131 | } |
9132 | |
9133 | /// Set each packed 16-bit integer in dst to all ones or all zeros based on the value of the corresponding bit in k. |
9134 | /// |
9135 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_movm_epi16&expand=3884) |
9136 | #[inline ] |
9137 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9138 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9139 | #[cfg_attr (test, assert_instr(vpmovm2w))] |
9140 | pub unsafe fn _mm_movm_epi16(k: __mmask8) -> __m128i { |
9141 | let one: i16x8 = _mm_set1_epi16__m128i( |
9142 | 1 << 15 |
9143 | | 1 << 14 |
9144 | | 1 << 13 |
9145 | | 1 << 12 |
9146 | | 1 << 11 |
9147 | | 1 << 10 |
9148 | | 1 << 9 |
9149 | | 1 << 8 |
9150 | | 1 << 7 |
9151 | | 1 << 6 |
9152 | | 1 << 5 |
9153 | | 1 << 4 |
9154 | | 1 << 3 |
9155 | | 1 << 2 |
9156 | | 1 << 1 |
9157 | | 1 << 0, |
9158 | ) |
9159 | .as_i16x8(); |
9160 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
9161 | transmute(src:simd_select_bitmask(m:k, yes:one, no:zero)) |
9162 | } |
9163 | |
9164 | /// Set each packed 8-bit integer in dst to all ones or all zeros based on the value of the corresponding bit in k. |
9165 | /// |
9166 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_movm_epi8&expand=3895) |
9167 | #[inline ] |
9168 | #[target_feature (enable = "avx512bw" )] |
9169 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9170 | #[cfg_attr (test, assert_instr(vpmovm2b))] |
9171 | pub unsafe fn _mm512_movm_epi8(k: __mmask64) -> __m512i { |
9172 | let one: i8x64 = |
9173 | _mm512_set1_epi8__m512i(1 << 7 | 1 << 6 | 1 << 5 | 1 << 4 | 1 << 3 | 1 << 2 | 1 << 1 | 1 << 0) |
9174 | .as_i8x64(); |
9175 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
9176 | transmute(src:simd_select_bitmask(m:k, yes:one, no:zero)) |
9177 | } |
9178 | |
9179 | /// Set each packed 8-bit integer in dst to all ones or all zeros based on the value of the corresponding bit in k. |
9180 | /// |
9181 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_movm_epi8&expand=3894) |
9182 | #[inline ] |
9183 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9184 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9185 | #[cfg_attr (test, assert_instr(vpmovm2b))] |
9186 | pub unsafe fn _mm256_movm_epi8(k: __mmask32) -> __m256i { |
9187 | let one: i8x32 = |
9188 | _mm256_set1_epi8__m256i(1 << 7 | 1 << 6 | 1 << 5 | 1 << 4 | 1 << 3 | 1 << 2 | 1 << 1 | 1 << 0) |
9189 | .as_i8x32(); |
9190 | let zero: i8x32 = _mm256_setzero_si256().as_i8x32(); |
9191 | transmute(src:simd_select_bitmask(m:k, yes:one, no:zero)) |
9192 | } |
9193 | |
9194 | /// Set each packed 8-bit integer in dst to all ones or all zeros based on the value of the corresponding bit in k. |
9195 | /// |
9196 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_movm_epi8&expand=3893) |
9197 | #[inline ] |
9198 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9199 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9200 | #[cfg_attr (test, assert_instr(vpmovm2b))] |
9201 | pub unsafe fn _mm_movm_epi8(k: __mmask16) -> __m128i { |
9202 | let one: i8x16 = _mm_set1_epi8__m128i(1 << 7 | 1 << 6 | 1 << 5 | 1 << 4 | 1 << 3 | 1 << 2 | 1 << 1 | 1 << 0) |
9203 | .as_i8x16(); |
9204 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
9205 | transmute(src:simd_select_bitmask(m:k, yes:one, no:zero)) |
9206 | } |
9207 | |
9208 | /// Add 32-bit masks in a and b, and store the result in k. |
9209 | /// |
9210 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kadd_mask32&expand=3207) |
9211 | #[inline ] |
9212 | #[target_feature (enable = "avx512bw" )] |
9213 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9214 | pub unsafe fn _kadd_mask32(a: __mmask32, b: __mmask32) -> __mmask32 { |
9215 | a + b |
9216 | } |
9217 | |
9218 | /// Add 64-bit masks in a and b, and store the result in k. |
9219 | /// |
9220 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kadd_mask64&expand=3208) |
9221 | #[inline ] |
9222 | #[target_feature (enable = "avx512bw" )] |
9223 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9224 | pub unsafe fn _kadd_mask64(a: __mmask64, b: __mmask64) -> __mmask64 { |
9225 | a + b |
9226 | } |
9227 | |
9228 | /// Compute the bitwise AND of 32-bit masks a and b, and store the result in k. |
9229 | /// |
9230 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kand_mask32&expand=3213) |
9231 | #[inline ] |
9232 | #[target_feature (enable = "avx512bw" )] |
9233 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9234 | pub unsafe fn _kand_mask32(a: __mmask32, b: __mmask32) -> __mmask32 { |
9235 | a & b |
9236 | } |
9237 | |
9238 | /// Compute the bitwise AND of 64-bit masks a and b, and store the result in k. |
9239 | /// |
9240 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kand_mask64&expand=3214) |
9241 | #[inline ] |
9242 | #[target_feature (enable = "avx512bw" )] |
9243 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9244 | pub unsafe fn _kand_mask64(a: __mmask64, b: __mmask64) -> __mmask64 { |
9245 | a & b |
9246 | } |
9247 | |
9248 | /// Compute the bitwise NOT of 32-bit mask a, and store the result in k. |
9249 | /// |
9250 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_knot_mask32&expand=3234) |
9251 | #[inline ] |
9252 | #[target_feature (enable = "avx512bw" )] |
9253 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9254 | pub unsafe fn _knot_mask32(a: __mmask32) -> __mmask32 { |
9255 | a ^ 0b11111111_11111111_11111111_11111111 |
9256 | } |
9257 | |
9258 | /// Compute the bitwise NOT of 64-bit mask a, and store the result in k. |
9259 | /// |
9260 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_knot_mask64&expand=3235) |
9261 | #[inline ] |
9262 | #[target_feature (enable = "avx512bw" )] |
9263 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9264 | pub unsafe fn _knot_mask64(a: __mmask64) -> __mmask64 { |
9265 | a ^ 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
9266 | } |
9267 | |
9268 | /// Compute the bitwise NOT of 32-bit masks a and then AND with b, and store the result in k. |
9269 | /// |
9270 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kandn_mask32&expand=3219) |
9271 | #[inline ] |
9272 | #[target_feature (enable = "avx512bw" )] |
9273 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9274 | pub unsafe fn _kandn_mask32(a: __mmask32, b: __mmask32) -> __mmask32 { |
9275 | _knot_mask32(a) & b |
9276 | } |
9277 | |
9278 | /// Compute the bitwise NOT of 64-bit masks a and then AND with b, and store the result in k. |
9279 | /// |
9280 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kandn_mask64&expand=3220) |
9281 | #[inline ] |
9282 | #[target_feature (enable = "avx512bw" )] |
9283 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9284 | pub unsafe fn _kandn_mask64(a: __mmask64, b: __mmask64) -> __mmask64 { |
9285 | _knot_mask64(a) & b |
9286 | } |
9287 | |
9288 | /// Compute the bitwise OR of 32-bit masks a and b, and store the result in k. |
9289 | /// |
9290 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kor_mask32&expand=3240) |
9291 | #[inline ] |
9292 | #[target_feature (enable = "avx512bw" )] |
9293 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9294 | pub unsafe fn _kor_mask32(a: __mmask32, b: __mmask32) -> __mmask32 { |
9295 | a | b |
9296 | } |
9297 | |
9298 | /// Compute the bitwise OR of 64-bit masks a and b, and store the result in k. |
9299 | /// |
9300 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kor_mask64&expand=3241) |
9301 | #[inline ] |
9302 | #[target_feature (enable = "avx512bw" )] |
9303 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9304 | pub unsafe fn _kor_mask64(a: __mmask64, b: __mmask64) -> __mmask64 { |
9305 | a | b |
9306 | } |
9307 | |
9308 | /// Compute the bitwise XOR of 32-bit masks a and b, and store the result in k. |
9309 | /// |
9310 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kxor_mask32&expand=3292) |
9311 | #[inline ] |
9312 | #[target_feature (enable = "avx512bw" )] |
9313 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9314 | pub unsafe fn _kxor_mask32(a: __mmask32, b: __mmask32) -> __mmask32 { |
9315 | a ^ b |
9316 | } |
9317 | |
9318 | /// Compute the bitwise XOR of 64-bit masks a and b, and store the result in k. |
9319 | /// |
9320 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kxor_mask64&expand=3293) |
9321 | #[inline ] |
9322 | #[target_feature (enable = "avx512bw" )] |
9323 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9324 | pub unsafe fn _kxor_mask64(a: __mmask64, b: __mmask64) -> __mmask64 { |
9325 | a ^ b |
9326 | } |
9327 | |
9328 | /// Compute the bitwise XNOR of 32-bit masks a and b, and store the result in k. |
9329 | /// |
9330 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kxnor_mask32&expand=3286) |
9331 | #[inline ] |
9332 | #[target_feature (enable = "avx512bw" )] |
9333 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9334 | pub unsafe fn _kxnor_mask32(a: __mmask32, b: __mmask32) -> __mmask32 { |
9335 | _knot_mask32(a ^ b) |
9336 | } |
9337 | |
9338 | /// Compute the bitwise XNOR of 64-bit masks a and b, and store the result in k. |
9339 | /// |
9340 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_kxnor_mask64&expand=3287) |
9341 | #[inline ] |
9342 | #[target_feature (enable = "avx512bw" )] |
9343 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9344 | pub unsafe fn _kxnor_mask64(a: __mmask64, b: __mmask64) -> __mmask64 { |
9345 | _knot_mask64(a ^ b) |
9346 | } |
9347 | |
9348 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst. |
9349 | /// |
9350 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cvtepi16_epi8&expand=1407) |
9351 | #[inline ] |
9352 | #[target_feature (enable = "avx512bw" )] |
9353 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9354 | #[cfg_attr (test, assert_instr(vpmovwb))] |
9355 | pub unsafe fn _mm512_cvtepi16_epi8(a: __m512i) -> __m256i { |
9356 | let a: i16x32 = a.as_i16x32(); |
9357 | transmute::<i8x32, _>(src:simd_cast(a)) |
9358 | } |
9359 | |
9360 | /// 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). |
9361 | /// |
9362 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtepi16_epi8&expand=1408) |
9363 | #[inline ] |
9364 | #[target_feature (enable = "avx512bw" )] |
9365 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9366 | #[cfg_attr (test, assert_instr(vpmovwb))] |
9367 | pub unsafe fn _mm512_mask_cvtepi16_epi8(src: __m256i, k: __mmask32, a: __m512i) -> __m256i { |
9368 | let convert: i8x32 = _mm512_cvtepi16_epi8(a).as_i8x32(); |
9369 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i8x32())) |
9370 | } |
9371 | |
9372 | /// 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). |
9373 | /// |
9374 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_cvtepi16_epi8&expand=1409) |
9375 | #[inline ] |
9376 | #[target_feature (enable = "avx512bw" )] |
9377 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9378 | #[cfg_attr (test, assert_instr(vpmovwb))] |
9379 | pub unsafe fn _mm512_maskz_cvtepi16_epi8(k: __mmask32, a: __m512i) -> __m256i { |
9380 | let convert: i8x32 = _mm512_cvtepi16_epi8(a).as_i8x32(); |
9381 | transmute(src:simd_select_bitmask( |
9382 | m:k, |
9383 | yes:convert, |
9384 | no:_mm256_setzero_si256().as_i8x32(), |
9385 | )) |
9386 | } |
9387 | |
9388 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst. |
9389 | /// |
9390 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cvtepi16_epi8&expand=1404) |
9391 | #[inline ] |
9392 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9393 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9394 | #[cfg_attr (test, assert_instr(vpmovwb))] |
9395 | pub unsafe fn _mm256_cvtepi16_epi8(a: __m256i) -> __m128i { |
9396 | let a: i16x16 = a.as_i16x16(); |
9397 | transmute::<i8x16, _>(src:simd_cast(a)) |
9398 | } |
9399 | |
9400 | /// 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). |
9401 | /// |
9402 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtepi16_epi8&expand=1405) |
9403 | #[inline ] |
9404 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9405 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9406 | #[cfg_attr (test, assert_instr(vpmovwb))] |
9407 | pub unsafe fn _mm256_mask_cvtepi16_epi8(src: __m128i, k: __mmask16, a: __m256i) -> __m128i { |
9408 | let convert: i8x16 = _mm256_cvtepi16_epi8(a).as_i8x16(); |
9409 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i8x16())) |
9410 | } |
9411 | |
9412 | /// 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). |
9413 | /// |
9414 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_cvtepi16_epi8&expand=1406) |
9415 | #[inline ] |
9416 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9417 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9418 | #[cfg_attr (test, assert_instr(vpmovwb))] |
9419 | pub unsafe fn _mm256_maskz_cvtepi16_epi8(k: __mmask16, a: __m256i) -> __m128i { |
9420 | let convert: i8x16 = _mm256_cvtepi16_epi8(a).as_i8x16(); |
9421 | transmute(src:simd_select_bitmask( |
9422 | m:k, |
9423 | yes:convert, |
9424 | no:_mm_setzero_si128().as_i8x16(), |
9425 | )) |
9426 | } |
9427 | |
9428 | /// Convert packed 16-bit integers in a to packed 8-bit integers with truncation, and store the results in dst. |
9429 | /// |
9430 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtepi16_epi8&expand=1401) |
9431 | #[inline ] |
9432 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9433 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9434 | #[cfg_attr (test, assert_instr(vpmovwb))] |
9435 | pub unsafe fn _mm_cvtepi16_epi8(a: __m128i) -> __m128i { |
9436 | let a: i16x8 = a.as_i16x8(); |
9437 | let zero: i16x8 = _mm_setzero_si128().as_i16x8(); |
9438 | let v256: i16x16 = simd_shuffle!(a, zero, [0, 1, 2, 3, 4, 5, 6, 7, 8, 8, 8, 8, 8, 8, 8, 8]); |
9439 | transmute::<i8x16, _>(src:simd_cast(v256)) |
9440 | } |
9441 | |
9442 | /// 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). |
9443 | /// |
9444 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtepi16_epi8&expand=1402) |
9445 | #[inline ] |
9446 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9447 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9448 | #[cfg_attr (test, assert_instr(vpmovwb))] |
9449 | pub unsafe fn _mm_mask_cvtepi16_epi8(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
9450 | let convert: i8x16 = _mm_cvtepi16_epi8(a).as_i8x16(); |
9451 | let k: __mmask16 = 0b11111111_11111111 & k as __mmask16; |
9452 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i8x16())) |
9453 | } |
9454 | |
9455 | /// 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). |
9456 | /// |
9457 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_cvtepi16_epi8&expand=1403) |
9458 | #[inline ] |
9459 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9460 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9461 | #[cfg_attr (test, assert_instr(vpmovwb))] |
9462 | pub unsafe fn _mm_maskz_cvtepi16_epi8(k: __mmask8, a: __m128i) -> __m128i { |
9463 | let convert: i8x16 = _mm_cvtepi16_epi8(a).as_i8x16(); |
9464 | let k: __mmask16 = 0b11111111_11111111 & k as __mmask16; |
9465 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
9466 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:zero)) |
9467 | } |
9468 | |
9469 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst. |
9470 | /// |
9471 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cvtsepi16_epi8&expand=1807) |
9472 | #[inline ] |
9473 | #[target_feature (enable = "avx512bw" )] |
9474 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9475 | #[cfg_attr (test, assert_instr(vpmovswb))] |
9476 | pub unsafe fn _mm512_cvtsepi16_epi8(a: __m512i) -> __m256i { |
9477 | transmute(src:vpmovswb( |
9478 | a:a.as_i16x32(), |
9479 | src:_mm256_setzero_si256().as_i8x32(), |
9480 | mask:0b11111111_11111111_11111111_11111111, |
9481 | )) |
9482 | } |
9483 | |
9484 | /// 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). |
9485 | /// |
9486 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtsepi16_epi8&expand=1808) |
9487 | #[inline ] |
9488 | #[target_feature (enable = "avx512bw" )] |
9489 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9490 | #[cfg_attr (test, assert_instr(vpmovswb))] |
9491 | pub unsafe fn _mm512_mask_cvtsepi16_epi8(src: __m256i, k: __mmask32, a: __m512i) -> __m256i { |
9492 | transmute(src:vpmovswb(a:a.as_i16x32(), src:src.as_i8x32(), mask:k)) |
9493 | } |
9494 | |
9495 | /// 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). |
9496 | /// |
9497 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_cvtsepi16_epi8&expand=1809) |
9498 | #[inline ] |
9499 | #[target_feature (enable = "avx512bw" )] |
9500 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9501 | #[cfg_attr (test, assert_instr(vpmovswb))] |
9502 | pub unsafe fn _mm512_maskz_cvtsepi16_epi8(k: __mmask32, a: __m512i) -> __m256i { |
9503 | transmute(src:vpmovswb( |
9504 | a:a.as_i16x32(), |
9505 | src:_mm256_setzero_si256().as_i8x32(), |
9506 | mask:k, |
9507 | )) |
9508 | } |
9509 | |
9510 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst. |
9511 | /// |
9512 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cvtsepi16_epi8&expand=1804) |
9513 | #[inline ] |
9514 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9515 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9516 | #[cfg_attr (test, assert_instr(vpmovswb))] |
9517 | pub unsafe fn _mm256_cvtsepi16_epi8(a: __m256i) -> __m128i { |
9518 | transmute(src:vpmovswb256( |
9519 | a:a.as_i16x16(), |
9520 | src:_mm_setzero_si128().as_i8x16(), |
9521 | mask:0b11111111_11111111, |
9522 | )) |
9523 | } |
9524 | |
9525 | /// 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). |
9526 | /// |
9527 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtsepi16_epi8&expand=1805) |
9528 | #[inline ] |
9529 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9530 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9531 | #[cfg_attr (test, assert_instr(vpmovswb))] |
9532 | pub unsafe fn _mm256_mask_cvtsepi16_epi8(src: __m128i, k: __mmask16, a: __m256i) -> __m128i { |
9533 | transmute(src:vpmovswb256(a:a.as_i16x16(), src:src.as_i8x16(), mask:k)) |
9534 | } |
9535 | |
9536 | /// 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). |
9537 | /// |
9538 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_cvtsepi16_epi8&expand=1806) |
9539 | #[inline ] |
9540 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9541 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9542 | #[cfg_attr (test, assert_instr(vpmovswb))] |
9543 | pub unsafe fn _mm256_maskz_cvtsepi16_epi8(k: __mmask16, a: __m256i) -> __m128i { |
9544 | transmute(src:vpmovswb256( |
9545 | a:a.as_i16x16(), |
9546 | src:_mm_setzero_si128().as_i8x16(), |
9547 | mask:k, |
9548 | )) |
9549 | } |
9550 | |
9551 | /// Convert packed signed 16-bit integers in a to packed 8-bit integers with signed saturation, and store the results in dst. |
9552 | /// |
9553 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtsepi16_epi8&expand=1801) |
9554 | #[inline ] |
9555 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9556 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9557 | #[cfg_attr (test, assert_instr(vpmovswb))] |
9558 | pub unsafe fn _mm_cvtsepi16_epi8(a: __m128i) -> __m128i { |
9559 | transmute(src:vpmovswb128( |
9560 | a:a.as_i16x8(), |
9561 | src:_mm_setzero_si128().as_i8x16(), |
9562 | mask:0b11111111, |
9563 | )) |
9564 | } |
9565 | |
9566 | /// 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). |
9567 | /// |
9568 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtsepi16_epi8&expand=1802) |
9569 | #[inline ] |
9570 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9571 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9572 | #[cfg_attr (test, assert_instr(vpmovswb))] |
9573 | pub unsafe fn _mm_mask_cvtsepi16_epi8(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
9574 | transmute(src:vpmovswb128(a:a.as_i16x8(), src:src.as_i8x16(), mask:k)) |
9575 | } |
9576 | |
9577 | /// 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). |
9578 | /// |
9579 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_cvtsepi16_epi8&expand=1803) |
9580 | #[inline ] |
9581 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9582 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9583 | #[cfg_attr (test, assert_instr(vpmovswb))] |
9584 | pub unsafe fn _mm_maskz_cvtsepi16_epi8(k: __mmask8, a: __m128i) -> __m128i { |
9585 | transmute(src:vpmovswb128(a:a.as_i16x8(), src:_mm_setzero_si128().as_i8x16(), mask:k)) |
9586 | } |
9587 | |
9588 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst. |
9589 | /// |
9590 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cvtusepi16_epi8&expand=2042) |
9591 | #[inline ] |
9592 | #[target_feature (enable = "avx512bw" )] |
9593 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9594 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
9595 | pub unsafe fn _mm512_cvtusepi16_epi8(a: __m512i) -> __m256i { |
9596 | transmute(src:vpmovuswb( |
9597 | a:a.as_u16x32(), |
9598 | src:_mm256_setzero_si256().as_u8x32(), |
9599 | mask:0b11111111_11111111_11111111_11111111, |
9600 | )) |
9601 | } |
9602 | |
9603 | /// 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). |
9604 | /// |
9605 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtusepi16_epi8&expand=2043) |
9606 | #[inline ] |
9607 | #[target_feature (enable = "avx512bw" )] |
9608 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9609 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
9610 | pub unsafe fn _mm512_mask_cvtusepi16_epi8(src: __m256i, k: __mmask32, a: __m512i) -> __m256i { |
9611 | transmute(src:vpmovuswb(a:a.as_u16x32(), src:src.as_u8x32(), mask:k)) |
9612 | } |
9613 | |
9614 | /// 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). |
9615 | /// |
9616 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_cvtusepi16_epi8&expand=2044) |
9617 | #[inline ] |
9618 | #[target_feature (enable = "avx512bw" )] |
9619 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9620 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
9621 | pub unsafe fn _mm512_maskz_cvtusepi16_epi8(k: __mmask32, a: __m512i) -> __m256i { |
9622 | transmute(src:vpmovuswb( |
9623 | a:a.as_u16x32(), |
9624 | src:_mm256_setzero_si256().as_u8x32(), |
9625 | mask:k, |
9626 | )) |
9627 | } |
9628 | |
9629 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst. |
9630 | /// |
9631 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_cvtusepi16_epi8&expand=2039) |
9632 | #[inline ] |
9633 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9634 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9635 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
9636 | pub unsafe fn _mm256_cvtusepi16_epi8(a: __m256i) -> __m128i { |
9637 | transmute(src:vpmovuswb256( |
9638 | a:a.as_u16x16(), |
9639 | src:_mm_setzero_si128().as_u8x16(), |
9640 | mask:0b11111111_11111111, |
9641 | )) |
9642 | } |
9643 | |
9644 | /// 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). |
9645 | /// |
9646 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtusepi16_epi8&expand=2040) |
9647 | #[inline ] |
9648 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9649 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9650 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
9651 | pub unsafe fn _mm256_mask_cvtusepi16_epi8(src: __m128i, k: __mmask16, a: __m256i) -> __m128i { |
9652 | transmute(src:vpmovuswb256(a:a.as_u16x16(), src:src.as_u8x16(), mask:k)) |
9653 | } |
9654 | |
9655 | /// 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). |
9656 | /// |
9657 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_cvtusepi16_epi8&expand=2041) |
9658 | #[inline ] |
9659 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9660 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9661 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
9662 | pub unsafe fn _mm256_maskz_cvtusepi16_epi8(k: __mmask16, a: __m256i) -> __m128i { |
9663 | transmute(src:vpmovuswb256( |
9664 | a:a.as_u16x16(), |
9665 | src:_mm_setzero_si128().as_u8x16(), |
9666 | mask:k, |
9667 | )) |
9668 | } |
9669 | |
9670 | /// Convert packed unsigned 16-bit integers in a to packed unsigned 8-bit integers with unsigned saturation, and store the results in dst. |
9671 | /// |
9672 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtusepi16_epi8&expand=2036) |
9673 | #[inline ] |
9674 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9675 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9676 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
9677 | pub unsafe fn _mm_cvtusepi16_epi8(a: __m128i) -> __m128i { |
9678 | transmute(src:vpmovuswb128( |
9679 | a:a.as_u16x8(), |
9680 | src:_mm_setzero_si128().as_u8x16(), |
9681 | mask:0b11111111, |
9682 | )) |
9683 | } |
9684 | |
9685 | /// 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). |
9686 | /// |
9687 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtusepi16_epi8&expand=2037) |
9688 | #[inline ] |
9689 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9690 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9691 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
9692 | pub unsafe fn _mm_mask_cvtusepi16_epi8(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
9693 | transmute(src:vpmovuswb128(a:a.as_u16x8(), src:src.as_u8x16(), mask:k)) |
9694 | } |
9695 | |
9696 | /// 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). |
9697 | /// |
9698 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_cvtusepi16_epi8&expand=2038) |
9699 | #[inline ] |
9700 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9701 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9702 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
9703 | pub unsafe fn _mm_maskz_cvtusepi16_epi8(k: __mmask8, a: __m128i) -> __m128i { |
9704 | transmute(src:vpmovuswb128( |
9705 | a:a.as_u16x8(), |
9706 | src:_mm_setzero_si128().as_u8x16(), |
9707 | mask:k, |
9708 | )) |
9709 | } |
9710 | |
9711 | /// Sign extend packed 8-bit integers in a to packed 16-bit integers, and store the results in dst. |
9712 | /// |
9713 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cvtepi8_epi16&expand=1526) |
9714 | #[inline ] |
9715 | #[target_feature (enable = "avx512bw" )] |
9716 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9717 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
9718 | pub unsafe fn _mm512_cvtepi8_epi16(a: __m256i) -> __m512i { |
9719 | let a: i8x32 = a.as_i8x32(); |
9720 | transmute::<i16x32, _>(src:simd_cast(a)) |
9721 | } |
9722 | |
9723 | /// 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). |
9724 | /// |
9725 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtepi8_epi16&expand=1527) |
9726 | #[inline ] |
9727 | #[target_feature (enable = "avx512bw" )] |
9728 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9729 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
9730 | pub unsafe fn _mm512_mask_cvtepi8_epi16(src: __m512i, k: __mmask32, a: __m256i) -> __m512i { |
9731 | let convert: i16x32 = _mm512_cvtepi8_epi16(a).as_i16x32(); |
9732 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i16x32())) |
9733 | } |
9734 | |
9735 | /// 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). |
9736 | /// |
9737 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_cvtepi8_epi16&expand=1528) |
9738 | #[inline ] |
9739 | #[target_feature (enable = "avx512bw" )] |
9740 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9741 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
9742 | pub unsafe fn _mm512_maskz_cvtepi8_epi16(k: __mmask32, a: __m256i) -> __m512i { |
9743 | let convert: i16x32 = _mm512_cvtepi8_epi16(a).as_i16x32(); |
9744 | transmute(src:simd_select_bitmask( |
9745 | m:k, |
9746 | yes:convert, |
9747 | no:_mm512_setzero_si512().as_i16x32(), |
9748 | )) |
9749 | } |
9750 | |
9751 | /// 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). |
9752 | /// |
9753 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtepi8_epi16&expand=1524) |
9754 | #[inline ] |
9755 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9756 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9757 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
9758 | pub unsafe fn _mm256_mask_cvtepi8_epi16(src: __m256i, k: __mmask16, a: __m128i) -> __m256i { |
9759 | let convert: i16x16 = _mm256_cvtepi8_epi16(a).as_i16x16(); |
9760 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i16x16())) |
9761 | } |
9762 | |
9763 | /// 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). |
9764 | /// |
9765 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_cvtepi8_epi16&expand=1525) |
9766 | #[inline ] |
9767 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9768 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9769 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
9770 | pub unsafe fn _mm256_maskz_cvtepi8_epi16(k: __mmask16, a: __m128i) -> __m256i { |
9771 | let convert: i16x16 = _mm256_cvtepi8_epi16(a).as_i16x16(); |
9772 | transmute(src:simd_select_bitmask( |
9773 | m:k, |
9774 | yes:convert, |
9775 | no:_mm256_setzero_si256().as_i16x16(), |
9776 | )) |
9777 | } |
9778 | |
9779 | /// 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). |
9780 | /// |
9781 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtepi8_epi16&expand=1521) |
9782 | #[inline ] |
9783 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9784 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9785 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
9786 | pub unsafe fn _mm_mask_cvtepi8_epi16(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
9787 | let convert: i16x8 = _mm_cvtepi8_epi16(a).as_i16x8(); |
9788 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i16x8())) |
9789 | } |
9790 | |
9791 | /// 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). |
9792 | /// |
9793 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_cvtepi8_epi16&expand=1522) |
9794 | #[inline ] |
9795 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9796 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9797 | #[cfg_attr (test, assert_instr(vpmovsxbw))] |
9798 | pub unsafe fn _mm_maskz_cvtepi8_epi16(k: __mmask8, a: __m128i) -> __m128i { |
9799 | let convert: i16x8 = _mm_cvtepi8_epi16(a).as_i16x8(); |
9800 | transmute(src:simd_select_bitmask( |
9801 | m:k, |
9802 | yes:convert, |
9803 | no:_mm_setzero_si128().as_i16x8(), |
9804 | )) |
9805 | } |
9806 | |
9807 | /// Zero extend packed unsigned 8-bit integers in a to packed 16-bit integers, and store the results in dst. |
9808 | /// |
9809 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_cvtepu8_epi16&expand=1612) |
9810 | #[inline ] |
9811 | #[target_feature (enable = "avx512bw" )] |
9812 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9813 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
9814 | pub unsafe fn _mm512_cvtepu8_epi16(a: __m256i) -> __m512i { |
9815 | let a: u8x32 = a.as_u8x32(); |
9816 | transmute::<i16x32, _>(src:simd_cast(a)) |
9817 | } |
9818 | |
9819 | /// 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). |
9820 | /// |
9821 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtepu8_epi16&expand=1613) |
9822 | #[inline ] |
9823 | #[target_feature (enable = "avx512bw" )] |
9824 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9825 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
9826 | pub unsafe fn _mm512_mask_cvtepu8_epi16(src: __m512i, k: __mmask32, a: __m256i) -> __m512i { |
9827 | let convert: i16x32 = _mm512_cvtepu8_epi16(a).as_i16x32(); |
9828 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i16x32())) |
9829 | } |
9830 | |
9831 | /// 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). |
9832 | /// |
9833 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_cvtepu8_epi16&expand=1614) |
9834 | #[inline ] |
9835 | #[target_feature (enable = "avx512bw" )] |
9836 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9837 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
9838 | pub unsafe fn _mm512_maskz_cvtepu8_epi16(k: __mmask32, a: __m256i) -> __m512i { |
9839 | let convert: i16x32 = _mm512_cvtepu8_epi16(a).as_i16x32(); |
9840 | transmute(src:simd_select_bitmask( |
9841 | m:k, |
9842 | yes:convert, |
9843 | no:_mm512_setzero_si512().as_i16x32(), |
9844 | )) |
9845 | } |
9846 | |
9847 | /// 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). |
9848 | /// |
9849 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtepu8_epi16&expand=1610) |
9850 | #[inline ] |
9851 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9852 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9853 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
9854 | pub unsafe fn _mm256_mask_cvtepu8_epi16(src: __m256i, k: __mmask16, a: __m128i) -> __m256i { |
9855 | let convert: i16x16 = _mm256_cvtepu8_epi16(a).as_i16x16(); |
9856 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i16x16())) |
9857 | } |
9858 | |
9859 | /// 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). |
9860 | /// |
9861 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_cvtepu8_epi16&expand=1611) |
9862 | #[inline ] |
9863 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9864 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9865 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
9866 | pub unsafe fn _mm256_maskz_cvtepu8_epi16(k: __mmask16, a: __m128i) -> __m256i { |
9867 | let convert: i16x16 = _mm256_cvtepu8_epi16(a).as_i16x16(); |
9868 | transmute(src:simd_select_bitmask( |
9869 | m:k, |
9870 | yes:convert, |
9871 | no:_mm256_setzero_si256().as_i16x16(), |
9872 | )) |
9873 | } |
9874 | |
9875 | /// 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). |
9876 | /// |
9877 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtepu8_epi16&expand=1607) |
9878 | #[inline ] |
9879 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9880 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9881 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
9882 | pub unsafe fn _mm_mask_cvtepu8_epi16(src: __m128i, k: __mmask8, a: __m128i) -> __m128i { |
9883 | let convert: i16x8 = _mm_cvtepu8_epi16(a).as_i16x8(); |
9884 | transmute(src:simd_select_bitmask(m:k, yes:convert, no:src.as_i16x8())) |
9885 | } |
9886 | |
9887 | /// 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). |
9888 | /// |
9889 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_cvtepu8_epi16&expand=1608) |
9890 | #[inline ] |
9891 | #[target_feature (enable = "avx512bw,avx512vl" )] |
9892 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9893 | #[cfg_attr (test, assert_instr(vpmovzxbw))] |
9894 | pub unsafe fn _mm_maskz_cvtepu8_epi16(k: __mmask8, a: __m128i) -> __m128i { |
9895 | let convert: i16x8 = _mm_cvtepu8_epi16(a).as_i16x8(); |
9896 | transmute(src:simd_select_bitmask( |
9897 | m:k, |
9898 | yes:convert, |
9899 | no:_mm_setzero_si128().as_i16x8(), |
9900 | )) |
9901 | } |
9902 | |
9903 | /// Shift 128-bit lanes in a left by imm8 bytes while shifting in zeros, and store the results in dst. |
9904 | /// |
9905 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_bslli_epi128&expand=591) |
9906 | #[inline ] |
9907 | #[target_feature (enable = "avx512bw" )] |
9908 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
9909 | #[cfg_attr (test, assert_instr(vpslldq, IMM8 = 3))] |
9910 | #[rustc_legacy_const_generics (1)] |
9911 | pub unsafe fn _mm512_bslli_epi128<const IMM8: i32>(a: __m512i) -> __m512i { |
9912 | static_assert_uimm_bits!(IMM8, 8); |
9913 | const fn mask(shift: i32, i: u32) -> u32 { |
9914 | let shift = shift as u32 & 0xff; |
9915 | if shift > 15 || i % 16 < shift { |
9916 | 0 |
9917 | } else { |
9918 | 64 + (i - shift) |
9919 | } |
9920 | } |
9921 | let a = a.as_i8x64(); |
9922 | let zero = _mm512_setzero_si512().as_i8x64(); |
9923 | let r: i8x64 = simd_shuffle!( |
9924 | zero, |
9925 | a, |
9926 | [ |
9927 | mask(IMM8, 0), |
9928 | mask(IMM8, 1), |
9929 | mask(IMM8, 2), |
9930 | mask(IMM8, 3), |
9931 | mask(IMM8, 4), |
9932 | mask(IMM8, 5), |
9933 | mask(IMM8, 6), |
9934 | mask(IMM8, 7), |
9935 | mask(IMM8, 8), |
9936 | mask(IMM8, 9), |
9937 | mask(IMM8, 10), |
9938 | mask(IMM8, 11), |
9939 | mask(IMM8, 12), |
9940 | mask(IMM8, 13), |
9941 | mask(IMM8, 14), |
9942 | mask(IMM8, 15), |
9943 | mask(IMM8, 16), |
9944 | mask(IMM8, 17), |
9945 | mask(IMM8, 18), |
9946 | mask(IMM8, 19), |
9947 | mask(IMM8, 20), |
9948 | mask(IMM8, 21), |
9949 | mask(IMM8, 22), |
9950 | mask(IMM8, 23), |
9951 | mask(IMM8, 24), |
9952 | mask(IMM8, 25), |
9953 | mask(IMM8, 26), |
9954 | mask(IMM8, 27), |
9955 | mask(IMM8, 28), |
9956 | mask(IMM8, 29), |
9957 | mask(IMM8, 30), |
9958 | mask(IMM8, 31), |
9959 | mask(IMM8, 32), |
9960 | mask(IMM8, 33), |
9961 | mask(IMM8, 34), |
9962 | mask(IMM8, 35), |
9963 | mask(IMM8, 36), |
9964 | mask(IMM8, 37), |
9965 | mask(IMM8, 38), |
9966 | mask(IMM8, 39), |
9967 | mask(IMM8, 40), |
9968 | mask(IMM8, 41), |
9969 | mask(IMM8, 42), |
9970 | mask(IMM8, 43), |
9971 | mask(IMM8, 44), |
9972 | mask(IMM8, 45), |
9973 | mask(IMM8, 46), |
9974 | mask(IMM8, 47), |
9975 | mask(IMM8, 48), |
9976 | mask(IMM8, 49), |
9977 | mask(IMM8, 50), |
9978 | mask(IMM8, 51), |
9979 | mask(IMM8, 52), |
9980 | mask(IMM8, 53), |
9981 | mask(IMM8, 54), |
9982 | mask(IMM8, 55), |
9983 | mask(IMM8, 56), |
9984 | mask(IMM8, 57), |
9985 | mask(IMM8, 58), |
9986 | mask(IMM8, 59), |
9987 | mask(IMM8, 60), |
9988 | mask(IMM8, 61), |
9989 | mask(IMM8, 62), |
9990 | mask(IMM8, 63), |
9991 | ], |
9992 | ); |
9993 | transmute(r) |
9994 | } |
9995 | |
9996 | /// Shift 128-bit lanes in a right by imm8 bytes while shifting in zeros, and store the results in dst. |
9997 | /// |
9998 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_bsrli_epi128&expand=594) |
9999 | #[inline ] |
10000 | #[target_feature (enable = "avx512bw" )] |
10001 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10002 | #[cfg_attr (test, assert_instr(vpsrldq, IMM8 = 3))] |
10003 | #[rustc_legacy_const_generics (1)] |
10004 | pub unsafe fn _mm512_bsrli_epi128<const IMM8: i32>(a: __m512i) -> __m512i { |
10005 | static_assert_uimm_bits!(IMM8, 8); |
10006 | let a = a.as_i8x64(); |
10007 | let zero = _mm512_setzero_si512().as_i8x64(); |
10008 | let r: i8x64 = match IMM8 % 16 { |
10009 | 0 => simd_shuffle!( |
10010 | a, |
10011 | zero, |
10012 | [ |
10013 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, |
10014 | 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, |
10015 | 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, |
10016 | ], |
10017 | ), |
10018 | 1 => simd_shuffle!( |
10019 | a, |
10020 | zero, |
10021 | [ |
10022 | 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 17, 18, 19, 20, 21, 22, 23, |
10023 | 24, 25, 26, 27, 28, 29, 30, 31, 80, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, |
10024 | 45, 46, 47, 96, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, |
10025 | ], |
10026 | ), |
10027 | 2 => simd_shuffle!( |
10028 | a, |
10029 | zero, |
10030 | [ |
10031 | 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 18, 19, 20, 21, 22, 23, 24, |
10032 | 25, 26, 27, 28, 29, 30, 31, 80, 81, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, |
10033 | 46, 47, 96, 97, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, |
10034 | ], |
10035 | ), |
10036 | 3 => simd_shuffle!( |
10037 | a, |
10038 | zero, |
10039 | [ |
10040 | 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 19, 20, 21, 22, 23, 24, |
10041 | 25, 26, 27, 28, 29, 30, 31, 80, 81, 82, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, |
10042 | 46, 47, 96, 97, 98, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, |
10043 | 114, |
10044 | ], |
10045 | ), |
10046 | 4 => simd_shuffle!( |
10047 | a, |
10048 | zero, |
10049 | [ |
10050 | 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 20, 21, 22, 23, 24, 25, |
10051 | 26, 27, 28, 29, 30, 31, 80, 81, 82, 83, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, |
10052 | 47, 96, 97, 98, 99, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, 114, |
10053 | 115, |
10054 | ], |
10055 | ), |
10056 | 5 => simd_shuffle!( |
10057 | a, |
10058 | zero, |
10059 | [ |
10060 | 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 21, 22, 23, 24, 25, 26, |
10061 | 27, 28, 29, 30, 31, 80, 81, 82, 83, 84, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, |
10062 | 96, 97, 98, 99, 100, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, 114, |
10063 | 115, 116, |
10064 | ], |
10065 | ), |
10066 | 6 => simd_shuffle!( |
10067 | a, |
10068 | zero, |
10069 | [ |
10070 | 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 22, 23, 24, 25, 26, 27, |
10071 | 28, 29, 30, 31, 80, 81, 82, 83, 84, 85, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 96, |
10072 | 97, 98, 99, 100, 101, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, 114, 115, |
10073 | 116, 117, |
10074 | ], |
10075 | ), |
10076 | 7 => simd_shuffle!( |
10077 | a, |
10078 | zero, |
10079 | [ |
10080 | 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 23, 24, 25, 26, 27, |
10081 | 28, 29, 30, 31, 80, 81, 82, 83, 84, 85, 86, 39, 40, 41, 42, 43, 44, 45, 46, 47, 96, |
10082 | 97, 98, 99, 100, 101, 102, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, 114, 115, |
10083 | 116, 117, 118, |
10084 | ], |
10085 | ), |
10086 | 8 => simd_shuffle!( |
10087 | a, |
10088 | zero, |
10089 | [ |
10090 | 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 24, 25, 26, 27, 28, |
10091 | 29, 30, 31, 80, 81, 82, 83, 84, 85, 86, 87, 40, 41, 42, 43, 44, 45, 46, 47, 96, 97, |
10092 | 98, 99, 100, 101, 102, 103, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, 114, 115, |
10093 | 116, 117, 118, 119, |
10094 | ], |
10095 | ), |
10096 | 9 => simd_shuffle!( |
10097 | a, |
10098 | zero, |
10099 | [ |
10100 | 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 25, 26, 27, 28, 29, |
10101 | 30, 31, 80, 81, 82, 83, 84, 85, 86, 87, 88, 41, 42, 43, 44, 45, 46, 47, 96, 97, 98, |
10102 | 99, 100, 101, 102, 103, 104, 57, 58, 59, 60, 61, 62, 63, 112, 113, 114, 115, 116, |
10103 | 117, 118, 119, 120, |
10104 | ], |
10105 | ), |
10106 | 10 => simd_shuffle!( |
10107 | a, |
10108 | zero, |
10109 | [ |
10110 | 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 26, 27, 28, 29, 30, |
10111 | 31, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 42, 43, 44, 45, 46, 47, 96, 97, 98, 99, |
10112 | 100, 101, 102, 103, 104, 105, 58, 59, 60, 61, 62, 63, 112, 113, 114, 115, 116, 117, |
10113 | 118, 119, 120, 121, |
10114 | ], |
10115 | ), |
10116 | 11 => simd_shuffle!( |
10117 | a, |
10118 | zero, |
10119 | [ |
10120 | 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 27, 28, 29, 30, 31, |
10121 | 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 43, 44, 45, 46, 47, 96, 97, 98, 99, |
10122 | 100, 101, 102, 103, 104, 105, 106, 59, 60, 61, 62, 63, 112, 113, 114, 115, 116, |
10123 | 117, 118, 119, 120, 121, 122, |
10124 | ], |
10125 | ), |
10126 | 12 => simd_shuffle!( |
10127 | a, |
10128 | zero, |
10129 | [ |
10130 | 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 28, 29, 30, 31, 80, |
10131 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 44, 45, 46, 47, 96, 97, 98, 99, 100, |
10132 | 101, 102, 103, 104, 105, 106, 107, 60, 61, 62, 63, 112, 113, 114, 115, 116, 117, |
10133 | 118, 119, 120, 121, 122, 123, |
10134 | ], |
10135 | ), |
10136 | 13 => simd_shuffle!( |
10137 | a, |
10138 | zero, |
10139 | [ |
10140 | 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 29, 30, 31, 80, 81, |
10141 | 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 45, 46, 47, 96, 97, 98, 99, 100, 101, |
10142 | 102, 103, 104, 105, 106, 107, 108, 61, 62, 63, 112, 113, 114, 115, 116, 117, 118, |
10143 | 119, 120, 121, 122, 123, 124, |
10144 | ], |
10145 | ), |
10146 | 14 => simd_shuffle!( |
10147 | a, |
10148 | zero, |
10149 | [ |
10150 | 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 30, 31, 80, 81, 82, |
10151 | 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 46, 47, 96, 97, 98, 99, 100, 101, 102, |
10152 | 103, 104, 105, 106, 107, 108, 109, 62, 63, 112, 113, 114, 115, 116, 117, 118, 119, |
10153 | 120, 121, 122, 123, 124, 125, |
10154 | ], |
10155 | ), |
10156 | 15 => simd_shuffle!( |
10157 | a, |
10158 | zero, |
10159 | [ |
10160 | 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 31, 80, 81, 82, 83, |
10161 | 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 47, 96, 97, 98, 99, 100, 101, 102, 103, |
10162 | 104, 105, 106, 107, 108, 109, 110, 63, 112, 113, 114, 115, 116, 117, 118, 119, 120, |
10163 | 121, 122, 123, 124, 125, 126, |
10164 | ], |
10165 | ), |
10166 | _ => zero, |
10167 | }; |
10168 | transmute(r) |
10169 | } |
10170 | |
10171 | /// 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. |
10172 | /// |
10173 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_alignr_epi8&expand=263) |
10174 | #[inline ] |
10175 | #[target_feature (enable = "avx512bw" )] |
10176 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10177 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 1))] |
10178 | #[rustc_legacy_const_generics (2)] |
10179 | pub unsafe fn _mm512_alignr_epi8<const IMM8: i32>(a: __m512i, b: __m512i) -> __m512i { |
10180 | // If palignr is shifting the pair of vectors more than the size of two |
10181 | // lanes, emit zero. |
10182 | if IMM8 > 32 { |
10183 | return _mm512_set1_epi8(0); |
10184 | } |
10185 | // If palignr is shifting the pair of input vectors more than one lane, |
10186 | // but less than two lanes, convert to shifting in zeroes. |
10187 | let (a, b) = if IMM8 > 16 { |
10188 | (_mm512_set1_epi8(0), a) |
10189 | } else { |
10190 | (a, b) |
10191 | }; |
10192 | let a = a.as_i8x64(); |
10193 | let b = b.as_i8x64(); |
10194 | |
10195 | let r: i8x64 = match IMM8 % 16 { |
10196 | 0 => simd_shuffle!( |
10197 | b, |
10198 | a, |
10199 | [ |
10200 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, |
10201 | 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, |
10202 | 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, |
10203 | ], |
10204 | ), |
10205 | 1 => simd_shuffle!( |
10206 | b, |
10207 | a, |
10208 | [ |
10209 | 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 17, 18, 19, 20, 21, 22, 23, |
10210 | 24, 25, 26, 27, 28, 29, 30, 31, 80, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, |
10211 | 45, 46, 47, 96, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, |
10212 | ], |
10213 | ), |
10214 | 2 => simd_shuffle!( |
10215 | b, |
10216 | a, |
10217 | [ |
10218 | 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 18, 19, 20, 21, 22, 23, 24, |
10219 | 25, 26, 27, 28, 29, 30, 31, 80, 81, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, |
10220 | 46, 47, 96, 97, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, |
10221 | ], |
10222 | ), |
10223 | 3 => simd_shuffle!( |
10224 | b, |
10225 | a, |
10226 | [ |
10227 | 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 19, 20, 21, 22, 23, 24, |
10228 | 25, 26, 27, 28, 29, 30, 31, 80, 81, 82, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, |
10229 | 46, 47, 96, 97, 98, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, |
10230 | 114, |
10231 | ], |
10232 | ), |
10233 | 4 => simd_shuffle!( |
10234 | b, |
10235 | a, |
10236 | [ |
10237 | 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 20, 21, 22, 23, 24, 25, |
10238 | 26, 27, 28, 29, 30, 31, 80, 81, 82, 83, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, |
10239 | 47, 96, 97, 98, 99, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, 114, |
10240 | 115, |
10241 | ], |
10242 | ), |
10243 | 5 => simd_shuffle!( |
10244 | b, |
10245 | a, |
10246 | [ |
10247 | 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 21, 22, 23, 24, 25, 26, |
10248 | 27, 28, 29, 30, 31, 80, 81, 82, 83, 84, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, |
10249 | 96, 97, 98, 99, 100, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, 114, |
10250 | 115, 116, |
10251 | ], |
10252 | ), |
10253 | 6 => simd_shuffle!( |
10254 | b, |
10255 | a, |
10256 | [ |
10257 | 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 22, 23, 24, 25, 26, 27, |
10258 | 28, 29, 30, 31, 80, 81, 82, 83, 84, 85, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 96, |
10259 | 97, 98, 99, 100, 101, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, 114, 115, |
10260 | 116, 117, |
10261 | ], |
10262 | ), |
10263 | 7 => simd_shuffle!( |
10264 | b, |
10265 | a, |
10266 | [ |
10267 | 7, 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 23, 24, 25, 26, 27, |
10268 | 28, 29, 30, 31, 80, 81, 82, 83, 84, 85, 86, 39, 40, 41, 42, 43, 44, 45, 46, 47, 96, |
10269 | 97, 98, 99, 100, 101, 102, 55, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, 114, 115, |
10270 | 116, 117, 118, |
10271 | ], |
10272 | ), |
10273 | 8 => simd_shuffle!( |
10274 | b, |
10275 | a, |
10276 | [ |
10277 | 8, 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 24, 25, 26, 27, 28, |
10278 | 29, 30, 31, 80, 81, 82, 83, 84, 85, 86, 87, 40, 41, 42, 43, 44, 45, 46, 47, 96, 97, |
10279 | 98, 99, 100, 101, 102, 103, 56, 57, 58, 59, 60, 61, 62, 63, 112, 113, 114, 115, |
10280 | 116, 117, 118, 119, |
10281 | ], |
10282 | ), |
10283 | 9 => simd_shuffle!( |
10284 | b, |
10285 | a, |
10286 | [ |
10287 | 9, 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 25, 26, 27, 28, 29, |
10288 | 30, 31, 80, 81, 82, 83, 84, 85, 86, 87, 88, 41, 42, 43, 44, 45, 46, 47, 96, 97, 98, |
10289 | 99, 100, 101, 102, 103, 104, 57, 58, 59, 60, 61, 62, 63, 112, 113, 114, 115, 116, |
10290 | 117, 118, 119, 120, |
10291 | ], |
10292 | ), |
10293 | 10 => simd_shuffle!( |
10294 | b, |
10295 | a, |
10296 | [ |
10297 | 10, 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 26, 27, 28, 29, 30, |
10298 | 31, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 42, 43, 44, 45, 46, 47, 96, 97, 98, 99, |
10299 | 100, 101, 102, 103, 104, 105, 58, 59, 60, 61, 62, 63, 112, 113, 114, 115, 116, 117, |
10300 | 118, 119, 120, 121, |
10301 | ], |
10302 | ), |
10303 | 11 => simd_shuffle!( |
10304 | b, |
10305 | a, |
10306 | [ |
10307 | 11, 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 27, 28, 29, 30, 31, |
10308 | 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 43, 44, 45, 46, 47, 96, 97, 98, 99, |
10309 | 100, 101, 102, 103, 104, 105, 106, 59, 60, 61, 62, 63, 112, 113, 114, 115, 116, |
10310 | 117, 118, 119, 120, 121, 122, |
10311 | ], |
10312 | ), |
10313 | 12 => simd_shuffle!( |
10314 | b, |
10315 | a, |
10316 | [ |
10317 | 12, 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 28, 29, 30, 31, 80, |
10318 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 44, 45, 46, 47, 96, 97, 98, 99, 100, |
10319 | 101, 102, 103, 104, 105, 106, 107, 60, 61, 62, 63, 112, 113, 114, 115, 116, 117, |
10320 | 118, 119, 120, 121, 122, 123, |
10321 | ], |
10322 | ), |
10323 | 13 => simd_shuffle!( |
10324 | b, |
10325 | a, |
10326 | [ |
10327 | 13, 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 29, 30, 31, 80, 81, |
10328 | 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 45, 46, 47, 96, 97, 98, 99, 100, 101, |
10329 | 102, 103, 104, 105, 106, 107, 108, 61, 62, 63, 112, 113, 114, 115, 116, 117, 118, |
10330 | 119, 120, 121, 122, 123, 124, |
10331 | ], |
10332 | ), |
10333 | 14 => simd_shuffle!( |
10334 | b, |
10335 | a, |
10336 | [ |
10337 | 14, 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 30, 31, 80, 81, 82, |
10338 | 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 46, 47, 96, 97, 98, 99, 100, 101, 102, |
10339 | 103, 104, 105, 106, 107, 108, 109, 62, 63, 112, 113, 114, 115, 116, 117, 118, 119, |
10340 | 120, 121, 122, 123, 124, 125, |
10341 | ], |
10342 | ), |
10343 | 15 => simd_shuffle!( |
10344 | b, |
10345 | a, |
10346 | [ |
10347 | 15, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 31, 80, 81, 82, 83, |
10348 | 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 47, 96, 97, 98, 99, 100, 101, 102, 103, |
10349 | 104, 105, 106, 107, 108, 109, 110, 63, 112, 113, 114, 115, 116, 117, 118, 119, 120, |
10350 | 121, 122, 123, 124, 125, 126, |
10351 | ], |
10352 | ), |
10353 | _ => b, |
10354 | }; |
10355 | transmute(r) |
10356 | } |
10357 | |
10358 | /// 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). |
10359 | /// |
10360 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_alignr_epi8&expand=264) |
10361 | #[inline ] |
10362 | #[target_feature (enable = "avx512bw" )] |
10363 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10364 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 1))] |
10365 | #[rustc_legacy_const_generics (4)] |
10366 | pub unsafe fn _mm512_mask_alignr_epi8<const IMM8: i32>( |
10367 | src: __m512i, |
10368 | k: __mmask64, |
10369 | a: __m512i, |
10370 | b: __m512i, |
10371 | ) -> __m512i { |
10372 | static_assert_uimm_bits!(IMM8, 8); |
10373 | let r: __m512i = _mm512_alignr_epi8::<IMM8>(a, b); |
10374 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i8x64(), no:src.as_i8x64())) |
10375 | } |
10376 | |
10377 | /// 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). |
10378 | /// |
10379 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_maskz_alignr_epi8&expand=265) |
10380 | #[inline ] |
10381 | #[target_feature (enable = "avx512bw" )] |
10382 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10383 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 1))] |
10384 | #[rustc_legacy_const_generics (3)] |
10385 | pub unsafe fn _mm512_maskz_alignr_epi8<const IMM8: i32>( |
10386 | k: __mmask64, |
10387 | a: __m512i, |
10388 | b: __m512i, |
10389 | ) -> __m512i { |
10390 | static_assert_uimm_bits!(IMM8, 8); |
10391 | let r: __m512i = _mm512_alignr_epi8::<IMM8>(a, b); |
10392 | let zero: i8x64 = _mm512_setzero_si512().as_i8x64(); |
10393 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i8x64(), no:zero)) |
10394 | } |
10395 | |
10396 | /// 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). |
10397 | /// |
10398 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_alignr_epi8&expand=261) |
10399 | #[inline ] |
10400 | #[target_feature (enable = "avx512bw,avx512vl" )] |
10401 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10402 | #[rustc_legacy_const_generics (4)] |
10403 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 5))] |
10404 | pub unsafe fn _mm256_mask_alignr_epi8<const IMM8: i32>( |
10405 | src: __m256i, |
10406 | k: __mmask32, |
10407 | a: __m256i, |
10408 | b: __m256i, |
10409 | ) -> __m256i { |
10410 | static_assert_uimm_bits!(IMM8, 8); |
10411 | let r: __m256i = _mm256_alignr_epi8::<IMM8>(a, b); |
10412 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i8x32(), no:src.as_i8x32())) |
10413 | } |
10414 | |
10415 | /// 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). |
10416 | /// |
10417 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_maskz_alignr_epi8&expand=262) |
10418 | #[inline ] |
10419 | #[target_feature (enable = "avx512bw,avx512vl" )] |
10420 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10421 | #[rustc_legacy_const_generics (3)] |
10422 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 5))] |
10423 | pub unsafe fn _mm256_maskz_alignr_epi8<const IMM8: i32>( |
10424 | k: __mmask32, |
10425 | a: __m256i, |
10426 | b: __m256i, |
10427 | ) -> __m256i { |
10428 | static_assert_uimm_bits!(IMM8, 8); |
10429 | let r: __m256i = _mm256_alignr_epi8::<IMM8>(a, b); |
10430 | transmute(src:simd_select_bitmask( |
10431 | m:k, |
10432 | yes:r.as_i8x32(), |
10433 | no:_mm256_setzero_si256().as_i8x32(), |
10434 | )) |
10435 | } |
10436 | |
10437 | /// 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). |
10438 | /// |
10439 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_alignr_epi8&expand=258) |
10440 | #[inline ] |
10441 | #[target_feature (enable = "avx512bw,avx512vl" )] |
10442 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10443 | #[rustc_legacy_const_generics (4)] |
10444 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 5))] |
10445 | pub unsafe fn _mm_mask_alignr_epi8<const IMM8: i32>( |
10446 | src: __m128i, |
10447 | k: __mmask16, |
10448 | a: __m128i, |
10449 | b: __m128i, |
10450 | ) -> __m128i { |
10451 | static_assert_uimm_bits!(IMM8, 8); |
10452 | let r: __m128i = _mm_alignr_epi8::<IMM8>(a, b); |
10453 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i8x16(), no:src.as_i8x16())) |
10454 | } |
10455 | |
10456 | /// 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). |
10457 | /// |
10458 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_maskz_alignr_epi8&expand=259) |
10459 | #[inline ] |
10460 | #[target_feature (enable = "avx512bw,avx512vl" )] |
10461 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10462 | #[rustc_legacy_const_generics (3)] |
10463 | #[cfg_attr (test, assert_instr(vpalignr, IMM8 = 5))] |
10464 | pub unsafe fn _mm_maskz_alignr_epi8<const IMM8: i32>( |
10465 | k: __mmask16, |
10466 | a: __m128i, |
10467 | b: __m128i, |
10468 | ) -> __m128i { |
10469 | static_assert_uimm_bits!(IMM8, 8); |
10470 | let r: __m128i = _mm_alignr_epi8::<IMM8>(a, b); |
10471 | let zero: i8x16 = _mm_setzero_si128().as_i8x16(); |
10472 | transmute(src:simd_select_bitmask(m:k, yes:r.as_i8x16(), no:zero)) |
10473 | } |
10474 | |
10475 | /// 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. |
10476 | /// |
10477 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtsepi16_storeu_epi8&expand=1812) |
10478 | #[inline ] |
10479 | #[target_feature (enable = "avx512bw" )] |
10480 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10481 | #[cfg_attr (test, assert_instr(vpmovswb))] |
10482 | pub unsafe fn _mm512_mask_cvtsepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask32, a: __m512i) { |
10483 | vpmovswbmem(mem_addr, a:a.as_i16x32(), mask:k); |
10484 | } |
10485 | |
10486 | /// 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. |
10487 | /// |
10488 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtsepi16_storeu_epi8&expand=1811) |
10489 | #[inline ] |
10490 | #[target_feature (enable = "avx512bw,avx512vl" )] |
10491 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10492 | #[cfg_attr (test, assert_instr(vpmovswb))] |
10493 | pub unsafe fn _mm256_mask_cvtsepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask16, a: __m256i) { |
10494 | vpmovswbmem256(mem_addr, a:a.as_i16x16(), mask:k); |
10495 | } |
10496 | |
10497 | /// 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. |
10498 | /// |
10499 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtsepi16_storeu_epi8&expand=1810) |
10500 | #[inline ] |
10501 | #[target_feature (enable = "avx512bw,avx512vl" )] |
10502 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10503 | #[cfg_attr (test, assert_instr(vpmovswb))] |
10504 | pub unsafe fn _mm_mask_cvtsepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask8, a: __m128i) { |
10505 | vpmovswbmem128(mem_addr, a:a.as_i16x8(), mask:k); |
10506 | } |
10507 | |
10508 | /// 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. |
10509 | /// |
10510 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtepi16_storeu_epi8&expand=1412) |
10511 | #[inline ] |
10512 | #[target_feature (enable = "avx512bw" )] |
10513 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10514 | #[cfg_attr (test, assert_instr(vpmovwb))] |
10515 | pub unsafe fn _mm512_mask_cvtepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask32, a: __m512i) { |
10516 | vpmovwbmem(mem_addr, a:a.as_i16x32(), mask:k); |
10517 | } |
10518 | |
10519 | /// 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. |
10520 | /// |
10521 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtepi16_storeu_epi8&expand=1411) |
10522 | #[inline ] |
10523 | #[target_feature (enable = "avx512bw,avx512vl" )] |
10524 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10525 | #[cfg_attr (test, assert_instr(vpmovwb))] |
10526 | pub unsafe fn _mm256_mask_cvtepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask16, a: __m256i) { |
10527 | vpmovwbmem256(mem_addr, a:a.as_i16x16(), mask:k); |
10528 | } |
10529 | |
10530 | /// 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. |
10531 | /// |
10532 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtepi16_storeu_epi8&expand=1410) |
10533 | #[inline ] |
10534 | #[target_feature (enable = "avx512bw,avx512vl" )] |
10535 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10536 | #[cfg_attr (test, assert_instr(vpmovwb))] |
10537 | pub unsafe fn _mm_mask_cvtepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask8, a: __m128i) { |
10538 | vpmovwbmem128(mem_addr, a:a.as_i16x8(), mask:k); |
10539 | } |
10540 | |
10541 | /// 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. |
10542 | /// |
10543 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm512_mask_cvtusepi16_storeu_epi8&expand=2047) |
10544 | #[inline ] |
10545 | #[target_feature (enable = "avx512bw" )] |
10546 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10547 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
10548 | pub unsafe fn _mm512_mask_cvtusepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask32, a: __m512i) { |
10549 | vpmovuswbmem(mem_addr, a:a.as_i16x32(), mask:k); |
10550 | } |
10551 | |
10552 | /// 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. |
10553 | /// |
10554 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_mask_cvtusepi16_storeu_epi8&expand=2046) |
10555 | #[inline ] |
10556 | #[target_feature (enable = "avx512bw,avx512vl" )] |
10557 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10558 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
10559 | pub unsafe fn _mm256_mask_cvtusepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask16, a: __m256i) { |
10560 | vpmovuswbmem256(mem_addr, a:a.as_i16x16(), mask:k); |
10561 | } |
10562 | |
10563 | /// 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. |
10564 | /// |
10565 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mask_cvtusepi16_storeu_epi8&expand=2045) |
10566 | #[inline ] |
10567 | #[target_feature (enable = "avx512bw,avx512vl" )] |
10568 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
10569 | #[cfg_attr (test, assert_instr(vpmovuswb))] |
10570 | pub unsafe fn _mm_mask_cvtusepi16_storeu_epi8(mem_addr: *mut i8, k: __mmask8, a: __m128i) { |
10571 | vpmovuswbmem128(mem_addr, a:a.as_i16x8(), mask:k); |
10572 | } |
10573 | |
10574 | #[allow (improper_ctypes)] |
10575 | extern "C" { |
10576 | #[link_name = "llvm.x86.avx512.mask.paddus.w.512" ] |
10577 | fn vpaddusw(a: u16x32, b: u16x32, src: u16x32, mask: u32) -> u16x32; |
10578 | #[link_name = "llvm.x86.avx512.mask.paddus.w.256" ] |
10579 | fn vpaddusw256(a: u16x16, b: u16x16, src: u16x16, mask: u16) -> u16x16; |
10580 | #[link_name = "llvm.x86.avx512.mask.paddus.w.128" ] |
10581 | fn vpaddusw128(a: u16x8, b: u16x8, src: u16x8, mask: u8) -> u16x8; |
10582 | |
10583 | #[link_name = "llvm.x86.avx512.mask.paddus.b.512" ] |
10584 | fn vpaddusb(a: u8x64, b: u8x64, src: u8x64, mask: u64) -> u8x64; |
10585 | #[link_name = "llvm.x86.avx512.mask.paddus.b.256" ] |
10586 | fn vpaddusb256(a: u8x32, b: u8x32, src: u8x32, mask: u32) -> u8x32; |
10587 | #[link_name = "llvm.x86.avx512.mask.paddus.b.128" ] |
10588 | fn vpaddusb128(a: u8x16, b: u8x16, src: u8x16, mask: u16) -> u8x16; |
10589 | |
10590 | #[link_name = "llvm.x86.avx512.mask.padds.w.512" ] |
10591 | fn vpaddsw(a: i16x32, b: i16x32, src: i16x32, mask: u32) -> i16x32; |
10592 | #[link_name = "llvm.x86.avx512.mask.padds.w.256" ] |
10593 | fn vpaddsw256(a: i16x16, b: i16x16, src: i16x16, mask: u16) -> i16x16; |
10594 | #[link_name = "llvm.x86.avx512.mask.padds.w.128" ] |
10595 | fn vpaddsw128(a: i16x8, b: i16x8, src: i16x8, mask: u8) -> i16x8; |
10596 | |
10597 | #[link_name = "llvm.x86.avx512.mask.padds.b.512" ] |
10598 | fn vpaddsb(a: i8x64, b: i8x64, src: i8x64, mask: u64) -> i8x64; |
10599 | #[link_name = "llvm.x86.avx512.mask.padds.b.256" ] |
10600 | fn vpaddsb256(a: i8x32, b: i8x32, src: i8x32, mask: u32) -> i8x32; |
10601 | #[link_name = "llvm.x86.avx512.mask.padds.b.128" ] |
10602 | fn vpaddsb128(a: i8x16, b: i8x16, src: i8x16, mask: u16) -> i8x16; |
10603 | |
10604 | #[link_name = "llvm.x86.avx512.mask.psubus.w.512" ] |
10605 | fn vpsubusw(a: u16x32, b: u16x32, src: u16x32, mask: u32) -> u16x32; |
10606 | #[link_name = "llvm.x86.avx512.mask.psubus.w.256" ] |
10607 | fn vpsubusw256(a: u16x16, b: u16x16, src: u16x16, mask: u16) -> u16x16; |
10608 | #[link_name = "llvm.x86.avx512.mask.psubus.w.128" ] |
10609 | fn vpsubusw128(a: u16x8, b: u16x8, src: u16x8, mask: u8) -> u16x8; |
10610 | |
10611 | #[link_name = "llvm.x86.avx512.mask.psubus.b.512" ] |
10612 | fn vpsubusb(a: u8x64, b: u8x64, src: u8x64, mask: u64) -> u8x64; |
10613 | #[link_name = "llvm.x86.avx512.mask.psubus.b.256" ] |
10614 | fn vpsubusb256(a: u8x32, b: u8x32, src: u8x32, mask: u32) -> u8x32; |
10615 | #[link_name = "llvm.x86.avx512.mask.psubus.b.128" ] |
10616 | fn vpsubusb128(a: u8x16, b: u8x16, src: u8x16, mask: u16) -> u8x16; |
10617 | |
10618 | #[link_name = "llvm.x86.avx512.mask.psubs.w.512" ] |
10619 | fn vpsubsw(a: i16x32, b: i16x32, src: i16x32, mask: u32) -> i16x32; |
10620 | #[link_name = "llvm.x86.avx512.mask.psubs.w.256" ] |
10621 | fn vpsubsw256(a: i16x16, b: i16x16, src: i16x16, mask: u16) -> i16x16; |
10622 | #[link_name = "llvm.x86.avx512.mask.psubs.w.128" ] |
10623 | fn vpsubsw128(a: i16x8, b: i16x8, src: i16x8, mask: u8) -> i16x8; |
10624 | |
10625 | #[link_name = "llvm.x86.avx512.mask.psubs.b.512" ] |
10626 | fn vpsubsb(a: i8x64, b: i8x64, src: i8x64, mask: u64) -> i8x64; |
10627 | #[link_name = "llvm.x86.avx512.mask.psubs.b.256" ] |
10628 | fn vpsubsb256(a: i8x32, b: i8x32, src: i8x32, mask: u32) -> i8x32; |
10629 | #[link_name = "llvm.x86.avx512.mask.psubs.b.128" ] |
10630 | fn vpsubsb128(a: i8x16, b: i8x16, src: i8x16, mask: u16) -> i8x16; |
10631 | |
10632 | #[link_name = "llvm.x86.avx512.pmulhu.w.512" ] |
10633 | fn vpmulhuw(a: u16x32, b: u16x32) -> u16x32; |
10634 | #[link_name = "llvm.x86.avx512.pmulh.w.512" ] |
10635 | fn vpmulhw(a: i16x32, b: i16x32) -> i16x32; |
10636 | #[link_name = "llvm.x86.avx512.pmul.hr.sw.512" ] |
10637 | fn vpmulhrsw(a: i16x32, b: i16x32) -> i16x32; |
10638 | |
10639 | #[link_name = "llvm.x86.avx512.mask.ucmp.w.512" ] |
10640 | fn vpcmpuw(a: u16x32, b: u16x32, op: i32, mask: u32) -> u32; |
10641 | #[link_name = "llvm.x86.avx512.mask.ucmp.w.256" ] |
10642 | fn vpcmpuw256(a: u16x16, b: u16x16, op: i32, mask: u16) -> u16; |
10643 | #[link_name = "llvm.x86.avx512.mask.ucmp.w.128" ] |
10644 | fn vpcmpuw128(a: u16x8, b: u16x8, op: i32, mask: u8) -> u8; |
10645 | |
10646 | #[link_name = "llvm.x86.avx512.mask.ucmp.b.512" ] |
10647 | fn vpcmpub(a: u8x64, b: u8x64, op: i32, mask: u64) -> u64; |
10648 | #[link_name = "llvm.x86.avx512.mask.ucmp.b.256" ] |
10649 | fn vpcmpub256(a: u8x32, b: u8x32, op: i32, mask: u32) -> u32; |
10650 | #[link_name = "llvm.x86.avx512.mask.ucmp.b.128" ] |
10651 | fn vpcmpub128(a: u8x16, b: u8x16, op: i32, mask: u16) -> u16; |
10652 | |
10653 | #[link_name = "llvm.x86.avx512.mask.cmp.w.512" ] |
10654 | fn vpcmpw(a: i16x32, b: i16x32, op: i32, mask: u32) -> u32; |
10655 | #[link_name = "llvm.x86.avx512.mask.cmp.w.256" ] |
10656 | fn vpcmpw256(a: i16x16, b: i16x16, op: i32, mask: u16) -> u16; |
10657 | #[link_name = "llvm.x86.avx512.mask.cmp.w.128" ] |
10658 | fn vpcmpw128(a: i16x8, b: i16x8, op: i32, mask: u8) -> u8; |
10659 | |
10660 | #[link_name = "llvm.x86.avx512.mask.cmp.b.512" ] |
10661 | fn vpcmpb(a: i8x64, b: i8x64, op: i32, mask: u64) -> u64; |
10662 | #[link_name = "llvm.x86.avx512.mask.cmp.b.256" ] |
10663 | fn vpcmpb256(a: i8x32, b: i8x32, op: i32, mask: u32) -> u32; |
10664 | #[link_name = "llvm.x86.avx512.mask.cmp.b.128" ] |
10665 | fn vpcmpb128(a: i8x16, b: i8x16, op: i32, mask: u16) -> u16; |
10666 | |
10667 | #[link_name = "llvm.x86.avx512.mask.pmaxu.w.512" ] |
10668 | fn vpmaxuw(a: u16x32, b: u16x32) -> u16x32; |
10669 | #[link_name = "llvm.x86.avx512.mask.pmaxu.b.512" ] |
10670 | fn vpmaxub(a: u8x64, b: u8x64) -> u8x64; |
10671 | #[link_name = "llvm.x86.avx512.mask.pmaxs.w.512" ] |
10672 | fn vpmaxsw(a: i16x32, b: i16x32) -> i16x32; |
10673 | #[link_name = "llvm.x86.avx512.mask.pmaxs.b.512" ] |
10674 | fn vpmaxsb(a: i8x64, b: i8x64) -> i8x64; |
10675 | |
10676 | #[link_name = "llvm.x86.avx512.mask.pminu.w.512" ] |
10677 | fn vpminuw(a: u16x32, b: u16x32) -> u16x32; |
10678 | #[link_name = "llvm.x86.avx512.mask.pminu.b.512" ] |
10679 | fn vpminub(a: u8x64, b: u8x64) -> u8x64; |
10680 | #[link_name = "llvm.x86.avx512.mask.pmins.w.512" ] |
10681 | fn vpminsw(a: i16x32, b: i16x32) -> i16x32; |
10682 | #[link_name = "llvm.x86.avx512.mask.pmins.b.512" ] |
10683 | fn vpminsb(a: i8x64, b: i8x64) -> i8x64; |
10684 | |
10685 | #[link_name = "llvm.x86.avx512.pmaddw.d.512" ] |
10686 | fn vpmaddwd(a: i16x32, b: i16x32) -> i32x16; |
10687 | #[link_name = "llvm.x86.avx512.pmaddubs.w.512" ] |
10688 | fn vpmaddubsw(a: i8x64, b: i8x64) -> i16x32; |
10689 | |
10690 | #[link_name = "llvm.x86.avx512.packssdw.512" ] |
10691 | fn vpackssdw(a: i32x16, b: i32x16) -> i16x32; |
10692 | #[link_name = "llvm.x86.avx512.packsswb.512" ] |
10693 | fn vpacksswb(a: i16x32, b: i16x32) -> i8x64; |
10694 | #[link_name = "llvm.x86.avx512.packusdw.512" ] |
10695 | fn vpackusdw(a: i32x16, b: i32x16) -> u16x32; |
10696 | #[link_name = "llvm.x86.avx512.packuswb.512" ] |
10697 | fn vpackuswb(a: i16x32, b: i16x32) -> u8x64; |
10698 | |
10699 | #[link_name = "llvm.x86.avx512.pavg.w.512" ] |
10700 | fn vpavgw(a: u16x32, b: u16x32) -> u16x32; |
10701 | #[link_name = "llvm.x86.avx512.pavg.b.512" ] |
10702 | fn vpavgb(a: u8x64, b: u8x64) -> u8x64; |
10703 | |
10704 | #[link_name = "llvm.x86.avx512.psll.w.512" ] |
10705 | fn vpsllw(a: i16x32, count: i16x8) -> i16x32; |
10706 | |
10707 | #[link_name = "llvm.x86.avx512.psllv.w.512" ] |
10708 | fn vpsllvw(a: i16x32, b: i16x32) -> i16x32; |
10709 | #[link_name = "llvm.x86.avx512.psllv.w.256" ] |
10710 | fn vpsllvw256(a: i16x16, b: i16x16) -> i16x16; |
10711 | #[link_name = "llvm.x86.avx512.psllv.w.128" ] |
10712 | fn vpsllvw128(a: i16x8, b: i16x8) -> i16x8; |
10713 | |
10714 | #[link_name = "llvm.x86.avx512.psrl.w.512" ] |
10715 | fn vpsrlw(a: i16x32, count: i16x8) -> i16x32; |
10716 | |
10717 | #[link_name = "llvm.x86.avx512.psrlv.w.512" ] |
10718 | fn vpsrlvw(a: i16x32, b: i16x32) -> i16x32; |
10719 | #[link_name = "llvm.x86.avx512.psrlv.w.256" ] |
10720 | fn vpsrlvw256(a: i16x16, b: i16x16) -> i16x16; |
10721 | #[link_name = "llvm.x86.avx512.psrlv.w.128" ] |
10722 | fn vpsrlvw128(a: i16x8, b: i16x8) -> i16x8; |
10723 | |
10724 | #[link_name = "llvm.x86.avx512.psra.w.512" ] |
10725 | fn vpsraw(a: i16x32, count: i16x8) -> i16x32; |
10726 | |
10727 | #[link_name = "llvm.x86.avx512.psrav.w.512" ] |
10728 | fn vpsravw(a: i16x32, count: i16x32) -> i16x32; |
10729 | #[link_name = "llvm.x86.avx512.psrav.w.256" ] |
10730 | fn vpsravw256(a: i16x16, count: i16x16) -> i16x16; |
10731 | #[link_name = "llvm.x86.avx512.psrav.w.128" ] |
10732 | fn vpsravw128(a: i16x8, count: i16x8) -> i16x8; |
10733 | |
10734 | #[link_name = "llvm.x86.avx512.vpermi2var.hi.512" ] |
10735 | fn vpermi2w(a: i16x32, idx: i16x32, b: i16x32) -> i16x32; |
10736 | #[link_name = "llvm.x86.avx512.vpermi2var.hi.256" ] |
10737 | fn vpermi2w256(a: i16x16, idx: i16x16, b: i16x16) -> i16x16; |
10738 | #[link_name = "llvm.x86.avx512.vpermi2var.hi.128" ] |
10739 | fn vpermi2w128(a: i16x8, idx: i16x8, b: i16x8) -> i16x8; |
10740 | |
10741 | #[link_name = "llvm.x86.avx512.permvar.hi.512" ] |
10742 | fn vpermw(a: i16x32, idx: i16x32) -> i16x32; |
10743 | #[link_name = "llvm.x86.avx512.permvar.hi.256" ] |
10744 | fn vpermw256(a: i16x16, idx: i16x16) -> i16x16; |
10745 | #[link_name = "llvm.x86.avx512.permvar.hi.128" ] |
10746 | fn vpermw128(a: i16x8, idx: i16x8) -> i16x8; |
10747 | |
10748 | #[link_name = "llvm.x86.avx512.pshuf.b.512" ] |
10749 | fn vpshufb(a: i8x64, b: i8x64) -> i8x64; |
10750 | |
10751 | #[link_name = "llvm.x86.avx512.psad.bw.512" ] |
10752 | fn vpsadbw(a: u8x64, b: u8x64) -> u64x8; |
10753 | |
10754 | #[link_name = "llvm.x86.avx512.dbpsadbw.512" ] |
10755 | fn vdbpsadbw(a: u8x64, b: u8x64, imm8: i32) -> u16x32; |
10756 | #[link_name = "llvm.x86.avx512.dbpsadbw.256" ] |
10757 | fn vdbpsadbw256(a: u8x32, b: u8x32, imm8: i32) -> u16x16; |
10758 | #[link_name = "llvm.x86.avx512.dbpsadbw.128" ] |
10759 | fn vdbpsadbw128(a: u8x16, b: u8x16, imm8: i32) -> u16x8; |
10760 | |
10761 | #[link_name = "llvm.x86.avx512.mask.pmovs.wb.512" ] |
10762 | fn vpmovswb(a: i16x32, src: i8x32, mask: u32) -> i8x32; |
10763 | #[link_name = "llvm.x86.avx512.mask.pmovs.wb.256" ] |
10764 | fn vpmovswb256(a: i16x16, src: i8x16, mask: u16) -> i8x16; |
10765 | #[link_name = "llvm.x86.avx512.mask.pmovs.wb.128" ] |
10766 | fn vpmovswb128(a: i16x8, src: i8x16, mask: u8) -> i8x16; |
10767 | |
10768 | #[link_name = "llvm.x86.avx512.mask.pmovus.wb.512" ] |
10769 | fn vpmovuswb(a: u16x32, src: u8x32, mask: u32) -> u8x32; |
10770 | #[link_name = "llvm.x86.avx512.mask.pmovus.wb.256" ] |
10771 | fn vpmovuswb256(a: u16x16, src: u8x16, mask: u16) -> u8x16; |
10772 | #[link_name = "llvm.x86.avx512.mask.pmovus.wb.128" ] |
10773 | fn vpmovuswb128(a: u16x8, src: u8x16, mask: u8) -> u8x16; |
10774 | |
10775 | #[link_name = "llvm.x86.avx512.mask.pmovs.wb.mem.512" ] |
10776 | fn vpmovswbmem(mem_addr: *mut i8, a: i16x32, mask: u32); |
10777 | #[link_name = "llvm.x86.avx512.mask.pmovs.wb.mem.256" ] |
10778 | fn vpmovswbmem256(mem_addr: *mut i8, a: i16x16, mask: u16); |
10779 | #[link_name = "llvm.x86.avx512.mask.pmovs.wb.mem.128" ] |
10780 | fn vpmovswbmem128(mem_addr: *mut i8, a: i16x8, mask: u8); |
10781 | |
10782 | #[link_name = "llvm.x86.avx512.mask.pmov.wb.mem.512" ] |
10783 | fn vpmovwbmem(mem_addr: *mut i8, a: i16x32, mask: u32); |
10784 | #[link_name = "llvm.x86.avx512.mask.pmov.wb.mem.256" ] |
10785 | fn vpmovwbmem256(mem_addr: *mut i8, a: i16x16, mask: u16); |
10786 | #[link_name = "llvm.x86.avx512.mask.pmov.wb.mem.128" ] |
10787 | fn vpmovwbmem128(mem_addr: *mut i8, a: i16x8, mask: u8); |
10788 | |
10789 | #[link_name = "llvm.x86.avx512.mask.pmovus.wb.mem.512" ] |
10790 | fn vpmovuswbmem(mem_addr: *mut i8, a: i16x32, mask: u32); |
10791 | #[link_name = "llvm.x86.avx512.mask.pmovus.wb.mem.256" ] |
10792 | fn vpmovuswbmem256(mem_addr: *mut i8, a: i16x16, mask: u16); |
10793 | #[link_name = "llvm.x86.avx512.mask.pmovus.wb.mem.128" ] |
10794 | fn vpmovuswbmem128(mem_addr: *mut i8, a: i16x8, mask: u8); |
10795 | } |
10796 | |
10797 | #[cfg (test)] |
10798 | mod tests { |
10799 | |
10800 | use stdarch_test::simd_test; |
10801 | |
10802 | use crate::core_arch::x86::*; |
10803 | use crate::hint::black_box; |
10804 | use crate::mem::{self}; |
10805 | |
10806 | #[simd_test(enable = "avx512bw" )] |
10807 | unsafe fn test_mm512_abs_epi16() { |
10808 | let a = _mm512_set1_epi16(-1); |
10809 | let r = _mm512_abs_epi16(a); |
10810 | let e = _mm512_set1_epi16(1); |
10811 | assert_eq_m512i(r, e); |
10812 | } |
10813 | |
10814 | #[simd_test(enable = "avx512bw" )] |
10815 | unsafe fn test_mm512_mask_abs_epi16() { |
10816 | let a = _mm512_set1_epi16(-1); |
10817 | let r = _mm512_mask_abs_epi16(a, 0, a); |
10818 | assert_eq_m512i(r, a); |
10819 | let r = _mm512_mask_abs_epi16(a, 0b00000000_11111111_00000000_11111111, a); |
10820 | #[rustfmt::skip] |
10821 | let e = _mm512_set_epi16(-1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1, |
10822 | -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1); |
10823 | assert_eq_m512i(r, e); |
10824 | } |
10825 | |
10826 | #[simd_test(enable = "avx512bw" )] |
10827 | unsafe fn test_mm512_maskz_abs_epi16() { |
10828 | let a = _mm512_set1_epi16(-1); |
10829 | let r = _mm512_maskz_abs_epi16(0, a); |
10830 | assert_eq_m512i(r, _mm512_setzero_si512()); |
10831 | let r = _mm512_maskz_abs_epi16(0b00000000_11111111_00000000_11111111, a); |
10832 | #[rustfmt::skip] |
10833 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, |
10834 | 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
10835 | assert_eq_m512i(r, e); |
10836 | } |
10837 | |
10838 | #[simd_test(enable = "avx512bw,avx512vl" )] |
10839 | unsafe fn test_mm256_mask_abs_epi16() { |
10840 | let a = _mm256_set1_epi16(-1); |
10841 | let r = _mm256_mask_abs_epi16(a, 0, a); |
10842 | assert_eq_m256i(r, a); |
10843 | let r = _mm256_mask_abs_epi16(a, 0b00000000_11111111, a); |
10844 | let e = _mm256_set_epi16(-1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1); |
10845 | assert_eq_m256i(r, e); |
10846 | } |
10847 | |
10848 | #[simd_test(enable = "avx512bw,avx512vl" )] |
10849 | unsafe fn test_mm256_maskz_abs_epi16() { |
10850 | let a = _mm256_set1_epi16(-1); |
10851 | let r = _mm256_maskz_abs_epi16(0, a); |
10852 | assert_eq_m256i(r, _mm256_setzero_si256()); |
10853 | let r = _mm256_maskz_abs_epi16(0b00000000_11111111, a); |
10854 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
10855 | assert_eq_m256i(r, e); |
10856 | } |
10857 | |
10858 | #[simd_test(enable = "avx512bw,avx512vl" )] |
10859 | unsafe fn test_mm_mask_abs_epi16() { |
10860 | let a = _mm_set1_epi16(-1); |
10861 | let r = _mm_mask_abs_epi16(a, 0, a); |
10862 | assert_eq_m128i(r, a); |
10863 | let r = _mm_mask_abs_epi16(a, 0b00001111, a); |
10864 | let e = _mm_set_epi16(-1, -1, -1, -1, 1, 1, 1, 1); |
10865 | assert_eq_m128i(r, e); |
10866 | } |
10867 | |
10868 | #[simd_test(enable = "avx512bw,avx512vl" )] |
10869 | unsafe fn test_mm_maskz_abs_epi16() { |
10870 | let a = _mm_set1_epi16(-1); |
10871 | let r = _mm_maskz_abs_epi16(0, a); |
10872 | assert_eq_m128i(r, _mm_setzero_si128()); |
10873 | let r = _mm_maskz_abs_epi16(0b00001111, a); |
10874 | let e = _mm_set_epi16(0, 0, 0, 0, 1, 1, 1, 1); |
10875 | assert_eq_m128i(r, e); |
10876 | } |
10877 | |
10878 | #[simd_test(enable = "avx512bw" )] |
10879 | unsafe fn test_mm512_abs_epi8() { |
10880 | let a = _mm512_set1_epi8(-1); |
10881 | let r = _mm512_abs_epi8(a); |
10882 | let e = _mm512_set1_epi8(1); |
10883 | assert_eq_m512i(r, e); |
10884 | } |
10885 | |
10886 | #[simd_test(enable = "avx512bw" )] |
10887 | unsafe fn test_mm512_mask_abs_epi8() { |
10888 | let a = _mm512_set1_epi8(-1); |
10889 | let r = _mm512_mask_abs_epi8(a, 0, a); |
10890 | assert_eq_m512i(r, a); |
10891 | let r = _mm512_mask_abs_epi8( |
10892 | a, |
10893 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
10894 | a, |
10895 | ); |
10896 | #[rustfmt::skip] |
10897 | let e = _mm512_set_epi8(-1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1, |
10898 | -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1, |
10899 | -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1, |
10900 | -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1); |
10901 | assert_eq_m512i(r, e); |
10902 | } |
10903 | |
10904 | #[simd_test(enable = "avx512bw" )] |
10905 | unsafe fn test_mm512_maskz_abs_epi8() { |
10906 | let a = _mm512_set1_epi8(-1); |
10907 | let r = _mm512_maskz_abs_epi8(0, a); |
10908 | assert_eq_m512i(r, _mm512_setzero_si512()); |
10909 | let r = _mm512_maskz_abs_epi8( |
10910 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
10911 | a, |
10912 | ); |
10913 | #[rustfmt::skip] |
10914 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, |
10915 | 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, |
10916 | 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, |
10917 | 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
10918 | assert_eq_m512i(r, e); |
10919 | } |
10920 | |
10921 | #[simd_test(enable = "avx512bw,avx512vl" )] |
10922 | unsafe fn test_mm256_mask_abs_epi8() { |
10923 | let a = _mm256_set1_epi8(-1); |
10924 | let r = _mm256_mask_abs_epi8(a, 0, a); |
10925 | assert_eq_m256i(r, a); |
10926 | let r = _mm256_mask_abs_epi8(a, 0b00000000_11111111_00000000_11111111, a); |
10927 | #[rustfmt::skip] |
10928 | let e = _mm256_set_epi8(-1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1, |
10929 | -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1); |
10930 | assert_eq_m256i(r, e); |
10931 | } |
10932 | |
10933 | #[simd_test(enable = "avx512bw,avx512vl" )] |
10934 | unsafe fn test_mm256_maskz_abs_epi8() { |
10935 | let a = _mm256_set1_epi8(-1); |
10936 | let r = _mm256_maskz_abs_epi8(0, a); |
10937 | assert_eq_m256i(r, _mm256_setzero_si256()); |
10938 | let r = _mm256_maskz_abs_epi8(0b00000000_11111111_00000000_11111111, a); |
10939 | #[rustfmt::skip] |
10940 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, |
10941 | 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
10942 | assert_eq_m256i(r, e); |
10943 | } |
10944 | |
10945 | #[simd_test(enable = "avx512bw,avx512vl" )] |
10946 | unsafe fn test_mm_mask_abs_epi8() { |
10947 | let a = _mm_set1_epi8(-1); |
10948 | let r = _mm_mask_abs_epi8(a, 0, a); |
10949 | assert_eq_m128i(r, a); |
10950 | let r = _mm_mask_abs_epi8(a, 0b00000000_11111111, a); |
10951 | let e = _mm_set_epi8(-1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1); |
10952 | assert_eq_m128i(r, e); |
10953 | } |
10954 | |
10955 | #[simd_test(enable = "avx512bw,avx512vl" )] |
10956 | unsafe fn test_mm_maskz_abs_epi8() { |
10957 | let a = _mm_set1_epi8(-1); |
10958 | let r = _mm_maskz_abs_epi8(0, a); |
10959 | assert_eq_m128i(r, _mm_setzero_si128()); |
10960 | let r = _mm_maskz_abs_epi8(0b00000000_11111111, a); |
10961 | #[rustfmt::skip] |
10962 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
10963 | assert_eq_m128i(r, e); |
10964 | } |
10965 | |
10966 | #[simd_test(enable = "avx512bw" )] |
10967 | unsafe fn test_mm512_add_epi16() { |
10968 | let a = _mm512_set1_epi16(1); |
10969 | let b = _mm512_set1_epi16(2); |
10970 | let r = _mm512_add_epi16(a, b); |
10971 | let e = _mm512_set1_epi16(3); |
10972 | assert_eq_m512i(r, e); |
10973 | } |
10974 | |
10975 | #[simd_test(enable = "avx512bw" )] |
10976 | unsafe fn test_mm512_mask_add_epi16() { |
10977 | let a = _mm512_set1_epi16(1); |
10978 | let b = _mm512_set1_epi16(2); |
10979 | let r = _mm512_mask_add_epi16(a, 0, a, b); |
10980 | assert_eq_m512i(r, a); |
10981 | let r = _mm512_mask_add_epi16(a, 0b00000000_11111111_00000000_11111111, a, b); |
10982 | #[rustfmt::skip] |
10983 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, |
10984 | 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3); |
10985 | assert_eq_m512i(r, e); |
10986 | } |
10987 | |
10988 | #[simd_test(enable = "avx512bw" )] |
10989 | unsafe fn test_mm512_maskz_add_epi16() { |
10990 | let a = _mm512_set1_epi16(1); |
10991 | let b = _mm512_set1_epi16(2); |
10992 | let r = _mm512_maskz_add_epi16(0, a, b); |
10993 | assert_eq_m512i(r, _mm512_setzero_si512()); |
10994 | let r = _mm512_maskz_add_epi16(0b00000000_11111111_00000000_11111111, a, b); |
10995 | #[rustfmt::skip] |
10996 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3, |
10997 | 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3); |
10998 | assert_eq_m512i(r, e); |
10999 | } |
11000 | |
11001 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11002 | unsafe fn test_mm256_mask_add_epi16() { |
11003 | let a = _mm256_set1_epi16(1); |
11004 | let b = _mm256_set1_epi16(2); |
11005 | let r = _mm256_mask_add_epi16(a, 0, a, b); |
11006 | assert_eq_m256i(r, a); |
11007 | let r = _mm256_mask_add_epi16(a, 0b00000000_11111111, a, b); |
11008 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3); |
11009 | assert_eq_m256i(r, e); |
11010 | } |
11011 | |
11012 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11013 | unsafe fn test_mm256_maskz_add_epi16() { |
11014 | let a = _mm256_set1_epi16(1); |
11015 | let b = _mm256_set1_epi16(2); |
11016 | let r = _mm256_maskz_add_epi16(0, a, b); |
11017 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11018 | let r = _mm256_maskz_add_epi16(0b00000000_11111111, a, b); |
11019 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3); |
11020 | assert_eq_m256i(r, e); |
11021 | } |
11022 | |
11023 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11024 | unsafe fn test_mm_mask_add_epi16() { |
11025 | let a = _mm_set1_epi16(1); |
11026 | let b = _mm_set1_epi16(2); |
11027 | let r = _mm_mask_add_epi16(a, 0, a, b); |
11028 | assert_eq_m128i(r, a); |
11029 | let r = _mm_mask_add_epi16(a, 0b00001111, a, b); |
11030 | let e = _mm_set_epi16(1, 1, 1, 1, 3, 3, 3, 3); |
11031 | assert_eq_m128i(r, e); |
11032 | } |
11033 | |
11034 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11035 | unsafe fn test_mm_maskz_add_epi16() { |
11036 | let a = _mm_set1_epi16(1); |
11037 | let b = _mm_set1_epi16(2); |
11038 | let r = _mm_maskz_add_epi16(0, a, b); |
11039 | assert_eq_m128i(r, _mm_setzero_si128()); |
11040 | let r = _mm_maskz_add_epi16(0b00001111, a, b); |
11041 | let e = _mm_set_epi16(0, 0, 0, 0, 3, 3, 3, 3); |
11042 | assert_eq_m128i(r, e); |
11043 | } |
11044 | |
11045 | #[simd_test(enable = "avx512bw" )] |
11046 | unsafe fn test_mm512_add_epi8() { |
11047 | let a = _mm512_set1_epi8(1); |
11048 | let b = _mm512_set1_epi8(2); |
11049 | let r = _mm512_add_epi8(a, b); |
11050 | let e = _mm512_set1_epi8(3); |
11051 | assert_eq_m512i(r, e); |
11052 | } |
11053 | |
11054 | #[simd_test(enable = "avx512bw" )] |
11055 | unsafe fn test_mm512_mask_add_epi8() { |
11056 | let a = _mm512_set1_epi8(1); |
11057 | let b = _mm512_set1_epi8(2); |
11058 | let r = _mm512_mask_add_epi8(a, 0, a, b); |
11059 | assert_eq_m512i(r, a); |
11060 | let r = _mm512_mask_add_epi8( |
11061 | a, |
11062 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
11063 | a, |
11064 | b, |
11065 | ); |
11066 | #[rustfmt::skip] |
11067 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, |
11068 | 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, |
11069 | 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, |
11070 | 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3); |
11071 | assert_eq_m512i(r, e); |
11072 | } |
11073 | |
11074 | #[simd_test(enable = "avx512bw" )] |
11075 | unsafe fn test_mm512_maskz_add_epi8() { |
11076 | let a = _mm512_set1_epi8(1); |
11077 | let b = _mm512_set1_epi8(2); |
11078 | let r = _mm512_maskz_add_epi8(0, a, b); |
11079 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11080 | let r = _mm512_maskz_add_epi8( |
11081 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
11082 | a, |
11083 | b, |
11084 | ); |
11085 | #[rustfmt::skip] |
11086 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3, |
11087 | 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3, |
11088 | 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3, |
11089 | 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3); |
11090 | assert_eq_m512i(r, e); |
11091 | } |
11092 | |
11093 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11094 | unsafe fn test_mm256_mask_add_epi8() { |
11095 | let a = _mm256_set1_epi8(1); |
11096 | let b = _mm256_set1_epi8(2); |
11097 | let r = _mm256_mask_add_epi8(a, 0, a, b); |
11098 | assert_eq_m256i(r, a); |
11099 | let r = _mm256_mask_add_epi8(a, 0b00000000_11111111_00000000_11111111, a, b); |
11100 | #[rustfmt::skip] |
11101 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, |
11102 | 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3); |
11103 | assert_eq_m256i(r, e); |
11104 | } |
11105 | |
11106 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11107 | unsafe fn test_mm256_maskz_add_epi8() { |
11108 | let a = _mm256_set1_epi8(1); |
11109 | let b = _mm256_set1_epi8(2); |
11110 | let r = _mm256_maskz_add_epi8(0, a, b); |
11111 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11112 | let r = _mm256_maskz_add_epi8(0b00000000_11111111_00000000_11111111, a, b); |
11113 | #[rustfmt::skip] |
11114 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3, |
11115 | 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3); |
11116 | assert_eq_m256i(r, e); |
11117 | } |
11118 | |
11119 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11120 | unsafe fn test_mm_mask_add_epi8() { |
11121 | let a = _mm_set1_epi8(1); |
11122 | let b = _mm_set1_epi8(2); |
11123 | let r = _mm_mask_add_epi8(a, 0, a, b); |
11124 | assert_eq_m128i(r, a); |
11125 | let r = _mm_mask_add_epi8(a, 0b00000000_11111111, a, b); |
11126 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3); |
11127 | assert_eq_m128i(r, e); |
11128 | } |
11129 | |
11130 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11131 | unsafe fn test_mm_maskz_add_epi8() { |
11132 | let a = _mm_set1_epi8(1); |
11133 | let b = _mm_set1_epi8(2); |
11134 | let r = _mm_maskz_add_epi8(0, a, b); |
11135 | assert_eq_m128i(r, _mm_setzero_si128()); |
11136 | let r = _mm_maskz_add_epi8(0b00000000_11111111, a, b); |
11137 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3); |
11138 | assert_eq_m128i(r, e); |
11139 | } |
11140 | |
11141 | #[simd_test(enable = "avx512bw" )] |
11142 | unsafe fn test_mm512_adds_epu16() { |
11143 | let a = _mm512_set1_epi16(1); |
11144 | let b = _mm512_set1_epi16(u16::MAX as i16); |
11145 | let r = _mm512_adds_epu16(a, b); |
11146 | let e = _mm512_set1_epi16(u16::MAX as i16); |
11147 | assert_eq_m512i(r, e); |
11148 | } |
11149 | |
11150 | #[simd_test(enable = "avx512bw" )] |
11151 | unsafe fn test_mm512_mask_adds_epu16() { |
11152 | let a = _mm512_set1_epi16(1); |
11153 | let b = _mm512_set1_epi16(u16::MAX as i16); |
11154 | let r = _mm512_mask_adds_epu16(a, 0, a, b); |
11155 | assert_eq_m512i(r, a); |
11156 | let r = _mm512_mask_adds_epu16(a, 0b00000000_00000000_00000000_00001111, a, b); |
11157 | #[rustfmt::skip] |
11158 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11159 | 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); |
11160 | assert_eq_m512i(r, e); |
11161 | } |
11162 | |
11163 | #[simd_test(enable = "avx512bw" )] |
11164 | unsafe fn test_mm512_maskz_adds_epu16() { |
11165 | let a = _mm512_set1_epi16(1); |
11166 | let b = _mm512_set1_epi16(u16::MAX as i16); |
11167 | let r = _mm512_maskz_adds_epu16(0, a, b); |
11168 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11169 | let r = _mm512_maskz_adds_epu16(0b00000000_00000000_00000000_00001111, a, b); |
11170 | #[rustfmt::skip] |
11171 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11172 | 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); |
11173 | assert_eq_m512i(r, e); |
11174 | } |
11175 | |
11176 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11177 | unsafe fn test_mm256_mask_adds_epu16() { |
11178 | let a = _mm256_set1_epi16(1); |
11179 | let b = _mm256_set1_epi16(u16::MAX as i16); |
11180 | let r = _mm256_mask_adds_epu16(a, 0, a, b); |
11181 | assert_eq_m256i(r, a); |
11182 | let r = _mm256_mask_adds_epu16(a, 0b00000000_00001111, a, b); |
11183 | #[rustfmt::skip] |
11184 | 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); |
11185 | assert_eq_m256i(r, e); |
11186 | } |
11187 | |
11188 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11189 | unsafe fn test_mm256_maskz_adds_epu16() { |
11190 | let a = _mm256_set1_epi16(1); |
11191 | let b = _mm256_set1_epi16(u16::MAX as i16); |
11192 | let r = _mm256_maskz_adds_epu16(0, a, b); |
11193 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11194 | let r = _mm256_maskz_adds_epu16(0b00000000_00001111, a, b); |
11195 | #[rustfmt::skip] |
11196 | 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); |
11197 | assert_eq_m256i(r, e); |
11198 | } |
11199 | |
11200 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11201 | unsafe fn test_mm_mask_adds_epu16() { |
11202 | let a = _mm_set1_epi16(1); |
11203 | let b = _mm_set1_epi16(u16::MAX as i16); |
11204 | let r = _mm_mask_adds_epu16(a, 0, a, b); |
11205 | assert_eq_m128i(r, a); |
11206 | let r = _mm_mask_adds_epu16(a, 0b00001111, a, b); |
11207 | #[rustfmt::skip] |
11208 | 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); |
11209 | assert_eq_m128i(r, e); |
11210 | } |
11211 | |
11212 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11213 | unsafe fn test_mm_maskz_adds_epu16() { |
11214 | let a = _mm_set1_epi16(1); |
11215 | let b = _mm_set1_epi16(u16::MAX as i16); |
11216 | let r = _mm_maskz_adds_epu16(0, a, b); |
11217 | assert_eq_m128i(r, _mm_setzero_si128()); |
11218 | let r = _mm_maskz_adds_epu16(0b00001111, a, b); |
11219 | #[rustfmt::skip] |
11220 | 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); |
11221 | assert_eq_m128i(r, e); |
11222 | } |
11223 | |
11224 | #[simd_test(enable = "avx512bw" )] |
11225 | unsafe fn test_mm512_adds_epu8() { |
11226 | let a = _mm512_set1_epi8(1); |
11227 | let b = _mm512_set1_epi8(u8::MAX as i8); |
11228 | let r = _mm512_adds_epu8(a, b); |
11229 | let e = _mm512_set1_epi8(u8::MAX as i8); |
11230 | assert_eq_m512i(r, e); |
11231 | } |
11232 | |
11233 | #[simd_test(enable = "avx512bw" )] |
11234 | unsafe fn test_mm512_mask_adds_epu8() { |
11235 | let a = _mm512_set1_epi8(1); |
11236 | let b = _mm512_set1_epi8(u8::MAX as i8); |
11237 | let r = _mm512_mask_adds_epu8(a, 0, a, b); |
11238 | assert_eq_m512i(r, a); |
11239 | let r = _mm512_mask_adds_epu8( |
11240 | a, |
11241 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
11242 | a, |
11243 | b, |
11244 | ); |
11245 | #[rustfmt::skip] |
11246 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11247 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11248 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11249 | 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); |
11250 | assert_eq_m512i(r, e); |
11251 | } |
11252 | |
11253 | #[simd_test(enable = "avx512bw" )] |
11254 | unsafe fn test_mm512_maskz_adds_epu8() { |
11255 | let a = _mm512_set1_epi8(1); |
11256 | let b = _mm512_set1_epi8(u8::MAX as i8); |
11257 | let r = _mm512_maskz_adds_epu8(0, a, b); |
11258 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11259 | let r = _mm512_maskz_adds_epu8( |
11260 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
11261 | a, |
11262 | b, |
11263 | ); |
11264 | #[rustfmt::skip] |
11265 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11266 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11267 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11268 | 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); |
11269 | assert_eq_m512i(r, e); |
11270 | } |
11271 | |
11272 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11273 | unsafe fn test_mm256_mask_adds_epu8() { |
11274 | let a = _mm256_set1_epi8(1); |
11275 | let b = _mm256_set1_epi8(u8::MAX as i8); |
11276 | let r = _mm256_mask_adds_epu8(a, 0, a, b); |
11277 | assert_eq_m256i(r, a); |
11278 | let r = _mm256_mask_adds_epu8(a, 0b00000000_00000000_00000000_00001111, a, b); |
11279 | #[rustfmt::skip] |
11280 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11281 | 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); |
11282 | assert_eq_m256i(r, e); |
11283 | } |
11284 | |
11285 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11286 | unsafe fn test_mm256_maskz_adds_epu8() { |
11287 | let a = _mm256_set1_epi8(1); |
11288 | let b = _mm256_set1_epi8(u8::MAX as i8); |
11289 | let r = _mm256_maskz_adds_epu8(0, a, b); |
11290 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11291 | let r = _mm256_maskz_adds_epu8(0b00000000_00000000_00000000_00001111, a, b); |
11292 | #[rustfmt::skip] |
11293 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11294 | 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); |
11295 | assert_eq_m256i(r, e); |
11296 | } |
11297 | |
11298 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11299 | unsafe fn test_mm_mask_adds_epu8() { |
11300 | let a = _mm_set1_epi8(1); |
11301 | let b = _mm_set1_epi8(u8::MAX as i8); |
11302 | let r = _mm_mask_adds_epu8(a, 0, a, b); |
11303 | assert_eq_m128i(r, a); |
11304 | let r = _mm_mask_adds_epu8(a, 0b00000000_00001111, a, b); |
11305 | #[rustfmt::skip] |
11306 | 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); |
11307 | assert_eq_m128i(r, e); |
11308 | } |
11309 | |
11310 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11311 | unsafe fn test_mm_maskz_adds_epu8() { |
11312 | let a = _mm_set1_epi8(1); |
11313 | let b = _mm_set1_epi8(u8::MAX as i8); |
11314 | let r = _mm_maskz_adds_epu8(0, a, b); |
11315 | assert_eq_m128i(r, _mm_setzero_si128()); |
11316 | let r = _mm_maskz_adds_epu8(0b00000000_00001111, a, b); |
11317 | #[rustfmt::skip] |
11318 | 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); |
11319 | assert_eq_m128i(r, e); |
11320 | } |
11321 | |
11322 | #[simd_test(enable = "avx512bw" )] |
11323 | unsafe fn test_mm512_adds_epi16() { |
11324 | let a = _mm512_set1_epi16(1); |
11325 | let b = _mm512_set1_epi16(i16::MAX); |
11326 | let r = _mm512_adds_epi16(a, b); |
11327 | let e = _mm512_set1_epi16(i16::MAX); |
11328 | assert_eq_m512i(r, e); |
11329 | } |
11330 | |
11331 | #[simd_test(enable = "avx512bw" )] |
11332 | unsafe fn test_mm512_mask_adds_epi16() { |
11333 | let a = _mm512_set1_epi16(1); |
11334 | let b = _mm512_set1_epi16(i16::MAX); |
11335 | let r = _mm512_mask_adds_epi16(a, 0, a, b); |
11336 | assert_eq_m512i(r, a); |
11337 | let r = _mm512_mask_adds_epi16(a, 0b00000000_00000000_00000000_00001111, a, b); |
11338 | #[rustfmt::skip] |
11339 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11340 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
11341 | assert_eq_m512i(r, e); |
11342 | } |
11343 | |
11344 | #[simd_test(enable = "avx512bw" )] |
11345 | unsafe fn test_mm512_maskz_adds_epi16() { |
11346 | let a = _mm512_set1_epi16(1); |
11347 | let b = _mm512_set1_epi16(i16::MAX); |
11348 | let r = _mm512_maskz_adds_epi16(0, a, b); |
11349 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11350 | let r = _mm512_maskz_adds_epi16(0b00000000_00000000_00000000_00001111, a, b); |
11351 | #[rustfmt::skip] |
11352 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11353 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
11354 | assert_eq_m512i(r, e); |
11355 | } |
11356 | |
11357 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11358 | unsafe fn test_mm256_mask_adds_epi16() { |
11359 | let a = _mm256_set1_epi16(1); |
11360 | let b = _mm256_set1_epi16(i16::MAX); |
11361 | let r = _mm256_mask_adds_epi16(a, 0, a, b); |
11362 | assert_eq_m256i(r, a); |
11363 | let r = _mm256_mask_adds_epi16(a, 0b00000000_00001111, a, b); |
11364 | #[rustfmt::skip] |
11365 | 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); |
11366 | assert_eq_m256i(r, e); |
11367 | } |
11368 | |
11369 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11370 | unsafe fn test_mm256_maskz_adds_epi16() { |
11371 | let a = _mm256_set1_epi16(1); |
11372 | let b = _mm256_set1_epi16(i16::MAX); |
11373 | let r = _mm256_maskz_adds_epi16(0, a, b); |
11374 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11375 | let r = _mm256_maskz_adds_epi16(0b00000000_00001111, a, b); |
11376 | #[rustfmt::skip] |
11377 | 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); |
11378 | assert_eq_m256i(r, e); |
11379 | } |
11380 | |
11381 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11382 | unsafe fn test_mm_mask_adds_epi16() { |
11383 | let a = _mm_set1_epi16(1); |
11384 | let b = _mm_set1_epi16(i16::MAX); |
11385 | let r = _mm_mask_adds_epi16(a, 0, a, b); |
11386 | assert_eq_m128i(r, a); |
11387 | let r = _mm_mask_adds_epi16(a, 0b00001111, a, b); |
11388 | let e = _mm_set_epi16(1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
11389 | assert_eq_m128i(r, e); |
11390 | } |
11391 | |
11392 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11393 | unsafe fn test_mm_maskz_adds_epi16() { |
11394 | let a = _mm_set1_epi16(1); |
11395 | let b = _mm_set1_epi16(i16::MAX); |
11396 | let r = _mm_maskz_adds_epi16(0, a, b); |
11397 | assert_eq_m128i(r, _mm_setzero_si128()); |
11398 | let r = _mm_maskz_adds_epi16(0b00001111, a, b); |
11399 | let e = _mm_set_epi16(0, 0, 0, 0, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
11400 | assert_eq_m128i(r, e); |
11401 | } |
11402 | |
11403 | #[simd_test(enable = "avx512bw" )] |
11404 | unsafe fn test_mm512_adds_epi8() { |
11405 | let a = _mm512_set1_epi8(1); |
11406 | let b = _mm512_set1_epi8(i8::MAX); |
11407 | let r = _mm512_adds_epi8(a, b); |
11408 | let e = _mm512_set1_epi8(i8::MAX); |
11409 | assert_eq_m512i(r, e); |
11410 | } |
11411 | |
11412 | #[simd_test(enable = "avx512bw" )] |
11413 | unsafe fn test_mm512_mask_adds_epi8() { |
11414 | let a = _mm512_set1_epi8(1); |
11415 | let b = _mm512_set1_epi8(i8::MAX); |
11416 | let r = _mm512_mask_adds_epi8(a, 0, a, b); |
11417 | assert_eq_m512i(r, a); |
11418 | let r = _mm512_mask_adds_epi8( |
11419 | a, |
11420 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
11421 | a, |
11422 | b, |
11423 | ); |
11424 | #[rustfmt::skip] |
11425 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11426 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11427 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11428 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
11429 | assert_eq_m512i(r, e); |
11430 | } |
11431 | |
11432 | #[simd_test(enable = "avx512bw" )] |
11433 | unsafe fn test_mm512_maskz_adds_epi8() { |
11434 | let a = _mm512_set1_epi8(1); |
11435 | let b = _mm512_set1_epi8(i8::MAX); |
11436 | let r = _mm512_maskz_adds_epi8(0, a, b); |
11437 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11438 | let r = _mm512_maskz_adds_epi8( |
11439 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
11440 | a, |
11441 | b, |
11442 | ); |
11443 | #[rustfmt::skip] |
11444 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11445 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11446 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11447 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
11448 | assert_eq_m512i(r, e); |
11449 | } |
11450 | |
11451 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11452 | unsafe fn test_mm256_mask_adds_epi8() { |
11453 | let a = _mm256_set1_epi8(1); |
11454 | let b = _mm256_set1_epi8(i8::MAX); |
11455 | let r = _mm256_mask_adds_epi8(a, 0, a, b); |
11456 | assert_eq_m256i(r, a); |
11457 | let r = _mm256_mask_adds_epi8(a, 0b00000000_00000000_00000000_00001111, a, b); |
11458 | #[rustfmt::skip] |
11459 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11460 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
11461 | assert_eq_m256i(r, e); |
11462 | } |
11463 | |
11464 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11465 | unsafe fn test_mm256_maskz_adds_epi8() { |
11466 | let a = _mm256_set1_epi8(1); |
11467 | let b = _mm256_set1_epi8(i8::MAX); |
11468 | let r = _mm256_maskz_adds_epi8(0, a, b); |
11469 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11470 | let r = _mm256_maskz_adds_epi8(0b00000000_00000000_00000000_00001111, a, b); |
11471 | #[rustfmt::skip] |
11472 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11473 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
11474 | assert_eq_m256i(r, e); |
11475 | } |
11476 | |
11477 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11478 | unsafe fn test_mm_mask_adds_epi8() { |
11479 | let a = _mm_set1_epi8(1); |
11480 | let b = _mm_set1_epi8(i8::MAX); |
11481 | let r = _mm_mask_adds_epi8(a, 0, a, b); |
11482 | assert_eq_m128i(r, a); |
11483 | let r = _mm_mask_adds_epi8(a, 0b00000000_00001111, a, b); |
11484 | #[rustfmt::skip] |
11485 | 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); |
11486 | assert_eq_m128i(r, e); |
11487 | } |
11488 | |
11489 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11490 | unsafe fn test_mm_maskz_adds_epi8() { |
11491 | let a = _mm_set1_epi8(1); |
11492 | let b = _mm_set1_epi8(i8::MAX); |
11493 | let r = _mm_maskz_adds_epi8(0, a, b); |
11494 | assert_eq_m128i(r, _mm_setzero_si128()); |
11495 | let r = _mm_maskz_adds_epi8(0b00000000_00001111, a, b); |
11496 | #[rustfmt::skip] |
11497 | 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); |
11498 | assert_eq_m128i(r, e); |
11499 | } |
11500 | |
11501 | #[simd_test(enable = "avx512bw" )] |
11502 | unsafe fn test_mm512_sub_epi16() { |
11503 | let a = _mm512_set1_epi16(1); |
11504 | let b = _mm512_set1_epi16(2); |
11505 | let r = _mm512_sub_epi16(a, b); |
11506 | let e = _mm512_set1_epi16(-1); |
11507 | assert_eq_m512i(r, e); |
11508 | } |
11509 | |
11510 | #[simd_test(enable = "avx512bw" )] |
11511 | unsafe fn test_mm512_mask_sub_epi16() { |
11512 | let a = _mm512_set1_epi16(1); |
11513 | let b = _mm512_set1_epi16(2); |
11514 | let r = _mm512_mask_sub_epi16(a, 0, a, b); |
11515 | assert_eq_m512i(r, a); |
11516 | let r = _mm512_mask_sub_epi16(a, 0b00000000_11111111_00000000_11111111, a, b); |
11517 | #[rustfmt::skip] |
11518 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, |
11519 | 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1); |
11520 | assert_eq_m512i(r, e); |
11521 | } |
11522 | |
11523 | #[simd_test(enable = "avx512bw" )] |
11524 | unsafe fn test_mm512_maskz_sub_epi16() { |
11525 | let a = _mm512_set1_epi16(1); |
11526 | let b = _mm512_set1_epi16(2); |
11527 | let r = _mm512_maskz_sub_epi16(0, a, b); |
11528 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11529 | let r = _mm512_maskz_sub_epi16(0b00000000_11111111_00000000_11111111, a, b); |
11530 | #[rustfmt::skip] |
11531 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1, |
11532 | 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
11533 | assert_eq_m512i(r, e); |
11534 | } |
11535 | |
11536 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11537 | unsafe fn test_mm256_mask_sub_epi16() { |
11538 | let a = _mm256_set1_epi16(1); |
11539 | let b = _mm256_set1_epi16(2); |
11540 | let r = _mm256_mask_sub_epi16(a, 0, a, b); |
11541 | assert_eq_m256i(r, a); |
11542 | let r = _mm256_mask_sub_epi16(a, 0b00000000_11111111, a, b); |
11543 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1); |
11544 | assert_eq_m256i(r, e); |
11545 | } |
11546 | |
11547 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11548 | unsafe fn test_mm256_maskz_sub_epi16() { |
11549 | let a = _mm256_set1_epi16(1); |
11550 | let b = _mm256_set1_epi16(2); |
11551 | let r = _mm256_maskz_sub_epi16(0, a, b); |
11552 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11553 | let r = _mm256_maskz_sub_epi16(0b00000000_11111111, a, b); |
11554 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
11555 | assert_eq_m256i(r, e); |
11556 | } |
11557 | |
11558 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11559 | unsafe fn test_mm_mask_sub_epi16() { |
11560 | let a = _mm_set1_epi16(1); |
11561 | let b = _mm_set1_epi16(2); |
11562 | let r = _mm_mask_sub_epi16(a, 0, a, b); |
11563 | assert_eq_m128i(r, a); |
11564 | let r = _mm_mask_sub_epi16(a, 0b00001111, a, b); |
11565 | let e = _mm_set_epi16(1, 1, 1, 1, -1, -1, -1, -1); |
11566 | assert_eq_m128i(r, e); |
11567 | } |
11568 | |
11569 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11570 | unsafe fn test_mm_maskz_sub_epi16() { |
11571 | let a = _mm_set1_epi16(1); |
11572 | let b = _mm_set1_epi16(2); |
11573 | let r = _mm_maskz_sub_epi16(0, a, b); |
11574 | assert_eq_m128i(r, _mm_setzero_si128()); |
11575 | let r = _mm_maskz_sub_epi16(0b00001111, a, b); |
11576 | let e = _mm_set_epi16(0, 0, 0, 0, -1, -1, -1, -1); |
11577 | assert_eq_m128i(r, e); |
11578 | } |
11579 | |
11580 | #[simd_test(enable = "avx512bw" )] |
11581 | unsafe fn test_mm512_sub_epi8() { |
11582 | let a = _mm512_set1_epi8(1); |
11583 | let b = _mm512_set1_epi8(2); |
11584 | let r = _mm512_sub_epi8(a, b); |
11585 | let e = _mm512_set1_epi8(-1); |
11586 | assert_eq_m512i(r, e); |
11587 | } |
11588 | |
11589 | #[simd_test(enable = "avx512bw" )] |
11590 | unsafe fn test_mm512_mask_sub_epi8() { |
11591 | let a = _mm512_set1_epi8(1); |
11592 | let b = _mm512_set1_epi8(2); |
11593 | let r = _mm512_mask_sub_epi8(a, 0, a, b); |
11594 | assert_eq_m512i(r, a); |
11595 | let r = _mm512_mask_sub_epi8( |
11596 | a, |
11597 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
11598 | a, |
11599 | b, |
11600 | ); |
11601 | #[rustfmt::skip] |
11602 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, |
11603 | 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, |
11604 | 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, |
11605 | 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1); |
11606 | assert_eq_m512i(r, e); |
11607 | } |
11608 | |
11609 | #[simd_test(enable = "avx512bw" )] |
11610 | unsafe fn test_mm512_maskz_sub_epi8() { |
11611 | let a = _mm512_set1_epi8(1); |
11612 | let b = _mm512_set1_epi8(2); |
11613 | let r = _mm512_maskz_sub_epi8(0, a, b); |
11614 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11615 | let r = _mm512_maskz_sub_epi8( |
11616 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
11617 | a, |
11618 | b, |
11619 | ); |
11620 | #[rustfmt::skip] |
11621 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1, |
11622 | 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1, |
11623 | 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1, |
11624 | 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
11625 | assert_eq_m512i(r, e); |
11626 | } |
11627 | |
11628 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11629 | unsafe fn test_mm256_mask_sub_epi8() { |
11630 | let a = _mm256_set1_epi8(1); |
11631 | let b = _mm256_set1_epi8(2); |
11632 | let r = _mm256_mask_sub_epi8(a, 0, a, b); |
11633 | assert_eq_m256i(r, a); |
11634 | let r = _mm256_mask_sub_epi8(a, 0b00000000_11111111_00000000_11111111, a, b); |
11635 | #[rustfmt::skip] |
11636 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, |
11637 | 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1); |
11638 | assert_eq_m256i(r, e); |
11639 | } |
11640 | |
11641 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11642 | unsafe fn test_mm256_maskz_sub_epi8() { |
11643 | let a = _mm256_set1_epi8(1); |
11644 | let b = _mm256_set1_epi8(2); |
11645 | let r = _mm256_maskz_sub_epi8(0, a, b); |
11646 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11647 | let r = _mm256_maskz_sub_epi8(0b00000000_11111111_00000000_11111111, a, b); |
11648 | #[rustfmt::skip] |
11649 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1, |
11650 | 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
11651 | assert_eq_m256i(r, e); |
11652 | } |
11653 | |
11654 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11655 | unsafe fn test_mm_mask_sub_epi8() { |
11656 | let a = _mm_set1_epi8(1); |
11657 | let b = _mm_set1_epi8(2); |
11658 | let r = _mm_mask_sub_epi8(a, 0, a, b); |
11659 | assert_eq_m128i(r, a); |
11660 | let r = _mm_mask_sub_epi8(a, 0b00000000_11111111, a, b); |
11661 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1); |
11662 | assert_eq_m128i(r, e); |
11663 | } |
11664 | |
11665 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11666 | unsafe fn test_mm_maskz_sub_epi8() { |
11667 | let a = _mm_set1_epi8(1); |
11668 | let b = _mm_set1_epi8(2); |
11669 | let r = _mm_maskz_sub_epi8(0, a, b); |
11670 | assert_eq_m128i(r, _mm_setzero_si128()); |
11671 | let r = _mm_maskz_sub_epi8(0b00000000_11111111, a, b); |
11672 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
11673 | assert_eq_m128i(r, e); |
11674 | } |
11675 | |
11676 | #[simd_test(enable = "avx512bw" )] |
11677 | unsafe fn test_mm512_subs_epu16() { |
11678 | let a = _mm512_set1_epi16(1); |
11679 | let b = _mm512_set1_epi16(u16::MAX as i16); |
11680 | let r = _mm512_subs_epu16(a, b); |
11681 | let e = _mm512_set1_epi16(0); |
11682 | assert_eq_m512i(r, e); |
11683 | } |
11684 | |
11685 | #[simd_test(enable = "avx512bw" )] |
11686 | unsafe fn test_mm512_mask_subs_epu16() { |
11687 | let a = _mm512_set1_epi16(1); |
11688 | let b = _mm512_set1_epi16(u16::MAX as i16); |
11689 | let r = _mm512_mask_subs_epu16(a, 0, a, b); |
11690 | assert_eq_m512i(r, a); |
11691 | let r = _mm512_mask_subs_epu16(a, 0b00000000_00000000_00000000_00001111, a, b); |
11692 | #[rustfmt::skip] |
11693 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11694 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
11695 | assert_eq_m512i(r, e); |
11696 | } |
11697 | |
11698 | #[simd_test(enable = "avx512bw" )] |
11699 | unsafe fn test_mm512_maskz_subs_epu16() { |
11700 | let a = _mm512_set1_epi16(1); |
11701 | let b = _mm512_set1_epi16(u16::MAX as i16); |
11702 | let r = _mm512_maskz_subs_epu16(0, a, b); |
11703 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11704 | let r = _mm512_maskz_subs_epu16(0b00000000_00000000_00000000_00001111, a, b); |
11705 | #[rustfmt::skip] |
11706 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11707 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
11708 | assert_eq_m512i(r, e); |
11709 | } |
11710 | |
11711 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11712 | unsafe fn test_mm256_mask_subs_epu16() { |
11713 | let a = _mm256_set1_epi16(1); |
11714 | let b = _mm256_set1_epi16(u16::MAX as i16); |
11715 | let r = _mm256_mask_subs_epu16(a, 0, a, b); |
11716 | assert_eq_m256i(r, a); |
11717 | let r = _mm256_mask_subs_epu16(a, 0b00000000_00001111, a, b); |
11718 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
11719 | assert_eq_m256i(r, e); |
11720 | } |
11721 | |
11722 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11723 | unsafe fn test_mm256_maskz_subs_epu16() { |
11724 | let a = _mm256_set1_epi16(1); |
11725 | let b = _mm256_set1_epi16(u16::MAX as i16); |
11726 | let r = _mm256_maskz_subs_epu16(0, a, b); |
11727 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11728 | let r = _mm256_maskz_subs_epu16(0b00000000_00001111, a, b); |
11729 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
11730 | assert_eq_m256i(r, e); |
11731 | } |
11732 | |
11733 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11734 | unsafe fn test_mm_mask_subs_epu16() { |
11735 | let a = _mm_set1_epi16(1); |
11736 | let b = _mm_set1_epi16(u16::MAX as i16); |
11737 | let r = _mm_mask_subs_epu16(a, 0, a, b); |
11738 | assert_eq_m128i(r, a); |
11739 | let r = _mm_mask_subs_epu16(a, 0b00001111, a, b); |
11740 | let e = _mm_set_epi16(1, 1, 1, 1, 0, 0, 0, 0); |
11741 | assert_eq_m128i(r, e); |
11742 | } |
11743 | |
11744 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11745 | unsafe fn test_mm_maskz_subs_epu16() { |
11746 | let a = _mm_set1_epi16(1); |
11747 | let b = _mm_set1_epi16(u16::MAX as i16); |
11748 | let r = _mm_maskz_subs_epu16(0, a, b); |
11749 | assert_eq_m128i(r, _mm_setzero_si128()); |
11750 | let r = _mm_maskz_subs_epu16(0b00001111, a, b); |
11751 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 0, 0); |
11752 | assert_eq_m128i(r, e); |
11753 | } |
11754 | |
11755 | #[simd_test(enable = "avx512bw" )] |
11756 | unsafe fn test_mm512_subs_epu8() { |
11757 | let a = _mm512_set1_epi8(1); |
11758 | let b = _mm512_set1_epi8(u8::MAX as i8); |
11759 | let r = _mm512_subs_epu8(a, b); |
11760 | let e = _mm512_set1_epi8(0); |
11761 | assert_eq_m512i(r, e); |
11762 | } |
11763 | |
11764 | #[simd_test(enable = "avx512bw" )] |
11765 | unsafe fn test_mm512_mask_subs_epu8() { |
11766 | let a = _mm512_set1_epi8(1); |
11767 | let b = _mm512_set1_epi8(u8::MAX as i8); |
11768 | let r = _mm512_mask_subs_epu8(a, 0, a, b); |
11769 | assert_eq_m512i(r, a); |
11770 | let r = _mm512_mask_subs_epu8( |
11771 | a, |
11772 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
11773 | a, |
11774 | b, |
11775 | ); |
11776 | #[rustfmt::skip] |
11777 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11778 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11779 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11780 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
11781 | assert_eq_m512i(r, e); |
11782 | } |
11783 | |
11784 | #[simd_test(enable = "avx512bw" )] |
11785 | unsafe fn test_mm512_maskz_subs_epu8() { |
11786 | let a = _mm512_set1_epi8(1); |
11787 | let b = _mm512_set1_epi8(u8::MAX as i8); |
11788 | let r = _mm512_maskz_subs_epu8(0, a, b); |
11789 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11790 | let r = _mm512_maskz_subs_epu8( |
11791 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
11792 | a, |
11793 | b, |
11794 | ); |
11795 | #[rustfmt::skip] |
11796 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11797 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11798 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11799 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
11800 | assert_eq_m512i(r, e); |
11801 | } |
11802 | |
11803 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11804 | unsafe fn test_mm256_mask_subs_epu8() { |
11805 | let a = _mm256_set1_epi8(1); |
11806 | let b = _mm256_set1_epi8(u8::MAX as i8); |
11807 | let r = _mm256_mask_subs_epu8(a, 0, a, b); |
11808 | assert_eq_m256i(r, a); |
11809 | let r = _mm256_mask_subs_epu8(a, 0b00000000_00000000_00000000_00001111, a, b); |
11810 | #[rustfmt::skip] |
11811 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
11812 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
11813 | assert_eq_m256i(r, e); |
11814 | } |
11815 | |
11816 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11817 | unsafe fn test_mm256_maskz_subs_epu8() { |
11818 | let a = _mm256_set1_epi8(1); |
11819 | let b = _mm256_set1_epi8(u8::MAX as i8); |
11820 | let r = _mm256_maskz_subs_epu8(0, a, b); |
11821 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11822 | let r = _mm256_maskz_subs_epu8(0b00000000_00000000_00000000_00001111, a, b); |
11823 | #[rustfmt::skip] |
11824 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11825 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
11826 | assert_eq_m256i(r, e); |
11827 | } |
11828 | |
11829 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11830 | unsafe fn test_mm_mask_subs_epu8() { |
11831 | let a = _mm_set1_epi8(1); |
11832 | let b = _mm_set1_epi8(u8::MAX as i8); |
11833 | let r = _mm_mask_subs_epu8(a, 0, a, b); |
11834 | assert_eq_m128i(r, a); |
11835 | let r = _mm_mask_subs_epu8(a, 0b00000000_00001111, a, b); |
11836 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
11837 | assert_eq_m128i(r, e); |
11838 | } |
11839 | |
11840 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11841 | unsafe fn test_mm_maskz_subs_epu8() { |
11842 | let a = _mm_set1_epi8(1); |
11843 | let b = _mm_set1_epi8(u8::MAX as i8); |
11844 | let r = _mm_maskz_subs_epu8(0, a, b); |
11845 | assert_eq_m128i(r, _mm_setzero_si128()); |
11846 | let r = _mm_maskz_subs_epu8(0b00000000_00001111, a, b); |
11847 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
11848 | assert_eq_m128i(r, e); |
11849 | } |
11850 | |
11851 | #[simd_test(enable = "avx512bw" )] |
11852 | unsafe fn test_mm512_subs_epi16() { |
11853 | let a = _mm512_set1_epi16(-1); |
11854 | let b = _mm512_set1_epi16(i16::MAX); |
11855 | let r = _mm512_subs_epi16(a, b); |
11856 | let e = _mm512_set1_epi16(i16::MIN); |
11857 | assert_eq_m512i(r, e); |
11858 | } |
11859 | |
11860 | #[simd_test(enable = "avx512bw" )] |
11861 | unsafe fn test_mm512_mask_subs_epi16() { |
11862 | let a = _mm512_set1_epi16(-1); |
11863 | let b = _mm512_set1_epi16(i16::MAX); |
11864 | let r = _mm512_mask_subs_epi16(a, 0, a, b); |
11865 | assert_eq_m512i(r, a); |
11866 | let r = _mm512_mask_subs_epi16(a, 0b00000000_00000000_00000000_00001111, a, b); |
11867 | #[rustfmt::skip] |
11868 | let e = _mm512_set_epi16(-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, |
11869 | -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, i16::MIN, i16::MIN, i16::MIN, i16::MIN); |
11870 | assert_eq_m512i(r, e); |
11871 | } |
11872 | |
11873 | #[simd_test(enable = "avx512bw" )] |
11874 | unsafe fn test_mm512_maskz_subs_epi16() { |
11875 | let a = _mm512_set1_epi16(-1); |
11876 | let b = _mm512_set1_epi16(i16::MAX); |
11877 | let r = _mm512_maskz_subs_epi16(0, a, b); |
11878 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11879 | let r = _mm512_maskz_subs_epi16(0b00000000_00000000_00000000_00001111, a, b); |
11880 | #[rustfmt::skip] |
11881 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11882 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i16::MIN, i16::MIN, i16::MIN, i16::MIN); |
11883 | assert_eq_m512i(r, e); |
11884 | } |
11885 | |
11886 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11887 | unsafe fn test_mm256_mask_subs_epi16() { |
11888 | let a = _mm256_set1_epi16(-1); |
11889 | let b = _mm256_set1_epi16(i16::MAX); |
11890 | let r = _mm256_mask_subs_epi16(a, 0, a, b); |
11891 | assert_eq_m256i(r, a); |
11892 | let r = _mm256_mask_subs_epi16(a, 0b00000000_00001111, a, b); |
11893 | #[rustfmt::skip] |
11894 | 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); |
11895 | assert_eq_m256i(r, e); |
11896 | } |
11897 | |
11898 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11899 | unsafe fn test_mm256_maskz_subs_epi16() { |
11900 | let a = _mm256_set1_epi16(-1); |
11901 | let b = _mm256_set1_epi16(i16::MAX); |
11902 | let r = _mm256_maskz_subs_epi16(0, a, b); |
11903 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11904 | let r = _mm256_maskz_subs_epi16(0b00000000_00001111, a, b); |
11905 | #[rustfmt::skip] |
11906 | 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); |
11907 | assert_eq_m256i(r, e); |
11908 | } |
11909 | |
11910 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11911 | unsafe fn test_mm_mask_subs_epi16() { |
11912 | let a = _mm_set1_epi16(-1); |
11913 | let b = _mm_set1_epi16(i16::MAX); |
11914 | let r = _mm_mask_subs_epi16(a, 0, a, b); |
11915 | assert_eq_m128i(r, a); |
11916 | let r = _mm_mask_subs_epi16(a, 0b00001111, a, b); |
11917 | let e = _mm_set_epi16(-1, -1, -1, -1, i16::MIN, i16::MIN, i16::MIN, i16::MIN); |
11918 | assert_eq_m128i(r, e); |
11919 | } |
11920 | |
11921 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11922 | unsafe fn test_mm_maskz_subs_epi16() { |
11923 | let a = _mm_set1_epi16(-1); |
11924 | let b = _mm_set1_epi16(i16::MAX); |
11925 | let r = _mm_maskz_subs_epi16(0, a, b); |
11926 | assert_eq_m128i(r, _mm_setzero_si128()); |
11927 | let r = _mm_maskz_subs_epi16(0b00001111, a, b); |
11928 | let e = _mm_set_epi16(0, 0, 0, 0, i16::MIN, i16::MIN, i16::MIN, i16::MIN); |
11929 | assert_eq_m128i(r, e); |
11930 | } |
11931 | |
11932 | #[simd_test(enable = "avx512bw" )] |
11933 | unsafe fn test_mm512_subs_epi8() { |
11934 | let a = _mm512_set1_epi8(-1); |
11935 | let b = _mm512_set1_epi8(i8::MAX); |
11936 | let r = _mm512_subs_epi8(a, b); |
11937 | let e = _mm512_set1_epi8(i8::MIN); |
11938 | assert_eq_m512i(r, e); |
11939 | } |
11940 | |
11941 | #[simd_test(enable = "avx512bw" )] |
11942 | unsafe fn test_mm512_mask_subs_epi8() { |
11943 | let a = _mm512_set1_epi8(-1); |
11944 | let b = _mm512_set1_epi8(i8::MAX); |
11945 | let r = _mm512_mask_subs_epi8(a, 0, a, b); |
11946 | assert_eq_m512i(r, a); |
11947 | let r = _mm512_mask_subs_epi8( |
11948 | a, |
11949 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
11950 | a, |
11951 | b, |
11952 | ); |
11953 | #[rustfmt::skip] |
11954 | let e = _mm512_set_epi8(-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, |
11955 | -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, |
11956 | -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, |
11957 | -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, i8::MIN, i8::MIN, i8::MIN, i8::MIN); |
11958 | assert_eq_m512i(r, e); |
11959 | } |
11960 | |
11961 | #[simd_test(enable = "avx512bw" )] |
11962 | unsafe fn test_mm512_maskz_subs_epi8() { |
11963 | let a = _mm512_set1_epi8(-1); |
11964 | let b = _mm512_set1_epi8(i8::MAX); |
11965 | let r = _mm512_maskz_subs_epi8(0, a, b); |
11966 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11967 | let r = _mm512_maskz_subs_epi8( |
11968 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
11969 | a, |
11970 | b, |
11971 | ); |
11972 | #[rustfmt::skip] |
11973 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11974 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11975 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
11976 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MIN, i8::MIN, i8::MIN, i8::MIN); |
11977 | assert_eq_m512i(r, e); |
11978 | } |
11979 | |
11980 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11981 | unsafe fn test_mm256_mask_subs_epi8() { |
11982 | let a = _mm256_set1_epi8(-1); |
11983 | let b = _mm256_set1_epi8(i8::MAX); |
11984 | let r = _mm256_mask_subs_epi8(a, 0, a, b); |
11985 | assert_eq_m256i(r, a); |
11986 | let r = _mm256_mask_subs_epi8(a, 0b00000000_00000000_00000000_00001111, a, b); |
11987 | #[rustfmt::skip] |
11988 | let e = _mm256_set_epi8(-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, |
11989 | -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, i8::MIN, i8::MIN, i8::MIN, i8::MIN); |
11990 | assert_eq_m256i(r, e); |
11991 | } |
11992 | |
11993 | #[simd_test(enable = "avx512bw,avx512vl" )] |
11994 | unsafe fn test_mm256_maskz_subs_epi8() { |
11995 | let a = _mm256_set1_epi8(-1); |
11996 | let b = _mm256_set1_epi8(i8::MAX); |
11997 | let r = _mm256_maskz_subs_epi8(0, a, b); |
11998 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11999 | let r = _mm256_maskz_subs_epi8(0b00000000_00000000_00000000_00001111, a, b); |
12000 | #[rustfmt::skip] |
12001 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
12002 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MIN, i8::MIN, i8::MIN, i8::MIN); |
12003 | assert_eq_m256i(r, e); |
12004 | } |
12005 | |
12006 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12007 | unsafe fn test_mm_mask_subs_epi8() { |
12008 | let a = _mm_set1_epi8(-1); |
12009 | let b = _mm_set1_epi8(i8::MAX); |
12010 | let r = _mm_mask_subs_epi8(a, 0, a, b); |
12011 | assert_eq_m128i(r, a); |
12012 | let r = _mm_mask_subs_epi8(a, 0b00000000_00001111, a, b); |
12013 | #[rustfmt::skip] |
12014 | 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); |
12015 | assert_eq_m128i(r, e); |
12016 | } |
12017 | |
12018 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12019 | unsafe fn test_mm_maskz_subs_epi8() { |
12020 | let a = _mm_set1_epi8(-1); |
12021 | let b = _mm_set1_epi8(i8::MAX); |
12022 | let r = _mm_maskz_subs_epi8(0, a, b); |
12023 | assert_eq_m128i(r, _mm_setzero_si128()); |
12024 | let r = _mm_maskz_subs_epi8(0b00000000_00001111, a, b); |
12025 | #[rustfmt::skip] |
12026 | 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); |
12027 | assert_eq_m128i(r, e); |
12028 | } |
12029 | |
12030 | #[simd_test(enable = "avx512bw" )] |
12031 | unsafe fn test_mm512_mulhi_epu16() { |
12032 | let a = _mm512_set1_epi16(1); |
12033 | let b = _mm512_set1_epi16(1); |
12034 | let r = _mm512_mulhi_epu16(a, b); |
12035 | let e = _mm512_set1_epi16(0); |
12036 | assert_eq_m512i(r, e); |
12037 | } |
12038 | |
12039 | #[simd_test(enable = "avx512bw" )] |
12040 | unsafe fn test_mm512_mask_mulhi_epu16() { |
12041 | let a = _mm512_set1_epi16(1); |
12042 | let b = _mm512_set1_epi16(1); |
12043 | let r = _mm512_mask_mulhi_epu16(a, 0, a, b); |
12044 | assert_eq_m512i(r, a); |
12045 | let r = _mm512_mask_mulhi_epu16(a, 0b00000000_00000000_00000000_00001111, a, b); |
12046 | #[rustfmt::skip] |
12047 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
12048 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
12049 | assert_eq_m512i(r, e); |
12050 | } |
12051 | |
12052 | #[simd_test(enable = "avx512bw" )] |
12053 | unsafe fn test_mm512_maskz_mulhi_epu16() { |
12054 | let a = _mm512_set1_epi16(1); |
12055 | let b = _mm512_set1_epi16(1); |
12056 | let r = _mm512_maskz_mulhi_epu16(0, a, b); |
12057 | assert_eq_m512i(r, _mm512_setzero_si512()); |
12058 | let r = _mm512_maskz_mulhi_epu16(0b00000000_00000000_00000000_00001111, a, b); |
12059 | #[rustfmt::skip] |
12060 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
12061 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
12062 | assert_eq_m512i(r, e); |
12063 | } |
12064 | |
12065 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12066 | unsafe fn test_mm256_mask_mulhi_epu16() { |
12067 | let a = _mm256_set1_epi16(1); |
12068 | let b = _mm256_set1_epi16(1); |
12069 | let r = _mm256_mask_mulhi_epu16(a, 0, a, b); |
12070 | assert_eq_m256i(r, a); |
12071 | let r = _mm256_mask_mulhi_epu16(a, 0b00000000_00001111, a, b); |
12072 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
12073 | assert_eq_m256i(r, e); |
12074 | } |
12075 | |
12076 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12077 | unsafe fn test_mm256_maskz_mulhi_epu16() { |
12078 | let a = _mm256_set1_epi16(1); |
12079 | let b = _mm256_set1_epi16(1); |
12080 | let r = _mm256_maskz_mulhi_epu16(0, a, b); |
12081 | assert_eq_m256i(r, _mm256_setzero_si256()); |
12082 | let r = _mm256_maskz_mulhi_epu16(0b00000000_00001111, a, b); |
12083 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
12084 | assert_eq_m256i(r, e); |
12085 | } |
12086 | |
12087 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12088 | unsafe fn test_mm_mask_mulhi_epu16() { |
12089 | let a = _mm_set1_epi16(1); |
12090 | let b = _mm_set1_epi16(1); |
12091 | let r = _mm_mask_mulhi_epu16(a, 0, a, b); |
12092 | assert_eq_m128i(r, a); |
12093 | let r = _mm_mask_mulhi_epu16(a, 0b00001111, a, b); |
12094 | let e = _mm_set_epi16(1, 1, 1, 1, 0, 0, 0, 0); |
12095 | assert_eq_m128i(r, e); |
12096 | } |
12097 | |
12098 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12099 | unsafe fn test_mm_maskz_mulhi_epu16() { |
12100 | let a = _mm_set1_epi16(1); |
12101 | let b = _mm_set1_epi16(1); |
12102 | let r = _mm_maskz_mulhi_epu16(0, a, b); |
12103 | assert_eq_m128i(r, _mm_setzero_si128()); |
12104 | let r = _mm_maskz_mulhi_epu16(0b00001111, a, b); |
12105 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 0, 0); |
12106 | assert_eq_m128i(r, e); |
12107 | } |
12108 | |
12109 | #[simd_test(enable = "avx512bw" )] |
12110 | unsafe fn test_mm512_mulhi_epi16() { |
12111 | let a = _mm512_set1_epi16(1); |
12112 | let b = _mm512_set1_epi16(1); |
12113 | let r = _mm512_mulhi_epi16(a, b); |
12114 | let e = _mm512_set1_epi16(0); |
12115 | assert_eq_m512i(r, e); |
12116 | } |
12117 | |
12118 | #[simd_test(enable = "avx512bw" )] |
12119 | unsafe fn test_mm512_mask_mulhi_epi16() { |
12120 | let a = _mm512_set1_epi16(1); |
12121 | let b = _mm512_set1_epi16(1); |
12122 | let r = _mm512_mask_mulhi_epi16(a, 0, a, b); |
12123 | assert_eq_m512i(r, a); |
12124 | let r = _mm512_mask_mulhi_epi16(a, 0b00000000_00000000_00000000_00001111, a, b); |
12125 | #[rustfmt::skip] |
12126 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
12127 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
12128 | assert_eq_m512i(r, e); |
12129 | } |
12130 | |
12131 | #[simd_test(enable = "avx512bw" )] |
12132 | unsafe fn test_mm512_maskz_mulhi_epi16() { |
12133 | let a = _mm512_set1_epi16(1); |
12134 | let b = _mm512_set1_epi16(1); |
12135 | let r = _mm512_maskz_mulhi_epi16(0, a, b); |
12136 | assert_eq_m512i(r, _mm512_setzero_si512()); |
12137 | let r = _mm512_maskz_mulhi_epi16(0b00000000_00000000_00000000_00001111, a, b); |
12138 | #[rustfmt::skip] |
12139 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
12140 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
12141 | assert_eq_m512i(r, e); |
12142 | } |
12143 | |
12144 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12145 | unsafe fn test_mm256_mask_mulhi_epi16() { |
12146 | let a = _mm256_set1_epi16(1); |
12147 | let b = _mm256_set1_epi16(1); |
12148 | let r = _mm256_mask_mulhi_epi16(a, 0, a, b); |
12149 | assert_eq_m256i(r, a); |
12150 | let r = _mm256_mask_mulhi_epi16(a, 0b00000000_00001111, a, b); |
12151 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
12152 | assert_eq_m256i(r, e); |
12153 | } |
12154 | |
12155 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12156 | unsafe fn test_mm256_maskz_mulhi_epi16() { |
12157 | let a = _mm256_set1_epi16(1); |
12158 | let b = _mm256_set1_epi16(1); |
12159 | let r = _mm256_maskz_mulhi_epi16(0, a, b); |
12160 | assert_eq_m256i(r, _mm256_setzero_si256()); |
12161 | let r = _mm256_maskz_mulhi_epi16(0b00000000_00001111, a, b); |
12162 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
12163 | assert_eq_m256i(r, e); |
12164 | } |
12165 | |
12166 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12167 | unsafe fn test_mm_mask_mulhi_epi16() { |
12168 | let a = _mm_set1_epi16(1); |
12169 | let b = _mm_set1_epi16(1); |
12170 | let r = _mm_mask_mulhi_epi16(a, 0, a, b); |
12171 | assert_eq_m128i(r, a); |
12172 | let r = _mm_mask_mulhi_epi16(a, 0b00001111, a, b); |
12173 | let e = _mm_set_epi16(1, 1, 1, 1, 0, 0, 0, 0); |
12174 | assert_eq_m128i(r, e); |
12175 | } |
12176 | |
12177 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12178 | unsafe fn test_mm_maskz_mulhi_epi16() { |
12179 | let a = _mm_set1_epi16(1); |
12180 | let b = _mm_set1_epi16(1); |
12181 | let r = _mm_maskz_mulhi_epi16(0, a, b); |
12182 | assert_eq_m128i(r, _mm_setzero_si128()); |
12183 | let r = _mm_maskz_mulhi_epi16(0b00001111, a, b); |
12184 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 0, 0); |
12185 | assert_eq_m128i(r, e); |
12186 | } |
12187 | |
12188 | #[simd_test(enable = "avx512bw" )] |
12189 | unsafe fn test_mm512_mulhrs_epi16() { |
12190 | let a = _mm512_set1_epi16(1); |
12191 | let b = _mm512_set1_epi16(1); |
12192 | let r = _mm512_mulhrs_epi16(a, b); |
12193 | let e = _mm512_set1_epi16(0); |
12194 | assert_eq_m512i(r, e); |
12195 | } |
12196 | |
12197 | #[simd_test(enable = "avx512bw" )] |
12198 | unsafe fn test_mm512_mask_mulhrs_epi16() { |
12199 | let a = _mm512_set1_epi16(1); |
12200 | let b = _mm512_set1_epi16(1); |
12201 | let r = _mm512_mask_mulhrs_epi16(a, 0, a, b); |
12202 | assert_eq_m512i(r, a); |
12203 | let r = _mm512_mask_mulhrs_epi16(a, 0b00000000_00000000_00000000_00001111, a, b); |
12204 | #[rustfmt::skip] |
12205 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
12206 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
12207 | assert_eq_m512i(r, e); |
12208 | } |
12209 | |
12210 | #[simd_test(enable = "avx512bw" )] |
12211 | unsafe fn test_mm512_maskz_mulhrs_epi16() { |
12212 | let a = _mm512_set1_epi16(1); |
12213 | let b = _mm512_set1_epi16(1); |
12214 | let r = _mm512_maskz_mulhrs_epi16(0, a, b); |
12215 | assert_eq_m512i(r, _mm512_setzero_si512()); |
12216 | let r = _mm512_maskz_mulhrs_epi16(0b00000000_00000000_00000000_00001111, a, b); |
12217 | #[rustfmt::skip] |
12218 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
12219 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
12220 | assert_eq_m512i(r, e); |
12221 | } |
12222 | |
12223 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12224 | unsafe fn test_mm256_mask_mulhrs_epi16() { |
12225 | let a = _mm256_set1_epi16(1); |
12226 | let b = _mm256_set1_epi16(1); |
12227 | let r = _mm256_mask_mulhrs_epi16(a, 0, a, b); |
12228 | assert_eq_m256i(r, a); |
12229 | let r = _mm256_mask_mulhrs_epi16(a, 0b00000000_00001111, a, b); |
12230 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
12231 | assert_eq_m256i(r, e); |
12232 | } |
12233 | |
12234 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12235 | unsafe fn test_mm256_maskz_mulhrs_epi16() { |
12236 | let a = _mm256_set1_epi16(1); |
12237 | let b = _mm256_set1_epi16(1); |
12238 | let r = _mm256_maskz_mulhrs_epi16(0, a, b); |
12239 | assert_eq_m256i(r, _mm256_setzero_si256()); |
12240 | let r = _mm256_maskz_mulhrs_epi16(0b00000000_00001111, a, b); |
12241 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
12242 | assert_eq_m256i(r, e); |
12243 | } |
12244 | |
12245 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12246 | unsafe fn test_mm_mask_mulhrs_epi16() { |
12247 | let a = _mm_set1_epi16(1); |
12248 | let b = _mm_set1_epi16(1); |
12249 | let r = _mm_mask_mulhrs_epi16(a, 0, a, b); |
12250 | assert_eq_m128i(r, a); |
12251 | let r = _mm_mask_mulhrs_epi16(a, 0b00001111, a, b); |
12252 | let e = _mm_set_epi16(1, 1, 1, 1, 0, 0, 0, 0); |
12253 | assert_eq_m128i(r, e); |
12254 | } |
12255 | |
12256 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12257 | unsafe fn test_mm_maskz_mulhrs_epi16() { |
12258 | let a = _mm_set1_epi16(1); |
12259 | let b = _mm_set1_epi16(1); |
12260 | let r = _mm_maskz_mulhrs_epi16(0, a, b); |
12261 | assert_eq_m128i(r, _mm_setzero_si128()); |
12262 | let r = _mm_maskz_mulhrs_epi16(0b00001111, a, b); |
12263 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 0, 0); |
12264 | assert_eq_m128i(r, e); |
12265 | } |
12266 | |
12267 | #[simd_test(enable = "avx512bw" )] |
12268 | unsafe fn test_mm512_mullo_epi16() { |
12269 | let a = _mm512_set1_epi16(1); |
12270 | let b = _mm512_set1_epi16(1); |
12271 | let r = _mm512_mullo_epi16(a, b); |
12272 | let e = _mm512_set1_epi16(1); |
12273 | assert_eq_m512i(r, e); |
12274 | } |
12275 | |
12276 | #[simd_test(enable = "avx512bw" )] |
12277 | unsafe fn test_mm512_mask_mullo_epi16() { |
12278 | let a = _mm512_set1_epi16(1); |
12279 | let b = _mm512_set1_epi16(1); |
12280 | let r = _mm512_mask_mullo_epi16(a, 0, a, b); |
12281 | assert_eq_m512i(r, a); |
12282 | let r = _mm512_mask_mullo_epi16(a, 0b00000000_00000000_00000000_00001111, a, b); |
12283 | #[rustfmt::skip] |
12284 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
12285 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
12286 | assert_eq_m512i(r, e); |
12287 | } |
12288 | |
12289 | #[simd_test(enable = "avx512bw" )] |
12290 | unsafe fn test_mm512_maskz_mullo_epi16() { |
12291 | let a = _mm512_set1_epi16(1); |
12292 | let b = _mm512_set1_epi16(1); |
12293 | let r = _mm512_maskz_mullo_epi16(0, a, b); |
12294 | assert_eq_m512i(r, _mm512_setzero_si512()); |
12295 | let r = _mm512_maskz_mullo_epi16(0b00000000_00000000_00000000_00001111, a, b); |
12296 | #[rustfmt::skip] |
12297 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
12298 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
12299 | assert_eq_m512i(r, e); |
12300 | } |
12301 | |
12302 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12303 | unsafe fn test_mm256_mask_mullo_epi16() { |
12304 | let a = _mm256_set1_epi16(1); |
12305 | let b = _mm256_set1_epi16(1); |
12306 | let r = _mm256_mask_mullo_epi16(a, 0, a, b); |
12307 | assert_eq_m256i(r, a); |
12308 | let r = _mm256_mask_mullo_epi16(a, 0b00000000_00001111, a, b); |
12309 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
12310 | assert_eq_m256i(r, e); |
12311 | } |
12312 | |
12313 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12314 | unsafe fn test_mm256_maskz_mullo_epi16() { |
12315 | let a = _mm256_set1_epi16(1); |
12316 | let b = _mm256_set1_epi16(1); |
12317 | let r = _mm256_maskz_mullo_epi16(0, a, b); |
12318 | assert_eq_m256i(r, _mm256_setzero_si256()); |
12319 | let r = _mm256_maskz_mullo_epi16(0b00000000_00001111, a, b); |
12320 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
12321 | assert_eq_m256i(r, e); |
12322 | } |
12323 | |
12324 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12325 | unsafe fn test_mm_mask_mullo_epi16() { |
12326 | let a = _mm_set1_epi16(1); |
12327 | let b = _mm_set1_epi16(1); |
12328 | let r = _mm_mask_mullo_epi16(a, 0, a, b); |
12329 | assert_eq_m128i(r, a); |
12330 | let r = _mm_mask_mullo_epi16(a, 0b00001111, a, b); |
12331 | let e = _mm_set_epi16(1, 1, 1, 1, 1, 1, 1, 1); |
12332 | assert_eq_m128i(r, e); |
12333 | } |
12334 | |
12335 | #[simd_test(enable = "avx512bw,avx512vl" )] |
12336 | unsafe fn test_mm_maskz_mullo_epi16() { |
12337 | let a = _mm_set1_epi16(1); |
12338 | let b = _mm_set1_epi16(1); |
12339 | let r = _mm_maskz_mullo_epi16(0, a, b); |
12340 | assert_eq_m128i(r, _mm_setzero_si128()); |
12341 | let r = _mm_maskz_mullo_epi16(0b00001111, a, b); |
12342 | let e = _mm_set_epi16(0, 0, 0, 0, 1, 1, 1, 1); |
12343 | assert_eq_m128i(r, e); |
12344 | } |
12345 | |
12346 | #[simd_test(enable = "avx512bw" )] |
12347 | unsafe fn test_mm512_max_epu16() { |
12348 | #[rustfmt::skip] |
12349 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12350 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12351 | #[rustfmt::skip] |
12352 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12353 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12354 | let r = _mm512_max_epu16(a, b); |
12355 | #[rustfmt::skip] |
12356 | let e = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
12357 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15); |
12358 | assert_eq_m512i(r, e); |
12359 | } |
12360 | |
12361 | #[simd_test(enable = "avx512f" )] |
12362 | unsafe fn test_mm512_mask_max_epu16() { |
12363 | #[rustfmt::skip] |
12364 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12365 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12366 | #[rustfmt::skip] |
12367 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12368 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12369 | let r = _mm512_mask_max_epu16(a, 0, a, b); |
12370 | assert_eq_m512i(r, a); |
12371 | let r = _mm512_mask_max_epu16(a, 0b00000000_11111111_00000000_11111111, a, b); |
12372 | #[rustfmt::skip] |
12373 | let e = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12374 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12375 | assert_eq_m512i(r, e); |
12376 | } |
12377 | |
12378 | #[simd_test(enable = "avx512f" )] |
12379 | unsafe fn test_mm512_maskz_max_epu16() { |
12380 | #[rustfmt::skip] |
12381 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12382 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12383 | #[rustfmt::skip] |
12384 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12385 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12386 | let r = _mm512_maskz_max_epu16(0, a, b); |
12387 | assert_eq_m512i(r, _mm512_setzero_si512()); |
12388 | let r = _mm512_maskz_max_epu16(0b00000000_11111111_00000000_11111111, a, b); |
12389 | #[rustfmt::skip] |
12390 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
12391 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
12392 | assert_eq_m512i(r, e); |
12393 | } |
12394 | |
12395 | #[simd_test(enable = "avx512f,avx512vl" )] |
12396 | unsafe fn test_mm256_mask_max_epu16() { |
12397 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12398 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12399 | let r = _mm256_mask_max_epu16(a, 0, a, b); |
12400 | assert_eq_m256i(r, a); |
12401 | let r = _mm256_mask_max_epu16(a, 0b00000000_11111111, a, b); |
12402 | let e = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12403 | assert_eq_m256i(r, e); |
12404 | } |
12405 | |
12406 | #[simd_test(enable = "avx512f,avx512vl" )] |
12407 | unsafe fn test_mm256_maskz_max_epu16() { |
12408 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12409 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12410 | let r = _mm256_maskz_max_epu16(0, a, b); |
12411 | assert_eq_m256i(r, _mm256_setzero_si256()); |
12412 | let r = _mm256_maskz_max_epu16(0b00000000_11111111, a, b); |
12413 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
12414 | assert_eq_m256i(r, e); |
12415 | } |
12416 | |
12417 | #[simd_test(enable = "avx512f,avx512vl" )] |
12418 | unsafe fn test_mm_mask_max_epu16() { |
12419 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
12420 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
12421 | let r = _mm_mask_max_epu16(a, 0, a, b); |
12422 | assert_eq_m128i(r, a); |
12423 | let r = _mm_mask_max_epu16(a, 0b00001111, a, b); |
12424 | let e = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
12425 | assert_eq_m128i(r, e); |
12426 | } |
12427 | |
12428 | #[simd_test(enable = "avx512f,avx512vl" )] |
12429 | unsafe fn test_mm_maskz_max_epu16() { |
12430 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
12431 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
12432 | let r = _mm_maskz_max_epu16(0, a, b); |
12433 | assert_eq_m128i(r, _mm_setzero_si128()); |
12434 | let r = _mm_maskz_max_epu16(0b00001111, a, b); |
12435 | let e = _mm_set_epi16(0, 0, 0, 0, 4, 5, 6, 7); |
12436 | assert_eq_m128i(r, e); |
12437 | } |
12438 | |
12439 | #[simd_test(enable = "avx512bw" )] |
12440 | unsafe fn test_mm512_max_epu8() { |
12441 | #[rustfmt::skip] |
12442 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12443 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12444 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12445 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12446 | #[rustfmt::skip] |
12447 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12448 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12449 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12450 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12451 | let r = _mm512_max_epu8(a, b); |
12452 | #[rustfmt::skip] |
12453 | let e = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
12454 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
12455 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
12456 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15); |
12457 | assert_eq_m512i(r, e); |
12458 | } |
12459 | |
12460 | #[simd_test(enable = "avx512f" )] |
12461 | unsafe fn test_mm512_mask_max_epu8() { |
12462 | #[rustfmt::skip] |
12463 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12464 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12465 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12466 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12467 | #[rustfmt::skip] |
12468 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12469 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12470 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12471 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12472 | let r = _mm512_mask_max_epu8(a, 0, a, b); |
12473 | assert_eq_m512i(r, a); |
12474 | let r = _mm512_mask_max_epu8( |
12475 | a, |
12476 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
12477 | a, |
12478 | b, |
12479 | ); |
12480 | #[rustfmt::skip] |
12481 | let e = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12482 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12483 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12484 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12485 | assert_eq_m512i(r, e); |
12486 | } |
12487 | |
12488 | #[simd_test(enable = "avx512f" )] |
12489 | unsafe fn test_mm512_maskz_max_epu8() { |
12490 | #[rustfmt::skip] |
12491 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12492 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12493 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12494 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12495 | #[rustfmt::skip] |
12496 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12497 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12498 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12499 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12500 | let r = _mm512_maskz_max_epu8(0, a, b); |
12501 | assert_eq_m512i(r, _mm512_setzero_si512()); |
12502 | let r = _mm512_maskz_max_epu8( |
12503 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
12504 | a, |
12505 | b, |
12506 | ); |
12507 | #[rustfmt::skip] |
12508 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
12509 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
12510 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
12511 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
12512 | assert_eq_m512i(r, e); |
12513 | } |
12514 | |
12515 | #[simd_test(enable = "avx512f,avx512vl" )] |
12516 | unsafe fn test_mm256_mask_max_epu8() { |
12517 | #[rustfmt::skip] |
12518 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12519 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12520 | #[rustfmt::skip] |
12521 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12522 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12523 | let r = _mm256_mask_max_epu8(a, 0, a, b); |
12524 | assert_eq_m256i(r, a); |
12525 | let r = _mm256_mask_max_epu8(a, 0b00000000_11111111_00000000_11111111, a, b); |
12526 | #[rustfmt::skip] |
12527 | let e = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12528 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12529 | assert_eq_m256i(r, e); |
12530 | } |
12531 | |
12532 | #[simd_test(enable = "avx512f,avx512vl" )] |
12533 | unsafe fn test_mm256_maskz_max_epu8() { |
12534 | #[rustfmt::skip] |
12535 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12536 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12537 | #[rustfmt::skip] |
12538 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12539 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12540 | let r = _mm256_maskz_max_epu8(0, a, b); |
12541 | assert_eq_m256i(r, _mm256_setzero_si256()); |
12542 | let r = _mm256_maskz_max_epu8(0b00000000_11111111_00000000_11111111, a, b); |
12543 | #[rustfmt::skip] |
12544 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
12545 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
12546 | assert_eq_m256i(r, e); |
12547 | } |
12548 | |
12549 | #[simd_test(enable = "avx512f,avx512vl" )] |
12550 | unsafe fn test_mm_mask_max_epu8() { |
12551 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12552 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12553 | let r = _mm_mask_max_epu8(a, 0, a, b); |
12554 | assert_eq_m128i(r, a); |
12555 | let r = _mm_mask_max_epu8(a, 0b00000000_11111111, a, b); |
12556 | let e = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12557 | assert_eq_m128i(r, e); |
12558 | } |
12559 | |
12560 | #[simd_test(enable = "avx512f,avx512vl" )] |
12561 | unsafe fn test_mm_maskz_max_epu8() { |
12562 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12563 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12564 | let r = _mm_maskz_max_epu8(0, a, b); |
12565 | assert_eq_m128i(r, _mm_setzero_si128()); |
12566 | let r = _mm_maskz_max_epu8(0b00000000_11111111, a, b); |
12567 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
12568 | assert_eq_m128i(r, e); |
12569 | } |
12570 | |
12571 | #[simd_test(enable = "avx512bw" )] |
12572 | unsafe fn test_mm512_max_epi16() { |
12573 | #[rustfmt::skip] |
12574 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12575 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12576 | #[rustfmt::skip] |
12577 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12578 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12579 | let r = _mm512_max_epi16(a, b); |
12580 | #[rustfmt::skip] |
12581 | let e = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
12582 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15); |
12583 | assert_eq_m512i(r, e); |
12584 | } |
12585 | |
12586 | #[simd_test(enable = "avx512f" )] |
12587 | unsafe fn test_mm512_mask_max_epi16() { |
12588 | #[rustfmt::skip] |
12589 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12590 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12591 | #[rustfmt::skip] |
12592 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12593 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12594 | let r = _mm512_mask_max_epi16(a, 0, a, b); |
12595 | assert_eq_m512i(r, a); |
12596 | let r = _mm512_mask_max_epi16(a, 0b00000000_11111111_00000000_11111111, a, b); |
12597 | #[rustfmt::skip] |
12598 | let e = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12599 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12600 | assert_eq_m512i(r, e); |
12601 | } |
12602 | |
12603 | #[simd_test(enable = "avx512f" )] |
12604 | unsafe fn test_mm512_maskz_max_epi16() { |
12605 | #[rustfmt::skip] |
12606 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12607 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12608 | #[rustfmt::skip] |
12609 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12610 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12611 | let r = _mm512_maskz_max_epi16(0, a, b); |
12612 | assert_eq_m512i(r, _mm512_setzero_si512()); |
12613 | let r = _mm512_maskz_max_epi16(0b00000000_11111111_00000000_11111111, a, b); |
12614 | #[rustfmt::skip] |
12615 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
12616 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
12617 | assert_eq_m512i(r, e); |
12618 | } |
12619 | |
12620 | #[simd_test(enable = "avx512f,avx512vl" )] |
12621 | unsafe fn test_mm256_mask_max_epi16() { |
12622 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12623 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12624 | let r = _mm256_mask_max_epi16(a, 0, a, b); |
12625 | assert_eq_m256i(r, a); |
12626 | let r = _mm256_mask_max_epi16(a, 0b00000000_11111111, a, b); |
12627 | let e = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12628 | assert_eq_m256i(r, e); |
12629 | } |
12630 | |
12631 | #[simd_test(enable = "avx512f,avx512vl" )] |
12632 | unsafe fn test_mm256_maskz_max_epi16() { |
12633 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12634 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12635 | let r = _mm256_maskz_max_epi16(0, a, b); |
12636 | assert_eq_m256i(r, _mm256_setzero_si256()); |
12637 | let r = _mm256_maskz_max_epi16(0b00000000_11111111, a, b); |
12638 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
12639 | assert_eq_m256i(r, e); |
12640 | } |
12641 | |
12642 | #[simd_test(enable = "avx512f,avx512vl" )] |
12643 | unsafe fn test_mm_mask_max_epi16() { |
12644 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
12645 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
12646 | let r = _mm_mask_max_epi16(a, 0, a, b); |
12647 | assert_eq_m128i(r, a); |
12648 | let r = _mm_mask_max_epi16(a, 0b00001111, a, b); |
12649 | let e = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
12650 | assert_eq_m128i(r, e); |
12651 | } |
12652 | |
12653 | #[simd_test(enable = "avx512f,avx512vl" )] |
12654 | unsafe fn test_mm_maskz_max_epi16() { |
12655 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
12656 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
12657 | let r = _mm_maskz_max_epi16(0, a, b); |
12658 | assert_eq_m128i(r, _mm_setzero_si128()); |
12659 | let r = _mm_maskz_max_epi16(0b00001111, a, b); |
12660 | let e = _mm_set_epi16(0, 0, 0, 0, 4, 5, 6, 7); |
12661 | assert_eq_m128i(r, e); |
12662 | } |
12663 | |
12664 | #[simd_test(enable = "avx512bw" )] |
12665 | unsafe fn test_mm512_max_epi8() { |
12666 | #[rustfmt::skip] |
12667 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12668 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12669 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12670 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12671 | #[rustfmt::skip] |
12672 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12673 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12674 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12675 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12676 | let r = _mm512_max_epi8(a, b); |
12677 | #[rustfmt::skip] |
12678 | let e = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
12679 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
12680 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15, |
12681 | 15, 14, 13, 12, 11, 10, 9, 8, 8, 9, 10, 11, 12, 13, 14, 15); |
12682 | assert_eq_m512i(r, e); |
12683 | } |
12684 | |
12685 | #[simd_test(enable = "avx512f" )] |
12686 | unsafe fn test_mm512_mask_max_epi8() { |
12687 | #[rustfmt::skip] |
12688 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12689 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12690 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12691 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12692 | #[rustfmt::skip] |
12693 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12694 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12695 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12696 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12697 | let r = _mm512_mask_max_epi8(a, 0, a, b); |
12698 | assert_eq_m512i(r, a); |
12699 | let r = _mm512_mask_max_epi8( |
12700 | a, |
12701 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
12702 | a, |
12703 | b, |
12704 | ); |
12705 | #[rustfmt::skip] |
12706 | let e = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12707 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12708 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12709 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12710 | assert_eq_m512i(r, e); |
12711 | } |
12712 | |
12713 | #[simd_test(enable = "avx512f" )] |
12714 | unsafe fn test_mm512_maskz_max_epi8() { |
12715 | #[rustfmt::skip] |
12716 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12717 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12718 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12719 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12720 | #[rustfmt::skip] |
12721 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12722 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12723 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12724 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12725 | let r = _mm512_maskz_max_epi8(0, a, b); |
12726 | assert_eq_m512i(r, _mm512_setzero_si512()); |
12727 | let r = _mm512_maskz_max_epi8( |
12728 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
12729 | a, |
12730 | b, |
12731 | ); |
12732 | #[rustfmt::skip] |
12733 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
12734 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
12735 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
12736 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
12737 | assert_eq_m512i(r, e); |
12738 | } |
12739 | |
12740 | #[simd_test(enable = "avx512f,avx512vl" )] |
12741 | unsafe fn test_mm256_mask_max_epi8() { |
12742 | #[rustfmt::skip] |
12743 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12744 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12745 | #[rustfmt::skip] |
12746 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12747 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12748 | let r = _mm256_mask_max_epi8(a, 0, a, b); |
12749 | assert_eq_m256i(r, a); |
12750 | let r = _mm256_mask_max_epi8(a, 0b00000000_11111111_00000000_11111111, a, b); |
12751 | #[rustfmt::skip] |
12752 | let e = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12753 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12754 | assert_eq_m256i(r, e); |
12755 | } |
12756 | |
12757 | #[simd_test(enable = "avx512f,avx512vl" )] |
12758 | unsafe fn test_mm256_maskz_max_epi8() { |
12759 | #[rustfmt::skip] |
12760 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12761 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12762 | #[rustfmt::skip] |
12763 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12764 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12765 | let r = _mm256_maskz_max_epi8(0, a, b); |
12766 | assert_eq_m256i(r, _mm256_setzero_si256()); |
12767 | let r = _mm256_maskz_max_epi8(0b00000000_11111111_00000000_11111111, a, b); |
12768 | #[rustfmt::skip] |
12769 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15, |
12770 | 0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
12771 | assert_eq_m256i(r, e); |
12772 | } |
12773 | |
12774 | #[simd_test(enable = "avx512f,avx512vl" )] |
12775 | unsafe fn test_mm_mask_max_epi8() { |
12776 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12777 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12778 | let r = _mm_mask_max_epi8(a, 0, a, b); |
12779 | assert_eq_m128i(r, a); |
12780 | let r = _mm_mask_max_epi8(a, 0b00000000_11111111, a, b); |
12781 | let e = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12782 | assert_eq_m128i(r, e); |
12783 | } |
12784 | |
12785 | #[simd_test(enable = "avx512f,avx512vl" )] |
12786 | unsafe fn test_mm_maskz_max_epi8() { |
12787 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12788 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12789 | let r = _mm_maskz_max_epi8(0, a, b); |
12790 | assert_eq_m128i(r, _mm_setzero_si128()); |
12791 | let r = _mm_maskz_max_epi8(0b00000000_11111111, a, b); |
12792 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
12793 | assert_eq_m128i(r, e); |
12794 | } |
12795 | |
12796 | #[simd_test(enable = "avx512bw" )] |
12797 | unsafe fn test_mm512_min_epu16() { |
12798 | #[rustfmt::skip] |
12799 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12800 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12801 | #[rustfmt::skip] |
12802 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12803 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12804 | let r = _mm512_min_epu16(a, b); |
12805 | #[rustfmt::skip] |
12806 | let e = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
12807 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
12808 | assert_eq_m512i(r, e); |
12809 | } |
12810 | |
12811 | #[simd_test(enable = "avx512f" )] |
12812 | unsafe fn test_mm512_mask_min_epu16() { |
12813 | #[rustfmt::skip] |
12814 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12815 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12816 | #[rustfmt::skip] |
12817 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12818 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12819 | let r = _mm512_mask_min_epu16(a, 0, a, b); |
12820 | assert_eq_m512i(r, a); |
12821 | let r = _mm512_mask_min_epu16(a, 0b00000000_11111111_00000000_11111111, a, b); |
12822 | #[rustfmt::skip] |
12823 | let e = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
12824 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
12825 | assert_eq_m512i(r, e); |
12826 | } |
12827 | |
12828 | #[simd_test(enable = "avx512f" )] |
12829 | unsafe fn test_mm512_maskz_min_epu16() { |
12830 | #[rustfmt::skip] |
12831 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12832 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12833 | #[rustfmt::skip] |
12834 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12835 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12836 | let r = _mm512_maskz_min_epu16(0, a, b); |
12837 | assert_eq_m512i(r, _mm512_setzero_si512()); |
12838 | let r = _mm512_maskz_min_epu16(0b00000000_11111111_00000000_11111111, a, b); |
12839 | #[rustfmt::skip] |
12840 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
12841 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
12842 | assert_eq_m512i(r, e); |
12843 | } |
12844 | |
12845 | #[simd_test(enable = "avx512f,avx512vl" )] |
12846 | unsafe fn test_mm256_mask_min_epu16() { |
12847 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12848 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12849 | let r = _mm256_mask_min_epu16(a, 0, a, b); |
12850 | assert_eq_m256i(r, a); |
12851 | let r = _mm256_mask_min_epu16(a, 0b00000000_11111111, a, b); |
12852 | let e = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
12853 | assert_eq_m256i(r, e); |
12854 | } |
12855 | |
12856 | #[simd_test(enable = "avx512f,avx512vl" )] |
12857 | unsafe fn test_mm256_maskz_min_epu16() { |
12858 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12859 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12860 | let r = _mm256_maskz_min_epu16(0, a, b); |
12861 | assert_eq_m256i(r, _mm256_setzero_si256()); |
12862 | let r = _mm256_maskz_min_epu16(0b00000000_11111111, a, b); |
12863 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
12864 | assert_eq_m256i(r, e); |
12865 | } |
12866 | |
12867 | #[simd_test(enable = "avx512f,avx512vl" )] |
12868 | unsafe fn test_mm_mask_min_epu16() { |
12869 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
12870 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
12871 | let r = _mm_mask_min_epu16(a, 0, a, b); |
12872 | assert_eq_m128i(r, a); |
12873 | let r = _mm_mask_min_epu16(a, 0b00001111, a, b); |
12874 | let e = _mm_set_epi16(0, 1, 2, 3, 3, 2, 1, 0); |
12875 | assert_eq_m128i(r, e); |
12876 | } |
12877 | |
12878 | #[simd_test(enable = "avx512f,avx512vl" )] |
12879 | unsafe fn test_mm_maskz_min_epu16() { |
12880 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
12881 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
12882 | let r = _mm_maskz_min_epu16(0, a, b); |
12883 | assert_eq_m128i(r, _mm_setzero_si128()); |
12884 | let r = _mm_maskz_min_epu16(0b00001111, a, b); |
12885 | let e = _mm_set_epi16(0, 0, 0, 0, 3, 2, 1, 0); |
12886 | assert_eq_m128i(r, e); |
12887 | } |
12888 | |
12889 | #[simd_test(enable = "avx512bw" )] |
12890 | unsafe fn test_mm512_min_epu8() { |
12891 | #[rustfmt::skip] |
12892 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12893 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12894 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12895 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12896 | #[rustfmt::skip] |
12897 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12898 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12899 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12900 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12901 | let r = _mm512_min_epu8(a, b); |
12902 | #[rustfmt::skip] |
12903 | let e = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
12904 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
12905 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
12906 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
12907 | assert_eq_m512i(r, e); |
12908 | } |
12909 | |
12910 | #[simd_test(enable = "avx512f" )] |
12911 | unsafe fn test_mm512_mask_min_epu8() { |
12912 | #[rustfmt::skip] |
12913 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12914 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12915 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12916 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12917 | #[rustfmt::skip] |
12918 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12919 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12920 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12921 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12922 | let r = _mm512_mask_min_epu8(a, 0, a, b); |
12923 | assert_eq_m512i(r, a); |
12924 | let r = _mm512_mask_min_epu8( |
12925 | a, |
12926 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
12927 | a, |
12928 | b, |
12929 | ); |
12930 | #[rustfmt::skip] |
12931 | let e = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
12932 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
12933 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
12934 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
12935 | assert_eq_m512i(r, e); |
12936 | } |
12937 | |
12938 | #[simd_test(enable = "avx512f" )] |
12939 | unsafe fn test_mm512_maskz_min_epu8() { |
12940 | #[rustfmt::skip] |
12941 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12942 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12943 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12944 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12945 | #[rustfmt::skip] |
12946 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12947 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12948 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12949 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12950 | let r = _mm512_maskz_min_epu8(0, a, b); |
12951 | assert_eq_m512i(r, _mm512_setzero_si512()); |
12952 | let r = _mm512_maskz_min_epu8( |
12953 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
12954 | a, |
12955 | b, |
12956 | ); |
12957 | #[rustfmt::skip] |
12958 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
12959 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
12960 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
12961 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
12962 | assert_eq_m512i(r, e); |
12963 | } |
12964 | |
12965 | #[simd_test(enable = "avx512f,avx512vl" )] |
12966 | unsafe fn test_mm256_mask_min_epu8() { |
12967 | #[rustfmt::skip] |
12968 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12969 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12970 | #[rustfmt::skip] |
12971 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12972 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12973 | let r = _mm256_mask_min_epu8(a, 0, a, b); |
12974 | assert_eq_m256i(r, a); |
12975 | let r = _mm256_mask_min_epu8(a, 0b00000000_11111111_00000000_11111111, a, b); |
12976 | #[rustfmt::skip] |
12977 | let e = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
12978 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
12979 | assert_eq_m256i(r, e); |
12980 | } |
12981 | |
12982 | #[simd_test(enable = "avx512f,avx512vl" )] |
12983 | unsafe fn test_mm256_maskz_min_epu8() { |
12984 | #[rustfmt::skip] |
12985 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
12986 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
12987 | #[rustfmt::skip] |
12988 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
12989 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
12990 | let r = _mm256_maskz_min_epu8(0, a, b); |
12991 | assert_eq_m256i(r, _mm256_setzero_si256()); |
12992 | let r = _mm256_maskz_min_epu8(0b00000000_11111111_00000000_11111111, a, b); |
12993 | #[rustfmt::skip] |
12994 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
12995 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
12996 | assert_eq_m256i(r, e); |
12997 | } |
12998 | |
12999 | #[simd_test(enable = "avx512f,avx512vl" )] |
13000 | unsafe fn test_mm_mask_min_epu8() { |
13001 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13002 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13003 | let r = _mm_mask_min_epu8(a, 0, a, b); |
13004 | assert_eq_m128i(r, a); |
13005 | let r = _mm_mask_min_epu8(a, 0b00000000_11111111, a, b); |
13006 | let e = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
13007 | assert_eq_m128i(r, e); |
13008 | } |
13009 | |
13010 | #[simd_test(enable = "avx512f,avx512vl" )] |
13011 | unsafe fn test_mm_maskz_min_epu8() { |
13012 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13013 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13014 | let r = _mm_maskz_min_epu8(0, a, b); |
13015 | assert_eq_m128i(r, _mm_setzero_si128()); |
13016 | let r = _mm_maskz_min_epu8(0b00000000_11111111, a, b); |
13017 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
13018 | assert_eq_m128i(r, e); |
13019 | } |
13020 | |
13021 | #[simd_test(enable = "avx512bw" )] |
13022 | unsafe fn test_mm512_min_epi16() { |
13023 | #[rustfmt::skip] |
13024 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13025 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13026 | #[rustfmt::skip] |
13027 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13028 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13029 | let r = _mm512_min_epi16(a, b); |
13030 | #[rustfmt::skip] |
13031 | let e = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
13032 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
13033 | assert_eq_m512i(r, e); |
13034 | } |
13035 | |
13036 | #[simd_test(enable = "avx512f" )] |
13037 | unsafe fn test_mm512_mask_min_epi16() { |
13038 | #[rustfmt::skip] |
13039 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13040 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13041 | #[rustfmt::skip] |
13042 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13043 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13044 | let r = _mm512_mask_min_epi16(a, 0, a, b); |
13045 | assert_eq_m512i(r, a); |
13046 | let r = _mm512_mask_min_epi16(a, 0b00000000_11111111_00000000_11111111, a, b); |
13047 | #[rustfmt::skip] |
13048 | let e = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
13049 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
13050 | assert_eq_m512i(r, e); |
13051 | } |
13052 | |
13053 | #[simd_test(enable = "avx512f" )] |
13054 | unsafe fn test_mm512_maskz_min_epi16() { |
13055 | #[rustfmt::skip] |
13056 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13057 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13058 | #[rustfmt::skip] |
13059 | let b = _mm512_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13060 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13061 | let r = _mm512_maskz_min_epi16(0, a, b); |
13062 | assert_eq_m512i(r, _mm512_setzero_si512()); |
13063 | let r = _mm512_maskz_min_epi16(0b00000000_11111111_00000000_11111111, a, b); |
13064 | #[rustfmt::skip] |
13065 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
13066 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
13067 | assert_eq_m512i(r, e); |
13068 | } |
13069 | |
13070 | #[simd_test(enable = "avx512f,avx512vl" )] |
13071 | unsafe fn test_mm256_mask_min_epi16() { |
13072 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13073 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13074 | let r = _mm256_mask_min_epi16(a, 0, a, b); |
13075 | assert_eq_m256i(r, a); |
13076 | let r = _mm256_mask_min_epi16(a, 0b00000000_11111111, a, b); |
13077 | let e = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
13078 | assert_eq_m256i(r, e); |
13079 | } |
13080 | |
13081 | #[simd_test(enable = "avx512f,avx512vl" )] |
13082 | unsafe fn test_mm256_maskz_min_epi16() { |
13083 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13084 | let b = _mm256_set_epi16(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13085 | let r = _mm256_maskz_min_epi16(0, a, b); |
13086 | assert_eq_m256i(r, _mm256_setzero_si256()); |
13087 | let r = _mm256_maskz_min_epi16(0b00000000_11111111, a, b); |
13088 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
13089 | assert_eq_m256i(r, e); |
13090 | } |
13091 | |
13092 | #[simd_test(enable = "avx512f,avx512vl" )] |
13093 | unsafe fn test_mm_mask_min_epi16() { |
13094 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
13095 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
13096 | let r = _mm_mask_min_epi16(a, 0, a, b); |
13097 | assert_eq_m128i(r, a); |
13098 | let r = _mm_mask_min_epi16(a, 0b00001111, a, b); |
13099 | let e = _mm_set_epi16(0, 1, 2, 3, 3, 2, 1, 0); |
13100 | assert_eq_m128i(r, e); |
13101 | } |
13102 | |
13103 | #[simd_test(enable = "avx512f,avx512vl" )] |
13104 | unsafe fn test_mm_maskz_min_epi16() { |
13105 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
13106 | let b = _mm_set_epi16(7, 6, 5, 4, 3, 2, 1, 0); |
13107 | let r = _mm_maskz_min_epi16(0, a, b); |
13108 | assert_eq_m128i(r, _mm_setzero_si128()); |
13109 | let r = _mm_maskz_min_epi16(0b00001111, a, b); |
13110 | let e = _mm_set_epi16(0, 0, 0, 0, 3, 2, 1, 0); |
13111 | assert_eq_m128i(r, e); |
13112 | } |
13113 | |
13114 | #[simd_test(enable = "avx512bw" )] |
13115 | unsafe fn test_mm512_min_epi8() { |
13116 | #[rustfmt::skip] |
13117 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13118 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13119 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13120 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13121 | #[rustfmt::skip] |
13122 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13123 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13124 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13125 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13126 | let r = _mm512_min_epi8(a, b); |
13127 | #[rustfmt::skip] |
13128 | let e = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
13129 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
13130 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
13131 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
13132 | assert_eq_m512i(r, e); |
13133 | } |
13134 | |
13135 | #[simd_test(enable = "avx512f" )] |
13136 | unsafe fn test_mm512_mask_min_epi8() { |
13137 | #[rustfmt::skip] |
13138 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13139 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13140 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13141 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13142 | #[rustfmt::skip] |
13143 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13144 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13145 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13146 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13147 | let r = _mm512_mask_min_epi8(a, 0, a, b); |
13148 | assert_eq_m512i(r, a); |
13149 | let r = _mm512_mask_min_epi8( |
13150 | a, |
13151 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
13152 | a, |
13153 | b, |
13154 | ); |
13155 | #[rustfmt::skip] |
13156 | let e = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
13157 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
13158 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
13159 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
13160 | assert_eq_m512i(r, e); |
13161 | } |
13162 | |
13163 | #[simd_test(enable = "avx512f" )] |
13164 | unsafe fn test_mm512_maskz_min_epi8() { |
13165 | #[rustfmt::skip] |
13166 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13167 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13168 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13169 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13170 | #[rustfmt::skip] |
13171 | let b = _mm512_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13172 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13173 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13174 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13175 | let r = _mm512_maskz_min_epi8(0, a, b); |
13176 | assert_eq_m512i(r, _mm512_setzero_si512()); |
13177 | let r = _mm512_maskz_min_epi8( |
13178 | 0b00000000_11111111_00000000_11111111_00000000_11111111_00000000_11111111, |
13179 | a, |
13180 | b, |
13181 | ); |
13182 | #[rustfmt::skip] |
13183 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
13184 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
13185 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
13186 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
13187 | assert_eq_m512i(r, e); |
13188 | } |
13189 | |
13190 | #[simd_test(enable = "avx512f,avx512vl" )] |
13191 | unsafe fn test_mm256_mask_min_epi8() { |
13192 | #[rustfmt::skip] |
13193 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13194 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13195 | #[rustfmt::skip] |
13196 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13197 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13198 | let r = _mm256_mask_min_epi8(a, 0, a, b); |
13199 | assert_eq_m256i(r, a); |
13200 | let r = _mm256_mask_min_epi8(a, 0b00000000_11111111_00000000_11111111, a, b); |
13201 | #[rustfmt::skip] |
13202 | let e = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0, |
13203 | 0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
13204 | assert_eq_m256i(r, e); |
13205 | } |
13206 | |
13207 | #[simd_test(enable = "avx512f,avx512vl" )] |
13208 | unsafe fn test_mm256_maskz_min_epi8() { |
13209 | #[rustfmt::skip] |
13210 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
13211 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13212 | #[rustfmt::skip] |
13213 | let b = _mm256_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, |
13214 | 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13215 | let r = _mm256_maskz_min_epi8(0, a, b); |
13216 | assert_eq_m256i(r, _mm256_setzero_si256()); |
13217 | let r = _mm256_maskz_min_epi8(0b00000000_11111111_00000000_11111111, a, b); |
13218 | #[rustfmt::skip] |
13219 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0, |
13220 | 0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
13221 | assert_eq_m256i(r, e); |
13222 | } |
13223 | |
13224 | #[simd_test(enable = "avx512f,avx512vl" )] |
13225 | unsafe fn test_mm_mask_min_epi8() { |
13226 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13227 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13228 | let r = _mm_mask_min_epi8(a, 0, a, b); |
13229 | assert_eq_m128i(r, a); |
13230 | let r = _mm_mask_min_epi8(a, 0b00000000_11111111, a, b); |
13231 | let e = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 7, 6, 5, 4, 3, 2, 1, 0); |
13232 | assert_eq_m128i(r, e); |
13233 | } |
13234 | |
13235 | #[simd_test(enable = "avx512f,avx512vl" )] |
13236 | unsafe fn test_mm_maskz_min_epi8() { |
13237 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
13238 | let b = _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); |
13239 | let r = _mm_maskz_min_epi8(0, a, b); |
13240 | assert_eq_m128i(r, _mm_setzero_si128()); |
13241 | let r = _mm_maskz_min_epi8(0b00000000_11111111, a, b); |
13242 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 7, 6, 5, 4, 3, 2, 1, 0); |
13243 | assert_eq_m128i(r, e); |
13244 | } |
13245 | |
13246 | #[simd_test(enable = "avx512bw" )] |
13247 | unsafe fn test_mm512_cmplt_epu16_mask() { |
13248 | let a = _mm512_set1_epi16(-2); |
13249 | let b = _mm512_set1_epi16(-1); |
13250 | let m = _mm512_cmplt_epu16_mask(a, b); |
13251 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13252 | } |
13253 | |
13254 | #[simd_test(enable = "avx512bw" )] |
13255 | unsafe fn test_mm512_mask_cmplt_epu16_mask() { |
13256 | let a = _mm512_set1_epi16(-2); |
13257 | let b = _mm512_set1_epi16(-1); |
13258 | let mask = 0b01010101_01010101_01010101_01010101; |
13259 | let r = _mm512_mask_cmplt_epu16_mask(mask, a, b); |
13260 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13261 | } |
13262 | |
13263 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13264 | unsafe fn test_mm256_cmplt_epu16_mask() { |
13265 | let a = _mm256_set1_epi16(-2); |
13266 | let b = _mm256_set1_epi16(-1); |
13267 | let m = _mm256_cmplt_epu16_mask(a, b); |
13268 | assert_eq!(m, 0b11111111_11111111); |
13269 | } |
13270 | |
13271 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13272 | unsafe fn test_mm256_mask_cmplt_epu16_mask() { |
13273 | let a = _mm256_set1_epi16(-2); |
13274 | let b = _mm256_set1_epi16(-1); |
13275 | let mask = 0b01010101_01010101; |
13276 | let r = _mm256_mask_cmplt_epu16_mask(mask, a, b); |
13277 | assert_eq!(r, 0b01010101_01010101); |
13278 | } |
13279 | |
13280 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13281 | unsafe fn test_mm_cmplt_epu16_mask() { |
13282 | let a = _mm_set1_epi16(-2); |
13283 | let b = _mm_set1_epi16(-1); |
13284 | let m = _mm_cmplt_epu16_mask(a, b); |
13285 | assert_eq!(m, 0b11111111); |
13286 | } |
13287 | |
13288 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13289 | unsafe fn test_mm_mask_cmplt_epu16_mask() { |
13290 | let a = _mm_set1_epi16(-2); |
13291 | let b = _mm_set1_epi16(-1); |
13292 | let mask = 0b01010101; |
13293 | let r = _mm_mask_cmplt_epu16_mask(mask, a, b); |
13294 | assert_eq!(r, 0b01010101); |
13295 | } |
13296 | |
13297 | #[simd_test(enable = "avx512bw" )] |
13298 | unsafe fn test_mm512_cmplt_epu8_mask() { |
13299 | let a = _mm512_set1_epi8(-2); |
13300 | let b = _mm512_set1_epi8(-1); |
13301 | let m = _mm512_cmplt_epu8_mask(a, b); |
13302 | assert_eq!( |
13303 | m, |
13304 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
13305 | ); |
13306 | } |
13307 | |
13308 | #[simd_test(enable = "avx512bw" )] |
13309 | unsafe fn test_mm512_mask_cmplt_epu8_mask() { |
13310 | let a = _mm512_set1_epi8(-2); |
13311 | let b = _mm512_set1_epi8(-1); |
13312 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
13313 | let r = _mm512_mask_cmplt_epu8_mask(mask, a, b); |
13314 | assert_eq!( |
13315 | r, |
13316 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
13317 | ); |
13318 | } |
13319 | |
13320 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13321 | unsafe fn test_mm256_cmplt_epu8_mask() { |
13322 | let a = _mm256_set1_epi8(-2); |
13323 | let b = _mm256_set1_epi8(-1); |
13324 | let m = _mm256_cmplt_epu8_mask(a, b); |
13325 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13326 | } |
13327 | |
13328 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13329 | unsafe fn test_mm256_mask_cmplt_epu8_mask() { |
13330 | let a = _mm256_set1_epi8(-2); |
13331 | let b = _mm256_set1_epi8(-1); |
13332 | let mask = 0b01010101_01010101_01010101_01010101; |
13333 | let r = _mm256_mask_cmplt_epu8_mask(mask, a, b); |
13334 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13335 | } |
13336 | |
13337 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13338 | unsafe fn test_mm_cmplt_epu8_mask() { |
13339 | let a = _mm_set1_epi8(-2); |
13340 | let b = _mm_set1_epi8(-1); |
13341 | let m = _mm_cmplt_epu8_mask(a, b); |
13342 | assert_eq!(m, 0b11111111_11111111); |
13343 | } |
13344 | |
13345 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13346 | unsafe fn test_mm_mask_cmplt_epu8_mask() { |
13347 | let a = _mm_set1_epi8(-2); |
13348 | let b = _mm_set1_epi8(-1); |
13349 | let mask = 0b01010101_01010101; |
13350 | let r = _mm_mask_cmplt_epu8_mask(mask, a, b); |
13351 | assert_eq!(r, 0b01010101_01010101); |
13352 | } |
13353 | |
13354 | #[simd_test(enable = "avx512bw" )] |
13355 | unsafe fn test_mm512_cmplt_epi16_mask() { |
13356 | let a = _mm512_set1_epi16(-2); |
13357 | let b = _mm512_set1_epi16(-1); |
13358 | let m = _mm512_cmplt_epi16_mask(a, b); |
13359 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13360 | } |
13361 | |
13362 | #[simd_test(enable = "avx512bw" )] |
13363 | unsafe fn test_mm512_mask_cmplt_epi16_mask() { |
13364 | let a = _mm512_set1_epi16(-2); |
13365 | let b = _mm512_set1_epi16(-1); |
13366 | let mask = 0b01010101_01010101_01010101_01010101; |
13367 | let r = _mm512_mask_cmplt_epi16_mask(mask, a, b); |
13368 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13369 | } |
13370 | |
13371 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13372 | unsafe fn test_mm256_cmplt_epi16_mask() { |
13373 | let a = _mm256_set1_epi16(-2); |
13374 | let b = _mm256_set1_epi16(-1); |
13375 | let m = _mm256_cmplt_epi16_mask(a, b); |
13376 | assert_eq!(m, 0b11111111_11111111); |
13377 | } |
13378 | |
13379 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13380 | unsafe fn test_mm256_mask_cmplt_epi16_mask() { |
13381 | let a = _mm256_set1_epi16(-2); |
13382 | let b = _mm256_set1_epi16(-1); |
13383 | let mask = 0b01010101_01010101; |
13384 | let r = _mm256_mask_cmplt_epi16_mask(mask, a, b); |
13385 | assert_eq!(r, 0b01010101_01010101); |
13386 | } |
13387 | |
13388 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13389 | unsafe fn test_mm_cmplt_epi16_mask() { |
13390 | let a = _mm_set1_epi16(-2); |
13391 | let b = _mm_set1_epi16(-1); |
13392 | let m = _mm_cmplt_epi16_mask(a, b); |
13393 | assert_eq!(m, 0b11111111); |
13394 | } |
13395 | |
13396 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13397 | unsafe fn test_mm_mask_cmplt_epi16_mask() { |
13398 | let a = _mm_set1_epi16(-2); |
13399 | let b = _mm_set1_epi16(-1); |
13400 | let mask = 0b01010101; |
13401 | let r = _mm_mask_cmplt_epi16_mask(mask, a, b); |
13402 | assert_eq!(r, 0b01010101); |
13403 | } |
13404 | |
13405 | #[simd_test(enable = "avx512bw" )] |
13406 | unsafe fn test_mm512_cmplt_epi8_mask() { |
13407 | let a = _mm512_set1_epi8(-2); |
13408 | let b = _mm512_set1_epi8(-1); |
13409 | let m = _mm512_cmplt_epi8_mask(a, b); |
13410 | assert_eq!( |
13411 | m, |
13412 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
13413 | ); |
13414 | } |
13415 | |
13416 | #[simd_test(enable = "avx512bw" )] |
13417 | unsafe fn test_mm512_mask_cmplt_epi8_mask() { |
13418 | let a = _mm512_set1_epi8(-2); |
13419 | let b = _mm512_set1_epi8(-1); |
13420 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
13421 | let r = _mm512_mask_cmplt_epi8_mask(mask, a, b); |
13422 | assert_eq!( |
13423 | r, |
13424 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
13425 | ); |
13426 | } |
13427 | |
13428 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13429 | unsafe fn test_mm256_cmplt_epi8_mask() { |
13430 | let a = _mm256_set1_epi8(-2); |
13431 | let b = _mm256_set1_epi8(-1); |
13432 | let m = _mm256_cmplt_epi8_mask(a, b); |
13433 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13434 | } |
13435 | |
13436 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13437 | unsafe fn test_mm256_mask_cmplt_epi8_mask() { |
13438 | let a = _mm256_set1_epi8(-2); |
13439 | let b = _mm256_set1_epi8(-1); |
13440 | let mask = 0b01010101_01010101_01010101_01010101; |
13441 | let r = _mm256_mask_cmplt_epi8_mask(mask, a, b); |
13442 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13443 | } |
13444 | |
13445 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13446 | unsafe fn test_mm_cmplt_epi8_mask() { |
13447 | let a = _mm_set1_epi8(-2); |
13448 | let b = _mm_set1_epi8(-1); |
13449 | let m = _mm_cmplt_epi8_mask(a, b); |
13450 | assert_eq!(m, 0b11111111_11111111); |
13451 | } |
13452 | |
13453 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13454 | unsafe fn test_mm_mask_cmplt_epi8_mask() { |
13455 | let a = _mm_set1_epi8(-2); |
13456 | let b = _mm_set1_epi8(-1); |
13457 | let mask = 0b01010101_01010101; |
13458 | let r = _mm_mask_cmplt_epi8_mask(mask, a, b); |
13459 | assert_eq!(r, 0b01010101_01010101); |
13460 | } |
13461 | |
13462 | #[simd_test(enable = "avx512bw" )] |
13463 | unsafe fn test_mm512_cmpgt_epu16_mask() { |
13464 | let a = _mm512_set1_epi16(2); |
13465 | let b = _mm512_set1_epi16(1); |
13466 | let m = _mm512_cmpgt_epu16_mask(a, b); |
13467 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13468 | } |
13469 | |
13470 | #[simd_test(enable = "avx512bw" )] |
13471 | unsafe fn test_mm512_mask_cmpgt_epu16_mask() { |
13472 | let a = _mm512_set1_epi16(2); |
13473 | let b = _mm512_set1_epi16(1); |
13474 | let mask = 0b01010101_01010101_01010101_01010101; |
13475 | let r = _mm512_mask_cmpgt_epu16_mask(mask, a, b); |
13476 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13477 | } |
13478 | |
13479 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13480 | unsafe fn test_mm256_cmpgt_epu16_mask() { |
13481 | let a = _mm256_set1_epi16(2); |
13482 | let b = _mm256_set1_epi16(1); |
13483 | let m = _mm256_cmpgt_epu16_mask(a, b); |
13484 | assert_eq!(m, 0b11111111_11111111); |
13485 | } |
13486 | |
13487 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13488 | unsafe fn test_mm256_mask_cmpgt_epu16_mask() { |
13489 | let a = _mm256_set1_epi16(2); |
13490 | let b = _mm256_set1_epi16(1); |
13491 | let mask = 0b01010101_01010101; |
13492 | let r = _mm256_mask_cmpgt_epu16_mask(mask, a, b); |
13493 | assert_eq!(r, 0b01010101_01010101); |
13494 | } |
13495 | |
13496 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13497 | unsafe fn test_mm_cmpgt_epu16_mask() { |
13498 | let a = _mm_set1_epi16(2); |
13499 | let b = _mm_set1_epi16(1); |
13500 | let m = _mm_cmpgt_epu16_mask(a, b); |
13501 | assert_eq!(m, 0b11111111); |
13502 | } |
13503 | |
13504 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13505 | unsafe fn test_mm_mask_cmpgt_epu16_mask() { |
13506 | let a = _mm_set1_epi16(2); |
13507 | let b = _mm_set1_epi16(1); |
13508 | let mask = 0b01010101; |
13509 | let r = _mm_mask_cmpgt_epu16_mask(mask, a, b); |
13510 | assert_eq!(r, 0b01010101); |
13511 | } |
13512 | |
13513 | #[simd_test(enable = "avx512bw" )] |
13514 | unsafe fn test_mm512_cmpgt_epu8_mask() { |
13515 | let a = _mm512_set1_epi8(2); |
13516 | let b = _mm512_set1_epi8(1); |
13517 | let m = _mm512_cmpgt_epu8_mask(a, b); |
13518 | assert_eq!( |
13519 | m, |
13520 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
13521 | ); |
13522 | } |
13523 | |
13524 | #[simd_test(enable = "avx512bw" )] |
13525 | unsafe fn test_mm512_mask_cmpgt_epu8_mask() { |
13526 | let a = _mm512_set1_epi8(2); |
13527 | let b = _mm512_set1_epi8(1); |
13528 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
13529 | let r = _mm512_mask_cmpgt_epu8_mask(mask, a, b); |
13530 | assert_eq!( |
13531 | r, |
13532 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
13533 | ); |
13534 | } |
13535 | |
13536 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13537 | unsafe fn test_mm256_cmpgt_epu8_mask() { |
13538 | let a = _mm256_set1_epi8(2); |
13539 | let b = _mm256_set1_epi8(1); |
13540 | let m = _mm256_cmpgt_epu8_mask(a, b); |
13541 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13542 | } |
13543 | |
13544 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13545 | unsafe fn test_mm256_mask_cmpgt_epu8_mask() { |
13546 | let a = _mm256_set1_epi8(2); |
13547 | let b = _mm256_set1_epi8(1); |
13548 | let mask = 0b01010101_01010101_01010101_01010101; |
13549 | let r = _mm256_mask_cmpgt_epu8_mask(mask, a, b); |
13550 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13551 | } |
13552 | |
13553 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13554 | unsafe fn test_mm_cmpgt_epu8_mask() { |
13555 | let a = _mm_set1_epi8(2); |
13556 | let b = _mm_set1_epi8(1); |
13557 | let m = _mm_cmpgt_epu8_mask(a, b); |
13558 | assert_eq!(m, 0b11111111_11111111); |
13559 | } |
13560 | |
13561 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13562 | unsafe fn test_mm_mask_cmpgt_epu8_mask() { |
13563 | let a = _mm_set1_epi8(2); |
13564 | let b = _mm_set1_epi8(1); |
13565 | let mask = 0b01010101_01010101; |
13566 | let r = _mm_mask_cmpgt_epu8_mask(mask, a, b); |
13567 | assert_eq!(r, 0b01010101_01010101); |
13568 | } |
13569 | |
13570 | #[simd_test(enable = "avx512bw" )] |
13571 | unsafe fn test_mm512_cmpgt_epi16_mask() { |
13572 | let a = _mm512_set1_epi16(2); |
13573 | let b = _mm512_set1_epi16(-1); |
13574 | let m = _mm512_cmpgt_epi16_mask(a, b); |
13575 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13576 | } |
13577 | |
13578 | #[simd_test(enable = "avx512bw" )] |
13579 | unsafe fn test_mm512_mask_cmpgt_epi16_mask() { |
13580 | let a = _mm512_set1_epi16(2); |
13581 | let b = _mm512_set1_epi16(-1); |
13582 | let mask = 0b01010101_01010101_01010101_01010101; |
13583 | let r = _mm512_mask_cmpgt_epi16_mask(mask, a, b); |
13584 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13585 | } |
13586 | |
13587 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13588 | unsafe fn test_mm256_cmpgt_epi16_mask() { |
13589 | let a = _mm256_set1_epi16(2); |
13590 | let b = _mm256_set1_epi16(-1); |
13591 | let m = _mm256_cmpgt_epi16_mask(a, b); |
13592 | assert_eq!(m, 0b11111111_11111111); |
13593 | } |
13594 | |
13595 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13596 | unsafe fn test_mm256_mask_cmpgt_epi16_mask() { |
13597 | let a = _mm256_set1_epi16(2); |
13598 | let b = _mm256_set1_epi16(-1); |
13599 | let mask = 0b001010101_01010101; |
13600 | let r = _mm256_mask_cmpgt_epi16_mask(mask, a, b); |
13601 | assert_eq!(r, 0b01010101_01010101); |
13602 | } |
13603 | |
13604 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13605 | unsafe fn test_mm_cmpgt_epi16_mask() { |
13606 | let a = _mm_set1_epi16(2); |
13607 | let b = _mm_set1_epi16(-1); |
13608 | let m = _mm_cmpgt_epi16_mask(a, b); |
13609 | assert_eq!(m, 0b11111111); |
13610 | } |
13611 | |
13612 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13613 | unsafe fn test_mm_mask_cmpgt_epi16_mask() { |
13614 | let a = _mm_set1_epi16(2); |
13615 | let b = _mm_set1_epi16(-1); |
13616 | let mask = 0b01010101; |
13617 | let r = _mm_mask_cmpgt_epi16_mask(mask, a, b); |
13618 | assert_eq!(r, 0b01010101); |
13619 | } |
13620 | |
13621 | #[simd_test(enable = "avx512bw" )] |
13622 | unsafe fn test_mm512_cmpgt_epi8_mask() { |
13623 | let a = _mm512_set1_epi8(2); |
13624 | let b = _mm512_set1_epi8(-1); |
13625 | let m = _mm512_cmpgt_epi8_mask(a, b); |
13626 | assert_eq!( |
13627 | m, |
13628 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
13629 | ); |
13630 | } |
13631 | |
13632 | #[simd_test(enable = "avx512bw" )] |
13633 | unsafe fn test_mm512_mask_cmpgt_epi8_mask() { |
13634 | let a = _mm512_set1_epi8(2); |
13635 | let b = _mm512_set1_epi8(-1); |
13636 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
13637 | let r = _mm512_mask_cmpgt_epi8_mask(mask, a, b); |
13638 | assert_eq!( |
13639 | r, |
13640 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
13641 | ); |
13642 | } |
13643 | |
13644 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13645 | unsafe fn test_mm256_cmpgt_epi8_mask() { |
13646 | let a = _mm256_set1_epi8(2); |
13647 | let b = _mm256_set1_epi8(-1); |
13648 | let m = _mm256_cmpgt_epi8_mask(a, b); |
13649 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13650 | } |
13651 | |
13652 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13653 | unsafe fn test_mm256_mask_cmpgt_epi8_mask() { |
13654 | let a = _mm256_set1_epi8(2); |
13655 | let b = _mm256_set1_epi8(-1); |
13656 | let mask = 0b01010101_01010101_01010101_01010101; |
13657 | let r = _mm256_mask_cmpgt_epi8_mask(mask, a, b); |
13658 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13659 | } |
13660 | |
13661 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13662 | unsafe fn test_mm_cmpgt_epi8_mask() { |
13663 | let a = _mm_set1_epi8(2); |
13664 | let b = _mm_set1_epi8(-1); |
13665 | let m = _mm_cmpgt_epi8_mask(a, b); |
13666 | assert_eq!(m, 0b11111111_11111111); |
13667 | } |
13668 | |
13669 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13670 | unsafe fn test_mm_mask_cmpgt_epi8_mask() { |
13671 | let a = _mm_set1_epi8(2); |
13672 | let b = _mm_set1_epi8(-1); |
13673 | let mask = 0b01010101_01010101; |
13674 | let r = _mm_mask_cmpgt_epi8_mask(mask, a, b); |
13675 | assert_eq!(r, 0b01010101_01010101); |
13676 | } |
13677 | |
13678 | #[simd_test(enable = "avx512bw" )] |
13679 | unsafe fn test_mm512_cmple_epu16_mask() { |
13680 | let a = _mm512_set1_epi16(-1); |
13681 | let b = _mm512_set1_epi16(-1); |
13682 | let m = _mm512_cmple_epu16_mask(a, b); |
13683 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13684 | } |
13685 | |
13686 | #[simd_test(enable = "avx512bw" )] |
13687 | unsafe fn test_mm512_mask_cmple_epu16_mask() { |
13688 | let a = _mm512_set1_epi16(-1); |
13689 | let b = _mm512_set1_epi16(-1); |
13690 | let mask = 0b01010101_01010101_01010101_01010101; |
13691 | let r = _mm512_mask_cmple_epu16_mask(mask, a, b); |
13692 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13693 | } |
13694 | |
13695 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13696 | unsafe fn test_mm256_cmple_epu16_mask() { |
13697 | let a = _mm256_set1_epi16(-1); |
13698 | let b = _mm256_set1_epi16(-1); |
13699 | let m = _mm256_cmple_epu16_mask(a, b); |
13700 | assert_eq!(m, 0b11111111_11111111); |
13701 | } |
13702 | |
13703 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13704 | unsafe fn test_mm256_mask_cmple_epu16_mask() { |
13705 | let a = _mm256_set1_epi16(-1); |
13706 | let b = _mm256_set1_epi16(-1); |
13707 | let mask = 0b01010101_01010101; |
13708 | let r = _mm256_mask_cmple_epu16_mask(mask, a, b); |
13709 | assert_eq!(r, 0b01010101_01010101); |
13710 | } |
13711 | |
13712 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13713 | unsafe fn test_mm_cmple_epu16_mask() { |
13714 | let a = _mm_set1_epi16(-1); |
13715 | let b = _mm_set1_epi16(-1); |
13716 | let m = _mm_cmple_epu16_mask(a, b); |
13717 | assert_eq!(m, 0b11111111); |
13718 | } |
13719 | |
13720 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13721 | unsafe fn test_mm_mask_cmple_epu16_mask() { |
13722 | let a = _mm_set1_epi16(-1); |
13723 | let b = _mm_set1_epi16(-1); |
13724 | let mask = 0b01010101; |
13725 | let r = _mm_mask_cmple_epu16_mask(mask, a, b); |
13726 | assert_eq!(r, 0b01010101); |
13727 | } |
13728 | |
13729 | #[simd_test(enable = "avx512bw" )] |
13730 | unsafe fn test_mm512_cmple_epu8_mask() { |
13731 | let a = _mm512_set1_epi8(-1); |
13732 | let b = _mm512_set1_epi8(-1); |
13733 | let m = _mm512_cmple_epu8_mask(a, b); |
13734 | assert_eq!( |
13735 | m, |
13736 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
13737 | ); |
13738 | } |
13739 | |
13740 | #[simd_test(enable = "avx512bw" )] |
13741 | unsafe fn test_mm512_mask_cmple_epu8_mask() { |
13742 | let a = _mm512_set1_epi8(-1); |
13743 | let b = _mm512_set1_epi8(-1); |
13744 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
13745 | let r = _mm512_mask_cmple_epu8_mask(mask, a, b); |
13746 | assert_eq!( |
13747 | r, |
13748 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
13749 | ); |
13750 | } |
13751 | |
13752 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13753 | unsafe fn test_mm256_cmple_epu8_mask() { |
13754 | let a = _mm256_set1_epi8(-1); |
13755 | let b = _mm256_set1_epi8(-1); |
13756 | let m = _mm256_cmple_epu8_mask(a, b); |
13757 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13758 | } |
13759 | |
13760 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13761 | unsafe fn test_mm256_mask_cmple_epu8_mask() { |
13762 | let a = _mm256_set1_epi8(-1); |
13763 | let b = _mm256_set1_epi8(-1); |
13764 | let mask = 0b01010101_01010101_01010101_01010101; |
13765 | let r = _mm256_mask_cmple_epu8_mask(mask, a, b); |
13766 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13767 | } |
13768 | |
13769 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13770 | unsafe fn test_mm_cmple_epu8_mask() { |
13771 | let a = _mm_set1_epi8(-1); |
13772 | let b = _mm_set1_epi8(-1); |
13773 | let m = _mm_cmple_epu8_mask(a, b); |
13774 | assert_eq!(m, 0b11111111_11111111); |
13775 | } |
13776 | |
13777 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13778 | unsafe fn test_mm_mask_cmple_epu8_mask() { |
13779 | let a = _mm_set1_epi8(-1); |
13780 | let b = _mm_set1_epi8(-1); |
13781 | let mask = 0b01010101_01010101; |
13782 | let r = _mm_mask_cmple_epu8_mask(mask, a, b); |
13783 | assert_eq!(r, 0b01010101_01010101); |
13784 | } |
13785 | |
13786 | #[simd_test(enable = "avx512bw" )] |
13787 | unsafe fn test_mm512_cmple_epi16_mask() { |
13788 | let a = _mm512_set1_epi16(-1); |
13789 | let b = _mm512_set1_epi16(-1); |
13790 | let m = _mm512_cmple_epi16_mask(a, b); |
13791 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13792 | } |
13793 | |
13794 | #[simd_test(enable = "avx512bw" )] |
13795 | unsafe fn test_mm512_mask_cmple_epi16_mask() { |
13796 | let a = _mm512_set1_epi16(-1); |
13797 | let b = _mm512_set1_epi16(-1); |
13798 | let mask = 0b01010101_01010101_01010101_01010101; |
13799 | let r = _mm512_mask_cmple_epi16_mask(mask, a, b); |
13800 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13801 | } |
13802 | |
13803 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13804 | unsafe fn test_mm256_cmple_epi16_mask() { |
13805 | let a = _mm256_set1_epi16(-1); |
13806 | let b = _mm256_set1_epi16(-1); |
13807 | let m = _mm256_cmple_epi16_mask(a, b); |
13808 | assert_eq!(m, 0b11111111_11111111); |
13809 | } |
13810 | |
13811 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13812 | unsafe fn test_mm256_mask_cmple_epi16_mask() { |
13813 | let a = _mm256_set1_epi16(-1); |
13814 | let b = _mm256_set1_epi16(-1); |
13815 | let mask = 0b01010101_01010101; |
13816 | let r = _mm256_mask_cmple_epi16_mask(mask, a, b); |
13817 | assert_eq!(r, 0b01010101_01010101); |
13818 | } |
13819 | |
13820 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13821 | unsafe fn test_mm_cmple_epi16_mask() { |
13822 | let a = _mm_set1_epi16(-1); |
13823 | let b = _mm_set1_epi16(-1); |
13824 | let m = _mm_cmple_epi16_mask(a, b); |
13825 | assert_eq!(m, 0b11111111); |
13826 | } |
13827 | |
13828 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13829 | unsafe fn test_mm_mask_cmple_epi16_mask() { |
13830 | let a = _mm_set1_epi16(-1); |
13831 | let b = _mm_set1_epi16(-1); |
13832 | let mask = 0b01010101; |
13833 | let r = _mm_mask_cmple_epi16_mask(mask, a, b); |
13834 | assert_eq!(r, 0b01010101); |
13835 | } |
13836 | |
13837 | #[simd_test(enable = "avx512bw" )] |
13838 | unsafe fn test_mm512_cmple_epi8_mask() { |
13839 | let a = _mm512_set1_epi8(-1); |
13840 | let b = _mm512_set1_epi8(-1); |
13841 | let m = _mm512_cmple_epi8_mask(a, b); |
13842 | assert_eq!( |
13843 | m, |
13844 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
13845 | ); |
13846 | } |
13847 | |
13848 | #[simd_test(enable = "avx512bw" )] |
13849 | unsafe fn test_mm512_mask_cmple_epi8_mask() { |
13850 | let a = _mm512_set1_epi8(-1); |
13851 | let b = _mm512_set1_epi8(-1); |
13852 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
13853 | let r = _mm512_mask_cmple_epi8_mask(mask, a, b); |
13854 | assert_eq!( |
13855 | r, |
13856 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
13857 | ); |
13858 | } |
13859 | |
13860 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13861 | unsafe fn test_mm256_cmple_epi8_mask() { |
13862 | let a = _mm256_set1_epi8(-1); |
13863 | let b = _mm256_set1_epi8(-1); |
13864 | let m = _mm256_cmple_epi8_mask(a, b); |
13865 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13866 | } |
13867 | |
13868 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13869 | unsafe fn test_mm256_mask_cmple_epi8_mask() { |
13870 | let a = _mm256_set1_epi8(-1); |
13871 | let b = _mm256_set1_epi8(-1); |
13872 | let mask = 0b01010101_01010101_01010101_01010101; |
13873 | let r = _mm256_mask_cmple_epi8_mask(mask, a, b); |
13874 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13875 | } |
13876 | |
13877 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13878 | unsafe fn test_mm_cmple_epi8_mask() { |
13879 | let a = _mm_set1_epi8(-1); |
13880 | let b = _mm_set1_epi8(-1); |
13881 | let m = _mm_cmple_epi8_mask(a, b); |
13882 | assert_eq!(m, 0b11111111_11111111); |
13883 | } |
13884 | |
13885 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13886 | unsafe fn test_mm_mask_cmple_epi8_mask() { |
13887 | let a = _mm_set1_epi8(-1); |
13888 | let b = _mm_set1_epi8(-1); |
13889 | let mask = 0b01010101_01010101; |
13890 | let r = _mm_mask_cmple_epi8_mask(mask, a, b); |
13891 | assert_eq!(r, 0b01010101_01010101); |
13892 | } |
13893 | |
13894 | #[simd_test(enable = "avx512bw" )] |
13895 | unsafe fn test_mm512_cmpge_epu16_mask() { |
13896 | let a = _mm512_set1_epi16(1); |
13897 | let b = _mm512_set1_epi16(1); |
13898 | let m = _mm512_cmpge_epu16_mask(a, b); |
13899 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13900 | } |
13901 | |
13902 | #[simd_test(enable = "avx512bw" )] |
13903 | unsafe fn test_mm512_mask_cmpge_epu16_mask() { |
13904 | let a = _mm512_set1_epi16(1); |
13905 | let b = _mm512_set1_epi16(1); |
13906 | let mask = 0b01010101_01010101_01010101_01010101; |
13907 | let r = _mm512_mask_cmpge_epu16_mask(mask, a, b); |
13908 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13909 | } |
13910 | |
13911 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13912 | unsafe fn test_mm256_cmpge_epu16_mask() { |
13913 | let a = _mm256_set1_epi16(1); |
13914 | let b = _mm256_set1_epi16(1); |
13915 | let m = _mm256_cmpge_epu16_mask(a, b); |
13916 | assert_eq!(m, 0b11111111_11111111); |
13917 | } |
13918 | |
13919 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13920 | unsafe fn test_mm256_mask_cmpge_epu16_mask() { |
13921 | let a = _mm256_set1_epi16(1); |
13922 | let b = _mm256_set1_epi16(1); |
13923 | let mask = 0b01010101_01010101; |
13924 | let r = _mm256_mask_cmpge_epu16_mask(mask, a, b); |
13925 | assert_eq!(r, 0b01010101_01010101); |
13926 | } |
13927 | |
13928 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13929 | unsafe fn test_mm_cmpge_epu16_mask() { |
13930 | let a = _mm_set1_epi16(1); |
13931 | let b = _mm_set1_epi16(1); |
13932 | let m = _mm_cmpge_epu16_mask(a, b); |
13933 | assert_eq!(m, 0b11111111); |
13934 | } |
13935 | |
13936 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13937 | unsafe fn test_mm_mask_cmpge_epu16_mask() { |
13938 | let a = _mm_set1_epi16(1); |
13939 | let b = _mm_set1_epi16(1); |
13940 | let mask = 0b01010101; |
13941 | let r = _mm_mask_cmpge_epu16_mask(mask, a, b); |
13942 | assert_eq!(r, 0b01010101); |
13943 | } |
13944 | |
13945 | #[simd_test(enable = "avx512bw" )] |
13946 | unsafe fn test_mm512_cmpge_epu8_mask() { |
13947 | let a = _mm512_set1_epi8(1); |
13948 | let b = _mm512_set1_epi8(1); |
13949 | let m = _mm512_cmpge_epu8_mask(a, b); |
13950 | assert_eq!( |
13951 | m, |
13952 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
13953 | ); |
13954 | } |
13955 | |
13956 | #[simd_test(enable = "avx512bw" )] |
13957 | unsafe fn test_mm512_mask_cmpge_epu8_mask() { |
13958 | let a = _mm512_set1_epi8(1); |
13959 | let b = _mm512_set1_epi8(1); |
13960 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
13961 | let r = _mm512_mask_cmpge_epu8_mask(mask, a, b); |
13962 | assert_eq!( |
13963 | r, |
13964 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
13965 | ); |
13966 | } |
13967 | |
13968 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13969 | unsafe fn test_mm256_cmpge_epu8_mask() { |
13970 | let a = _mm256_set1_epi8(1); |
13971 | let b = _mm256_set1_epi8(1); |
13972 | let m = _mm256_cmpge_epu8_mask(a, b); |
13973 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
13974 | } |
13975 | |
13976 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13977 | unsafe fn test_mm256_mask_cmpge_epu8_mask() { |
13978 | let a = _mm256_set1_epi8(1); |
13979 | let b = _mm256_set1_epi8(1); |
13980 | let mask = 0b01010101_01010101_01010101_01010101; |
13981 | let r = _mm256_mask_cmpge_epu8_mask(mask, a, b); |
13982 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
13983 | } |
13984 | |
13985 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13986 | unsafe fn test_mm_cmpge_epu8_mask() { |
13987 | let a = _mm_set1_epi8(1); |
13988 | let b = _mm_set1_epi8(1); |
13989 | let m = _mm_cmpge_epu8_mask(a, b); |
13990 | assert_eq!(m, 0b11111111_11111111); |
13991 | } |
13992 | |
13993 | #[simd_test(enable = "avx512bw,avx512vl" )] |
13994 | unsafe fn test_mm_mask_cmpge_epu8_mask() { |
13995 | let a = _mm_set1_epi8(1); |
13996 | let b = _mm_set1_epi8(1); |
13997 | let mask = 0b01010101_01010101; |
13998 | let r = _mm_mask_cmpge_epu8_mask(mask, a, b); |
13999 | assert_eq!(r, 0b01010101_01010101); |
14000 | } |
14001 | |
14002 | #[simd_test(enable = "avx512bw" )] |
14003 | unsafe fn test_mm512_cmpge_epi16_mask() { |
14004 | let a = _mm512_set1_epi16(-1); |
14005 | let b = _mm512_set1_epi16(-1); |
14006 | let m = _mm512_cmpge_epi16_mask(a, b); |
14007 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14008 | } |
14009 | |
14010 | #[simd_test(enable = "avx512bw" )] |
14011 | unsafe fn test_mm512_mask_cmpge_epi16_mask() { |
14012 | let a = _mm512_set1_epi16(-1); |
14013 | let b = _mm512_set1_epi16(-1); |
14014 | let mask = 0b01010101_01010101_01010101_01010101; |
14015 | let r = _mm512_mask_cmpge_epi16_mask(mask, a, b); |
14016 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14017 | } |
14018 | |
14019 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14020 | unsafe fn test_mm256_cmpge_epi16_mask() { |
14021 | let a = _mm256_set1_epi16(-1); |
14022 | let b = _mm256_set1_epi16(-1); |
14023 | let m = _mm256_cmpge_epi16_mask(a, b); |
14024 | assert_eq!(m, 0b11111111_11111111); |
14025 | } |
14026 | |
14027 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14028 | unsafe fn test_mm256_mask_cmpge_epi16_mask() { |
14029 | let a = _mm256_set1_epi16(-1); |
14030 | let b = _mm256_set1_epi16(-1); |
14031 | let mask = 0b01010101_01010101; |
14032 | let r = _mm256_mask_cmpge_epi16_mask(mask, a, b); |
14033 | assert_eq!(r, 0b01010101_01010101); |
14034 | } |
14035 | |
14036 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14037 | unsafe fn test_mm_cmpge_epi16_mask() { |
14038 | let a = _mm_set1_epi16(-1); |
14039 | let b = _mm_set1_epi16(-1); |
14040 | let m = _mm_cmpge_epi16_mask(a, b); |
14041 | assert_eq!(m, 0b11111111); |
14042 | } |
14043 | |
14044 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14045 | unsafe fn test_mm_mask_cmpge_epi16_mask() { |
14046 | let a = _mm_set1_epi16(-1); |
14047 | let b = _mm_set1_epi16(-1); |
14048 | let mask = 0b01010101; |
14049 | let r = _mm_mask_cmpge_epi16_mask(mask, a, b); |
14050 | assert_eq!(r, 0b01010101); |
14051 | } |
14052 | |
14053 | #[simd_test(enable = "avx512bw" )] |
14054 | unsafe fn test_mm512_cmpge_epi8_mask() { |
14055 | let a = _mm512_set1_epi8(-1); |
14056 | let b = _mm512_set1_epi8(-1); |
14057 | let m = _mm512_cmpge_epi8_mask(a, b); |
14058 | assert_eq!( |
14059 | m, |
14060 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
14061 | ); |
14062 | } |
14063 | |
14064 | #[simd_test(enable = "avx512bw" )] |
14065 | unsafe fn test_mm512_mask_cmpge_epi8_mask() { |
14066 | let a = _mm512_set1_epi8(-1); |
14067 | let b = _mm512_set1_epi8(-1); |
14068 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
14069 | let r = _mm512_mask_cmpge_epi8_mask(mask, a, b); |
14070 | assert_eq!( |
14071 | r, |
14072 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
14073 | ); |
14074 | } |
14075 | |
14076 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14077 | unsafe fn test_mm256_cmpge_epi8_mask() { |
14078 | let a = _mm256_set1_epi8(-1); |
14079 | let b = _mm256_set1_epi8(-1); |
14080 | let m = _mm256_cmpge_epi8_mask(a, b); |
14081 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14082 | } |
14083 | |
14084 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14085 | unsafe fn test_mm256_mask_cmpge_epi8_mask() { |
14086 | let a = _mm256_set1_epi8(-1); |
14087 | let b = _mm256_set1_epi8(-1); |
14088 | let mask = 0b01010101_01010101_01010101_01010101; |
14089 | let r = _mm256_mask_cmpge_epi8_mask(mask, a, b); |
14090 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14091 | } |
14092 | |
14093 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14094 | unsafe fn test_mm_cmpge_epi8_mask() { |
14095 | let a = _mm_set1_epi8(-1); |
14096 | let b = _mm_set1_epi8(-1); |
14097 | let m = _mm_cmpge_epi8_mask(a, b); |
14098 | assert_eq!(m, 0b11111111_11111111); |
14099 | } |
14100 | |
14101 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14102 | unsafe fn test_mm_mask_cmpge_epi8_mask() { |
14103 | let a = _mm_set1_epi8(-1); |
14104 | let b = _mm_set1_epi8(-1); |
14105 | let mask = 0b01010101_01010101; |
14106 | let r = _mm_mask_cmpge_epi8_mask(mask, a, b); |
14107 | assert_eq!(r, 0b01010101_01010101); |
14108 | } |
14109 | |
14110 | #[simd_test(enable = "avx512bw" )] |
14111 | unsafe fn test_mm512_cmpeq_epu16_mask() { |
14112 | let a = _mm512_set1_epi16(1); |
14113 | let b = _mm512_set1_epi16(1); |
14114 | let m = _mm512_cmpeq_epu16_mask(a, b); |
14115 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14116 | } |
14117 | |
14118 | #[simd_test(enable = "avx512bw" )] |
14119 | unsafe fn test_mm512_mask_cmpeq_epu16_mask() { |
14120 | let a = _mm512_set1_epi16(1); |
14121 | let b = _mm512_set1_epi16(1); |
14122 | let mask = 0b01010101_01010101_01010101_01010101; |
14123 | let r = _mm512_mask_cmpeq_epu16_mask(mask, a, b); |
14124 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14125 | } |
14126 | |
14127 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14128 | unsafe fn test_mm256_cmpeq_epu16_mask() { |
14129 | let a = _mm256_set1_epi16(1); |
14130 | let b = _mm256_set1_epi16(1); |
14131 | let m = _mm256_cmpeq_epu16_mask(a, b); |
14132 | assert_eq!(m, 0b11111111_11111111); |
14133 | } |
14134 | |
14135 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14136 | unsafe fn test_mm256_mask_cmpeq_epu16_mask() { |
14137 | let a = _mm256_set1_epi16(1); |
14138 | let b = _mm256_set1_epi16(1); |
14139 | let mask = 0b01010101_01010101; |
14140 | let r = _mm256_mask_cmpeq_epu16_mask(mask, a, b); |
14141 | assert_eq!(r, 0b01010101_01010101); |
14142 | } |
14143 | |
14144 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14145 | unsafe fn test_mm_cmpeq_epu16_mask() { |
14146 | let a = _mm_set1_epi16(1); |
14147 | let b = _mm_set1_epi16(1); |
14148 | let m = _mm_cmpeq_epu16_mask(a, b); |
14149 | assert_eq!(m, 0b11111111); |
14150 | } |
14151 | |
14152 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14153 | unsafe fn test_mm_mask_cmpeq_epu16_mask() { |
14154 | let a = _mm_set1_epi16(1); |
14155 | let b = _mm_set1_epi16(1); |
14156 | let mask = 0b01010101; |
14157 | let r = _mm_mask_cmpeq_epu16_mask(mask, a, b); |
14158 | assert_eq!(r, 0b01010101); |
14159 | } |
14160 | |
14161 | #[simd_test(enable = "avx512bw" )] |
14162 | unsafe fn test_mm512_cmpeq_epu8_mask() { |
14163 | let a = _mm512_set1_epi8(1); |
14164 | let b = _mm512_set1_epi8(1); |
14165 | let m = _mm512_cmpeq_epu8_mask(a, b); |
14166 | assert_eq!( |
14167 | m, |
14168 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
14169 | ); |
14170 | } |
14171 | |
14172 | #[simd_test(enable = "avx512bw" )] |
14173 | unsafe fn test_mm512_mask_cmpeq_epu8_mask() { |
14174 | let a = _mm512_set1_epi8(1); |
14175 | let b = _mm512_set1_epi8(1); |
14176 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
14177 | let r = _mm512_mask_cmpeq_epu8_mask(mask, a, b); |
14178 | assert_eq!( |
14179 | r, |
14180 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
14181 | ); |
14182 | } |
14183 | |
14184 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14185 | unsafe fn test_mm256_cmpeq_epu8_mask() { |
14186 | let a = _mm256_set1_epi8(1); |
14187 | let b = _mm256_set1_epi8(1); |
14188 | let m = _mm256_cmpeq_epu8_mask(a, b); |
14189 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14190 | } |
14191 | |
14192 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14193 | unsafe fn test_mm256_mask_cmpeq_epu8_mask() { |
14194 | let a = _mm256_set1_epi8(1); |
14195 | let b = _mm256_set1_epi8(1); |
14196 | let mask = 0b01010101_01010101_01010101_01010101; |
14197 | let r = _mm256_mask_cmpeq_epu8_mask(mask, a, b); |
14198 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14199 | } |
14200 | |
14201 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14202 | unsafe fn test_mm_cmpeq_epu8_mask() { |
14203 | let a = _mm_set1_epi8(1); |
14204 | let b = _mm_set1_epi8(1); |
14205 | let m = _mm_cmpeq_epu8_mask(a, b); |
14206 | assert_eq!(m, 0b11111111_11111111); |
14207 | } |
14208 | |
14209 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14210 | unsafe fn test_mm_mask_cmpeq_epu8_mask() { |
14211 | let a = _mm_set1_epi8(1); |
14212 | let b = _mm_set1_epi8(1); |
14213 | let mask = 0b01010101_01010101; |
14214 | let r = _mm_mask_cmpeq_epu8_mask(mask, a, b); |
14215 | assert_eq!(r, 0b01010101_01010101); |
14216 | } |
14217 | |
14218 | #[simd_test(enable = "avx512bw" )] |
14219 | unsafe fn test_mm512_cmpeq_epi16_mask() { |
14220 | let a = _mm512_set1_epi16(-1); |
14221 | let b = _mm512_set1_epi16(-1); |
14222 | let m = _mm512_cmpeq_epi16_mask(a, b); |
14223 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14224 | } |
14225 | |
14226 | #[simd_test(enable = "avx512bw" )] |
14227 | unsafe fn test_mm512_mask_cmpeq_epi16_mask() { |
14228 | let a = _mm512_set1_epi16(-1); |
14229 | let b = _mm512_set1_epi16(-1); |
14230 | let mask = 0b01010101_01010101_01010101_01010101; |
14231 | let r = _mm512_mask_cmpeq_epi16_mask(mask, a, b); |
14232 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14233 | } |
14234 | |
14235 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14236 | unsafe fn test_mm256_cmpeq_epi16_mask() { |
14237 | let a = _mm256_set1_epi16(-1); |
14238 | let b = _mm256_set1_epi16(-1); |
14239 | let m = _mm256_cmpeq_epi16_mask(a, b); |
14240 | assert_eq!(m, 0b11111111_11111111); |
14241 | } |
14242 | |
14243 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14244 | unsafe fn test_mm256_mask_cmpeq_epi16_mask() { |
14245 | let a = _mm256_set1_epi16(-1); |
14246 | let b = _mm256_set1_epi16(-1); |
14247 | let mask = 0b01010101_01010101; |
14248 | let r = _mm256_mask_cmpeq_epi16_mask(mask, a, b); |
14249 | assert_eq!(r, 0b01010101_01010101); |
14250 | } |
14251 | |
14252 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14253 | unsafe fn test_mm_cmpeq_epi16_mask() { |
14254 | let a = _mm_set1_epi16(-1); |
14255 | let b = _mm_set1_epi16(-1); |
14256 | let m = _mm_cmpeq_epi16_mask(a, b); |
14257 | assert_eq!(m, 0b11111111); |
14258 | } |
14259 | |
14260 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14261 | unsafe fn test_mm_mask_cmpeq_epi16_mask() { |
14262 | let a = _mm_set1_epi16(-1); |
14263 | let b = _mm_set1_epi16(-1); |
14264 | let mask = 0b01010101; |
14265 | let r = _mm_mask_cmpeq_epi16_mask(mask, a, b); |
14266 | assert_eq!(r, 0b01010101); |
14267 | } |
14268 | |
14269 | #[simd_test(enable = "avx512bw" )] |
14270 | unsafe fn test_mm512_cmpeq_epi8_mask() { |
14271 | let a = _mm512_set1_epi8(-1); |
14272 | let b = _mm512_set1_epi8(-1); |
14273 | let m = _mm512_cmpeq_epi8_mask(a, b); |
14274 | assert_eq!( |
14275 | m, |
14276 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
14277 | ); |
14278 | } |
14279 | |
14280 | #[simd_test(enable = "avx512bw" )] |
14281 | unsafe fn test_mm512_mask_cmpeq_epi8_mask() { |
14282 | let a = _mm512_set1_epi8(-1); |
14283 | let b = _mm512_set1_epi8(-1); |
14284 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
14285 | let r = _mm512_mask_cmpeq_epi8_mask(mask, a, b); |
14286 | assert_eq!( |
14287 | r, |
14288 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
14289 | ); |
14290 | } |
14291 | |
14292 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14293 | unsafe fn test_mm256_cmpeq_epi8_mask() { |
14294 | let a = _mm256_set1_epi8(-1); |
14295 | let b = _mm256_set1_epi8(-1); |
14296 | let m = _mm256_cmpeq_epi8_mask(a, b); |
14297 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14298 | } |
14299 | |
14300 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14301 | unsafe fn test_mm256_mask_cmpeq_epi8_mask() { |
14302 | let a = _mm256_set1_epi8(-1); |
14303 | let b = _mm256_set1_epi8(-1); |
14304 | let mask = 0b01010101_01010101_01010101_01010101; |
14305 | let r = _mm256_mask_cmpeq_epi8_mask(mask, a, b); |
14306 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14307 | } |
14308 | |
14309 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14310 | unsafe fn test_mm_cmpeq_epi8_mask() { |
14311 | let a = _mm_set1_epi8(-1); |
14312 | let b = _mm_set1_epi8(-1); |
14313 | let m = _mm_cmpeq_epi8_mask(a, b); |
14314 | assert_eq!(m, 0b11111111_11111111); |
14315 | } |
14316 | |
14317 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14318 | unsafe fn test_mm_mask_cmpeq_epi8_mask() { |
14319 | let a = _mm_set1_epi8(-1); |
14320 | let b = _mm_set1_epi8(-1); |
14321 | let mask = 0b01010101_01010101; |
14322 | let r = _mm_mask_cmpeq_epi8_mask(mask, a, b); |
14323 | assert_eq!(r, 0b01010101_01010101); |
14324 | } |
14325 | |
14326 | #[simd_test(enable = "avx512bw" )] |
14327 | unsafe fn test_mm512_cmpneq_epu16_mask() { |
14328 | let a = _mm512_set1_epi16(2); |
14329 | let b = _mm512_set1_epi16(1); |
14330 | let m = _mm512_cmpneq_epu16_mask(a, b); |
14331 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14332 | } |
14333 | |
14334 | #[simd_test(enable = "avx512bw" )] |
14335 | unsafe fn test_mm512_mask_cmpneq_epu16_mask() { |
14336 | let a = _mm512_set1_epi16(2); |
14337 | let b = _mm512_set1_epi16(1); |
14338 | let mask = 0b01010101_01010101_01010101_01010101; |
14339 | let r = _mm512_mask_cmpneq_epu16_mask(mask, a, b); |
14340 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14341 | } |
14342 | |
14343 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14344 | unsafe fn test_mm256_cmpneq_epu16_mask() { |
14345 | let a = _mm256_set1_epi16(2); |
14346 | let b = _mm256_set1_epi16(1); |
14347 | let m = _mm256_cmpneq_epu16_mask(a, b); |
14348 | assert_eq!(m, 0b11111111_11111111); |
14349 | } |
14350 | |
14351 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14352 | unsafe fn test_mm256_mask_cmpneq_epu16_mask() { |
14353 | let a = _mm256_set1_epi16(2); |
14354 | let b = _mm256_set1_epi16(1); |
14355 | let mask = 0b01010101_01010101; |
14356 | let r = _mm256_mask_cmpneq_epu16_mask(mask, a, b); |
14357 | assert_eq!(r, 0b01010101_01010101); |
14358 | } |
14359 | |
14360 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14361 | unsafe fn test_mm_cmpneq_epu16_mask() { |
14362 | let a = _mm_set1_epi16(2); |
14363 | let b = _mm_set1_epi16(1); |
14364 | let m = _mm_cmpneq_epu16_mask(a, b); |
14365 | assert_eq!(m, 0b11111111); |
14366 | } |
14367 | |
14368 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14369 | unsafe fn test_mm_mask_cmpneq_epu16_mask() { |
14370 | let a = _mm_set1_epi16(2); |
14371 | let b = _mm_set1_epi16(1); |
14372 | let mask = 0b01010101; |
14373 | let r = _mm_mask_cmpneq_epu16_mask(mask, a, b); |
14374 | assert_eq!(r, 0b01010101); |
14375 | } |
14376 | |
14377 | #[simd_test(enable = "avx512bw" )] |
14378 | unsafe fn test_mm512_cmpneq_epu8_mask() { |
14379 | let a = _mm512_set1_epi8(2); |
14380 | let b = _mm512_set1_epi8(1); |
14381 | let m = _mm512_cmpneq_epu8_mask(a, b); |
14382 | assert_eq!( |
14383 | m, |
14384 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
14385 | ); |
14386 | } |
14387 | |
14388 | #[simd_test(enable = "avx512bw" )] |
14389 | unsafe fn test_mm512_mask_cmpneq_epu8_mask() { |
14390 | let a = _mm512_set1_epi8(2); |
14391 | let b = _mm512_set1_epi8(1); |
14392 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
14393 | let r = _mm512_mask_cmpneq_epu8_mask(mask, a, b); |
14394 | assert_eq!( |
14395 | r, |
14396 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
14397 | ); |
14398 | } |
14399 | |
14400 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14401 | unsafe fn test_mm256_cmpneq_epu8_mask() { |
14402 | let a = _mm256_set1_epi8(2); |
14403 | let b = _mm256_set1_epi8(1); |
14404 | let m = _mm256_cmpneq_epu8_mask(a, b); |
14405 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14406 | } |
14407 | |
14408 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14409 | unsafe fn test_mm256_mask_cmpneq_epu8_mask() { |
14410 | let a = _mm256_set1_epi8(2); |
14411 | let b = _mm256_set1_epi8(1); |
14412 | let mask = 0b01010101_01010101_01010101_01010101; |
14413 | let r = _mm256_mask_cmpneq_epu8_mask(mask, a, b); |
14414 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14415 | } |
14416 | |
14417 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14418 | unsafe fn test_mm_cmpneq_epu8_mask() { |
14419 | let a = _mm_set1_epi8(2); |
14420 | let b = _mm_set1_epi8(1); |
14421 | let m = _mm_cmpneq_epu8_mask(a, b); |
14422 | assert_eq!(m, 0b11111111_11111111); |
14423 | } |
14424 | |
14425 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14426 | unsafe fn test_mm_mask_cmpneq_epu8_mask() { |
14427 | let a = _mm_set1_epi8(2); |
14428 | let b = _mm_set1_epi8(1); |
14429 | let mask = 0b01010101_01010101; |
14430 | let r = _mm_mask_cmpneq_epu8_mask(mask, a, b); |
14431 | assert_eq!(r, 0b01010101_01010101); |
14432 | } |
14433 | |
14434 | #[simd_test(enable = "avx512bw" )] |
14435 | unsafe fn test_mm512_cmpneq_epi16_mask() { |
14436 | let a = _mm512_set1_epi16(1); |
14437 | let b = _mm512_set1_epi16(-1); |
14438 | let m = _mm512_cmpneq_epi16_mask(a, b); |
14439 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14440 | } |
14441 | |
14442 | #[simd_test(enable = "avx512bw" )] |
14443 | unsafe fn test_mm512_mask_cmpneq_epi16_mask() { |
14444 | let a = _mm512_set1_epi16(1); |
14445 | let b = _mm512_set1_epi16(-1); |
14446 | let mask = 0b01010101_01010101_01010101_01010101; |
14447 | let r = _mm512_mask_cmpneq_epi16_mask(mask, a, b); |
14448 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14449 | } |
14450 | |
14451 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14452 | unsafe fn test_mm256_cmpneq_epi16_mask() { |
14453 | let a = _mm256_set1_epi16(1); |
14454 | let b = _mm256_set1_epi16(-1); |
14455 | let m = _mm256_cmpneq_epi16_mask(a, b); |
14456 | assert_eq!(m, 0b11111111_11111111); |
14457 | } |
14458 | |
14459 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14460 | unsafe fn test_mm256_mask_cmpneq_epi16_mask() { |
14461 | let a = _mm256_set1_epi16(1); |
14462 | let b = _mm256_set1_epi16(-1); |
14463 | let mask = 0b01010101_01010101; |
14464 | let r = _mm256_mask_cmpneq_epi16_mask(mask, a, b); |
14465 | assert_eq!(r, 0b01010101_01010101); |
14466 | } |
14467 | |
14468 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14469 | unsafe fn test_mm_cmpneq_epi16_mask() { |
14470 | let a = _mm_set1_epi16(1); |
14471 | let b = _mm_set1_epi16(-1); |
14472 | let m = _mm_cmpneq_epi16_mask(a, b); |
14473 | assert_eq!(m, 0b11111111); |
14474 | } |
14475 | |
14476 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14477 | unsafe fn test_mm_mask_cmpneq_epi16_mask() { |
14478 | let a = _mm_set1_epi16(1); |
14479 | let b = _mm_set1_epi16(-1); |
14480 | let mask = 0b01010101; |
14481 | let r = _mm_mask_cmpneq_epi16_mask(mask, a, b); |
14482 | assert_eq!(r, 0b01010101); |
14483 | } |
14484 | |
14485 | #[simd_test(enable = "avx512bw" )] |
14486 | unsafe fn test_mm512_cmpneq_epi8_mask() { |
14487 | let a = _mm512_set1_epi8(1); |
14488 | let b = _mm512_set1_epi8(-1); |
14489 | let m = _mm512_cmpneq_epi8_mask(a, b); |
14490 | assert_eq!( |
14491 | m, |
14492 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
14493 | ); |
14494 | } |
14495 | |
14496 | #[simd_test(enable = "avx512bw" )] |
14497 | unsafe fn test_mm512_mask_cmpneq_epi8_mask() { |
14498 | let a = _mm512_set1_epi8(1); |
14499 | let b = _mm512_set1_epi8(-1); |
14500 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
14501 | let r = _mm512_mask_cmpneq_epi8_mask(mask, a, b); |
14502 | assert_eq!( |
14503 | r, |
14504 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
14505 | ); |
14506 | } |
14507 | |
14508 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14509 | unsafe fn test_mm256_cmpneq_epi8_mask() { |
14510 | let a = _mm256_set1_epi8(1); |
14511 | let b = _mm256_set1_epi8(-1); |
14512 | let m = _mm256_cmpneq_epi8_mask(a, b); |
14513 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14514 | } |
14515 | |
14516 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14517 | unsafe fn test_mm256_mask_cmpneq_epi8_mask() { |
14518 | let a = _mm256_set1_epi8(1); |
14519 | let b = _mm256_set1_epi8(-1); |
14520 | let mask = 0b01010101_01010101_01010101_01010101; |
14521 | let r = _mm256_mask_cmpneq_epi8_mask(mask, a, b); |
14522 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14523 | } |
14524 | |
14525 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14526 | unsafe fn test_mm_cmpneq_epi8_mask() { |
14527 | let a = _mm_set1_epi8(1); |
14528 | let b = _mm_set1_epi8(-1); |
14529 | let m = _mm_cmpneq_epi8_mask(a, b); |
14530 | assert_eq!(m, 0b11111111_11111111); |
14531 | } |
14532 | |
14533 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14534 | unsafe fn test_mm_mask_cmpneq_epi8_mask() { |
14535 | let a = _mm_set1_epi8(1); |
14536 | let b = _mm_set1_epi8(-1); |
14537 | let mask = 0b01010101_01010101; |
14538 | let r = _mm_mask_cmpneq_epi8_mask(mask, a, b); |
14539 | assert_eq!(r, 0b01010101_01010101); |
14540 | } |
14541 | |
14542 | #[simd_test(enable = "avx512bw" )] |
14543 | unsafe fn test_mm512_cmp_epu16_mask() { |
14544 | let a = _mm512_set1_epi16(0); |
14545 | let b = _mm512_set1_epi16(1); |
14546 | let m = _mm512_cmp_epu16_mask::<_MM_CMPINT_LT>(a, b); |
14547 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14548 | } |
14549 | |
14550 | #[simd_test(enable = "avx512bw" )] |
14551 | unsafe fn test_mm512_mask_cmp_epu16_mask() { |
14552 | let a = _mm512_set1_epi16(0); |
14553 | let b = _mm512_set1_epi16(1); |
14554 | let mask = 0b01010101_01010101_01010101_01010101; |
14555 | let r = _mm512_mask_cmp_epu16_mask::<_MM_CMPINT_LT>(mask, a, b); |
14556 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14557 | } |
14558 | |
14559 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14560 | unsafe fn test_mm256_cmp_epu16_mask() { |
14561 | let a = _mm256_set1_epi16(0); |
14562 | let b = _mm256_set1_epi16(1); |
14563 | let m = _mm256_cmp_epu16_mask::<_MM_CMPINT_LT>(a, b); |
14564 | assert_eq!(m, 0b11111111_11111111); |
14565 | } |
14566 | |
14567 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14568 | unsafe fn test_mm256_mask_cmp_epu16_mask() { |
14569 | let a = _mm256_set1_epi16(0); |
14570 | let b = _mm256_set1_epi16(1); |
14571 | let mask = 0b01010101_01010101; |
14572 | let r = _mm256_mask_cmp_epu16_mask::<_MM_CMPINT_LT>(mask, a, b); |
14573 | assert_eq!(r, 0b01010101_01010101); |
14574 | } |
14575 | |
14576 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14577 | unsafe fn test_mm_cmp_epu16_mask() { |
14578 | let a = _mm_set1_epi16(0); |
14579 | let b = _mm_set1_epi16(1); |
14580 | let m = _mm_cmp_epu16_mask::<_MM_CMPINT_LT>(a, b); |
14581 | assert_eq!(m, 0b11111111); |
14582 | } |
14583 | |
14584 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14585 | unsafe fn test_mm_mask_cmp_epu16_mask() { |
14586 | let a = _mm_set1_epi16(0); |
14587 | let b = _mm_set1_epi16(1); |
14588 | let mask = 0b01010101; |
14589 | let r = _mm_mask_cmp_epu16_mask::<_MM_CMPINT_LT>(mask, a, b); |
14590 | assert_eq!(r, 0b01010101); |
14591 | } |
14592 | |
14593 | #[simd_test(enable = "avx512bw" )] |
14594 | unsafe fn test_mm512_cmp_epu8_mask() { |
14595 | let a = _mm512_set1_epi8(0); |
14596 | let b = _mm512_set1_epi8(1); |
14597 | let m = _mm512_cmp_epu8_mask::<_MM_CMPINT_LT>(a, b); |
14598 | assert_eq!( |
14599 | m, |
14600 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
14601 | ); |
14602 | } |
14603 | |
14604 | #[simd_test(enable = "avx512bw" )] |
14605 | unsafe fn test_mm512_mask_cmp_epu8_mask() { |
14606 | let a = _mm512_set1_epi8(0); |
14607 | let b = _mm512_set1_epi8(1); |
14608 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
14609 | let r = _mm512_mask_cmp_epu8_mask::<_MM_CMPINT_LT>(mask, a, b); |
14610 | assert_eq!( |
14611 | r, |
14612 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
14613 | ); |
14614 | } |
14615 | |
14616 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14617 | unsafe fn test_mm256_cmp_epu8_mask() { |
14618 | let a = _mm256_set1_epi8(0); |
14619 | let b = _mm256_set1_epi8(1); |
14620 | let m = _mm256_cmp_epu8_mask::<_MM_CMPINT_LT>(a, b); |
14621 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14622 | } |
14623 | |
14624 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14625 | unsafe fn test_mm256_mask_cmp_epu8_mask() { |
14626 | let a = _mm256_set1_epi8(0); |
14627 | let b = _mm256_set1_epi8(1); |
14628 | let mask = 0b01010101_01010101_01010101_01010101; |
14629 | let r = _mm256_mask_cmp_epu8_mask::<_MM_CMPINT_LT>(mask, a, b); |
14630 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14631 | } |
14632 | |
14633 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14634 | unsafe fn test_mm_cmp_epu8_mask() { |
14635 | let a = _mm_set1_epi8(0); |
14636 | let b = _mm_set1_epi8(1); |
14637 | let m = _mm_cmp_epu8_mask::<_MM_CMPINT_LT>(a, b); |
14638 | assert_eq!(m, 0b11111111_11111111); |
14639 | } |
14640 | |
14641 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14642 | unsafe fn test_mm_mask_cmp_epu8_mask() { |
14643 | let a = _mm_set1_epi8(0); |
14644 | let b = _mm_set1_epi8(1); |
14645 | let mask = 0b01010101_01010101; |
14646 | let r = _mm_mask_cmp_epu8_mask::<_MM_CMPINT_LT>(mask, a, b); |
14647 | assert_eq!(r, 0b01010101_01010101); |
14648 | } |
14649 | |
14650 | #[simd_test(enable = "avx512bw" )] |
14651 | unsafe fn test_mm512_cmp_epi16_mask() { |
14652 | let a = _mm512_set1_epi16(0); |
14653 | let b = _mm512_set1_epi16(1); |
14654 | let m = _mm512_cmp_epi16_mask::<_MM_CMPINT_LT>(a, b); |
14655 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14656 | } |
14657 | |
14658 | #[simd_test(enable = "avx512bw" )] |
14659 | unsafe fn test_mm512_mask_cmp_epi16_mask() { |
14660 | let a = _mm512_set1_epi16(0); |
14661 | let b = _mm512_set1_epi16(1); |
14662 | let mask = 0b01010101_01010101_01010101_01010101; |
14663 | let r = _mm512_mask_cmp_epi16_mask::<_MM_CMPINT_LT>(mask, a, b); |
14664 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14665 | } |
14666 | |
14667 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14668 | unsafe fn test_mm256_cmp_epi16_mask() { |
14669 | let a = _mm256_set1_epi16(0); |
14670 | let b = _mm256_set1_epi16(1); |
14671 | let m = _mm256_cmp_epi16_mask::<_MM_CMPINT_LT>(a, b); |
14672 | assert_eq!(m, 0b11111111_11111111); |
14673 | } |
14674 | |
14675 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14676 | unsafe fn test_mm256_mask_cmp_epi16_mask() { |
14677 | let a = _mm256_set1_epi16(0); |
14678 | let b = _mm256_set1_epi16(1); |
14679 | let mask = 0b01010101_01010101; |
14680 | let r = _mm256_mask_cmp_epi16_mask::<_MM_CMPINT_LT>(mask, a, b); |
14681 | assert_eq!(r, 0b01010101_01010101); |
14682 | } |
14683 | |
14684 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14685 | unsafe fn test_mm_cmp_epi16_mask() { |
14686 | let a = _mm_set1_epi16(0); |
14687 | let b = _mm_set1_epi16(1); |
14688 | let m = _mm_cmp_epi16_mask::<_MM_CMPINT_LT>(a, b); |
14689 | assert_eq!(m, 0b11111111); |
14690 | } |
14691 | |
14692 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14693 | unsafe fn test_mm_mask_cmp_epi16_mask() { |
14694 | let a = _mm_set1_epi16(0); |
14695 | let b = _mm_set1_epi16(1); |
14696 | let mask = 0b01010101; |
14697 | let r = _mm_mask_cmp_epi16_mask::<_MM_CMPINT_LT>(mask, a, b); |
14698 | assert_eq!(r, 0b01010101); |
14699 | } |
14700 | |
14701 | #[simd_test(enable = "avx512bw" )] |
14702 | unsafe fn test_mm512_cmp_epi8_mask() { |
14703 | let a = _mm512_set1_epi8(0); |
14704 | let b = _mm512_set1_epi8(1); |
14705 | let m = _mm512_cmp_epi8_mask::<_MM_CMPINT_LT>(a, b); |
14706 | assert_eq!( |
14707 | m, |
14708 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111 |
14709 | ); |
14710 | } |
14711 | |
14712 | #[simd_test(enable = "avx512bw" )] |
14713 | unsafe fn test_mm512_mask_cmp_epi8_mask() { |
14714 | let a = _mm512_set1_epi8(0); |
14715 | let b = _mm512_set1_epi8(1); |
14716 | let mask = 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101; |
14717 | let r = _mm512_mask_cmp_epi8_mask::<_MM_CMPINT_LT>(mask, a, b); |
14718 | assert_eq!( |
14719 | r, |
14720 | 0b01010101_01010101_01010101_01010101_01010101_01010101_01010101_01010101 |
14721 | ); |
14722 | } |
14723 | |
14724 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14725 | unsafe fn test_mm256_cmp_epi8_mask() { |
14726 | let a = _mm256_set1_epi8(0); |
14727 | let b = _mm256_set1_epi8(1); |
14728 | let m = _mm256_cmp_epi8_mask::<_MM_CMPINT_LT>(a, b); |
14729 | assert_eq!(m, 0b11111111_11111111_11111111_11111111); |
14730 | } |
14731 | |
14732 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14733 | unsafe fn test_mm256_mask_cmp_epi8_mask() { |
14734 | let a = _mm256_set1_epi8(0); |
14735 | let b = _mm256_set1_epi8(1); |
14736 | let mask = 0b01010101_01010101_01010101_01010101; |
14737 | let r = _mm256_mask_cmp_epi8_mask::<_MM_CMPINT_LT>(mask, a, b); |
14738 | assert_eq!(r, 0b01010101_01010101_01010101_01010101); |
14739 | } |
14740 | |
14741 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14742 | unsafe fn test_mm_cmp_epi8_mask() { |
14743 | let a = _mm_set1_epi8(0); |
14744 | let b = _mm_set1_epi8(1); |
14745 | let m = _mm_cmp_epi8_mask::<_MM_CMPINT_LT>(a, b); |
14746 | assert_eq!(m, 0b11111111_11111111); |
14747 | } |
14748 | |
14749 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14750 | unsafe fn test_mm_mask_cmp_epi8_mask() { |
14751 | let a = _mm_set1_epi8(0); |
14752 | let b = _mm_set1_epi8(1); |
14753 | let mask = 0b01010101_01010101; |
14754 | let r = _mm_mask_cmp_epi8_mask::<_MM_CMPINT_LT>(mask, a, b); |
14755 | assert_eq!(r, 0b01010101_01010101); |
14756 | } |
14757 | |
14758 | #[simd_test(enable = "avx512bw" )] |
14759 | unsafe fn test_mm512_loadu_epi16() { |
14760 | #[rustfmt::skip] |
14761 | 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]; |
14762 | let r = _mm512_loadu_epi16(&a[0]); |
14763 | #[rustfmt::skip] |
14764 | 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); |
14765 | assert_eq_m512i(r, e); |
14766 | } |
14767 | |
14768 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14769 | unsafe fn test_mm256_loadu_epi16() { |
14770 | let a: [i16; 16] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
14771 | let r = _mm256_loadu_epi16(&a[0]); |
14772 | let e = _mm256_set_epi16(16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1); |
14773 | assert_eq_m256i(r, e); |
14774 | } |
14775 | |
14776 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14777 | unsafe fn test_mm_loadu_epi16() { |
14778 | let a: [i16; 8] = [1, 2, 3, 4, 5, 6, 7, 8]; |
14779 | let r = _mm_loadu_epi16(&a[0]); |
14780 | let e = _mm_set_epi16(8, 7, 6, 5, 4, 3, 2, 1); |
14781 | assert_eq_m128i(r, e); |
14782 | } |
14783 | |
14784 | #[simd_test(enable = "avx512bw" )] |
14785 | unsafe fn test_mm512_loadu_epi8() { |
14786 | #[rustfmt::skip] |
14787 | 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, |
14788 | 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]; |
14789 | let r = _mm512_loadu_epi8(&a[0]); |
14790 | #[rustfmt::skip] |
14791 | 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, |
14792 | 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); |
14793 | assert_eq_m512i(r, e); |
14794 | } |
14795 | |
14796 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14797 | unsafe fn test_mm256_loadu_epi8() { |
14798 | #[rustfmt::skip] |
14799 | 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]; |
14800 | let r = _mm256_loadu_epi8(&a[0]); |
14801 | #[rustfmt::skip] |
14802 | 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); |
14803 | assert_eq_m256i(r, e); |
14804 | } |
14805 | |
14806 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14807 | unsafe fn test_mm_loadu_epi8() { |
14808 | let a: [i8; 16] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
14809 | let r = _mm_loadu_epi8(&a[0]); |
14810 | let e = _mm_set_epi8(16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1); |
14811 | assert_eq_m128i(r, e); |
14812 | } |
14813 | |
14814 | #[simd_test(enable = "avx512bw" )] |
14815 | unsafe fn test_mm512_storeu_epi16() { |
14816 | let a = _mm512_set1_epi16(9); |
14817 | let mut r = _mm512_undefined_epi32(); |
14818 | _mm512_storeu_epi16(&mut r as *mut _ as *mut i16, a); |
14819 | assert_eq_m512i(r, a); |
14820 | } |
14821 | |
14822 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14823 | unsafe fn test_mm256_storeu_epi16() { |
14824 | let a = _mm256_set1_epi16(9); |
14825 | let mut r = _mm256_set1_epi32(0); |
14826 | _mm256_storeu_epi16(&mut r as *mut _ as *mut i16, a); |
14827 | assert_eq_m256i(r, a); |
14828 | } |
14829 | |
14830 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14831 | unsafe fn test_mm_storeu_epi16() { |
14832 | let a = _mm_set1_epi16(9); |
14833 | let mut r = _mm_set1_epi32(0); |
14834 | _mm_storeu_epi16(&mut r as *mut _ as *mut i16, a); |
14835 | assert_eq_m128i(r, a); |
14836 | } |
14837 | |
14838 | #[simd_test(enable = "avx512bw" )] |
14839 | unsafe fn test_mm512_storeu_epi8() { |
14840 | let a = _mm512_set1_epi8(9); |
14841 | let mut r = _mm512_undefined_epi32(); |
14842 | _mm512_storeu_epi8(&mut r as *mut _ as *mut i8, a); |
14843 | assert_eq_m512i(r, a); |
14844 | } |
14845 | |
14846 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14847 | unsafe fn test_mm256_storeu_epi8() { |
14848 | let a = _mm256_set1_epi8(9); |
14849 | let mut r = _mm256_set1_epi32(0); |
14850 | _mm256_storeu_epi8(&mut r as *mut _ as *mut i8, a); |
14851 | assert_eq_m256i(r, a); |
14852 | } |
14853 | |
14854 | #[simd_test(enable = "avx512bw,avx512vl" )] |
14855 | unsafe fn test_mm_storeu_epi8() { |
14856 | let a = _mm_set1_epi8(9); |
14857 | let mut r = _mm_set1_epi32(0); |
14858 | _mm_storeu_epi8(&mut r as *mut _ as *mut i8, a); |
14859 | assert_eq_m128i(r, a); |
14860 | } |
14861 | |
14862 | #[simd_test(enable = "avx512f,avx512bw" )] |
14863 | unsafe fn test_mm512_mask_loadu_epi16() { |
14864 | let src = _mm512_set1_epi16(42); |
14865 | let a = &[ |
14866 | 1_i16, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
14867 | 24, 25, 26, 27, 28, 29, 30, 31, 32, |
14868 | ]; |
14869 | let p = a.as_ptr(); |
14870 | let m = 0b10101010_11001100_11101000_11001010; |
14871 | let r = _mm512_mask_loadu_epi16(src, m, black_box(p)); |
14872 | let e = &[ |
14873 | 42_i16, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, 42, 42, 19, 20, 42, 42, |
14874 | 23, 24, 42, 26, 42, 28, 42, 30, 42, 32, |
14875 | ]; |
14876 | let e = _mm512_loadu_epi16(e.as_ptr()); |
14877 | assert_eq_m512i(r, e); |
14878 | } |
14879 | |
14880 | #[simd_test(enable = "avx512f,avx512bw" )] |
14881 | unsafe fn test_mm512_maskz_loadu_epi16() { |
14882 | let a = &[ |
14883 | 1_i16, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
14884 | 24, 25, 26, 27, 28, 29, 30, 31, 32, |
14885 | ]; |
14886 | let p = a.as_ptr(); |
14887 | let m = 0b10101010_11001100_11101000_11001010; |
14888 | let r = _mm512_maskz_loadu_epi16(m, black_box(p)); |
14889 | let e = &[ |
14890 | 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, |
14891 | 26, 0, 28, 0, 30, 0, 32, |
14892 | ]; |
14893 | let e = _mm512_loadu_epi16(e.as_ptr()); |
14894 | assert_eq_m512i(r, e); |
14895 | } |
14896 | |
14897 | #[simd_test(enable = "avx512f,avx512bw" )] |
14898 | unsafe fn test_mm512_mask_storeu_epi16() { |
14899 | let mut r = [42_i16; 32]; |
14900 | let a = &[ |
14901 | 1_i16, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
14902 | 24, 25, 26, 27, 28, 29, 30, 31, 32, |
14903 | ]; |
14904 | let a = _mm512_loadu_epi16(a.as_ptr()); |
14905 | let m = 0b10101010_11001100_11101000_11001010; |
14906 | _mm512_mask_storeu_epi16(r.as_mut_ptr(), m, a); |
14907 | let e = &[ |
14908 | 42_i16, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, 42, 42, 19, 20, 42, 42, |
14909 | 23, 24, 42, 26, 42, 28, 42, 30, 42, 32, |
14910 | ]; |
14911 | let e = _mm512_loadu_epi16(e.as_ptr()); |
14912 | assert_eq_m512i(_mm512_loadu_epi16(r.as_ptr()), e); |
14913 | } |
14914 | |
14915 | #[simd_test(enable = "avx512f,avx512bw" )] |
14916 | unsafe fn test_mm512_mask_loadu_epi8() { |
14917 | let src = _mm512_set1_epi8(42); |
14918 | let a = &[ |
14919 | 1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
14920 | 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, |
14921 | 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, |
14922 | ]; |
14923 | let p = a.as_ptr(); |
14924 | let m = 0b00000000_11111111_11111111_00000000_10101010_11001100_11101000_11001010; |
14925 | let r = _mm512_mask_loadu_epi8(src, m, black_box(p)); |
14926 | let e = &[ |
14927 | 42_i8, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, 42, 42, 19, 20, 42, 42, |
14928 | 23, 24, 42, 26, 42, 28, 42, 30, 42, 32, 42, 42, 42, 42, 42, 42, 42, 42, 41, 42, 43, 44, |
14929 | 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 42, 42, 42, 42, 42, 42, 42, 42, |
14930 | ]; |
14931 | let e = _mm512_loadu_epi8(e.as_ptr()); |
14932 | assert_eq_m512i(r, e); |
14933 | } |
14934 | |
14935 | #[simd_test(enable = "avx512f,avx512bw" )] |
14936 | unsafe fn test_mm512_maskz_loadu_epi8() { |
14937 | let a = &[ |
14938 | 1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
14939 | 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, |
14940 | 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, |
14941 | ]; |
14942 | let p = a.as_ptr(); |
14943 | let m = 0b00000000_11111111_11111111_00000000_10101010_11001100_11101000_11001010; |
14944 | let r = _mm512_maskz_loadu_epi8(m, black_box(p)); |
14945 | let e = &[ |
14946 | 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, |
14947 | 26, 0, 28, 0, 30, 0, 32, 0, 0, 0, 0, 0, 0, 0, 0, 41, 42, 43, 44, 45, 46, 47, 48, 49, |
14948 | 50, 51, 52, 53, 54, 55, 56, 0, 0, 0, 0, 0, 0, 0, 0, |
14949 | ]; |
14950 | let e = _mm512_loadu_epi8(e.as_ptr()); |
14951 | assert_eq_m512i(r, e); |
14952 | } |
14953 | |
14954 | #[simd_test(enable = "avx512f,avx512bw" )] |
14955 | unsafe fn test_mm512_mask_storeu_epi8() { |
14956 | let mut r = [42_i8; 64]; |
14957 | let a = &[ |
14958 | 1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
14959 | 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, |
14960 | 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, |
14961 | ]; |
14962 | let a = _mm512_loadu_epi8(a.as_ptr()); |
14963 | let m = 0b00000000_11111111_11111111_00000000_10101010_11001100_11101000_11001010; |
14964 | _mm512_mask_storeu_epi8(r.as_mut_ptr(), m, a); |
14965 | let e = &[ |
14966 | 42_i8, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, 42, 42, 19, 20, 42, 42, |
14967 | 23, 24, 42, 26, 42, 28, 42, 30, 42, 32, 42, 42, 42, 42, 42, 42, 42, 42, 41, 42, 43, 44, |
14968 | 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 42, 42, 42, 42, 42, 42, 42, 42, |
14969 | ]; |
14970 | let e = _mm512_loadu_epi8(e.as_ptr()); |
14971 | assert_eq_m512i(_mm512_loadu_epi8(r.as_ptr()), e); |
14972 | } |
14973 | |
14974 | #[simd_test(enable = "avx512f,avx512bw,avx512vl" )] |
14975 | unsafe fn test_mm256_mask_loadu_epi16() { |
14976 | let src = _mm256_set1_epi16(42); |
14977 | let a = &[1_i16, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
14978 | let p = a.as_ptr(); |
14979 | let m = 0b11101000_11001010; |
14980 | let r = _mm256_mask_loadu_epi16(src, m, black_box(p)); |
14981 | let e = &[ |
14982 | 42_i16, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, |
14983 | ]; |
14984 | let e = _mm256_loadu_epi16(e.as_ptr()); |
14985 | assert_eq_m256i(r, e); |
14986 | } |
14987 | |
14988 | #[simd_test(enable = "avx512f,avx512bw,avx512vl" )] |
14989 | unsafe fn test_mm256_maskz_loadu_epi16() { |
14990 | let a = &[1_i16, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
14991 | let p = a.as_ptr(); |
14992 | let m = 0b11101000_11001010; |
14993 | let r = _mm256_maskz_loadu_epi16(m, black_box(p)); |
14994 | let e = &[0_i16, 2, 0, 4, 0, 0, 7, 8, 0, 0, 0, 12, 0, 14, 15, 16]; |
14995 | let e = _mm256_loadu_epi16(e.as_ptr()); |
14996 | assert_eq_m256i(r, e); |
14997 | } |
14998 | |
14999 | #[simd_test(enable = "avx512f,avx512bw,avx512vl" )] |
15000 | unsafe fn test_mm256_mask_storeu_epi16() { |
15001 | let mut r = [42_i16; 16]; |
15002 | let a = &[1_i16, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
15003 | let a = _mm256_loadu_epi16(a.as_ptr()); |
15004 | let m = 0b11101000_11001010; |
15005 | _mm256_mask_storeu_epi16(r.as_mut_ptr(), m, a); |
15006 | let e = &[ |
15007 | 42_i16, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, |
15008 | ]; |
15009 | let e = _mm256_loadu_epi16(e.as_ptr()); |
15010 | assert_eq_m256i(_mm256_loadu_epi16(r.as_ptr()), e); |
15011 | } |
15012 | |
15013 | #[simd_test(enable = "avx512f,avx512bw,avx512vl" )] |
15014 | unsafe fn test_mm256_mask_loadu_epi8() { |
15015 | let src = _mm256_set1_epi8(42); |
15016 | let a = &[ |
15017 | 1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
15018 | 24, 25, 26, 27, 28, 29, 30, 31, 32, |
15019 | ]; |
15020 | let p = a.as_ptr(); |
15021 | let m = 0b10101010_11001100_11101000_11001010; |
15022 | let r = _mm256_mask_loadu_epi8(src, m, black_box(p)); |
15023 | let e = &[ |
15024 | 42_i8, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, 42, 42, 19, 20, 42, 42, |
15025 | 23, 24, 42, 26, 42, 28, 42, 30, 42, 32, |
15026 | ]; |
15027 | let e = _mm256_loadu_epi8(e.as_ptr()); |
15028 | assert_eq_m256i(r, e); |
15029 | } |
15030 | |
15031 | #[simd_test(enable = "avx512f,avx512bw,avx512vl" )] |
15032 | unsafe fn test_mm256_maskz_loadu_epi8() { |
15033 | let a = &[ |
15034 | 1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
15035 | 24, 25, 26, 27, 28, 29, 30, 31, 32, |
15036 | ]; |
15037 | let p = a.as_ptr(); |
15038 | let m = 0b10101010_11001100_11101000_11001010; |
15039 | let r = _mm256_maskz_loadu_epi8(m, black_box(p)); |
15040 | let e = &[ |
15041 | 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, |
15042 | 26, 0, 28, 0, 30, 0, 32, |
15043 | ]; |
15044 | let e = _mm256_loadu_epi8(e.as_ptr()); |
15045 | assert_eq_m256i(r, e); |
15046 | } |
15047 | |
15048 | #[simd_test(enable = "avx512f,avx512bw,avx512vl" )] |
15049 | unsafe fn test_mm256_mask_storeu_epi8() { |
15050 | let mut r = [42_i8; 32]; |
15051 | let a = &[ |
15052 | 1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, |
15053 | 24, 25, 26, 27, 28, 29, 30, 31, 32, |
15054 | ]; |
15055 | let a = _mm256_loadu_epi8(a.as_ptr()); |
15056 | let m = 0b10101010_11001100_11101000_11001010; |
15057 | _mm256_mask_storeu_epi8(r.as_mut_ptr(), m, a); |
15058 | let e = &[ |
15059 | 42_i8, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, 42, 42, 19, 20, 42, 42, |
15060 | 23, 24, 42, 26, 42, 28, 42, 30, 42, 32, |
15061 | ]; |
15062 | let e = _mm256_loadu_epi8(e.as_ptr()); |
15063 | assert_eq_m256i(_mm256_loadu_epi8(r.as_ptr()), e); |
15064 | } |
15065 | |
15066 | #[simd_test(enable = "avx512f,avx512bw,avx512vl" )] |
15067 | unsafe fn test_mm_mask_loadu_epi16() { |
15068 | let src = _mm_set1_epi16(42); |
15069 | let a = &[1_i16, 2, 3, 4, 5, 6, 7, 8]; |
15070 | let p = a.as_ptr(); |
15071 | let m = 0b11001010; |
15072 | let r = _mm_mask_loadu_epi16(src, m, black_box(p)); |
15073 | let e = &[42_i16, 2, 42, 4, 42, 42, 7, 8]; |
15074 | let e = _mm_loadu_epi16(e.as_ptr()); |
15075 | assert_eq_m128i(r, e); |
15076 | } |
15077 | |
15078 | #[simd_test(enable = "avx512f,avx512bw,avx512vl" )] |
15079 | unsafe fn test_mm_maskz_loadu_epi16() { |
15080 | let a = &[1_i16, 2, 3, 4, 5, 6, 7, 8]; |
15081 | let p = a.as_ptr(); |
15082 | let m = 0b11001010; |
15083 | let r = _mm_maskz_loadu_epi16(m, black_box(p)); |
15084 | let e = &[0_i16, 2, 0, 4, 0, 0, 7, 8]; |
15085 | let e = _mm_loadu_epi16(e.as_ptr()); |
15086 | assert_eq_m128i(r, e); |
15087 | } |
15088 | |
15089 | #[simd_test(enable = "avx512f,avx512bw,avx512vl" )] |
15090 | unsafe fn test_mm_mask_storeu_epi16() { |
15091 | let mut r = [42_i16; 8]; |
15092 | let a = &[1_i16, 2, 3, 4, 5, 6, 7, 8]; |
15093 | let a = _mm_loadu_epi16(a.as_ptr()); |
15094 | let m = 0b11001010; |
15095 | _mm_mask_storeu_epi16(r.as_mut_ptr(), m, a); |
15096 | let e = &[42_i16, 2, 42, 4, 42, 42, 7, 8]; |
15097 | let e = _mm_loadu_epi16(e.as_ptr()); |
15098 | assert_eq_m128i(_mm_loadu_epi16(r.as_ptr()), e); |
15099 | } |
15100 | |
15101 | #[simd_test(enable = "avx512f,avx512bw,avx512vl" )] |
15102 | unsafe fn test_mm_mask_loadu_epi8() { |
15103 | let src = _mm_set1_epi8(42); |
15104 | let a = &[1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
15105 | let p = a.as_ptr(); |
15106 | let m = 0b11101000_11001010; |
15107 | let r = _mm_mask_loadu_epi8(src, m, black_box(p)); |
15108 | let e = &[ |
15109 | 42_i8, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, |
15110 | ]; |
15111 | let e = _mm_loadu_epi8(e.as_ptr()); |
15112 | assert_eq_m128i(r, e); |
15113 | } |
15114 | |
15115 | #[simd_test(enable = "avx512f,avx512bw,avx512vl" )] |
15116 | unsafe fn test_mm_maskz_loadu_epi8() { |
15117 | let a = &[1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
15118 | let p = a.as_ptr(); |
15119 | let m = 0b11101000_11001010; |
15120 | let r = _mm_maskz_loadu_epi8(m, black_box(p)); |
15121 | let e = &[0_i8, 2, 0, 4, 0, 0, 7, 8, 0, 0, 0, 12, 0, 14, 15, 16]; |
15122 | let e = _mm_loadu_epi8(e.as_ptr()); |
15123 | assert_eq_m128i(r, e); |
15124 | } |
15125 | |
15126 | #[simd_test(enable = "avx512f,avx512bw,avx512vl" )] |
15127 | unsafe fn test_mm_mask_storeu_epi8() { |
15128 | let mut r = [42_i8; 16]; |
15129 | let a = &[1_i8, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]; |
15130 | let a = _mm_loadu_epi8(a.as_ptr()); |
15131 | let m = 0b11101000_11001010; |
15132 | _mm_mask_storeu_epi8(r.as_mut_ptr(), m, a); |
15133 | let e = &[ |
15134 | 42_i8, 2, 42, 4, 42, 42, 7, 8, 42, 42, 42, 12, 42, 14, 15, 16, |
15135 | ]; |
15136 | let e = _mm_loadu_epi8(e.as_ptr()); |
15137 | assert_eq_m128i(_mm_loadu_epi8(r.as_ptr()), e); |
15138 | } |
15139 | |
15140 | #[simd_test(enable = "avx512bw" )] |
15141 | unsafe fn test_mm512_madd_epi16() { |
15142 | let a = _mm512_set1_epi16(1); |
15143 | let b = _mm512_set1_epi16(1); |
15144 | let r = _mm512_madd_epi16(a, b); |
15145 | let e = _mm512_set1_epi32(2); |
15146 | assert_eq_m512i(r, e); |
15147 | } |
15148 | |
15149 | #[simd_test(enable = "avx512bw" )] |
15150 | unsafe fn test_mm512_mask_madd_epi16() { |
15151 | let a = _mm512_set1_epi16(1); |
15152 | let b = _mm512_set1_epi16(1); |
15153 | let r = _mm512_mask_madd_epi16(a, 0, a, b); |
15154 | assert_eq_m512i(r, a); |
15155 | let r = _mm512_mask_madd_epi16(a, 0b00000000_00001111, a, b); |
15156 | let e = _mm512_set_epi32( |
15157 | 1 << 16 | 1, |
15158 | 1 << 16 | 1, |
15159 | 1 << 16 | 1, |
15160 | 1 << 16 | 1, |
15161 | 1 << 16 | 1, |
15162 | 1 << 16 | 1, |
15163 | 1 << 16 | 1, |
15164 | 1 << 16 | 1, |
15165 | 1 << 16 | 1, |
15166 | 1 << 16 | 1, |
15167 | 1 << 16 | 1, |
15168 | 1 << 16 | 1, |
15169 | 2, |
15170 | 2, |
15171 | 2, |
15172 | 2, |
15173 | ); |
15174 | assert_eq_m512i(r, e); |
15175 | } |
15176 | |
15177 | #[simd_test(enable = "avx512bw" )] |
15178 | unsafe fn test_mm512_maskz_madd_epi16() { |
15179 | let a = _mm512_set1_epi16(1); |
15180 | let b = _mm512_set1_epi16(1); |
15181 | let r = _mm512_maskz_madd_epi16(0, a, b); |
15182 | assert_eq_m512i(r, _mm512_setzero_si512()); |
15183 | let r = _mm512_maskz_madd_epi16(0b00000000_00001111, a, b); |
15184 | let e = _mm512_set_epi32(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2); |
15185 | assert_eq_m512i(r, e); |
15186 | } |
15187 | |
15188 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15189 | unsafe fn test_mm256_mask_madd_epi16() { |
15190 | let a = _mm256_set1_epi16(1); |
15191 | let b = _mm256_set1_epi16(1); |
15192 | let r = _mm256_mask_madd_epi16(a, 0, a, b); |
15193 | assert_eq_m256i(r, a); |
15194 | let r = _mm256_mask_madd_epi16(a, 0b00001111, a, b); |
15195 | let e = _mm256_set_epi32( |
15196 | 1 << 16 | 1, |
15197 | 1 << 16 | 1, |
15198 | 1 << 16 | 1, |
15199 | 1 << 16 | 1, |
15200 | 2, |
15201 | 2, |
15202 | 2, |
15203 | 2, |
15204 | ); |
15205 | assert_eq_m256i(r, e); |
15206 | } |
15207 | |
15208 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15209 | unsafe fn test_mm256_maskz_madd_epi16() { |
15210 | let a = _mm256_set1_epi16(1); |
15211 | let b = _mm256_set1_epi16(1); |
15212 | let r = _mm256_maskz_madd_epi16(0, a, b); |
15213 | assert_eq_m256i(r, _mm256_setzero_si256()); |
15214 | let r = _mm256_maskz_madd_epi16(0b00001111, a, b); |
15215 | let e = _mm256_set_epi32(0, 0, 0, 0, 2, 2, 2, 2); |
15216 | assert_eq_m256i(r, e); |
15217 | } |
15218 | |
15219 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15220 | unsafe fn test_mm_mask_madd_epi16() { |
15221 | let a = _mm_set1_epi16(1); |
15222 | let b = _mm_set1_epi16(1); |
15223 | let r = _mm_mask_madd_epi16(a, 0, a, b); |
15224 | assert_eq_m128i(r, a); |
15225 | let r = _mm_mask_madd_epi16(a, 0b00001111, a, b); |
15226 | let e = _mm_set_epi32(2, 2, 2, 2); |
15227 | assert_eq_m128i(r, e); |
15228 | } |
15229 | |
15230 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15231 | unsafe fn test_mm_maskz_madd_epi16() { |
15232 | let a = _mm_set1_epi16(1); |
15233 | let b = _mm_set1_epi16(1); |
15234 | let r = _mm_maskz_madd_epi16(0, a, b); |
15235 | assert_eq_m128i(r, _mm_setzero_si128()); |
15236 | let r = _mm_maskz_madd_epi16(0b00001111, a, b); |
15237 | let e = _mm_set_epi32(2, 2, 2, 2); |
15238 | assert_eq_m128i(r, e); |
15239 | } |
15240 | |
15241 | #[simd_test(enable = "avx512bw" )] |
15242 | unsafe fn test_mm512_maddubs_epi16() { |
15243 | let a = _mm512_set1_epi8(1); |
15244 | let b = _mm512_set1_epi8(1); |
15245 | let r = _mm512_maddubs_epi16(a, b); |
15246 | let e = _mm512_set1_epi16(2); |
15247 | assert_eq_m512i(r, e); |
15248 | } |
15249 | |
15250 | #[simd_test(enable = "avx512bw" )] |
15251 | unsafe fn test_mm512_mask_maddubs_epi16() { |
15252 | let a = _mm512_set1_epi8(1); |
15253 | let b = _mm512_set1_epi8(1); |
15254 | let src = _mm512_set1_epi16(1); |
15255 | let r = _mm512_mask_maddubs_epi16(src, 0, a, b); |
15256 | assert_eq_m512i(r, src); |
15257 | let r = _mm512_mask_add_epi16(src, 0b00000000_00000000_00000000_00000001, a, b); |
15258 | #[rustfmt::skip] |
15259 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15260 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1<<9|2); |
15261 | assert_eq_m512i(r, e); |
15262 | } |
15263 | |
15264 | #[simd_test(enable = "avx512bw" )] |
15265 | unsafe fn test_mm512_maskz_maddubs_epi16() { |
15266 | let a = _mm512_set1_epi8(1); |
15267 | let b = _mm512_set1_epi8(1); |
15268 | let r = _mm512_maskz_maddubs_epi16(0, a, b); |
15269 | assert_eq_m512i(r, _mm512_setzero_si512()); |
15270 | let r = _mm512_maskz_maddubs_epi16(0b00000000_11111111_00000000_11111111, a, b); |
15271 | #[rustfmt::skip] |
15272 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, |
15273 | 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2); |
15274 | assert_eq_m512i(r, e); |
15275 | } |
15276 | |
15277 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15278 | unsafe fn test_mm256_mask_maddubs_epi16() { |
15279 | let a = _mm256_set1_epi8(1); |
15280 | let b = _mm256_set1_epi8(1); |
15281 | let src = _mm256_set1_epi16(1); |
15282 | let r = _mm256_mask_maddubs_epi16(src, 0, a, b); |
15283 | assert_eq_m256i(r, src); |
15284 | let r = _mm256_mask_add_epi16(src, 0b00000000_00000001, a, b); |
15285 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 << 9 | 2); |
15286 | assert_eq_m256i(r, e); |
15287 | } |
15288 | |
15289 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15290 | unsafe fn test_mm256_maskz_maddubs_epi16() { |
15291 | let a = _mm256_set1_epi8(1); |
15292 | let b = _mm256_set1_epi8(1); |
15293 | let r = _mm256_maskz_maddubs_epi16(0, a, b); |
15294 | assert_eq_m256i(r, _mm256_setzero_si256()); |
15295 | let r = _mm256_maskz_maddubs_epi16(0b00000000_11111111, a, b); |
15296 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2); |
15297 | assert_eq_m256i(r, e); |
15298 | } |
15299 | |
15300 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15301 | unsafe fn test_mm_mask_maddubs_epi16() { |
15302 | let a = _mm_set1_epi8(1); |
15303 | let b = _mm_set1_epi8(1); |
15304 | let src = _mm_set1_epi16(1); |
15305 | let r = _mm_mask_maddubs_epi16(src, 0, a, b); |
15306 | assert_eq_m128i(r, src); |
15307 | let r = _mm_mask_add_epi16(src, 0b00000001, a, b); |
15308 | let e = _mm_set_epi16(1, 1, 1, 1, 1, 1, 1, 1 << 9 | 2); |
15309 | assert_eq_m128i(r, e); |
15310 | } |
15311 | |
15312 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15313 | unsafe fn test_mm_maskz_maddubs_epi16() { |
15314 | let a = _mm_set1_epi8(1); |
15315 | let b = _mm_set1_epi8(1); |
15316 | let r = _mm_maskz_maddubs_epi16(0, a, b); |
15317 | assert_eq_m128i(r, _mm_setzero_si128()); |
15318 | let r = _mm_maskz_maddubs_epi16(0b00001111, a, b); |
15319 | let e = _mm_set_epi16(0, 0, 0, 0, 2, 2, 2, 2); |
15320 | assert_eq_m128i(r, e); |
15321 | } |
15322 | |
15323 | #[simd_test(enable = "avx512bw" )] |
15324 | unsafe fn test_mm512_packs_epi32() { |
15325 | let a = _mm512_set1_epi32(i32::MAX); |
15326 | let b = _mm512_set1_epi32(1); |
15327 | let r = _mm512_packs_epi32(a, b); |
15328 | #[rustfmt::skip] |
15329 | 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, |
15330 | 1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX, 1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
15331 | assert_eq_m512i(r, e); |
15332 | } |
15333 | |
15334 | #[simd_test(enable = "avx512bw" )] |
15335 | unsafe fn test_mm512_mask_packs_epi32() { |
15336 | let a = _mm512_set1_epi32(i32::MAX); |
15337 | let b = _mm512_set1_epi32(1 << 16 | 1); |
15338 | let r = _mm512_mask_packs_epi32(a, 0, a, b); |
15339 | assert_eq_m512i(r, a); |
15340 | let r = _mm512_mask_packs_epi32(b, 0b00000000_00000000_00000000_00001111, a, b); |
15341 | #[rustfmt::skip] |
15342 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15343 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
15344 | assert_eq_m512i(r, e); |
15345 | } |
15346 | |
15347 | #[simd_test(enable = "avx512bw" )] |
15348 | unsafe fn test_mm512_maskz_packs_epi32() { |
15349 | let a = _mm512_set1_epi32(i32::MAX); |
15350 | let b = _mm512_set1_epi32(1); |
15351 | let r = _mm512_maskz_packs_epi32(0, a, b); |
15352 | assert_eq_m512i(r, _mm512_setzero_si512()); |
15353 | let r = _mm512_maskz_packs_epi32(0b00000000_00000000_00000000_00001111, a, b); |
15354 | #[rustfmt::skip] |
15355 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15356 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
15357 | assert_eq_m512i(r, e); |
15358 | } |
15359 | |
15360 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15361 | unsafe fn test_mm256_mask_packs_epi32() { |
15362 | let a = _mm256_set1_epi32(i32::MAX); |
15363 | let b = _mm256_set1_epi32(1 << 16 | 1); |
15364 | let r = _mm256_mask_packs_epi32(a, 0, a, b); |
15365 | assert_eq_m256i(r, a); |
15366 | let r = _mm256_mask_packs_epi32(b, 0b00000000_00001111, a, b); |
15367 | #[rustfmt::skip] |
15368 | 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); |
15369 | assert_eq_m256i(r, e); |
15370 | } |
15371 | |
15372 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15373 | unsafe fn test_mm256_maskz_packs_epi32() { |
15374 | let a = _mm256_set1_epi32(i32::MAX); |
15375 | let b = _mm256_set1_epi32(1); |
15376 | let r = _mm256_maskz_packs_epi32(0, a, b); |
15377 | assert_eq_m256i(r, _mm256_setzero_si256()); |
15378 | let r = _mm256_maskz_packs_epi32(0b00000000_00001111, a, b); |
15379 | #[rustfmt::skip] |
15380 | 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); |
15381 | assert_eq_m256i(r, e); |
15382 | } |
15383 | |
15384 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15385 | unsafe fn test_mm_mask_packs_epi32() { |
15386 | let a = _mm_set1_epi32(i32::MAX); |
15387 | let b = _mm_set1_epi32(1 << 16 | 1); |
15388 | let r = _mm_mask_packs_epi32(a, 0, a, b); |
15389 | assert_eq_m128i(r, a); |
15390 | let r = _mm_mask_packs_epi32(b, 0b00001111, a, b); |
15391 | let e = _mm_set_epi16(1, 1, 1, 1, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
15392 | assert_eq_m128i(r, e); |
15393 | } |
15394 | |
15395 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15396 | unsafe fn test_mm_maskz_packs_epi32() { |
15397 | let a = _mm_set1_epi32(i32::MAX); |
15398 | let b = _mm_set1_epi32(1); |
15399 | let r = _mm_maskz_packs_epi32(0, a, b); |
15400 | assert_eq_m128i(r, _mm_setzero_si128()); |
15401 | let r = _mm_maskz_packs_epi32(0b00001111, a, b); |
15402 | let e = _mm_set_epi16(0, 0, 0, 0, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
15403 | assert_eq_m128i(r, e); |
15404 | } |
15405 | |
15406 | #[simd_test(enable = "avx512bw" )] |
15407 | unsafe fn test_mm512_packs_epi16() { |
15408 | let a = _mm512_set1_epi16(i16::MAX); |
15409 | let b = _mm512_set1_epi16(1); |
15410 | let r = _mm512_packs_epi16(a, b); |
15411 | #[rustfmt::skip] |
15412 | 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, |
15413 | 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, |
15414 | 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, |
15415 | 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
15416 | assert_eq_m512i(r, e); |
15417 | } |
15418 | |
15419 | #[simd_test(enable = "avx512bw" )] |
15420 | unsafe fn test_mm512_mask_packs_epi16() { |
15421 | let a = _mm512_set1_epi16(i16::MAX); |
15422 | let b = _mm512_set1_epi16(1 << 8 | 1); |
15423 | let r = _mm512_mask_packs_epi16(a, 0, a, b); |
15424 | assert_eq_m512i(r, a); |
15425 | let r = _mm512_mask_packs_epi16( |
15426 | b, |
15427 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
15428 | a, |
15429 | b, |
15430 | ); |
15431 | #[rustfmt::skip] |
15432 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15433 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15434 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15435 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
15436 | assert_eq_m512i(r, e); |
15437 | } |
15438 | |
15439 | #[simd_test(enable = "avx512bw" )] |
15440 | unsafe fn test_mm512_maskz_packs_epi16() { |
15441 | let a = _mm512_set1_epi16(i16::MAX); |
15442 | let b = _mm512_set1_epi16(1); |
15443 | let r = _mm512_maskz_packs_epi16(0, a, b); |
15444 | assert_eq_m512i(r, _mm512_setzero_si512()); |
15445 | let r = _mm512_maskz_packs_epi16( |
15446 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
15447 | a, |
15448 | b, |
15449 | ); |
15450 | #[rustfmt::skip] |
15451 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15452 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15453 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15454 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
15455 | assert_eq_m512i(r, e); |
15456 | } |
15457 | |
15458 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15459 | unsafe fn test_mm256_mask_packs_epi16() { |
15460 | let a = _mm256_set1_epi16(i16::MAX); |
15461 | let b = _mm256_set1_epi16(1 << 8 | 1); |
15462 | let r = _mm256_mask_packs_epi16(a, 0, a, b); |
15463 | assert_eq_m256i(r, a); |
15464 | let r = _mm256_mask_packs_epi16(b, 0b00000000_00000000_00000000_00001111, a, b); |
15465 | #[rustfmt::skip] |
15466 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15467 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
15468 | assert_eq_m256i(r, e); |
15469 | } |
15470 | |
15471 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15472 | unsafe fn test_mm256_maskz_packs_epi16() { |
15473 | let a = _mm256_set1_epi16(i16::MAX); |
15474 | let b = _mm256_set1_epi16(1); |
15475 | let r = _mm256_maskz_packs_epi16(0, a, b); |
15476 | assert_eq_m256i(r, _mm256_setzero_si256()); |
15477 | let r = _mm256_maskz_packs_epi16(0b00000000_00000000_00000000_00001111, a, b); |
15478 | #[rustfmt::skip] |
15479 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15480 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX, i8::MAX, i8::MAX); |
15481 | assert_eq_m256i(r, e); |
15482 | } |
15483 | |
15484 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15485 | unsafe fn test_mm_mask_packs_epi16() { |
15486 | let a = _mm_set1_epi16(i16::MAX); |
15487 | let b = _mm_set1_epi16(1 << 8 | 1); |
15488 | let r = _mm_mask_packs_epi16(a, 0, a, b); |
15489 | assert_eq_m128i(r, a); |
15490 | let r = _mm_mask_packs_epi16(b, 0b00000000_00001111, a, b); |
15491 | #[rustfmt::skip] |
15492 | 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); |
15493 | assert_eq_m128i(r, e); |
15494 | } |
15495 | |
15496 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15497 | unsafe fn test_mm_maskz_packs_epi16() { |
15498 | let a = _mm_set1_epi16(i16::MAX); |
15499 | let b = _mm_set1_epi16(1); |
15500 | let r = _mm_maskz_packs_epi16(0, a, b); |
15501 | assert_eq_m128i(r, _mm_setzero_si128()); |
15502 | let r = _mm_maskz_packs_epi16(0b00000000_00001111, a, b); |
15503 | #[rustfmt::skip] |
15504 | 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); |
15505 | assert_eq_m128i(r, e); |
15506 | } |
15507 | |
15508 | #[simd_test(enable = "avx512bw" )] |
15509 | unsafe fn test_mm512_packus_epi32() { |
15510 | let a = _mm512_set1_epi32(-1); |
15511 | let b = _mm512_set1_epi32(1); |
15512 | let r = _mm512_packus_epi32(a, b); |
15513 | #[rustfmt::skip] |
15514 | let e = _mm512_set_epi16(1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, |
15515 | 1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0); |
15516 | assert_eq_m512i(r, e); |
15517 | } |
15518 | |
15519 | #[simd_test(enable = "avx512bw" )] |
15520 | unsafe fn test_mm512_mask_packus_epi32() { |
15521 | let a = _mm512_set1_epi32(-1); |
15522 | let b = _mm512_set1_epi32(1 << 16 | 1); |
15523 | let r = _mm512_mask_packus_epi32(a, 0, a, b); |
15524 | assert_eq_m512i(r, a); |
15525 | let r = _mm512_mask_packus_epi32(b, 0b00000000_00000000_00000000_00001111, a, b); |
15526 | #[rustfmt::skip] |
15527 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15528 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
15529 | assert_eq_m512i(r, e); |
15530 | } |
15531 | |
15532 | #[simd_test(enable = "avx512bw" )] |
15533 | unsafe fn test_mm512_maskz_packus_epi32() { |
15534 | let a = _mm512_set1_epi32(-1); |
15535 | let b = _mm512_set1_epi32(1); |
15536 | let r = _mm512_maskz_packus_epi32(0, a, b); |
15537 | assert_eq_m512i(r, _mm512_setzero_si512()); |
15538 | let r = _mm512_maskz_packus_epi32(0b00000000_00000000_00000000_00001111, a, b); |
15539 | #[rustfmt::skip] |
15540 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15541 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
15542 | assert_eq_m512i(r, e); |
15543 | } |
15544 | |
15545 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15546 | unsafe fn test_mm256_mask_packus_epi32() { |
15547 | let a = _mm256_set1_epi32(-1); |
15548 | let b = _mm256_set1_epi32(1 << 16 | 1); |
15549 | let r = _mm256_mask_packus_epi32(a, 0, a, b); |
15550 | assert_eq_m256i(r, a); |
15551 | let r = _mm256_mask_packus_epi32(b, 0b00000000_00001111, a, b); |
15552 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
15553 | assert_eq_m256i(r, e); |
15554 | } |
15555 | |
15556 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15557 | unsafe fn test_mm256_maskz_packus_epi32() { |
15558 | let a = _mm256_set1_epi32(-1); |
15559 | let b = _mm256_set1_epi32(1); |
15560 | let r = _mm256_maskz_packus_epi32(0, a, b); |
15561 | assert_eq_m256i(r, _mm256_setzero_si256()); |
15562 | let r = _mm256_maskz_packus_epi32(0b00000000_00001111, a, b); |
15563 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
15564 | assert_eq_m256i(r, e); |
15565 | } |
15566 | |
15567 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15568 | unsafe fn test_mm_mask_packus_epi32() { |
15569 | let a = _mm_set1_epi32(-1); |
15570 | let b = _mm_set1_epi32(1 << 16 | 1); |
15571 | let r = _mm_mask_packus_epi32(a, 0, a, b); |
15572 | assert_eq_m128i(r, a); |
15573 | let r = _mm_mask_packus_epi32(b, 0b00001111, a, b); |
15574 | let e = _mm_set_epi16(1, 1, 1, 1, 0, 0, 0, 0); |
15575 | assert_eq_m128i(r, e); |
15576 | } |
15577 | |
15578 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15579 | unsafe fn test_mm_maskz_packus_epi32() { |
15580 | let a = _mm_set1_epi32(-1); |
15581 | let b = _mm_set1_epi32(1); |
15582 | let r = _mm_maskz_packus_epi32(0, a, b); |
15583 | assert_eq_m128i(r, _mm_setzero_si128()); |
15584 | let r = _mm_maskz_packus_epi32(0b00001111, a, b); |
15585 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 0, 0); |
15586 | assert_eq_m128i(r, e); |
15587 | } |
15588 | |
15589 | #[simd_test(enable = "avx512bw" )] |
15590 | unsafe fn test_mm512_packus_epi16() { |
15591 | let a = _mm512_set1_epi16(-1); |
15592 | let b = _mm512_set1_epi16(1); |
15593 | let r = _mm512_packus_epi16(a, b); |
15594 | #[rustfmt::skip] |
15595 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, |
15596 | 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, |
15597 | 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, |
15598 | 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0); |
15599 | assert_eq_m512i(r, e); |
15600 | } |
15601 | |
15602 | #[simd_test(enable = "avx512bw" )] |
15603 | unsafe fn test_mm512_mask_packus_epi16() { |
15604 | let a = _mm512_set1_epi16(-1); |
15605 | let b = _mm512_set1_epi16(1 << 8 | 1); |
15606 | let r = _mm512_mask_packus_epi16(a, 0, a, b); |
15607 | assert_eq_m512i(r, a); |
15608 | let r = _mm512_mask_packus_epi16( |
15609 | b, |
15610 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
15611 | a, |
15612 | b, |
15613 | ); |
15614 | #[rustfmt::skip] |
15615 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15616 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15617 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15618 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
15619 | assert_eq_m512i(r, e); |
15620 | } |
15621 | |
15622 | #[simd_test(enable = "avx512bw" )] |
15623 | unsafe fn test_mm512_maskz_packus_epi16() { |
15624 | let a = _mm512_set1_epi16(-1); |
15625 | let b = _mm512_set1_epi16(1); |
15626 | let r = _mm512_maskz_packus_epi16(0, a, b); |
15627 | assert_eq_m512i(r, _mm512_setzero_si512()); |
15628 | let r = _mm512_maskz_packus_epi16( |
15629 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
15630 | a, |
15631 | b, |
15632 | ); |
15633 | #[rustfmt::skip] |
15634 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15635 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15636 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15637 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
15638 | assert_eq_m512i(r, e); |
15639 | } |
15640 | |
15641 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15642 | unsafe fn test_mm256_mask_packus_epi16() { |
15643 | let a = _mm256_set1_epi16(-1); |
15644 | let b = _mm256_set1_epi16(1 << 8 | 1); |
15645 | let r = _mm256_mask_packus_epi16(a, 0, a, b); |
15646 | assert_eq_m256i(r, a); |
15647 | let r = _mm256_mask_packus_epi16(b, 0b00000000_00000000_00000000_00001111, a, b); |
15648 | #[rustfmt::skip] |
15649 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15650 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
15651 | assert_eq_m256i(r, e); |
15652 | } |
15653 | |
15654 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15655 | unsafe fn test_mm256_maskz_packus_epi16() { |
15656 | let a = _mm256_set1_epi16(-1); |
15657 | let b = _mm256_set1_epi16(1); |
15658 | let r = _mm256_maskz_packus_epi16(0, a, b); |
15659 | assert_eq_m256i(r, _mm256_setzero_si256()); |
15660 | let r = _mm256_maskz_packus_epi16(0b00000000_00000000_00000000_00001111, a, b); |
15661 | #[rustfmt::skip] |
15662 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15663 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
15664 | assert_eq_m256i(r, e); |
15665 | } |
15666 | |
15667 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15668 | unsafe fn test_mm_mask_packus_epi16() { |
15669 | let a = _mm_set1_epi16(-1); |
15670 | let b = _mm_set1_epi16(1 << 8 | 1); |
15671 | let r = _mm_mask_packus_epi16(a, 0, a, b); |
15672 | assert_eq_m128i(r, a); |
15673 | let r = _mm_mask_packus_epi16(b, 0b00000000_00001111, a, b); |
15674 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0); |
15675 | assert_eq_m128i(r, e); |
15676 | } |
15677 | |
15678 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15679 | unsafe fn test_mm_maskz_packus_epi16() { |
15680 | let a = _mm_set1_epi16(-1); |
15681 | let b = _mm_set1_epi16(1); |
15682 | let r = _mm_maskz_packus_epi16(0, a, b); |
15683 | assert_eq_m128i(r, _mm_setzero_si128()); |
15684 | let r = _mm_maskz_packus_epi16(0b00000000_00001111, a, b); |
15685 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
15686 | assert_eq_m128i(r, e); |
15687 | } |
15688 | |
15689 | #[simd_test(enable = "avx512bw" )] |
15690 | unsafe fn test_mm512_avg_epu16() { |
15691 | let a = _mm512_set1_epi16(1); |
15692 | let b = _mm512_set1_epi16(1); |
15693 | let r = _mm512_avg_epu16(a, b); |
15694 | let e = _mm512_set1_epi16(1); |
15695 | assert_eq_m512i(r, e); |
15696 | } |
15697 | |
15698 | #[simd_test(enable = "avx512bw" )] |
15699 | unsafe fn test_mm512_mask_avg_epu16() { |
15700 | let a = _mm512_set1_epi16(1); |
15701 | let b = _mm512_set1_epi16(1); |
15702 | let r = _mm512_mask_avg_epu16(a, 0, a, b); |
15703 | assert_eq_m512i(r, a); |
15704 | let r = _mm512_mask_avg_epu16(a, 0b00000000_00000000_00000000_00001111, a, b); |
15705 | #[rustfmt::skip] |
15706 | let e = _mm512_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15707 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
15708 | assert_eq_m512i(r, e); |
15709 | } |
15710 | |
15711 | #[simd_test(enable = "avx512bw" )] |
15712 | unsafe fn test_mm512_maskz_avg_epu16() { |
15713 | let a = _mm512_set1_epi16(1); |
15714 | let b = _mm512_set1_epi16(1); |
15715 | let r = _mm512_maskz_avg_epu16(0, a, b); |
15716 | assert_eq_m512i(r, _mm512_setzero_si512()); |
15717 | let r = _mm512_maskz_avg_epu16(0b00000000_00000000_00000000_00001111, a, b); |
15718 | #[rustfmt::skip] |
15719 | let e = _mm512_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15720 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
15721 | assert_eq_m512i(r, e); |
15722 | } |
15723 | |
15724 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15725 | unsafe fn test_mm256_mask_avg_epu16() { |
15726 | let a = _mm256_set1_epi16(1); |
15727 | let b = _mm256_set1_epi16(1); |
15728 | let r = _mm256_mask_avg_epu16(a, 0, a, b); |
15729 | assert_eq_m256i(r, a); |
15730 | let r = _mm256_mask_avg_epu16(a, 0b00000000_00001111, a, b); |
15731 | let e = _mm256_set_epi16(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
15732 | assert_eq_m256i(r, e); |
15733 | } |
15734 | |
15735 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15736 | unsafe fn test_mm256_maskz_avg_epu16() { |
15737 | let a = _mm256_set1_epi16(1); |
15738 | let b = _mm256_set1_epi16(1); |
15739 | let r = _mm256_maskz_avg_epu16(0, a, b); |
15740 | assert_eq_m256i(r, _mm256_setzero_si256()); |
15741 | let r = _mm256_maskz_avg_epu16(0b00000000_00001111, a, b); |
15742 | let e = _mm256_set_epi16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
15743 | assert_eq_m256i(r, e); |
15744 | } |
15745 | |
15746 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15747 | unsafe fn test_mm_mask_avg_epu16() { |
15748 | let a = _mm_set1_epi16(1); |
15749 | let b = _mm_set1_epi16(1); |
15750 | let r = _mm_mask_avg_epu16(a, 0, a, b); |
15751 | assert_eq_m128i(r, a); |
15752 | let r = _mm_mask_avg_epu16(a, 0b00001111, a, b); |
15753 | let e = _mm_set_epi16(1, 1, 1, 1, 1, 1, 1, 1); |
15754 | assert_eq_m128i(r, e); |
15755 | } |
15756 | |
15757 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15758 | unsafe fn test_mm_maskz_avg_epu16() { |
15759 | let a = _mm_set1_epi16(1); |
15760 | let b = _mm_set1_epi16(1); |
15761 | let r = _mm_maskz_avg_epu16(0, a, b); |
15762 | assert_eq_m128i(r, _mm_setzero_si128()); |
15763 | let r = _mm_maskz_avg_epu16(0b00001111, a, b); |
15764 | let e = _mm_set_epi16(0, 0, 0, 0, 1, 1, 1, 1); |
15765 | assert_eq_m128i(r, e); |
15766 | } |
15767 | |
15768 | #[simd_test(enable = "avx512bw" )] |
15769 | unsafe fn test_mm512_avg_epu8() { |
15770 | let a = _mm512_set1_epi8(1); |
15771 | let b = _mm512_set1_epi8(1); |
15772 | let r = _mm512_avg_epu8(a, b); |
15773 | let e = _mm512_set1_epi8(1); |
15774 | assert_eq_m512i(r, e); |
15775 | } |
15776 | |
15777 | #[simd_test(enable = "avx512bw" )] |
15778 | unsafe fn test_mm512_mask_avg_epu8() { |
15779 | let a = _mm512_set1_epi8(1); |
15780 | let b = _mm512_set1_epi8(1); |
15781 | let r = _mm512_mask_avg_epu8(a, 0, a, b); |
15782 | assert_eq_m512i(r, a); |
15783 | let r = _mm512_mask_avg_epu8( |
15784 | a, |
15785 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00001111, |
15786 | a, |
15787 | b, |
15788 | ); |
15789 | #[rustfmt::skip] |
15790 | let e = _mm512_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15791 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15792 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15793 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
15794 | assert_eq_m512i(r, e); |
15795 | } |
15796 | |
15797 | #[simd_test(enable = "avx512bw" )] |
15798 | unsafe fn test_mm512_maskz_avg_epu8() { |
15799 | let a = _mm512_set1_epi8(1); |
15800 | let b = _mm512_set1_epi8(1); |
15801 | let r = _mm512_maskz_avg_epu8(0, a, b); |
15802 | assert_eq_m512i(r, _mm512_setzero_si512()); |
15803 | let r = _mm512_maskz_avg_epu8( |
15804 | 0b00000000_000000000_00000000_00000000_00000000_0000000_00000000_00001111, |
15805 | a, |
15806 | b, |
15807 | ); |
15808 | #[rustfmt::skip] |
15809 | let e = _mm512_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15810 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15811 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15812 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
15813 | assert_eq_m512i(r, e); |
15814 | } |
15815 | |
15816 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15817 | unsafe fn test_mm256_mask_avg_epu8() { |
15818 | let a = _mm256_set1_epi8(1); |
15819 | let b = _mm256_set1_epi8(1); |
15820 | let r = _mm256_mask_avg_epu8(a, 0, a, b); |
15821 | assert_eq_m256i(r, a); |
15822 | let r = _mm256_mask_avg_epu8(a, 0b00000000_00000000_00000000_00001111, a, b); |
15823 | #[rustfmt::skip] |
15824 | let e = _mm256_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
15825 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
15826 | assert_eq_m256i(r, e); |
15827 | } |
15828 | |
15829 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15830 | unsafe fn test_mm256_maskz_avg_epu8() { |
15831 | let a = _mm256_set1_epi8(1); |
15832 | let b = _mm256_set1_epi8(1); |
15833 | let r = _mm256_maskz_avg_epu8(0, a, b); |
15834 | assert_eq_m256i(r, _mm256_setzero_si256()); |
15835 | let r = _mm256_maskz_avg_epu8(0b00000000_0000000_00000000_00001111, a, b); |
15836 | #[rustfmt::skip] |
15837 | let e = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
15838 | 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
15839 | assert_eq_m256i(r, e); |
15840 | } |
15841 | |
15842 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15843 | unsafe fn test_mm_mask_avg_epu8() { |
15844 | let a = _mm_set1_epi8(1); |
15845 | let b = _mm_set1_epi8(1); |
15846 | let r = _mm_mask_avg_epu8(a, 0, a, b); |
15847 | assert_eq_m128i(r, a); |
15848 | let r = _mm_mask_avg_epu8(a, 0b00000000_00001111, a, b); |
15849 | let e = _mm_set_epi8(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1); |
15850 | assert_eq_m128i(r, e); |
15851 | } |
15852 | |
15853 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15854 | unsafe fn test_mm_maskz_avg_epu8() { |
15855 | let a = _mm_set1_epi8(1); |
15856 | let b = _mm_set1_epi8(1); |
15857 | let r = _mm_maskz_avg_epu8(0, a, b); |
15858 | assert_eq_m128i(r, _mm_setzero_si128()); |
15859 | let r = _mm_maskz_avg_epu8(0b00000000_00001111, a, b); |
15860 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1); |
15861 | assert_eq_m128i(r, e); |
15862 | } |
15863 | |
15864 | #[simd_test(enable = "avx512bw" )] |
15865 | unsafe fn test_mm512_sll_epi16() { |
15866 | let a = _mm512_set1_epi16(1 << 15); |
15867 | let count = _mm_set1_epi16(2); |
15868 | let r = _mm512_sll_epi16(a, count); |
15869 | let e = _mm512_set1_epi16(0); |
15870 | assert_eq_m512i(r, e); |
15871 | } |
15872 | |
15873 | #[simd_test(enable = "avx512bw" )] |
15874 | unsafe fn test_mm512_mask_sll_epi16() { |
15875 | let a = _mm512_set1_epi16(1 << 15); |
15876 | let count = _mm_set1_epi16(2); |
15877 | let r = _mm512_mask_sll_epi16(a, 0, a, count); |
15878 | assert_eq_m512i(r, a); |
15879 | let r = _mm512_mask_sll_epi16(a, 0b11111111_11111111_11111111_11111111, a, count); |
15880 | let e = _mm512_set1_epi16(0); |
15881 | assert_eq_m512i(r, e); |
15882 | } |
15883 | |
15884 | #[simd_test(enable = "avx512bw" )] |
15885 | unsafe fn test_mm512_maskz_sll_epi16() { |
15886 | let a = _mm512_set1_epi16(1 << 15); |
15887 | let count = _mm_set1_epi16(2); |
15888 | let r = _mm512_maskz_sll_epi16(0, a, count); |
15889 | assert_eq_m512i(r, _mm512_setzero_si512()); |
15890 | let r = _mm512_maskz_sll_epi16(0b11111111_11111111_11111111_11111111, a, count); |
15891 | let e = _mm512_set1_epi16(0); |
15892 | assert_eq_m512i(r, e); |
15893 | } |
15894 | |
15895 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15896 | unsafe fn test_mm256_mask_sll_epi16() { |
15897 | let a = _mm256_set1_epi16(1 << 15); |
15898 | let count = _mm_set1_epi16(2); |
15899 | let r = _mm256_mask_sll_epi16(a, 0, a, count); |
15900 | assert_eq_m256i(r, a); |
15901 | let r = _mm256_mask_sll_epi16(a, 0b11111111_11111111, a, count); |
15902 | let e = _mm256_set1_epi16(0); |
15903 | assert_eq_m256i(r, e); |
15904 | } |
15905 | |
15906 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15907 | unsafe fn test_mm256_maskz_sll_epi16() { |
15908 | let a = _mm256_set1_epi16(1 << 15); |
15909 | let count = _mm_set1_epi16(2); |
15910 | let r = _mm256_maskz_sll_epi16(0, a, count); |
15911 | assert_eq_m256i(r, _mm256_setzero_si256()); |
15912 | let r = _mm256_maskz_sll_epi16(0b11111111_11111111, a, count); |
15913 | let e = _mm256_set1_epi16(0); |
15914 | assert_eq_m256i(r, e); |
15915 | } |
15916 | |
15917 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15918 | unsafe fn test_mm_mask_sll_epi16() { |
15919 | let a = _mm_set1_epi16(1 << 15); |
15920 | let count = _mm_set1_epi16(2); |
15921 | let r = _mm_mask_sll_epi16(a, 0, a, count); |
15922 | assert_eq_m128i(r, a); |
15923 | let r = _mm_mask_sll_epi16(a, 0b11111111, a, count); |
15924 | let e = _mm_set1_epi16(0); |
15925 | assert_eq_m128i(r, e); |
15926 | } |
15927 | |
15928 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15929 | unsafe fn test_mm_maskz_sll_epi16() { |
15930 | let a = _mm_set1_epi16(1 << 15); |
15931 | let count = _mm_set1_epi16(2); |
15932 | let r = _mm_maskz_sll_epi16(0, a, count); |
15933 | assert_eq_m128i(r, _mm_setzero_si128()); |
15934 | let r = _mm_maskz_sll_epi16(0b11111111, a, count); |
15935 | let e = _mm_set1_epi16(0); |
15936 | assert_eq_m128i(r, e); |
15937 | } |
15938 | |
15939 | #[simd_test(enable = "avx512bw" )] |
15940 | unsafe fn test_mm512_slli_epi16() { |
15941 | let a = _mm512_set1_epi16(1 << 15); |
15942 | let r = _mm512_slli_epi16::<1>(a); |
15943 | let e = _mm512_set1_epi16(0); |
15944 | assert_eq_m512i(r, e); |
15945 | } |
15946 | |
15947 | #[simd_test(enable = "avx512bw" )] |
15948 | unsafe fn test_mm512_mask_slli_epi16() { |
15949 | let a = _mm512_set1_epi16(1 << 15); |
15950 | let r = _mm512_mask_slli_epi16::<1>(a, 0, a); |
15951 | assert_eq_m512i(r, a); |
15952 | let r = _mm512_mask_slli_epi16::<1>(a, 0b11111111_11111111_11111111_11111111, a); |
15953 | let e = _mm512_set1_epi16(0); |
15954 | assert_eq_m512i(r, e); |
15955 | } |
15956 | |
15957 | #[simd_test(enable = "avx512bw" )] |
15958 | unsafe fn test_mm512_maskz_slli_epi16() { |
15959 | let a = _mm512_set1_epi16(1 << 15); |
15960 | let r = _mm512_maskz_slli_epi16::<1>(0, a); |
15961 | assert_eq_m512i(r, _mm512_setzero_si512()); |
15962 | let r = _mm512_maskz_slli_epi16::<1>(0b11111111_11111111_11111111_11111111, a); |
15963 | let e = _mm512_set1_epi16(0); |
15964 | assert_eq_m512i(r, e); |
15965 | } |
15966 | |
15967 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15968 | unsafe fn test_mm256_mask_slli_epi16() { |
15969 | let a = _mm256_set1_epi16(1 << 15); |
15970 | let r = _mm256_mask_slli_epi16::<1>(a, 0, a); |
15971 | assert_eq_m256i(r, a); |
15972 | let r = _mm256_mask_slli_epi16::<1>(a, 0b11111111_11111111, a); |
15973 | let e = _mm256_set1_epi16(0); |
15974 | assert_eq_m256i(r, e); |
15975 | } |
15976 | |
15977 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15978 | unsafe fn test_mm256_maskz_slli_epi16() { |
15979 | let a = _mm256_set1_epi16(1 << 15); |
15980 | let r = _mm256_maskz_slli_epi16::<1>(0, a); |
15981 | assert_eq_m256i(r, _mm256_setzero_si256()); |
15982 | let r = _mm256_maskz_slli_epi16::<1>(0b11111111_11111111, a); |
15983 | let e = _mm256_set1_epi16(0); |
15984 | assert_eq_m256i(r, e); |
15985 | } |
15986 | |
15987 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15988 | unsafe fn test_mm_mask_slli_epi16() { |
15989 | let a = _mm_set1_epi16(1 << 15); |
15990 | let r = _mm_mask_slli_epi16::<1>(a, 0, a); |
15991 | assert_eq_m128i(r, a); |
15992 | let r = _mm_mask_slli_epi16::<1>(a, 0b11111111, a); |
15993 | let e = _mm_set1_epi16(0); |
15994 | assert_eq_m128i(r, e); |
15995 | } |
15996 | |
15997 | #[simd_test(enable = "avx512bw,avx512vl" )] |
15998 | unsafe fn test_mm_maskz_slli_epi16() { |
15999 | let a = _mm_set1_epi16(1 << 15); |
16000 | let r = _mm_maskz_slli_epi16::<1>(0, a); |
16001 | assert_eq_m128i(r, _mm_setzero_si128()); |
16002 | let r = _mm_maskz_slli_epi16::<1>(0b11111111, a); |
16003 | let e = _mm_set1_epi16(0); |
16004 | assert_eq_m128i(r, e); |
16005 | } |
16006 | |
16007 | #[simd_test(enable = "avx512bw" )] |
16008 | unsafe fn test_mm512_sllv_epi16() { |
16009 | let a = _mm512_set1_epi16(1 << 15); |
16010 | let count = _mm512_set1_epi16(2); |
16011 | let r = _mm512_sllv_epi16(a, count); |
16012 | let e = _mm512_set1_epi16(0); |
16013 | assert_eq_m512i(r, e); |
16014 | } |
16015 | |
16016 | #[simd_test(enable = "avx512bw" )] |
16017 | unsafe fn test_mm512_mask_sllv_epi16() { |
16018 | let a = _mm512_set1_epi16(1 << 15); |
16019 | let count = _mm512_set1_epi16(2); |
16020 | let r = _mm512_mask_sllv_epi16(a, 0, a, count); |
16021 | assert_eq_m512i(r, a); |
16022 | let r = _mm512_mask_sllv_epi16(a, 0b11111111_11111111_11111111_11111111, a, count); |
16023 | let e = _mm512_set1_epi16(0); |
16024 | assert_eq_m512i(r, e); |
16025 | } |
16026 | |
16027 | #[simd_test(enable = "avx512bw" )] |
16028 | unsafe fn test_mm512_maskz_sllv_epi16() { |
16029 | let a = _mm512_set1_epi16(1 << 15); |
16030 | let count = _mm512_set1_epi16(2); |
16031 | let r = _mm512_maskz_sllv_epi16(0, a, count); |
16032 | assert_eq_m512i(r, _mm512_setzero_si512()); |
16033 | let r = _mm512_maskz_sllv_epi16(0b11111111_11111111_11111111_11111111, a, count); |
16034 | let e = _mm512_set1_epi16(0); |
16035 | assert_eq_m512i(r, e); |
16036 | } |
16037 | |
16038 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16039 | unsafe fn test_mm256_sllv_epi16() { |
16040 | let a = _mm256_set1_epi16(1 << 15); |
16041 | let count = _mm256_set1_epi16(2); |
16042 | let r = _mm256_sllv_epi16(a, count); |
16043 | let e = _mm256_set1_epi16(0); |
16044 | assert_eq_m256i(r, e); |
16045 | } |
16046 | |
16047 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16048 | unsafe fn test_mm256_mask_sllv_epi16() { |
16049 | let a = _mm256_set1_epi16(1 << 15); |
16050 | let count = _mm256_set1_epi16(2); |
16051 | let r = _mm256_mask_sllv_epi16(a, 0, a, count); |
16052 | assert_eq_m256i(r, a); |
16053 | let r = _mm256_mask_sllv_epi16(a, 0b11111111_11111111, a, count); |
16054 | let e = _mm256_set1_epi16(0); |
16055 | assert_eq_m256i(r, e); |
16056 | } |
16057 | |
16058 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16059 | unsafe fn test_mm256_maskz_sllv_epi16() { |
16060 | let a = _mm256_set1_epi16(1 << 15); |
16061 | let count = _mm256_set1_epi16(2); |
16062 | let r = _mm256_maskz_sllv_epi16(0, a, count); |
16063 | assert_eq_m256i(r, _mm256_setzero_si256()); |
16064 | let r = _mm256_maskz_sllv_epi16(0b11111111_11111111, a, count); |
16065 | let e = _mm256_set1_epi16(0); |
16066 | assert_eq_m256i(r, e); |
16067 | } |
16068 | |
16069 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16070 | unsafe fn test_mm_sllv_epi16() { |
16071 | let a = _mm_set1_epi16(1 << 15); |
16072 | let count = _mm_set1_epi16(2); |
16073 | let r = _mm_sllv_epi16(a, count); |
16074 | let e = _mm_set1_epi16(0); |
16075 | assert_eq_m128i(r, e); |
16076 | } |
16077 | |
16078 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16079 | unsafe fn test_mm_mask_sllv_epi16() { |
16080 | let a = _mm_set1_epi16(1 << 15); |
16081 | let count = _mm_set1_epi16(2); |
16082 | let r = _mm_mask_sllv_epi16(a, 0, a, count); |
16083 | assert_eq_m128i(r, a); |
16084 | let r = _mm_mask_sllv_epi16(a, 0b11111111, a, count); |
16085 | let e = _mm_set1_epi16(0); |
16086 | assert_eq_m128i(r, e); |
16087 | } |
16088 | |
16089 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16090 | unsafe fn test_mm_maskz_sllv_epi16() { |
16091 | let a = _mm_set1_epi16(1 << 15); |
16092 | let count = _mm_set1_epi16(2); |
16093 | let r = _mm_maskz_sllv_epi16(0, a, count); |
16094 | assert_eq_m128i(r, _mm_setzero_si128()); |
16095 | let r = _mm_maskz_sllv_epi16(0b11111111, a, count); |
16096 | let e = _mm_set1_epi16(0); |
16097 | assert_eq_m128i(r, e); |
16098 | } |
16099 | |
16100 | #[simd_test(enable = "avx512bw" )] |
16101 | unsafe fn test_mm512_srl_epi16() { |
16102 | let a = _mm512_set1_epi16(1 << 1); |
16103 | let count = _mm_set1_epi16(2); |
16104 | let r = _mm512_srl_epi16(a, count); |
16105 | let e = _mm512_set1_epi16(0); |
16106 | assert_eq_m512i(r, e); |
16107 | } |
16108 | |
16109 | #[simd_test(enable = "avx512bw" )] |
16110 | unsafe fn test_mm512_mask_srl_epi16() { |
16111 | let a = _mm512_set1_epi16(1 << 1); |
16112 | let count = _mm_set1_epi16(2); |
16113 | let r = _mm512_mask_srl_epi16(a, 0, a, count); |
16114 | assert_eq_m512i(r, a); |
16115 | let r = _mm512_mask_srl_epi16(a, 0b11111111_11111111_11111111_11111111, a, count); |
16116 | let e = _mm512_set1_epi16(0); |
16117 | assert_eq_m512i(r, e); |
16118 | } |
16119 | |
16120 | #[simd_test(enable = "avx512bw" )] |
16121 | unsafe fn test_mm512_maskz_srl_epi16() { |
16122 | let a = _mm512_set1_epi16(1 << 1); |
16123 | let count = _mm_set1_epi16(2); |
16124 | let r = _mm512_maskz_srl_epi16(0, a, count); |
16125 | assert_eq_m512i(r, _mm512_setzero_si512()); |
16126 | let r = _mm512_maskz_srl_epi16(0b11111111_11111111_11111111_11111111, a, count); |
16127 | let e = _mm512_set1_epi16(0); |
16128 | assert_eq_m512i(r, e); |
16129 | } |
16130 | |
16131 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16132 | unsafe fn test_mm256_mask_srl_epi16() { |
16133 | let a = _mm256_set1_epi16(1 << 1); |
16134 | let count = _mm_set1_epi16(2); |
16135 | let r = _mm256_mask_srl_epi16(a, 0, a, count); |
16136 | assert_eq_m256i(r, a); |
16137 | let r = _mm256_mask_srl_epi16(a, 0b11111111_11111111, a, count); |
16138 | let e = _mm256_set1_epi16(0); |
16139 | assert_eq_m256i(r, e); |
16140 | } |
16141 | |
16142 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16143 | unsafe fn test_mm256_maskz_srl_epi16() { |
16144 | let a = _mm256_set1_epi16(1 << 1); |
16145 | let count = _mm_set1_epi16(2); |
16146 | let r = _mm256_maskz_srl_epi16(0, a, count); |
16147 | assert_eq_m256i(r, _mm256_setzero_si256()); |
16148 | let r = _mm256_maskz_srl_epi16(0b11111111_11111111, a, count); |
16149 | let e = _mm256_set1_epi16(0); |
16150 | assert_eq_m256i(r, e); |
16151 | } |
16152 | |
16153 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16154 | unsafe fn test_mm_mask_srl_epi16() { |
16155 | let a = _mm_set1_epi16(1 << 1); |
16156 | let count = _mm_set1_epi16(2); |
16157 | let r = _mm_mask_srl_epi16(a, 0, a, count); |
16158 | assert_eq_m128i(r, a); |
16159 | let r = _mm_mask_srl_epi16(a, 0b11111111, a, count); |
16160 | let e = _mm_set1_epi16(0); |
16161 | assert_eq_m128i(r, e); |
16162 | } |
16163 | |
16164 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16165 | unsafe fn test_mm_maskz_srl_epi16() { |
16166 | let a = _mm_set1_epi16(1 << 1); |
16167 | let count = _mm_set1_epi16(2); |
16168 | let r = _mm_maskz_srl_epi16(0, a, count); |
16169 | assert_eq_m128i(r, _mm_setzero_si128()); |
16170 | let r = _mm_maskz_srl_epi16(0b11111111, a, count); |
16171 | let e = _mm_set1_epi16(0); |
16172 | assert_eq_m128i(r, e); |
16173 | } |
16174 | |
16175 | #[simd_test(enable = "avx512bw" )] |
16176 | unsafe fn test_mm512_srli_epi16() { |
16177 | let a = _mm512_set1_epi16(1 << 1); |
16178 | let r = _mm512_srli_epi16::<2>(a); |
16179 | let e = _mm512_set1_epi16(0); |
16180 | assert_eq_m512i(r, e); |
16181 | } |
16182 | |
16183 | #[simd_test(enable = "avx512bw" )] |
16184 | unsafe fn test_mm512_mask_srli_epi16() { |
16185 | let a = _mm512_set1_epi16(1 << 1); |
16186 | let r = _mm512_mask_srli_epi16::<2>(a, 0, a); |
16187 | assert_eq_m512i(r, a); |
16188 | let r = _mm512_mask_srli_epi16::<2>(a, 0b11111111_11111111_11111111_11111111, a); |
16189 | let e = _mm512_set1_epi16(0); |
16190 | assert_eq_m512i(r, e); |
16191 | } |
16192 | |
16193 | #[simd_test(enable = "avx512bw" )] |
16194 | unsafe fn test_mm512_maskz_srli_epi16() { |
16195 | let a = _mm512_set1_epi16(1 << 1); |
16196 | let r = _mm512_maskz_srli_epi16::<2>(0, a); |
16197 | assert_eq_m512i(r, _mm512_setzero_si512()); |
16198 | let r = _mm512_maskz_srli_epi16::<2>(0b11111111_11111111_11111111_11111111, a); |
16199 | let e = _mm512_set1_epi16(0); |
16200 | assert_eq_m512i(r, e); |
16201 | } |
16202 | |
16203 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16204 | unsafe fn test_mm256_mask_srli_epi16() { |
16205 | let a = _mm256_set1_epi16(1 << 1); |
16206 | let r = _mm256_mask_srli_epi16::<2>(a, 0, a); |
16207 | assert_eq_m256i(r, a); |
16208 | let r = _mm256_mask_srli_epi16::<2>(a, 0b11111111_11111111, a); |
16209 | let e = _mm256_set1_epi16(0); |
16210 | assert_eq_m256i(r, e); |
16211 | } |
16212 | |
16213 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16214 | unsafe fn test_mm256_maskz_srli_epi16() { |
16215 | let a = _mm256_set1_epi16(1 << 1); |
16216 | let r = _mm256_maskz_srli_epi16::<2>(0, a); |
16217 | assert_eq_m256i(r, _mm256_setzero_si256()); |
16218 | let r = _mm256_maskz_srli_epi16::<2>(0b11111111_11111111, a); |
16219 | let e = _mm256_set1_epi16(0); |
16220 | assert_eq_m256i(r, e); |
16221 | } |
16222 | |
16223 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16224 | unsafe fn test_mm_mask_srli_epi16() { |
16225 | let a = _mm_set1_epi16(1 << 1); |
16226 | let r = _mm_mask_srli_epi16::<2>(a, 0, a); |
16227 | assert_eq_m128i(r, a); |
16228 | let r = _mm_mask_srli_epi16::<2>(a, 0b11111111, a); |
16229 | let e = _mm_set1_epi16(0); |
16230 | assert_eq_m128i(r, e); |
16231 | } |
16232 | |
16233 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16234 | unsafe fn test_mm_maskz_srli_epi16() { |
16235 | let a = _mm_set1_epi16(1 << 1); |
16236 | let r = _mm_maskz_srli_epi16::<2>(0, a); |
16237 | assert_eq_m128i(r, _mm_setzero_si128()); |
16238 | let r = _mm_maskz_srli_epi16::<2>(0b11111111, a); |
16239 | let e = _mm_set1_epi16(0); |
16240 | assert_eq_m128i(r, e); |
16241 | } |
16242 | |
16243 | #[simd_test(enable = "avx512bw" )] |
16244 | unsafe fn test_mm512_srlv_epi16() { |
16245 | let a = _mm512_set1_epi16(1 << 1); |
16246 | let count = _mm512_set1_epi16(2); |
16247 | let r = _mm512_srlv_epi16(a, count); |
16248 | let e = _mm512_set1_epi16(0); |
16249 | assert_eq_m512i(r, e); |
16250 | } |
16251 | |
16252 | #[simd_test(enable = "avx512bw" )] |
16253 | unsafe fn test_mm512_mask_srlv_epi16() { |
16254 | let a = _mm512_set1_epi16(1 << 1); |
16255 | let count = _mm512_set1_epi16(2); |
16256 | let r = _mm512_mask_srlv_epi16(a, 0, a, count); |
16257 | assert_eq_m512i(r, a); |
16258 | let r = _mm512_mask_srlv_epi16(a, 0b11111111_11111111_11111111_11111111, a, count); |
16259 | let e = _mm512_set1_epi16(0); |
16260 | assert_eq_m512i(r, e); |
16261 | } |
16262 | |
16263 | #[simd_test(enable = "avx512bw" )] |
16264 | unsafe fn test_mm512_maskz_srlv_epi16() { |
16265 | let a = _mm512_set1_epi16(1 << 1); |
16266 | let count = _mm512_set1_epi16(2); |
16267 | let r = _mm512_maskz_srlv_epi16(0, a, count); |
16268 | assert_eq_m512i(r, _mm512_setzero_si512()); |
16269 | let r = _mm512_maskz_srlv_epi16(0b11111111_11111111_11111111_11111111, a, count); |
16270 | let e = _mm512_set1_epi16(0); |
16271 | assert_eq_m512i(r, e); |
16272 | } |
16273 | |
16274 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16275 | unsafe fn test_mm256_srlv_epi16() { |
16276 | let a = _mm256_set1_epi16(1 << 1); |
16277 | let count = _mm256_set1_epi16(2); |
16278 | let r = _mm256_srlv_epi16(a, count); |
16279 | let e = _mm256_set1_epi16(0); |
16280 | assert_eq_m256i(r, e); |
16281 | } |
16282 | |
16283 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16284 | unsafe fn test_mm256_mask_srlv_epi16() { |
16285 | let a = _mm256_set1_epi16(1 << 1); |
16286 | let count = _mm256_set1_epi16(2); |
16287 | let r = _mm256_mask_srlv_epi16(a, 0, a, count); |
16288 | assert_eq_m256i(r, a); |
16289 | let r = _mm256_mask_srlv_epi16(a, 0b11111111_11111111, a, count); |
16290 | let e = _mm256_set1_epi16(0); |
16291 | assert_eq_m256i(r, e); |
16292 | } |
16293 | |
16294 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16295 | unsafe fn test_mm256_maskz_srlv_epi16() { |
16296 | let a = _mm256_set1_epi16(1 << 1); |
16297 | let count = _mm256_set1_epi16(2); |
16298 | let r = _mm256_maskz_srlv_epi16(0, a, count); |
16299 | assert_eq_m256i(r, _mm256_setzero_si256()); |
16300 | let r = _mm256_maskz_srlv_epi16(0b11111111_11111111, a, count); |
16301 | let e = _mm256_set1_epi16(0); |
16302 | assert_eq_m256i(r, e); |
16303 | } |
16304 | |
16305 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16306 | unsafe fn test_mm_srlv_epi16() { |
16307 | let a = _mm_set1_epi16(1 << 1); |
16308 | let count = _mm_set1_epi16(2); |
16309 | let r = _mm_srlv_epi16(a, count); |
16310 | let e = _mm_set1_epi16(0); |
16311 | assert_eq_m128i(r, e); |
16312 | } |
16313 | |
16314 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16315 | unsafe fn test_mm_mask_srlv_epi16() { |
16316 | let a = _mm_set1_epi16(1 << 1); |
16317 | let count = _mm_set1_epi16(2); |
16318 | let r = _mm_mask_srlv_epi16(a, 0, a, count); |
16319 | assert_eq_m128i(r, a); |
16320 | let r = _mm_mask_srlv_epi16(a, 0b11111111, a, count); |
16321 | let e = _mm_set1_epi16(0); |
16322 | assert_eq_m128i(r, e); |
16323 | } |
16324 | |
16325 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16326 | unsafe fn test_mm_maskz_srlv_epi16() { |
16327 | let a = _mm_set1_epi16(1 << 1); |
16328 | let count = _mm_set1_epi16(2); |
16329 | let r = _mm_maskz_srlv_epi16(0, a, count); |
16330 | assert_eq_m128i(r, _mm_setzero_si128()); |
16331 | let r = _mm_maskz_srlv_epi16(0b11111111, a, count); |
16332 | let e = _mm_set1_epi16(0); |
16333 | assert_eq_m128i(r, e); |
16334 | } |
16335 | |
16336 | #[simd_test(enable = "avx512bw" )] |
16337 | unsafe fn test_mm512_sra_epi16() { |
16338 | let a = _mm512_set1_epi16(8); |
16339 | let count = _mm_set1_epi16(1); |
16340 | let r = _mm512_sra_epi16(a, count); |
16341 | let e = _mm512_set1_epi16(0); |
16342 | assert_eq_m512i(r, e); |
16343 | } |
16344 | |
16345 | #[simd_test(enable = "avx512bw" )] |
16346 | unsafe fn test_mm512_mask_sra_epi16() { |
16347 | let a = _mm512_set1_epi16(8); |
16348 | let count = _mm_set1_epi16(1); |
16349 | let r = _mm512_mask_sra_epi16(a, 0, a, count); |
16350 | assert_eq_m512i(r, a); |
16351 | let r = _mm512_mask_sra_epi16(a, 0b11111111_11111111_11111111_11111111, a, count); |
16352 | let e = _mm512_set1_epi16(0); |
16353 | assert_eq_m512i(r, e); |
16354 | } |
16355 | |
16356 | #[simd_test(enable = "avx512bw" )] |
16357 | unsafe fn test_mm512_maskz_sra_epi16() { |
16358 | let a = _mm512_set1_epi16(8); |
16359 | let count = _mm_set1_epi16(1); |
16360 | let r = _mm512_maskz_sra_epi16(0, a, count); |
16361 | assert_eq_m512i(r, _mm512_setzero_si512()); |
16362 | let r = _mm512_maskz_sra_epi16(0b11111111_11111111_11111111_11111111, a, count); |
16363 | let e = _mm512_set1_epi16(0); |
16364 | assert_eq_m512i(r, e); |
16365 | } |
16366 | |
16367 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16368 | unsafe fn test_mm256_mask_sra_epi16() { |
16369 | let a = _mm256_set1_epi16(8); |
16370 | let count = _mm_set1_epi16(1); |
16371 | let r = _mm256_mask_sra_epi16(a, 0, a, count); |
16372 | assert_eq_m256i(r, a); |
16373 | let r = _mm256_mask_sra_epi16(a, 0b11111111_11111111, a, count); |
16374 | let e = _mm256_set1_epi16(0); |
16375 | assert_eq_m256i(r, e); |
16376 | } |
16377 | |
16378 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16379 | unsafe fn test_mm256_maskz_sra_epi16() { |
16380 | let a = _mm256_set1_epi16(8); |
16381 | let count = _mm_set1_epi16(1); |
16382 | let r = _mm256_maskz_sra_epi16(0, a, count); |
16383 | assert_eq_m256i(r, _mm256_setzero_si256()); |
16384 | let r = _mm256_maskz_sra_epi16(0b11111111_11111111, a, count); |
16385 | let e = _mm256_set1_epi16(0); |
16386 | assert_eq_m256i(r, e); |
16387 | } |
16388 | |
16389 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16390 | unsafe fn test_mm_mask_sra_epi16() { |
16391 | let a = _mm_set1_epi16(8); |
16392 | let count = _mm_set1_epi16(1); |
16393 | let r = _mm_mask_sra_epi16(a, 0, a, count); |
16394 | assert_eq_m128i(r, a); |
16395 | let r = _mm_mask_sra_epi16(a, 0b11111111, a, count); |
16396 | let e = _mm_set1_epi16(0); |
16397 | assert_eq_m128i(r, e); |
16398 | } |
16399 | |
16400 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16401 | unsafe fn test_mm_maskz_sra_epi16() { |
16402 | let a = _mm_set1_epi16(8); |
16403 | let count = _mm_set1_epi16(1); |
16404 | let r = _mm_maskz_sra_epi16(0, a, count); |
16405 | assert_eq_m128i(r, _mm_setzero_si128()); |
16406 | let r = _mm_maskz_sra_epi16(0b11111111, a, count); |
16407 | let e = _mm_set1_epi16(0); |
16408 | assert_eq_m128i(r, e); |
16409 | } |
16410 | |
16411 | #[simd_test(enable = "avx512bw" )] |
16412 | unsafe fn test_mm512_srai_epi16() { |
16413 | let a = _mm512_set1_epi16(8); |
16414 | let r = _mm512_srai_epi16::<2>(a); |
16415 | let e = _mm512_set1_epi16(2); |
16416 | assert_eq_m512i(r, e); |
16417 | } |
16418 | |
16419 | #[simd_test(enable = "avx512bw" )] |
16420 | unsafe fn test_mm512_mask_srai_epi16() { |
16421 | let a = _mm512_set1_epi16(8); |
16422 | let r = _mm512_mask_srai_epi16::<2>(a, 0, a); |
16423 | assert_eq_m512i(r, a); |
16424 | let r = _mm512_mask_srai_epi16::<2>(a, 0b11111111_11111111_11111111_11111111, a); |
16425 | let e = _mm512_set1_epi16(2); |
16426 | assert_eq_m512i(r, e); |
16427 | } |
16428 | |
16429 | #[simd_test(enable = "avx512bw" )] |
16430 | unsafe fn test_mm512_maskz_srai_epi16() { |
16431 | let a = _mm512_set1_epi16(8); |
16432 | let r = _mm512_maskz_srai_epi16::<2>(0, a); |
16433 | assert_eq_m512i(r, _mm512_setzero_si512()); |
16434 | let r = _mm512_maskz_srai_epi16::<2>(0b11111111_11111111_11111111_11111111, a); |
16435 | let e = _mm512_set1_epi16(2); |
16436 | assert_eq_m512i(r, e); |
16437 | } |
16438 | |
16439 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16440 | unsafe fn test_mm256_mask_srai_epi16() { |
16441 | let a = _mm256_set1_epi16(8); |
16442 | let r = _mm256_mask_srai_epi16::<2>(a, 0, a); |
16443 | assert_eq_m256i(r, a); |
16444 | let r = _mm256_mask_srai_epi16::<2>(a, 0b11111111_11111111, a); |
16445 | let e = _mm256_set1_epi16(2); |
16446 | assert_eq_m256i(r, e); |
16447 | } |
16448 | |
16449 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16450 | unsafe fn test_mm256_maskz_srai_epi16() { |
16451 | let a = _mm256_set1_epi16(8); |
16452 | let r = _mm256_maskz_srai_epi16::<2>(0, a); |
16453 | assert_eq_m256i(r, _mm256_setzero_si256()); |
16454 | let r = _mm256_maskz_srai_epi16::<2>(0b11111111_11111111, a); |
16455 | let e = _mm256_set1_epi16(2); |
16456 | assert_eq_m256i(r, e); |
16457 | } |
16458 | |
16459 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16460 | unsafe fn test_mm_mask_srai_epi16() { |
16461 | let a = _mm_set1_epi16(8); |
16462 | let r = _mm_mask_srai_epi16::<2>(a, 0, a); |
16463 | assert_eq_m128i(r, a); |
16464 | let r = _mm_mask_srai_epi16::<2>(a, 0b11111111, a); |
16465 | let e = _mm_set1_epi16(2); |
16466 | assert_eq_m128i(r, e); |
16467 | } |
16468 | |
16469 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16470 | unsafe fn test_mm_maskz_srai_epi16() { |
16471 | let a = _mm_set1_epi16(8); |
16472 | let r = _mm_maskz_srai_epi16::<2>(0, a); |
16473 | assert_eq_m128i(r, _mm_setzero_si128()); |
16474 | let r = _mm_maskz_srai_epi16::<2>(0b11111111, a); |
16475 | let e = _mm_set1_epi16(2); |
16476 | assert_eq_m128i(r, e); |
16477 | } |
16478 | |
16479 | #[simd_test(enable = "avx512bw" )] |
16480 | unsafe fn test_mm512_srav_epi16() { |
16481 | let a = _mm512_set1_epi16(8); |
16482 | let count = _mm512_set1_epi16(2); |
16483 | let r = _mm512_srav_epi16(a, count); |
16484 | let e = _mm512_set1_epi16(2); |
16485 | assert_eq_m512i(r, e); |
16486 | } |
16487 | |
16488 | #[simd_test(enable = "avx512bw" )] |
16489 | unsafe fn test_mm512_mask_srav_epi16() { |
16490 | let a = _mm512_set1_epi16(8); |
16491 | let count = _mm512_set1_epi16(2); |
16492 | let r = _mm512_mask_srav_epi16(a, 0, a, count); |
16493 | assert_eq_m512i(r, a); |
16494 | let r = _mm512_mask_srav_epi16(a, 0b11111111_11111111_11111111_11111111, a, count); |
16495 | let e = _mm512_set1_epi16(2); |
16496 | assert_eq_m512i(r, e); |
16497 | } |
16498 | |
16499 | #[simd_test(enable = "avx512bw" )] |
16500 | unsafe fn test_mm512_maskz_srav_epi16() { |
16501 | let a = _mm512_set1_epi16(8); |
16502 | let count = _mm512_set1_epi16(2); |
16503 | let r = _mm512_maskz_srav_epi16(0, a, count); |
16504 | assert_eq_m512i(r, _mm512_setzero_si512()); |
16505 | let r = _mm512_maskz_srav_epi16(0b11111111_11111111_11111111_11111111, a, count); |
16506 | let e = _mm512_set1_epi16(2); |
16507 | assert_eq_m512i(r, e); |
16508 | } |
16509 | |
16510 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16511 | unsafe fn test_mm256_srav_epi16() { |
16512 | let a = _mm256_set1_epi16(8); |
16513 | let count = _mm256_set1_epi16(2); |
16514 | let r = _mm256_srav_epi16(a, count); |
16515 | let e = _mm256_set1_epi16(2); |
16516 | assert_eq_m256i(r, e); |
16517 | } |
16518 | |
16519 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16520 | unsafe fn test_mm256_mask_srav_epi16() { |
16521 | let a = _mm256_set1_epi16(8); |
16522 | let count = _mm256_set1_epi16(2); |
16523 | let r = _mm256_mask_srav_epi16(a, 0, a, count); |
16524 | assert_eq_m256i(r, a); |
16525 | let r = _mm256_mask_srav_epi16(a, 0b11111111_11111111, a, count); |
16526 | let e = _mm256_set1_epi16(2); |
16527 | assert_eq_m256i(r, e); |
16528 | } |
16529 | |
16530 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16531 | unsafe fn test_mm256_maskz_srav_epi16() { |
16532 | let a = _mm256_set1_epi16(8); |
16533 | let count = _mm256_set1_epi16(2); |
16534 | let r = _mm256_maskz_srav_epi16(0, a, count); |
16535 | assert_eq_m256i(r, _mm256_setzero_si256()); |
16536 | let r = _mm256_maskz_srav_epi16(0b11111111_11111111, a, count); |
16537 | let e = _mm256_set1_epi16(2); |
16538 | assert_eq_m256i(r, e); |
16539 | } |
16540 | |
16541 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16542 | unsafe fn test_mm_srav_epi16() { |
16543 | let a = _mm_set1_epi16(8); |
16544 | let count = _mm_set1_epi16(2); |
16545 | let r = _mm_srav_epi16(a, count); |
16546 | let e = _mm_set1_epi16(2); |
16547 | assert_eq_m128i(r, e); |
16548 | } |
16549 | |
16550 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16551 | unsafe fn test_mm_mask_srav_epi16() { |
16552 | let a = _mm_set1_epi16(8); |
16553 | let count = _mm_set1_epi16(2); |
16554 | let r = _mm_mask_srav_epi16(a, 0, a, count); |
16555 | assert_eq_m128i(r, a); |
16556 | let r = _mm_mask_srav_epi16(a, 0b11111111, a, count); |
16557 | let e = _mm_set1_epi16(2); |
16558 | assert_eq_m128i(r, e); |
16559 | } |
16560 | |
16561 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16562 | unsafe fn test_mm_maskz_srav_epi16() { |
16563 | let a = _mm_set1_epi16(8); |
16564 | let count = _mm_set1_epi16(2); |
16565 | let r = _mm_maskz_srav_epi16(0, a, count); |
16566 | assert_eq_m128i(r, _mm_setzero_si128()); |
16567 | let r = _mm_maskz_srav_epi16(0b11111111, a, count); |
16568 | let e = _mm_set1_epi16(2); |
16569 | assert_eq_m128i(r, e); |
16570 | } |
16571 | |
16572 | #[simd_test(enable = "avx512bw" )] |
16573 | unsafe fn test_mm512_permutex2var_epi16() { |
16574 | #[rustfmt::skip] |
16575 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
16576 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
16577 | #[rustfmt::skip] |
16578 | 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, |
16579 | 9, 1<<5, 10, 1<<5, 11, 1<<5, 12, 1<<5, 13, 1<<5, 14, 1<<5, 15, 1<<5, 16, 1<<5); |
16580 | let b = _mm512_set1_epi16(100); |
16581 | let r = _mm512_permutex2var_epi16(a, idx, b); |
16582 | #[rustfmt::skip] |
16583 | let e = _mm512_set_epi16( |
16584 | 30, 100, 29, 100, 28, 100, 27, 100, 26, 100, 25, 100, 24, 100, 23, 100, |
16585 | 22, 100, 21, 100, 20, 100, 19, 100, 18, 100, 17, 100, 16, 100, 15, 100, |
16586 | ); |
16587 | assert_eq_m512i(r, e); |
16588 | } |
16589 | |
16590 | #[simd_test(enable = "avx512bw" )] |
16591 | unsafe fn test_mm512_mask_permutex2var_epi16() { |
16592 | #[rustfmt::skip] |
16593 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
16594 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
16595 | #[rustfmt::skip] |
16596 | 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, |
16597 | 9, 1<<5, 10, 1<<5, 11, 1<<5, 12, 1<<5, 13, 1<<5, 14, 1<<5, 15, 1<<5, 16, 1<<5); |
16598 | let b = _mm512_set1_epi16(100); |
16599 | let r = _mm512_mask_permutex2var_epi16(a, 0, idx, b); |
16600 | assert_eq_m512i(r, a); |
16601 | let r = _mm512_mask_permutex2var_epi16(a, 0b11111111_11111111_11111111_11111111, idx, b); |
16602 | #[rustfmt::skip] |
16603 | let e = _mm512_set_epi16( |
16604 | 30, 100, 29, 100, 28, 100, 27, 100, 26, 100, 25, 100, 24, 100, 23, 100, |
16605 | 22, 100, 21, 100, 20, 100, 19, 100, 18, 100, 17, 100, 16, 100, 15, 100, |
16606 | ); |
16607 | assert_eq_m512i(r, e); |
16608 | } |
16609 | |
16610 | #[simd_test(enable = "avx512bw" )] |
16611 | unsafe fn test_mm512_maskz_permutex2var_epi16() { |
16612 | #[rustfmt::skip] |
16613 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
16614 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
16615 | #[rustfmt::skip] |
16616 | 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, |
16617 | 9, 1<<5, 10, 1<<5, 11, 1<<5, 12, 1<<5, 13, 1<<5, 14, 1<<5, 15, 1<<5, 16, 1<<5); |
16618 | let b = _mm512_set1_epi16(100); |
16619 | let r = _mm512_maskz_permutex2var_epi16(0, a, idx, b); |
16620 | assert_eq_m512i(r, _mm512_setzero_si512()); |
16621 | let r = _mm512_maskz_permutex2var_epi16(0b11111111_11111111_11111111_11111111, a, idx, b); |
16622 | #[rustfmt::skip] |
16623 | let e = _mm512_set_epi16( |
16624 | 30, 100, 29, 100, 28, 100, 27, 100, 26, 100, 25, 100, 24, 100, 23, 100, |
16625 | 22, 100, 21, 100, 20, 100, 19, 100, 18, 100, 17, 100, 16, 100, 15, 100, |
16626 | ); |
16627 | assert_eq_m512i(r, e); |
16628 | } |
16629 | |
16630 | #[simd_test(enable = "avx512bw" )] |
16631 | unsafe fn test_mm512_mask2_permutex2var_epi16() { |
16632 | #[rustfmt::skip] |
16633 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
16634 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
16635 | #[rustfmt::skip] |
16636 | 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, |
16637 | 9, 1<<5, 10, 1<<5, 11, 1<<5, 12, 1<<5, 13, 1<<5, 14, 1<<5, 15, 1<<5, 16, 1<<5); |
16638 | let b = _mm512_set1_epi16(100); |
16639 | let r = _mm512_mask2_permutex2var_epi16(a, idx, 0, b); |
16640 | assert_eq_m512i(r, idx); |
16641 | let r = _mm512_mask2_permutex2var_epi16(a, idx, 0b11111111_11111111_11111111_11111111, b); |
16642 | #[rustfmt::skip] |
16643 | let e = _mm512_set_epi16( |
16644 | 30, 100, 29, 100, 28, 100, 27, 100, 26, 100, 25, 100, 24, 100, 23, 100, |
16645 | 22, 100, 21, 100, 20, 100, 19, 100, 18, 100, 17, 100, 16, 100, 15, 100, |
16646 | ); |
16647 | assert_eq_m512i(r, e); |
16648 | } |
16649 | |
16650 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16651 | unsafe fn test_mm256_permutex2var_epi16() { |
16652 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
16653 | #[rustfmt::skip] |
16654 | 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); |
16655 | let b = _mm256_set1_epi16(100); |
16656 | let r = _mm256_permutex2var_epi16(a, idx, b); |
16657 | let e = _mm256_set_epi16( |
16658 | 14, 100, 13, 100, 12, 100, 11, 100, 10, 100, 9, 100, 8, 100, 7, 100, |
16659 | ); |
16660 | assert_eq_m256i(r, e); |
16661 | } |
16662 | |
16663 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16664 | unsafe fn test_mm256_mask_permutex2var_epi16() { |
16665 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
16666 | #[rustfmt::skip] |
16667 | 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); |
16668 | let b = _mm256_set1_epi16(100); |
16669 | let r = _mm256_mask_permutex2var_epi16(a, 0, idx, b); |
16670 | assert_eq_m256i(r, a); |
16671 | let r = _mm256_mask_permutex2var_epi16(a, 0b11111111_11111111, idx, b); |
16672 | let e = _mm256_set_epi16( |
16673 | 14, 100, 13, 100, 12, 100, 11, 100, 10, 100, 9, 100, 8, 100, 7, 100, |
16674 | ); |
16675 | assert_eq_m256i(r, e); |
16676 | } |
16677 | |
16678 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16679 | unsafe fn test_mm256_maskz_permutex2var_epi16() { |
16680 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
16681 | #[rustfmt::skip] |
16682 | 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); |
16683 | let b = _mm256_set1_epi16(100); |
16684 | let r = _mm256_maskz_permutex2var_epi16(0, a, idx, b); |
16685 | assert_eq_m256i(r, _mm256_setzero_si256()); |
16686 | let r = _mm256_maskz_permutex2var_epi16(0b11111111_11111111, a, idx, b); |
16687 | let e = _mm256_set_epi16( |
16688 | 14, 100, 13, 100, 12, 100, 11, 100, 10, 100, 9, 100, 8, 100, 7, 100, |
16689 | ); |
16690 | assert_eq_m256i(r, e); |
16691 | } |
16692 | |
16693 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16694 | unsafe fn test_mm256_mask2_permutex2var_epi16() { |
16695 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
16696 | #[rustfmt::skip] |
16697 | 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); |
16698 | let b = _mm256_set1_epi16(100); |
16699 | let r = _mm256_mask2_permutex2var_epi16(a, idx, 0, b); |
16700 | assert_eq_m256i(r, idx); |
16701 | let r = _mm256_mask2_permutex2var_epi16(a, idx, 0b11111111_11111111, b); |
16702 | #[rustfmt::skip] |
16703 | let e = _mm256_set_epi16( |
16704 | 14, 100, 13, 100, 12, 100, 11, 100, 10, 100, 9, 100, 8, 100, 7, 100, |
16705 | ); |
16706 | assert_eq_m256i(r, e); |
16707 | } |
16708 | |
16709 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16710 | unsafe fn test_mm_permutex2var_epi16() { |
16711 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
16712 | let idx = _mm_set_epi16(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
16713 | let b = _mm_set1_epi16(100); |
16714 | let r = _mm_permutex2var_epi16(a, idx, b); |
16715 | let e = _mm_set_epi16(6, 100, 5, 100, 4, 100, 3, 100); |
16716 | assert_eq_m128i(r, e); |
16717 | } |
16718 | |
16719 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16720 | unsafe fn test_mm_mask_permutex2var_epi16() { |
16721 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
16722 | let idx = _mm_set_epi16(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
16723 | let b = _mm_set1_epi16(100); |
16724 | let r = _mm_mask_permutex2var_epi16(a, 0, idx, b); |
16725 | assert_eq_m128i(r, a); |
16726 | let r = _mm_mask_permutex2var_epi16(a, 0b11111111, idx, b); |
16727 | let e = _mm_set_epi16(6, 100, 5, 100, 4, 100, 3, 100); |
16728 | assert_eq_m128i(r, e); |
16729 | } |
16730 | |
16731 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16732 | unsafe fn test_mm_maskz_permutex2var_epi16() { |
16733 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
16734 | let idx = _mm_set_epi16(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
16735 | let b = _mm_set1_epi16(100); |
16736 | let r = _mm_maskz_permutex2var_epi16(0, a, idx, b); |
16737 | assert_eq_m128i(r, _mm_setzero_si128()); |
16738 | let r = _mm_maskz_permutex2var_epi16(0b11111111, a, idx, b); |
16739 | let e = _mm_set_epi16(6, 100, 5, 100, 4, 100, 3, 100); |
16740 | assert_eq_m128i(r, e); |
16741 | } |
16742 | |
16743 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16744 | unsafe fn test_mm_mask2_permutex2var_epi16() { |
16745 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
16746 | let idx = _mm_set_epi16(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
16747 | let b = _mm_set1_epi16(100); |
16748 | let r = _mm_mask2_permutex2var_epi16(a, idx, 0, b); |
16749 | assert_eq_m128i(r, idx); |
16750 | let r = _mm_mask2_permutex2var_epi16(a, idx, 0b11111111, b); |
16751 | let e = _mm_set_epi16(6, 100, 5, 100, 4, 100, 3, 100); |
16752 | assert_eq_m128i(r, e); |
16753 | } |
16754 | |
16755 | #[simd_test(enable = "avx512bw" )] |
16756 | unsafe fn test_mm512_permutexvar_epi16() { |
16757 | let idx = _mm512_set1_epi16(1); |
16758 | #[rustfmt::skip] |
16759 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
16760 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
16761 | let r = _mm512_permutexvar_epi16(idx, a); |
16762 | let e = _mm512_set1_epi16(30); |
16763 | assert_eq_m512i(r, e); |
16764 | } |
16765 | |
16766 | #[simd_test(enable = "avx512bw" )] |
16767 | unsafe fn test_mm512_mask_permutexvar_epi16() { |
16768 | let idx = _mm512_set1_epi16(1); |
16769 | #[rustfmt::skip] |
16770 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
16771 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
16772 | let r = _mm512_mask_permutexvar_epi16(a, 0, idx, a); |
16773 | assert_eq_m512i(r, a); |
16774 | let r = _mm512_mask_permutexvar_epi16(a, 0b11111111_11111111_11111111_11111111, idx, a); |
16775 | let e = _mm512_set1_epi16(30); |
16776 | assert_eq_m512i(r, e); |
16777 | } |
16778 | |
16779 | #[simd_test(enable = "avx512bw" )] |
16780 | unsafe fn test_mm512_maskz_permutexvar_epi16() { |
16781 | let idx = _mm512_set1_epi16(1); |
16782 | #[rustfmt::skip] |
16783 | let a = _mm512_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
16784 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
16785 | let r = _mm512_maskz_permutexvar_epi16(0, idx, a); |
16786 | assert_eq_m512i(r, _mm512_setzero_si512()); |
16787 | let r = _mm512_maskz_permutexvar_epi16(0b11111111_11111111_11111111_11111111, idx, a); |
16788 | let e = _mm512_set1_epi16(30); |
16789 | assert_eq_m512i(r, e); |
16790 | } |
16791 | |
16792 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16793 | unsafe fn test_mm256_permutexvar_epi16() { |
16794 | let idx = _mm256_set1_epi16(1); |
16795 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
16796 | let r = _mm256_permutexvar_epi16(idx, a); |
16797 | let e = _mm256_set1_epi16(14); |
16798 | assert_eq_m256i(r, e); |
16799 | } |
16800 | |
16801 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16802 | unsafe fn test_mm256_mask_permutexvar_epi16() { |
16803 | let idx = _mm256_set1_epi16(1); |
16804 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
16805 | let r = _mm256_mask_permutexvar_epi16(a, 0, idx, a); |
16806 | assert_eq_m256i(r, a); |
16807 | let r = _mm256_mask_permutexvar_epi16(a, 0b11111111_11111111, idx, a); |
16808 | let e = _mm256_set1_epi16(14); |
16809 | assert_eq_m256i(r, e); |
16810 | } |
16811 | |
16812 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16813 | unsafe fn test_mm256_maskz_permutexvar_epi16() { |
16814 | let idx = _mm256_set1_epi16(1); |
16815 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
16816 | let r = _mm256_maskz_permutexvar_epi16(0, idx, a); |
16817 | assert_eq_m256i(r, _mm256_setzero_si256()); |
16818 | let r = _mm256_maskz_permutexvar_epi16(0b11111111_11111111, idx, a); |
16819 | let e = _mm256_set1_epi16(14); |
16820 | assert_eq_m256i(r, e); |
16821 | } |
16822 | |
16823 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16824 | unsafe fn test_mm_permutexvar_epi16() { |
16825 | let idx = _mm_set1_epi16(1); |
16826 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
16827 | let r = _mm_permutexvar_epi16(idx, a); |
16828 | let e = _mm_set1_epi16(6); |
16829 | assert_eq_m128i(r, e); |
16830 | } |
16831 | |
16832 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16833 | unsafe fn test_mm_mask_permutexvar_epi16() { |
16834 | let idx = _mm_set1_epi16(1); |
16835 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
16836 | let r = _mm_mask_permutexvar_epi16(a, 0, idx, a); |
16837 | assert_eq_m128i(r, a); |
16838 | let r = _mm_mask_permutexvar_epi16(a, 0b11111111, idx, a); |
16839 | let e = _mm_set1_epi16(6); |
16840 | assert_eq_m128i(r, e); |
16841 | } |
16842 | |
16843 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16844 | unsafe fn test_mm_maskz_permutexvar_epi16() { |
16845 | let idx = _mm_set1_epi16(1); |
16846 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
16847 | let r = _mm_maskz_permutexvar_epi16(0, idx, a); |
16848 | assert_eq_m128i(r, _mm_setzero_si128()); |
16849 | let r = _mm_maskz_permutexvar_epi16(0b11111111, idx, a); |
16850 | let e = _mm_set1_epi16(6); |
16851 | assert_eq_m128i(r, e); |
16852 | } |
16853 | |
16854 | #[simd_test(enable = "avx512bw" )] |
16855 | unsafe fn test_mm512_mask_blend_epi16() { |
16856 | let a = _mm512_set1_epi16(1); |
16857 | let b = _mm512_set1_epi16(2); |
16858 | let r = _mm512_mask_blend_epi16(0b11111111_00000000_11111111_00000000, a, b); |
16859 | #[rustfmt::skip] |
16860 | let e = _mm512_set_epi16(2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, |
16861 | 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1); |
16862 | assert_eq_m512i(r, e); |
16863 | } |
16864 | |
16865 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16866 | unsafe fn test_mm256_mask_blend_epi16() { |
16867 | let a = _mm256_set1_epi16(1); |
16868 | let b = _mm256_set1_epi16(2); |
16869 | let r = _mm256_mask_blend_epi16(0b11111111_00000000, a, b); |
16870 | let e = _mm256_set_epi16(2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1); |
16871 | assert_eq_m256i(r, e); |
16872 | } |
16873 | |
16874 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16875 | unsafe fn test_mm_mask_blend_epi16() { |
16876 | let a = _mm_set1_epi16(1); |
16877 | let b = _mm_set1_epi16(2); |
16878 | let r = _mm_mask_blend_epi16(0b11110000, a, b); |
16879 | let e = _mm_set_epi16(2, 2, 2, 2, 1, 1, 1, 1); |
16880 | assert_eq_m128i(r, e); |
16881 | } |
16882 | |
16883 | #[simd_test(enable = "avx512bw" )] |
16884 | unsafe fn test_mm512_mask_blend_epi8() { |
16885 | let a = _mm512_set1_epi8(1); |
16886 | let b = _mm512_set1_epi8(2); |
16887 | let r = _mm512_mask_blend_epi8( |
16888 | 0b11111111_00000000_11111111_00000000_11111111_00000000_11111111_00000000, |
16889 | a, |
16890 | b, |
16891 | ); |
16892 | #[rustfmt::skip] |
16893 | let e = _mm512_set_epi8(2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, |
16894 | 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, |
16895 | 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, |
16896 | 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1); |
16897 | assert_eq_m512i(r, e); |
16898 | } |
16899 | |
16900 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16901 | unsafe fn test_mm256_mask_blend_epi8() { |
16902 | let a = _mm256_set1_epi8(1); |
16903 | let b = _mm256_set1_epi8(2); |
16904 | let r = _mm256_mask_blend_epi8(0b11111111_00000000_11111111_00000000, a, b); |
16905 | #[rustfmt::skip] |
16906 | let e = _mm256_set_epi8(2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, |
16907 | 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1); |
16908 | assert_eq_m256i(r, e); |
16909 | } |
16910 | |
16911 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16912 | unsafe fn test_mm_mask_blend_epi8() { |
16913 | let a = _mm_set1_epi8(1); |
16914 | let b = _mm_set1_epi8(2); |
16915 | let r = _mm_mask_blend_epi8(0b11111111_00000000, a, b); |
16916 | let e = _mm_set_epi8(2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1); |
16917 | assert_eq_m128i(r, e); |
16918 | } |
16919 | |
16920 | #[simd_test(enable = "avx512bw" )] |
16921 | unsafe fn test_mm512_broadcastw_epi16() { |
16922 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
16923 | let r = _mm512_broadcastw_epi16(a); |
16924 | let e = _mm512_set1_epi16(24); |
16925 | assert_eq_m512i(r, e); |
16926 | } |
16927 | |
16928 | #[simd_test(enable = "avx512bw" )] |
16929 | unsafe fn test_mm512_mask_broadcastw_epi16() { |
16930 | let src = _mm512_set1_epi16(1); |
16931 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
16932 | let r = _mm512_mask_broadcastw_epi16(src, 0, a); |
16933 | assert_eq_m512i(r, src); |
16934 | let r = _mm512_mask_broadcastw_epi16(src, 0b11111111_11111111_11111111_11111111, a); |
16935 | let e = _mm512_set1_epi16(24); |
16936 | assert_eq_m512i(r, e); |
16937 | } |
16938 | |
16939 | #[simd_test(enable = "avx512bw" )] |
16940 | unsafe fn test_mm512_maskz_broadcastw_epi16() { |
16941 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
16942 | let r = _mm512_maskz_broadcastw_epi16(0, a); |
16943 | assert_eq_m512i(r, _mm512_setzero_si512()); |
16944 | let r = _mm512_maskz_broadcastw_epi16(0b11111111_11111111_11111111_11111111, a); |
16945 | let e = _mm512_set1_epi16(24); |
16946 | assert_eq_m512i(r, e); |
16947 | } |
16948 | |
16949 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16950 | unsafe fn test_mm256_mask_broadcastw_epi16() { |
16951 | let src = _mm256_set1_epi16(1); |
16952 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
16953 | let r = _mm256_mask_broadcastw_epi16(src, 0, a); |
16954 | assert_eq_m256i(r, src); |
16955 | let r = _mm256_mask_broadcastw_epi16(src, 0b11111111_11111111, a); |
16956 | let e = _mm256_set1_epi16(24); |
16957 | assert_eq_m256i(r, e); |
16958 | } |
16959 | |
16960 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16961 | unsafe fn test_mm256_maskz_broadcastw_epi16() { |
16962 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
16963 | let r = _mm256_maskz_broadcastw_epi16(0, a); |
16964 | assert_eq_m256i(r, _mm256_setzero_si256()); |
16965 | let r = _mm256_maskz_broadcastw_epi16(0b11111111_11111111, a); |
16966 | let e = _mm256_set1_epi16(24); |
16967 | assert_eq_m256i(r, e); |
16968 | } |
16969 | |
16970 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16971 | unsafe fn test_mm_mask_broadcastw_epi16() { |
16972 | let src = _mm_set1_epi16(1); |
16973 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
16974 | let r = _mm_mask_broadcastw_epi16(src, 0, a); |
16975 | assert_eq_m128i(r, src); |
16976 | let r = _mm_mask_broadcastw_epi16(src, 0b11111111, a); |
16977 | let e = _mm_set1_epi16(24); |
16978 | assert_eq_m128i(r, e); |
16979 | } |
16980 | |
16981 | #[simd_test(enable = "avx512bw,avx512vl" )] |
16982 | unsafe fn test_mm_maskz_broadcastw_epi16() { |
16983 | let a = _mm_set_epi16(17, 18, 19, 20, 21, 22, 23, 24); |
16984 | let r = _mm_maskz_broadcastw_epi16(0, a); |
16985 | assert_eq_m128i(r, _mm_setzero_si128()); |
16986 | let r = _mm_maskz_broadcastw_epi16(0b11111111, a); |
16987 | let e = _mm_set1_epi16(24); |
16988 | assert_eq_m128i(r, e); |
16989 | } |
16990 | |
16991 | #[simd_test(enable = "avx512bw" )] |
16992 | unsafe fn test_mm512_broadcastb_epi8() { |
16993 | let a = _mm_set_epi8( |
16994 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
16995 | ); |
16996 | let r = _mm512_broadcastb_epi8(a); |
16997 | let e = _mm512_set1_epi8(32); |
16998 | assert_eq_m512i(r, e); |
16999 | } |
17000 | |
17001 | #[simd_test(enable = "avx512bw" )] |
17002 | unsafe fn test_mm512_mask_broadcastb_epi8() { |
17003 | let src = _mm512_set1_epi8(1); |
17004 | let a = _mm_set_epi8( |
17005 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
17006 | ); |
17007 | let r = _mm512_mask_broadcastb_epi8(src, 0, a); |
17008 | assert_eq_m512i(r, src); |
17009 | let r = _mm512_mask_broadcastb_epi8( |
17010 | src, |
17011 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
17012 | a, |
17013 | ); |
17014 | let e = _mm512_set1_epi8(32); |
17015 | assert_eq_m512i(r, e); |
17016 | } |
17017 | |
17018 | #[simd_test(enable = "avx512bw" )] |
17019 | unsafe fn test_mm512_maskz_broadcastb_epi8() { |
17020 | let a = _mm_set_epi8( |
17021 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
17022 | ); |
17023 | let r = _mm512_maskz_broadcastb_epi8(0, a); |
17024 | assert_eq_m512i(r, _mm512_setzero_si512()); |
17025 | let r = _mm512_maskz_broadcastb_epi8( |
17026 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
17027 | a, |
17028 | ); |
17029 | let e = _mm512_set1_epi8(32); |
17030 | assert_eq_m512i(r, e); |
17031 | } |
17032 | |
17033 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17034 | unsafe fn test_mm256_mask_broadcastb_epi8() { |
17035 | let src = _mm256_set1_epi8(1); |
17036 | let a = _mm_set_epi8( |
17037 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
17038 | ); |
17039 | let r = _mm256_mask_broadcastb_epi8(src, 0, a); |
17040 | assert_eq_m256i(r, src); |
17041 | let r = _mm256_mask_broadcastb_epi8(src, 0b11111111_11111111_11111111_11111111, a); |
17042 | let e = _mm256_set1_epi8(32); |
17043 | assert_eq_m256i(r, e); |
17044 | } |
17045 | |
17046 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17047 | unsafe fn test_mm256_maskz_broadcastb_epi8() { |
17048 | let a = _mm_set_epi8( |
17049 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
17050 | ); |
17051 | let r = _mm256_maskz_broadcastb_epi8(0, a); |
17052 | assert_eq_m256i(r, _mm256_setzero_si256()); |
17053 | let r = _mm256_maskz_broadcastb_epi8(0b11111111_11111111_11111111_11111111, a); |
17054 | let e = _mm256_set1_epi8(32); |
17055 | assert_eq_m256i(r, e); |
17056 | } |
17057 | |
17058 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17059 | unsafe fn test_mm_mask_broadcastb_epi8() { |
17060 | let src = _mm_set1_epi8(1); |
17061 | let a = _mm_set_epi8( |
17062 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
17063 | ); |
17064 | let r = _mm_mask_broadcastb_epi8(src, 0, a); |
17065 | assert_eq_m128i(r, src); |
17066 | let r = _mm_mask_broadcastb_epi8(src, 0b11111111_11111111, a); |
17067 | let e = _mm_set1_epi8(32); |
17068 | assert_eq_m128i(r, e); |
17069 | } |
17070 | |
17071 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17072 | unsafe fn test_mm_maskz_broadcastb_epi8() { |
17073 | let a = _mm_set_epi8( |
17074 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
17075 | ); |
17076 | let r = _mm_maskz_broadcastb_epi8(0, a); |
17077 | assert_eq_m128i(r, _mm_setzero_si128()); |
17078 | let r = _mm_maskz_broadcastb_epi8(0b11111111_11111111, a); |
17079 | let e = _mm_set1_epi8(32); |
17080 | assert_eq_m128i(r, e); |
17081 | } |
17082 | |
17083 | #[simd_test(enable = "avx512bw" )] |
17084 | unsafe fn test_mm512_unpackhi_epi16() { |
17085 | #[rustfmt::skip] |
17086 | let a = _mm512_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17087 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
17088 | #[rustfmt::skip] |
17089 | let b = _mm512_set_epi16(33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17090 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
17091 | let r = _mm512_unpackhi_epi16(a, b); |
17092 | #[rustfmt::skip] |
17093 | let e = _mm512_set_epi16(33, 1, 34, 2, 35, 3, 36, 4, 41, 9, 42, 10, 43, 11, 44, 12, |
17094 | 49, 17, 50, 18, 51, 19, 52, 20, 57, 25, 58, 26, 59, 27, 60, 28); |
17095 | assert_eq_m512i(r, e); |
17096 | } |
17097 | |
17098 | #[simd_test(enable = "avx512bw" )] |
17099 | unsafe fn test_mm512_mask_unpackhi_epi16() { |
17100 | #[rustfmt::skip] |
17101 | let a = _mm512_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17102 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
17103 | #[rustfmt::skip] |
17104 | let b = _mm512_set_epi16(33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17105 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
17106 | let r = _mm512_mask_unpackhi_epi16(a, 0, a, b); |
17107 | assert_eq_m512i(r, a); |
17108 | let r = _mm512_mask_unpackhi_epi16(a, 0b11111111_11111111_11111111_11111111, a, b); |
17109 | #[rustfmt::skip] |
17110 | let e = _mm512_set_epi16(33, 1, 34, 2, 35, 3, 36, 4, 41, 9, 42, 10, 43, 11, 44, 12, |
17111 | 49, 17, 50, 18, 51, 19, 52, 20, 57, 25, 58, 26, 59, 27, 60, 28); |
17112 | assert_eq_m512i(r, e); |
17113 | } |
17114 | |
17115 | #[simd_test(enable = "avx512bw" )] |
17116 | unsafe fn test_mm512_maskz_unpackhi_epi16() { |
17117 | #[rustfmt::skip] |
17118 | let a = _mm512_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17119 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
17120 | #[rustfmt::skip] |
17121 | let b = _mm512_set_epi16(33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17122 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
17123 | let r = _mm512_maskz_unpackhi_epi16(0, a, b); |
17124 | assert_eq_m512i(r, _mm512_setzero_si512()); |
17125 | let r = _mm512_maskz_unpackhi_epi16(0b11111111_11111111_11111111_11111111, a, b); |
17126 | #[rustfmt::skip] |
17127 | let e = _mm512_set_epi16(33, 1, 34, 2, 35, 3, 36, 4, 41, 9, 42, 10, 43, 11, 44, 12, |
17128 | 49, 17, 50, 18, 51, 19, 52, 20, 57, 25, 58, 26, 59, 27, 60, 28); |
17129 | assert_eq_m512i(r, e); |
17130 | } |
17131 | |
17132 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17133 | unsafe fn test_mm256_mask_unpackhi_epi16() { |
17134 | let a = _mm256_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
17135 | let b = _mm256_set_epi16( |
17136 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17137 | ); |
17138 | let r = _mm256_mask_unpackhi_epi16(a, 0, a, b); |
17139 | assert_eq_m256i(r, a); |
17140 | let r = _mm256_mask_unpackhi_epi16(a, 0b11111111_11111111, a, b); |
17141 | let e = _mm256_set_epi16(33, 1, 34, 2, 35, 3, 36, 4, 41, 9, 42, 10, 43, 11, 44, 12); |
17142 | assert_eq_m256i(r, e); |
17143 | } |
17144 | |
17145 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17146 | unsafe fn test_mm256_maskz_unpackhi_epi16() { |
17147 | let a = _mm256_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
17148 | let b = _mm256_set_epi16( |
17149 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17150 | ); |
17151 | let r = _mm256_maskz_unpackhi_epi16(0, a, b); |
17152 | assert_eq_m256i(r, _mm256_setzero_si256()); |
17153 | let r = _mm256_maskz_unpackhi_epi16(0b11111111_11111111, a, b); |
17154 | let e = _mm256_set_epi16(33, 1, 34, 2, 35, 3, 36, 4, 41, 9, 42, 10, 43, 11, 44, 12); |
17155 | assert_eq_m256i(r, e); |
17156 | } |
17157 | |
17158 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17159 | unsafe fn test_mm_mask_unpackhi_epi16() { |
17160 | let a = _mm_set_epi16(1, 2, 3, 4, 5, 6, 7, 8); |
17161 | let b = _mm_set_epi16(33, 34, 35, 36, 37, 38, 39, 40); |
17162 | let r = _mm_mask_unpackhi_epi16(a, 0, a, b); |
17163 | assert_eq_m128i(r, a); |
17164 | let r = _mm_mask_unpackhi_epi16(a, 0b11111111, a, b); |
17165 | let e = _mm_set_epi16(33, 1, 34, 2, 35, 3, 36, 4); |
17166 | assert_eq_m128i(r, e); |
17167 | } |
17168 | |
17169 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17170 | unsafe fn test_mm_maskz_unpackhi_epi16() { |
17171 | let a = _mm_set_epi16(1, 2, 3, 4, 5, 6, 7, 8); |
17172 | let b = _mm_set_epi16(33, 34, 35, 36, 37, 38, 39, 40); |
17173 | let r = _mm_maskz_unpackhi_epi16(0, a, b); |
17174 | assert_eq_m128i(r, _mm_setzero_si128()); |
17175 | let r = _mm_maskz_unpackhi_epi16(0b11111111, a, b); |
17176 | let e = _mm_set_epi16(33, 1, 34, 2, 35, 3, 36, 4); |
17177 | assert_eq_m128i(r, e); |
17178 | } |
17179 | |
17180 | #[simd_test(enable = "avx512bw" )] |
17181 | unsafe fn test_mm512_unpackhi_epi8() { |
17182 | #[rustfmt::skip] |
17183 | let a = _mm512_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17184 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
17185 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17186 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
17187 | #[rustfmt::skip] |
17188 | let b = _mm512_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17189 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, |
17190 | 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, |
17191 | 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 0); |
17192 | let r = _mm512_unpackhi_epi8(a, b); |
17193 | #[rustfmt::skip] |
17194 | let e = _mm512_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8, |
17195 | 81, 17, 82, 18, 83, 19, 84, 20, 85, 21, 86, 22, 87, 23, 88, 24, |
17196 | 97, 33, 98, 34, 99, 35, 100, 36, 101, 37, 102, 38, 103, 39, 104, 40, |
17197 | 113, 49, 114, 50, 115, 51, 116, 52, 117, 53, 118, 54, 119, 55, 120, 56); |
17198 | assert_eq_m512i(r, e); |
17199 | } |
17200 | |
17201 | #[simd_test(enable = "avx512bw" )] |
17202 | unsafe fn test_mm512_mask_unpackhi_epi8() { |
17203 | #[rustfmt::skip] |
17204 | let a = _mm512_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17205 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
17206 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17207 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
17208 | #[rustfmt::skip] |
17209 | let b = _mm512_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17210 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, |
17211 | 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, |
17212 | 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 0); |
17213 | let r = _mm512_mask_unpackhi_epi8(a, 0, a, b); |
17214 | assert_eq_m512i(r, a); |
17215 | let r = _mm512_mask_unpackhi_epi8( |
17216 | a, |
17217 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
17218 | a, |
17219 | b, |
17220 | ); |
17221 | #[rustfmt::skip] |
17222 | let e = _mm512_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8, |
17223 | 81, 17, 82, 18, 83, 19, 84, 20, 85, 21, 86, 22, 87, 23, 88, 24, |
17224 | 97, 33, 98, 34, 99, 35, 100, 36, 101, 37, 102, 38, 103, 39, 104, 40, |
17225 | 113, 49, 114, 50, 115, 51, 116, 52, 117, 53, 118, 54, 119, 55, 120, 56); |
17226 | assert_eq_m512i(r, e); |
17227 | } |
17228 | |
17229 | #[simd_test(enable = "avx512bw" )] |
17230 | unsafe fn test_mm512_maskz_unpackhi_epi8() { |
17231 | #[rustfmt::skip] |
17232 | let a = _mm512_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17233 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
17234 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17235 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
17236 | #[rustfmt::skip] |
17237 | let b = _mm512_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17238 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, |
17239 | 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, |
17240 | 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 0); |
17241 | let r = _mm512_maskz_unpackhi_epi8(0, a, b); |
17242 | assert_eq_m512i(r, _mm512_setzero_si512()); |
17243 | let r = _mm512_maskz_unpackhi_epi8( |
17244 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
17245 | a, |
17246 | b, |
17247 | ); |
17248 | #[rustfmt::skip] |
17249 | let e = _mm512_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8, |
17250 | 81, 17, 82, 18, 83, 19, 84, 20, 85, 21, 86, 22, 87, 23, 88, 24, |
17251 | 97, 33, 98, 34, 99, 35, 100, 36, 101, 37, 102, 38, 103, 39, 104, 40, |
17252 | 113, 49, 114, 50, 115, 51, 116, 52, 117, 53, 118, 54, 119, 55, 120, 56); |
17253 | assert_eq_m512i(r, e); |
17254 | } |
17255 | |
17256 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17257 | unsafe fn test_mm256_mask_unpackhi_epi8() { |
17258 | #[rustfmt::skip] |
17259 | let a = _mm256_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17260 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
17261 | #[rustfmt::skip] |
17262 | let b = _mm256_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17263 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96); |
17264 | let r = _mm256_mask_unpackhi_epi8(a, 0, a, b); |
17265 | assert_eq_m256i(r, a); |
17266 | let r = _mm256_mask_unpackhi_epi8(a, 0b11111111_11111111_11111111_11111111, a, b); |
17267 | #[rustfmt::skip] |
17268 | let e = _mm256_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8, |
17269 | 81, 17, 82, 18, 83, 19, 84, 20, 85, 21, 86, 22, 87, 23, 88, 24); |
17270 | assert_eq_m256i(r, e); |
17271 | } |
17272 | |
17273 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17274 | unsafe fn test_mm256_maskz_unpackhi_epi8() { |
17275 | #[rustfmt::skip] |
17276 | let a = _mm256_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17277 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
17278 | #[rustfmt::skip] |
17279 | let b = _mm256_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17280 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96); |
17281 | let r = _mm256_maskz_unpackhi_epi8(0, a, b); |
17282 | assert_eq_m256i(r, _mm256_setzero_si256()); |
17283 | let r = _mm256_maskz_unpackhi_epi8(0b11111111_11111111_11111111_11111111, a, b); |
17284 | #[rustfmt::skip] |
17285 | let e = _mm256_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8, |
17286 | 81, 17, 82, 18, 83, 19, 84, 20, 85, 21, 86, 22, 87, 23, 88, 24); |
17287 | assert_eq_m256i(r, e); |
17288 | } |
17289 | |
17290 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17291 | unsafe fn test_mm_mask_unpackhi_epi8() { |
17292 | let a = _mm_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
17293 | let b = _mm_set_epi8( |
17294 | 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17295 | ); |
17296 | let r = _mm_mask_unpackhi_epi8(a, 0, a, b); |
17297 | assert_eq_m128i(r, a); |
17298 | let r = _mm_mask_unpackhi_epi8(a, 0b11111111_11111111, a, b); |
17299 | let e = _mm_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8); |
17300 | assert_eq_m128i(r, e); |
17301 | } |
17302 | |
17303 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17304 | unsafe fn test_mm_maskz_unpackhi_epi8() { |
17305 | let a = _mm_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
17306 | let b = _mm_set_epi8( |
17307 | 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17308 | ); |
17309 | let r = _mm_maskz_unpackhi_epi8(0, a, b); |
17310 | assert_eq_m128i(r, _mm_setzero_si128()); |
17311 | let r = _mm_maskz_unpackhi_epi8(0b11111111_11111111, a, b); |
17312 | let e = _mm_set_epi8(65, 1, 66, 2, 67, 3, 68, 4, 69, 5, 70, 6, 71, 7, 72, 8); |
17313 | assert_eq_m128i(r, e); |
17314 | } |
17315 | |
17316 | #[simd_test(enable = "avx512bw" )] |
17317 | unsafe fn test_mm512_unpacklo_epi16() { |
17318 | #[rustfmt::skip] |
17319 | let a = _mm512_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17320 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
17321 | #[rustfmt::skip] |
17322 | let b = _mm512_set_epi16(33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17323 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
17324 | let r = _mm512_unpacklo_epi16(a, b); |
17325 | #[rustfmt::skip] |
17326 | let e = _mm512_set_epi16(37, 5, 38, 6, 39, 7, 40, 8, 45, 13, 46, 14, 47, 15, 48, 16, |
17327 | 53, 21, 54, 22, 55, 23, 56, 24, 61, 29, 62, 30, 63, 31, 64, 32); |
17328 | assert_eq_m512i(r, e); |
17329 | } |
17330 | |
17331 | #[simd_test(enable = "avx512bw" )] |
17332 | unsafe fn test_mm512_mask_unpacklo_epi16() { |
17333 | #[rustfmt::skip] |
17334 | let a = _mm512_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17335 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
17336 | #[rustfmt::skip] |
17337 | let b = _mm512_set_epi16(33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17338 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
17339 | let r = _mm512_mask_unpacklo_epi16(a, 0, a, b); |
17340 | assert_eq_m512i(r, a); |
17341 | let r = _mm512_mask_unpacklo_epi16(a, 0b11111111_11111111_11111111_11111111, a, b); |
17342 | #[rustfmt::skip] |
17343 | let e = _mm512_set_epi16(37, 5, 38, 6, 39, 7, 40, 8, 45, 13, 46, 14, 47, 15, 48, 16, |
17344 | 53, 21, 54, 22, 55, 23, 56, 24, 61, 29, 62, 30, 63, 31, 64, 32); |
17345 | assert_eq_m512i(r, e); |
17346 | } |
17347 | |
17348 | #[simd_test(enable = "avx512bw" )] |
17349 | unsafe fn test_mm512_maskz_unpacklo_epi16() { |
17350 | #[rustfmt::skip] |
17351 | let a = _mm512_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17352 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
17353 | #[rustfmt::skip] |
17354 | let b = _mm512_set_epi16(33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17355 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
17356 | let r = _mm512_maskz_unpacklo_epi16(0, a, b); |
17357 | assert_eq_m512i(r, _mm512_setzero_si512()); |
17358 | let r = _mm512_maskz_unpacklo_epi16(0b11111111_11111111_11111111_11111111, a, b); |
17359 | #[rustfmt::skip] |
17360 | let e = _mm512_set_epi16(37, 5, 38, 6, 39, 7, 40, 8, 45, 13, 46, 14, 47, 15, 48, 16, |
17361 | 53, 21, 54, 22, 55, 23, 56, 24, 61, 29, 62, 30, 63, 31, 64, 32); |
17362 | assert_eq_m512i(r, e); |
17363 | } |
17364 | |
17365 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17366 | unsafe fn test_mm256_mask_unpacklo_epi16() { |
17367 | let a = _mm256_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
17368 | let b = _mm256_set_epi16( |
17369 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17370 | ); |
17371 | let r = _mm256_mask_unpacklo_epi16(a, 0, a, b); |
17372 | assert_eq_m256i(r, a); |
17373 | let r = _mm256_mask_unpacklo_epi16(a, 0b11111111_11111111, a, b); |
17374 | let e = _mm256_set_epi16(37, 5, 38, 6, 39, 7, 40, 8, 45, 13, 46, 14, 47, 15, 48, 16); |
17375 | assert_eq_m256i(r, e); |
17376 | } |
17377 | |
17378 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17379 | unsafe fn test_mm256_maskz_unpacklo_epi16() { |
17380 | let a = _mm256_set_epi16(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
17381 | let b = _mm256_set_epi16( |
17382 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17383 | ); |
17384 | let r = _mm256_maskz_unpacklo_epi16(0, a, b); |
17385 | assert_eq_m256i(r, _mm256_setzero_si256()); |
17386 | let r = _mm256_maskz_unpacklo_epi16(0b11111111_11111111, a, b); |
17387 | let e = _mm256_set_epi16(37, 5, 38, 6, 39, 7, 40, 8, 45, 13, 46, 14, 47, 15, 48, 16); |
17388 | assert_eq_m256i(r, e); |
17389 | } |
17390 | |
17391 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17392 | unsafe fn test_mm_mask_unpacklo_epi16() { |
17393 | let a = _mm_set_epi16(1, 2, 3, 4, 5, 6, 7, 8); |
17394 | let b = _mm_set_epi16(33, 34, 35, 36, 37, 38, 39, 40); |
17395 | let r = _mm_mask_unpacklo_epi16(a, 0, a, b); |
17396 | assert_eq_m128i(r, a); |
17397 | let r = _mm_mask_unpacklo_epi16(a, 0b11111111, a, b); |
17398 | let e = _mm_set_epi16(37, 5, 38, 6, 39, 7, 40, 8); |
17399 | assert_eq_m128i(r, e); |
17400 | } |
17401 | |
17402 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17403 | unsafe fn test_mm_maskz_unpacklo_epi16() { |
17404 | let a = _mm_set_epi16(1, 2, 3, 4, 5, 6, 7, 8); |
17405 | let b = _mm_set_epi16(33, 34, 35, 36, 37, 38, 39, 40); |
17406 | let r = _mm_maskz_unpacklo_epi16(0, a, b); |
17407 | assert_eq_m128i(r, _mm_setzero_si128()); |
17408 | let r = _mm_maskz_unpacklo_epi16(0b11111111, a, b); |
17409 | let e = _mm_set_epi16(37, 5, 38, 6, 39, 7, 40, 8); |
17410 | assert_eq_m128i(r, e); |
17411 | } |
17412 | |
17413 | #[simd_test(enable = "avx512bw" )] |
17414 | unsafe fn test_mm512_unpacklo_epi8() { |
17415 | #[rustfmt::skip] |
17416 | let a = _mm512_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17417 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
17418 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17419 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
17420 | #[rustfmt::skip] |
17421 | let b = _mm512_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17422 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, |
17423 | 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, |
17424 | 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 0); |
17425 | let r = _mm512_unpacklo_epi8(a, b); |
17426 | #[rustfmt::skip] |
17427 | let e = _mm512_set_epi8(73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
17428 | 89, 25, 90, 26, 91, 27, 92, 28, 93, 29, 94, 30, 95, 31, 96, 32, |
17429 | 105, 41, 106, 42, 107, 43, 108, 44, 109, 45, 110, 46, 111, 47, 112, 48, |
17430 | 121, 57, 122, 58, 123, 59, 124, 60, 125, 61, 126, 62, 127, 63, 0, 64); |
17431 | assert_eq_m512i(r, e); |
17432 | } |
17433 | |
17434 | #[simd_test(enable = "avx512bw" )] |
17435 | unsafe fn test_mm512_mask_unpacklo_epi8() { |
17436 | #[rustfmt::skip] |
17437 | let a = _mm512_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17438 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
17439 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17440 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
17441 | #[rustfmt::skip] |
17442 | let b = _mm512_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17443 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, |
17444 | 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, |
17445 | 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 0); |
17446 | let r = _mm512_mask_unpacklo_epi8(a, 0, a, b); |
17447 | assert_eq_m512i(r, a); |
17448 | let r = _mm512_mask_unpacklo_epi8( |
17449 | a, |
17450 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
17451 | a, |
17452 | b, |
17453 | ); |
17454 | #[rustfmt::skip] |
17455 | let e = _mm512_set_epi8(73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
17456 | 89, 25, 90, 26, 91, 27, 92, 28, 93, 29, 94, 30, 95, 31, 96, 32, |
17457 | 105, 41, 106, 42, 107, 43, 108, 44, 109, 45, 110, 46, 111, 47, 112, 48, |
17458 | 121, 57, 122, 58, 123, 59, 124, 60, 125, 61, 126, 62, 127, 63, 0, 64); |
17459 | assert_eq_m512i(r, e); |
17460 | } |
17461 | |
17462 | #[simd_test(enable = "avx512bw" )] |
17463 | unsafe fn test_mm512_maskz_unpacklo_epi8() { |
17464 | #[rustfmt::skip] |
17465 | let a = _mm512_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17466 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
17467 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
17468 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64); |
17469 | #[rustfmt::skip] |
17470 | let b = _mm512_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17471 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, |
17472 | 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, |
17473 | 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 0); |
17474 | let r = _mm512_maskz_unpacklo_epi8(0, a, b); |
17475 | assert_eq_m512i(r, _mm512_setzero_si512()); |
17476 | let r = _mm512_maskz_unpacklo_epi8( |
17477 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
17478 | a, |
17479 | b, |
17480 | ); |
17481 | #[rustfmt::skip] |
17482 | let e = _mm512_set_epi8(73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
17483 | 89, 25, 90, 26, 91, 27, 92, 28, 93, 29, 94, 30, 95, 31, 96, 32, |
17484 | 105, 41, 106, 42, 107, 43, 108, 44, 109, 45, 110, 46, 111, 47, 112, 48, |
17485 | 121, 57, 122, 58, 123, 59, 124, 60, 125, 61, 126, 62, 127, 63, 0, 64); |
17486 | assert_eq_m512i(r, e); |
17487 | } |
17488 | |
17489 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17490 | unsafe fn test_mm256_mask_unpacklo_epi8() { |
17491 | #[rustfmt::skip] |
17492 | let a = _mm256_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17493 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
17494 | #[rustfmt::skip] |
17495 | let b = _mm256_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17496 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96); |
17497 | let r = _mm256_mask_unpacklo_epi8(a, 0, a, b); |
17498 | assert_eq_m256i(r, a); |
17499 | let r = _mm256_mask_unpacklo_epi8(a, 0b11111111_11111111_11111111_11111111, a, b); |
17500 | #[rustfmt::skip] |
17501 | let e = _mm256_set_epi8(73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
17502 | 89, 25, 90, 26, 91, 27, 92, 28, 93, 29, 94, 30, 95, 31, 96, 32); |
17503 | assert_eq_m256i(r, e); |
17504 | } |
17505 | |
17506 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17507 | unsafe fn test_mm256_maskz_unpacklo_epi8() { |
17508 | #[rustfmt::skip] |
17509 | let a = _mm256_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
17510 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32); |
17511 | #[rustfmt::skip] |
17512 | let b = _mm256_set_epi8(65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17513 | 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96); |
17514 | let r = _mm256_maskz_unpacklo_epi8(0, a, b); |
17515 | assert_eq_m256i(r, _mm256_setzero_si256()); |
17516 | let r = _mm256_maskz_unpacklo_epi8(0b11111111_11111111_11111111_11111111, a, b); |
17517 | #[rustfmt::skip] |
17518 | let e = _mm256_set_epi8(73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
17519 | 89, 25, 90, 26, 91, 27, 92, 28, 93, 29, 94, 30, 95, 31, 96, 32); |
17520 | assert_eq_m256i(r, e); |
17521 | } |
17522 | |
17523 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17524 | unsafe fn test_mm_mask_unpacklo_epi8() { |
17525 | let a = _mm_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
17526 | let b = _mm_set_epi8( |
17527 | 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17528 | ); |
17529 | let r = _mm_mask_unpacklo_epi8(a, 0, a, b); |
17530 | assert_eq_m128i(r, a); |
17531 | let r = _mm_mask_unpacklo_epi8(a, 0b11111111_11111111, a, b); |
17532 | let e = _mm_set_epi8( |
17533 | 73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
17534 | ); |
17535 | assert_eq_m128i(r, e); |
17536 | } |
17537 | |
17538 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17539 | unsafe fn test_mm_maskz_unpacklo_epi8() { |
17540 | let a = _mm_set_epi8(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
17541 | let b = _mm_set_epi8( |
17542 | 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, |
17543 | ); |
17544 | let r = _mm_maskz_unpacklo_epi8(0, a, b); |
17545 | assert_eq_m128i(r, _mm_setzero_si128()); |
17546 | let r = _mm_maskz_unpacklo_epi8(0b11111111_11111111, a, b); |
17547 | let e = _mm_set_epi8( |
17548 | 73, 9, 74, 10, 75, 11, 76, 12, 77, 13, 78, 14, 79, 15, 80, 16, |
17549 | ); |
17550 | assert_eq_m128i(r, e); |
17551 | } |
17552 | |
17553 | #[simd_test(enable = "avx512bw" )] |
17554 | unsafe fn test_mm512_mask_mov_epi16() { |
17555 | let src = _mm512_set1_epi16(1); |
17556 | let a = _mm512_set1_epi16(2); |
17557 | let r = _mm512_mask_mov_epi16(src, 0, a); |
17558 | assert_eq_m512i(r, src); |
17559 | let r = _mm512_mask_mov_epi16(src, 0b11111111_11111111_11111111_11111111, a); |
17560 | assert_eq_m512i(r, a); |
17561 | } |
17562 | |
17563 | #[simd_test(enable = "avx512bw" )] |
17564 | unsafe fn test_mm512_maskz_mov_epi16() { |
17565 | let a = _mm512_set1_epi16(2); |
17566 | let r = _mm512_maskz_mov_epi16(0, a); |
17567 | assert_eq_m512i(r, _mm512_setzero_si512()); |
17568 | let r = _mm512_maskz_mov_epi16(0b11111111_11111111_11111111_11111111, a); |
17569 | assert_eq_m512i(r, a); |
17570 | } |
17571 | |
17572 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17573 | unsafe fn test_mm256_mask_mov_epi16() { |
17574 | let src = _mm256_set1_epi16(1); |
17575 | let a = _mm256_set1_epi16(2); |
17576 | let r = _mm256_mask_mov_epi16(src, 0, a); |
17577 | assert_eq_m256i(r, src); |
17578 | let r = _mm256_mask_mov_epi16(src, 0b11111111_11111111, a); |
17579 | assert_eq_m256i(r, a); |
17580 | } |
17581 | |
17582 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17583 | unsafe fn test_mm256_maskz_mov_epi16() { |
17584 | let a = _mm256_set1_epi16(2); |
17585 | let r = _mm256_maskz_mov_epi16(0, a); |
17586 | assert_eq_m256i(r, _mm256_setzero_si256()); |
17587 | let r = _mm256_maskz_mov_epi16(0b11111111_11111111, a); |
17588 | assert_eq_m256i(r, a); |
17589 | } |
17590 | |
17591 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17592 | unsafe fn test_mm_mask_mov_epi16() { |
17593 | let src = _mm_set1_epi16(1); |
17594 | let a = _mm_set1_epi16(2); |
17595 | let r = _mm_mask_mov_epi16(src, 0, a); |
17596 | assert_eq_m128i(r, src); |
17597 | let r = _mm_mask_mov_epi16(src, 0b11111111, a); |
17598 | assert_eq_m128i(r, a); |
17599 | } |
17600 | |
17601 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17602 | unsafe fn test_mm_maskz_mov_epi16() { |
17603 | let a = _mm_set1_epi16(2); |
17604 | let r = _mm_maskz_mov_epi16(0, a); |
17605 | assert_eq_m128i(r, _mm_setzero_si128()); |
17606 | let r = _mm_maskz_mov_epi16(0b11111111, a); |
17607 | assert_eq_m128i(r, a); |
17608 | } |
17609 | |
17610 | #[simd_test(enable = "avx512bw" )] |
17611 | unsafe fn test_mm512_mask_mov_epi8() { |
17612 | let src = _mm512_set1_epi8(1); |
17613 | let a = _mm512_set1_epi8(2); |
17614 | let r = _mm512_mask_mov_epi8(src, 0, a); |
17615 | assert_eq_m512i(r, src); |
17616 | let r = _mm512_mask_mov_epi8( |
17617 | src, |
17618 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
17619 | a, |
17620 | ); |
17621 | assert_eq_m512i(r, a); |
17622 | } |
17623 | |
17624 | #[simd_test(enable = "avx512bw" )] |
17625 | unsafe fn test_mm512_maskz_mov_epi8() { |
17626 | let a = _mm512_set1_epi8(2); |
17627 | let r = _mm512_maskz_mov_epi8(0, a); |
17628 | assert_eq_m512i(r, _mm512_setzero_si512()); |
17629 | let r = _mm512_maskz_mov_epi8( |
17630 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
17631 | a, |
17632 | ); |
17633 | assert_eq_m512i(r, a); |
17634 | } |
17635 | |
17636 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17637 | unsafe fn test_mm256_mask_mov_epi8() { |
17638 | let src = _mm256_set1_epi8(1); |
17639 | let a = _mm256_set1_epi8(2); |
17640 | let r = _mm256_mask_mov_epi8(src, 0, a); |
17641 | assert_eq_m256i(r, src); |
17642 | let r = _mm256_mask_mov_epi8(src, 0b11111111_11111111_11111111_11111111, a); |
17643 | assert_eq_m256i(r, a); |
17644 | } |
17645 | |
17646 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17647 | unsafe fn test_mm256_maskz_mov_epi8() { |
17648 | let a = _mm256_set1_epi8(2); |
17649 | let r = _mm256_maskz_mov_epi8(0, a); |
17650 | assert_eq_m256i(r, _mm256_setzero_si256()); |
17651 | let r = _mm256_maskz_mov_epi8(0b11111111_11111111_11111111_11111111, a); |
17652 | assert_eq_m256i(r, a); |
17653 | } |
17654 | |
17655 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17656 | unsafe fn test_mm_mask_mov_epi8() { |
17657 | let src = _mm_set1_epi8(1); |
17658 | let a = _mm_set1_epi8(2); |
17659 | let r = _mm_mask_mov_epi8(src, 0, a); |
17660 | assert_eq_m128i(r, src); |
17661 | let r = _mm_mask_mov_epi8(src, 0b11111111_11111111, a); |
17662 | assert_eq_m128i(r, a); |
17663 | } |
17664 | |
17665 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17666 | unsafe fn test_mm_maskz_mov_epi8() { |
17667 | let a = _mm_set1_epi8(2); |
17668 | let r = _mm_maskz_mov_epi8(0, a); |
17669 | assert_eq_m128i(r, _mm_setzero_si128()); |
17670 | let r = _mm_maskz_mov_epi8(0b11111111_11111111, a); |
17671 | assert_eq_m128i(r, a); |
17672 | } |
17673 | |
17674 | #[simd_test(enable = "avx512bw" )] |
17675 | unsafe fn test_mm512_mask_set1_epi16() { |
17676 | let src = _mm512_set1_epi16(2); |
17677 | let a: i16 = 11; |
17678 | let r = _mm512_mask_set1_epi16(src, 0, a); |
17679 | assert_eq_m512i(r, src); |
17680 | let r = _mm512_mask_set1_epi16(src, 0b11111111_11111111_11111111_11111111, a); |
17681 | let e = _mm512_set1_epi16(11); |
17682 | assert_eq_m512i(r, e); |
17683 | } |
17684 | |
17685 | #[simd_test(enable = "avx512bw" )] |
17686 | unsafe fn test_mm512_maskz_set1_epi16() { |
17687 | let a: i16 = 11; |
17688 | let r = _mm512_maskz_set1_epi16(0, a); |
17689 | assert_eq_m512i(r, _mm512_setzero_si512()); |
17690 | let r = _mm512_maskz_set1_epi16(0b11111111_11111111_11111111_11111111, a); |
17691 | let e = _mm512_set1_epi16(11); |
17692 | assert_eq_m512i(r, e); |
17693 | } |
17694 | |
17695 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17696 | unsafe fn test_mm256_mask_set1_epi16() { |
17697 | let src = _mm256_set1_epi16(2); |
17698 | let a: i16 = 11; |
17699 | let r = _mm256_mask_set1_epi16(src, 0, a); |
17700 | assert_eq_m256i(r, src); |
17701 | let r = _mm256_mask_set1_epi16(src, 0b11111111_11111111, a); |
17702 | let e = _mm256_set1_epi16(11); |
17703 | assert_eq_m256i(r, e); |
17704 | } |
17705 | |
17706 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17707 | unsafe fn test_mm256_maskz_set1_epi16() { |
17708 | let a: i16 = 11; |
17709 | let r = _mm256_maskz_set1_epi16(0, a); |
17710 | assert_eq_m256i(r, _mm256_setzero_si256()); |
17711 | let r = _mm256_maskz_set1_epi16(0b11111111_11111111, a); |
17712 | let e = _mm256_set1_epi16(11); |
17713 | assert_eq_m256i(r, e); |
17714 | } |
17715 | |
17716 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17717 | unsafe fn test_mm_mask_set1_epi16() { |
17718 | let src = _mm_set1_epi16(2); |
17719 | let a: i16 = 11; |
17720 | let r = _mm_mask_set1_epi16(src, 0, a); |
17721 | assert_eq_m128i(r, src); |
17722 | let r = _mm_mask_set1_epi16(src, 0b11111111, a); |
17723 | let e = _mm_set1_epi16(11); |
17724 | assert_eq_m128i(r, e); |
17725 | } |
17726 | |
17727 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17728 | unsafe fn test_mm_maskz_set1_epi16() { |
17729 | let a: i16 = 11; |
17730 | let r = _mm_maskz_set1_epi16(0, a); |
17731 | assert_eq_m128i(r, _mm_setzero_si128()); |
17732 | let r = _mm_maskz_set1_epi16(0b11111111, a); |
17733 | let e = _mm_set1_epi16(11); |
17734 | assert_eq_m128i(r, e); |
17735 | } |
17736 | |
17737 | #[simd_test(enable = "avx512bw" )] |
17738 | unsafe fn test_mm512_mask_set1_epi8() { |
17739 | let src = _mm512_set1_epi8(2); |
17740 | let a: i8 = 11; |
17741 | let r = _mm512_mask_set1_epi8(src, 0, a); |
17742 | assert_eq_m512i(r, src); |
17743 | let r = _mm512_mask_set1_epi8( |
17744 | src, |
17745 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
17746 | a, |
17747 | ); |
17748 | let e = _mm512_set1_epi8(11); |
17749 | assert_eq_m512i(r, e); |
17750 | } |
17751 | |
17752 | #[simd_test(enable = "avx512bw" )] |
17753 | unsafe fn test_mm512_maskz_set1_epi8() { |
17754 | let a: i8 = 11; |
17755 | let r = _mm512_maskz_set1_epi8(0, a); |
17756 | assert_eq_m512i(r, _mm512_setzero_si512()); |
17757 | let r = _mm512_maskz_set1_epi8( |
17758 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
17759 | a, |
17760 | ); |
17761 | let e = _mm512_set1_epi8(11); |
17762 | assert_eq_m512i(r, e); |
17763 | } |
17764 | |
17765 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17766 | unsafe fn test_mm256_mask_set1_epi8() { |
17767 | let src = _mm256_set1_epi8(2); |
17768 | let a: i8 = 11; |
17769 | let r = _mm256_mask_set1_epi8(src, 0, a); |
17770 | assert_eq_m256i(r, src); |
17771 | let r = _mm256_mask_set1_epi8(src, 0b11111111_11111111_11111111_11111111, a); |
17772 | let e = _mm256_set1_epi8(11); |
17773 | assert_eq_m256i(r, e); |
17774 | } |
17775 | |
17776 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17777 | unsafe fn test_mm256_maskz_set1_epi8() { |
17778 | let a: i8 = 11; |
17779 | let r = _mm256_maskz_set1_epi8(0, a); |
17780 | assert_eq_m256i(r, _mm256_setzero_si256()); |
17781 | let r = _mm256_maskz_set1_epi8(0b11111111_11111111_11111111_11111111, a); |
17782 | let e = _mm256_set1_epi8(11); |
17783 | assert_eq_m256i(r, e); |
17784 | } |
17785 | |
17786 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17787 | unsafe fn test_mm_mask_set1_epi8() { |
17788 | let src = _mm_set1_epi8(2); |
17789 | let a: i8 = 11; |
17790 | let r = _mm_mask_set1_epi8(src, 0, a); |
17791 | assert_eq_m128i(r, src); |
17792 | let r = _mm_mask_set1_epi8(src, 0b11111111_11111111, a); |
17793 | let e = _mm_set1_epi8(11); |
17794 | assert_eq_m128i(r, e); |
17795 | } |
17796 | |
17797 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17798 | unsafe fn test_mm_maskz_set1_epi8() { |
17799 | let a: i8 = 11; |
17800 | let r = _mm_maskz_set1_epi8(0, a); |
17801 | assert_eq_m128i(r, _mm_setzero_si128()); |
17802 | let r = _mm_maskz_set1_epi8(0b11111111_11111111, a); |
17803 | let e = _mm_set1_epi8(11); |
17804 | assert_eq_m128i(r, e); |
17805 | } |
17806 | |
17807 | #[simd_test(enable = "avx512bw" )] |
17808 | unsafe fn test_mm512_shufflelo_epi16() { |
17809 | #[rustfmt::skip] |
17810 | let a = _mm512_set_epi16( |
17811 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
17812 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
17813 | ); |
17814 | #[rustfmt::skip] |
17815 | let e = _mm512_set_epi16( |
17816 | 0, 1, 2, 3, 7, 6, 6, 4, 8, 9, 10, 11, 15, 14, 14, 12, |
17817 | 16, 17, 18, 19, 23, 22, 22, 20, 24, 25, 26, 27, 31, 30, 30, 28, |
17818 | ); |
17819 | let r = _mm512_shufflelo_epi16::<0b00_01_01_11>(a); |
17820 | assert_eq_m512i(r, e); |
17821 | } |
17822 | |
17823 | #[simd_test(enable = "avx512bw" )] |
17824 | unsafe fn test_mm512_mask_shufflelo_epi16() { |
17825 | #[rustfmt::skip] |
17826 | let a = _mm512_set_epi16( |
17827 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
17828 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
17829 | ); |
17830 | let r = _mm512_mask_shufflelo_epi16::<0b00_01_01_11>(a, 0, a); |
17831 | assert_eq_m512i(r, a); |
17832 | let r = _mm512_mask_shufflelo_epi16::<0b00_01_01_11>( |
17833 | a, |
17834 | 0b11111111_11111111_11111111_11111111, |
17835 | a, |
17836 | ); |
17837 | #[rustfmt::skip] |
17838 | let e = _mm512_set_epi16( |
17839 | 0, 1, 2, 3, 7, 6, 6, 4, 8, 9, 10, 11, 15, 14, 14, 12, |
17840 | 16, 17, 18, 19, 23, 22, 22, 20, 24, 25, 26, 27, 31, 30, 30, 28, |
17841 | ); |
17842 | assert_eq_m512i(r, e); |
17843 | } |
17844 | |
17845 | #[simd_test(enable = "avx512bw" )] |
17846 | unsafe fn test_mm512_maskz_shufflelo_epi16() { |
17847 | #[rustfmt::skip] |
17848 | let a = _mm512_set_epi16( |
17849 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
17850 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
17851 | ); |
17852 | let r = _mm512_maskz_shufflelo_epi16::<0b00_01_01_11>(0, a); |
17853 | assert_eq_m512i(r, _mm512_setzero_si512()); |
17854 | let r = |
17855 | _mm512_maskz_shufflelo_epi16::<0b00_01_01_11>(0b11111111_11111111_11111111_11111111, a); |
17856 | #[rustfmt::skip] |
17857 | let e = _mm512_set_epi16( |
17858 | 0, 1, 2, 3, 7, 6, 6, 4, 8, 9, 10, 11, 15, 14, 14, 12, |
17859 | 16, 17, 18, 19, 23, 22, 22, 20, 24, 25, 26, 27, 31, 30, 30, 28, |
17860 | ); |
17861 | assert_eq_m512i(r, e); |
17862 | } |
17863 | |
17864 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17865 | unsafe fn test_mm256_mask_shufflelo_epi16() { |
17866 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
17867 | let r = _mm256_mask_shufflelo_epi16::<0b00_01_01_11>(a, 0, a); |
17868 | assert_eq_m256i(r, a); |
17869 | let r = _mm256_mask_shufflelo_epi16::<0b00_01_01_11>(a, 0b11111111_11111111, a); |
17870 | let e = _mm256_set_epi16(0, 1, 2, 3, 7, 6, 6, 4, 8, 9, 10, 11, 15, 14, 14, 12); |
17871 | assert_eq_m256i(r, e); |
17872 | } |
17873 | |
17874 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17875 | unsafe fn test_mm256_maskz_shufflelo_epi16() { |
17876 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
17877 | let r = _mm256_maskz_shufflelo_epi16::<0b00_01_01_11>(0, a); |
17878 | assert_eq_m256i(r, _mm256_setzero_si256()); |
17879 | let r = _mm256_maskz_shufflelo_epi16::<0b00_01_01_11>(0b11111111_11111111, a); |
17880 | let e = _mm256_set_epi16(0, 1, 2, 3, 7, 6, 6, 4, 8, 9, 10, 11, 15, 14, 14, 12); |
17881 | assert_eq_m256i(r, e); |
17882 | } |
17883 | |
17884 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17885 | unsafe fn test_mm_mask_shufflelo_epi16() { |
17886 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
17887 | let r = _mm_mask_shufflelo_epi16::<0b00_01_01_11>(a, 0, a); |
17888 | assert_eq_m128i(r, a); |
17889 | let r = _mm_mask_shufflelo_epi16::<0b00_01_01_11>(a, 0b11111111, a); |
17890 | let e = _mm_set_epi16(0, 1, 2, 3, 7, 6, 6, 4); |
17891 | assert_eq_m128i(r, e); |
17892 | } |
17893 | |
17894 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17895 | unsafe fn test_mm_maskz_shufflelo_epi16() { |
17896 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
17897 | let r = _mm_maskz_shufflelo_epi16::<0b00_01_01_11>(0, a); |
17898 | assert_eq_m128i(r, _mm_setzero_si128()); |
17899 | let r = _mm_maskz_shufflelo_epi16::<0b00_01_01_11>(0b11111111, a); |
17900 | let e = _mm_set_epi16(0, 1, 2, 3, 7, 6, 6, 4); |
17901 | assert_eq_m128i(r, e); |
17902 | } |
17903 | |
17904 | #[simd_test(enable = "avx512bw" )] |
17905 | unsafe fn test_mm512_shufflehi_epi16() { |
17906 | #[rustfmt::skip] |
17907 | let a = _mm512_set_epi16( |
17908 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
17909 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
17910 | ); |
17911 | #[rustfmt::skip] |
17912 | let e = _mm512_set_epi16( |
17913 | 3, 2, 2, 0, 4, 5, 6, 7, 11, 10, 10, 8, 12, 13, 14, 15, |
17914 | 19, 18, 18, 16, 20, 21, 22, 23, 27, 26, 26, 24, 28, 29, 30, 31, |
17915 | ); |
17916 | let r = _mm512_shufflehi_epi16::<0b00_01_01_11>(a); |
17917 | assert_eq_m512i(r, e); |
17918 | } |
17919 | |
17920 | #[simd_test(enable = "avx512bw" )] |
17921 | unsafe fn test_mm512_mask_shufflehi_epi16() { |
17922 | #[rustfmt::skip] |
17923 | let a = _mm512_set_epi16( |
17924 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
17925 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
17926 | ); |
17927 | let r = _mm512_mask_shufflehi_epi16::<0b00_01_01_11>(a, 0, a); |
17928 | assert_eq_m512i(r, a); |
17929 | let r = _mm512_mask_shufflehi_epi16::<0b00_01_01_11>( |
17930 | a, |
17931 | 0b11111111_11111111_11111111_11111111, |
17932 | a, |
17933 | ); |
17934 | #[rustfmt::skip] |
17935 | let e = _mm512_set_epi16( |
17936 | 3, 2, 2, 0, 4, 5, 6, 7, 11, 10, 10, 8, 12, 13, 14, 15, |
17937 | 19, 18, 18, 16, 20, 21, 22, 23, 27, 26, 26, 24, 28, 29, 30, 31, |
17938 | ); |
17939 | assert_eq_m512i(r, e); |
17940 | } |
17941 | |
17942 | #[simd_test(enable = "avx512bw" )] |
17943 | unsafe fn test_mm512_maskz_shufflehi_epi16() { |
17944 | #[rustfmt::skip] |
17945 | let a = _mm512_set_epi16( |
17946 | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
17947 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
17948 | ); |
17949 | let r = _mm512_maskz_shufflehi_epi16::<0b00_01_01_11>(0, a); |
17950 | assert_eq_m512i(r, _mm512_setzero_si512()); |
17951 | let r = |
17952 | _mm512_maskz_shufflehi_epi16::<0b00_01_01_11>(0b11111111_11111111_11111111_11111111, a); |
17953 | #[rustfmt::skip] |
17954 | let e = _mm512_set_epi16( |
17955 | 3, 2, 2, 0, 4, 5, 6, 7, 11, 10, 10, 8, 12, 13, 14, 15, |
17956 | 19, 18, 18, 16, 20, 21, 22, 23, 27, 26, 26, 24, 28, 29, 30, 31, |
17957 | ); |
17958 | assert_eq_m512i(r, e); |
17959 | } |
17960 | |
17961 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17962 | unsafe fn test_mm256_mask_shufflehi_epi16() { |
17963 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
17964 | let r = _mm256_mask_shufflehi_epi16::<0b00_01_01_11>(a, 0, a); |
17965 | assert_eq_m256i(r, a); |
17966 | let r = _mm256_mask_shufflehi_epi16::<0b00_01_01_11>(a, 0b11111111_11111111, a); |
17967 | let e = _mm256_set_epi16(3, 2, 2, 0, 4, 5, 6, 7, 11, 10, 10, 8, 12, 13, 14, 15); |
17968 | assert_eq_m256i(r, e); |
17969 | } |
17970 | |
17971 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17972 | unsafe fn test_mm256_maskz_shufflehi_epi16() { |
17973 | let a = _mm256_set_epi16(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
17974 | let r = _mm256_maskz_shufflehi_epi16::<0b00_01_01_11>(0, a); |
17975 | assert_eq_m256i(r, _mm256_setzero_si256()); |
17976 | let r = _mm256_maskz_shufflehi_epi16::<0b00_01_01_11>(0b11111111_11111111, a); |
17977 | let e = _mm256_set_epi16(3, 2, 2, 0, 4, 5, 6, 7, 11, 10, 10, 8, 12, 13, 14, 15); |
17978 | assert_eq_m256i(r, e); |
17979 | } |
17980 | |
17981 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17982 | unsafe fn test_mm_mask_shufflehi_epi16() { |
17983 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
17984 | let r = _mm_mask_shufflehi_epi16::<0b00_01_01_11>(a, 0, a); |
17985 | assert_eq_m128i(r, a); |
17986 | let r = _mm_mask_shufflehi_epi16::<0b00_01_01_11>(a, 0b11111111, a); |
17987 | let e = _mm_set_epi16(3, 2, 2, 0, 4, 5, 6, 7); |
17988 | assert_eq_m128i(r, e); |
17989 | } |
17990 | |
17991 | #[simd_test(enable = "avx512bw,avx512vl" )] |
17992 | unsafe fn test_mm_maskz_shufflehi_epi16() { |
17993 | let a = _mm_set_epi16(0, 1, 2, 3, 4, 5, 6, 7); |
17994 | let r = _mm_maskz_shufflehi_epi16::<0b00_01_01_11>(0, a); |
17995 | assert_eq_m128i(r, _mm_setzero_si128()); |
17996 | let r = _mm_maskz_shufflehi_epi16::<0b00_01_01_11>(0b11111111, a); |
17997 | let e = _mm_set_epi16(3, 2, 2, 0, 4, 5, 6, 7); |
17998 | assert_eq_m128i(r, e); |
17999 | } |
18000 | |
18001 | #[simd_test(enable = "avx512bw" )] |
18002 | unsafe fn test_mm512_shuffle_epi8() { |
18003 | #[rustfmt::skip] |
18004 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
18005 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
18006 | 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, |
18007 | 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63); |
18008 | let b = _mm512_set1_epi8(1); |
18009 | let r = _mm512_shuffle_epi8(a, b); |
18010 | #[rustfmt::skip] |
18011 | let e = _mm512_set_epi8(14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
18012 | 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, |
18013 | 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, |
18014 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62); |
18015 | assert_eq_m512i(r, e); |
18016 | } |
18017 | |
18018 | #[simd_test(enable = "avx512bw" )] |
18019 | unsafe fn test_mm512_mask_shuffle_epi8() { |
18020 | #[rustfmt::skip] |
18021 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
18022 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
18023 | 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, |
18024 | 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63); |
18025 | let b = _mm512_set1_epi8(1); |
18026 | let r = _mm512_mask_shuffle_epi8(a, 0, a, b); |
18027 | assert_eq_m512i(r, a); |
18028 | let r = _mm512_mask_shuffle_epi8( |
18029 | a, |
18030 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
18031 | a, |
18032 | b, |
18033 | ); |
18034 | #[rustfmt::skip] |
18035 | let e = _mm512_set_epi8(14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
18036 | 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, |
18037 | 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, |
18038 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62); |
18039 | assert_eq_m512i(r, e); |
18040 | } |
18041 | |
18042 | #[simd_test(enable = "avx512bw" )] |
18043 | unsafe fn test_mm512_maskz_shuffle_epi8() { |
18044 | #[rustfmt::skip] |
18045 | let a = _mm512_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
18046 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
18047 | 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, |
18048 | 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63); |
18049 | let b = _mm512_set1_epi8(1); |
18050 | let r = _mm512_maskz_shuffle_epi8(0, a, b); |
18051 | assert_eq_m512i(r, _mm512_setzero_si512()); |
18052 | let r = _mm512_maskz_shuffle_epi8( |
18053 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
18054 | a, |
18055 | b, |
18056 | ); |
18057 | #[rustfmt::skip] |
18058 | let e = _mm512_set_epi8(14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
18059 | 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, |
18060 | 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, |
18061 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, 62); |
18062 | assert_eq_m512i(r, e); |
18063 | } |
18064 | |
18065 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18066 | unsafe fn test_mm256_mask_shuffle_epi8() { |
18067 | #[rustfmt::skip] |
18068 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
18069 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
18070 | let b = _mm256_set1_epi8(1); |
18071 | let r = _mm256_mask_shuffle_epi8(a, 0, a, b); |
18072 | assert_eq_m256i(r, a); |
18073 | let r = _mm256_mask_shuffle_epi8(a, 0b11111111_11111111_11111111_11111111, a, b); |
18074 | #[rustfmt::skip] |
18075 | let e = _mm256_set_epi8(14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
18076 | 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30); |
18077 | assert_eq_m256i(r, e); |
18078 | } |
18079 | |
18080 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18081 | unsafe fn test_mm256_maskz_shuffle_epi8() { |
18082 | #[rustfmt::skip] |
18083 | let a = _mm256_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
18084 | 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31); |
18085 | let b = _mm256_set1_epi8(1); |
18086 | let r = _mm256_maskz_shuffle_epi8(0, a, b); |
18087 | assert_eq_m256i(r, _mm256_setzero_si256()); |
18088 | let r = _mm256_maskz_shuffle_epi8(0b11111111_11111111_11111111_11111111, a, b); |
18089 | #[rustfmt::skip] |
18090 | let e = _mm256_set_epi8(14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
18091 | 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30); |
18092 | assert_eq_m256i(r, e); |
18093 | } |
18094 | |
18095 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18096 | unsafe fn test_mm_mask_shuffle_epi8() { |
18097 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
18098 | let b = _mm_set1_epi8(1); |
18099 | let r = _mm_mask_shuffle_epi8(a, 0, a, b); |
18100 | assert_eq_m128i(r, a); |
18101 | let r = _mm_mask_shuffle_epi8(a, 0b11111111_11111111, a, b); |
18102 | let e = _mm_set_epi8( |
18103 | 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
18104 | ); |
18105 | assert_eq_m128i(r, e); |
18106 | } |
18107 | |
18108 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18109 | unsafe fn test_mm_maskz_shuffle_epi8() { |
18110 | #[rustfmt::skip] |
18111 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
18112 | let b = _mm_set1_epi8(1); |
18113 | let r = _mm_maskz_shuffle_epi8(0, a, b); |
18114 | assert_eq_m128i(r, _mm_setzero_si128()); |
18115 | let r = _mm_maskz_shuffle_epi8(0b11111111_11111111, a, b); |
18116 | let e = _mm_set_epi8( |
18117 | 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, |
18118 | ); |
18119 | assert_eq_m128i(r, e); |
18120 | } |
18121 | |
18122 | #[simd_test(enable = "avx512bw" )] |
18123 | unsafe fn test_mm512_test_epi16_mask() { |
18124 | let a = _mm512_set1_epi16(1 << 0); |
18125 | let b = _mm512_set1_epi16(1 << 0 | 1 << 1); |
18126 | let r = _mm512_test_epi16_mask(a, b); |
18127 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
18128 | assert_eq!(r, e); |
18129 | } |
18130 | |
18131 | #[simd_test(enable = "avx512bw" )] |
18132 | unsafe fn test_mm512_mask_test_epi16_mask() { |
18133 | let a = _mm512_set1_epi16(1 << 0); |
18134 | let b = _mm512_set1_epi16(1 << 0 | 1 << 1); |
18135 | let r = _mm512_mask_test_epi16_mask(0, a, b); |
18136 | assert_eq!(r, 0); |
18137 | let r = _mm512_mask_test_epi16_mask(0b11111111_11111111_11111111_11111111, a, b); |
18138 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
18139 | assert_eq!(r, e); |
18140 | } |
18141 | |
18142 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18143 | unsafe fn test_mm256_test_epi16_mask() { |
18144 | let a = _mm256_set1_epi16(1 << 0); |
18145 | let b = _mm256_set1_epi16(1 << 0 | 1 << 1); |
18146 | let r = _mm256_test_epi16_mask(a, b); |
18147 | let e: __mmask16 = 0b11111111_11111111; |
18148 | assert_eq!(r, e); |
18149 | } |
18150 | |
18151 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18152 | unsafe fn test_mm256_mask_test_epi16_mask() { |
18153 | let a = _mm256_set1_epi16(1 << 0); |
18154 | let b = _mm256_set1_epi16(1 << 0 | 1 << 1); |
18155 | let r = _mm256_mask_test_epi16_mask(0, a, b); |
18156 | assert_eq!(r, 0); |
18157 | let r = _mm256_mask_test_epi16_mask(0b11111111_11111111, a, b); |
18158 | let e: __mmask16 = 0b11111111_11111111; |
18159 | assert_eq!(r, e); |
18160 | } |
18161 | |
18162 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18163 | unsafe fn test_mm_test_epi16_mask() { |
18164 | let a = _mm_set1_epi16(1 << 0); |
18165 | let b = _mm_set1_epi16(1 << 0 | 1 << 1); |
18166 | let r = _mm_test_epi16_mask(a, b); |
18167 | let e: __mmask8 = 0b11111111; |
18168 | assert_eq!(r, e); |
18169 | } |
18170 | |
18171 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18172 | unsafe fn test_mm_mask_test_epi16_mask() { |
18173 | let a = _mm_set1_epi16(1 << 0); |
18174 | let b = _mm_set1_epi16(1 << 0 | 1 << 1); |
18175 | let r = _mm_mask_test_epi16_mask(0, a, b); |
18176 | assert_eq!(r, 0); |
18177 | let r = _mm_mask_test_epi16_mask(0b11111111, a, b); |
18178 | let e: __mmask8 = 0b11111111; |
18179 | assert_eq!(r, e); |
18180 | } |
18181 | |
18182 | #[simd_test(enable = "avx512bw" )] |
18183 | unsafe fn test_mm512_test_epi8_mask() { |
18184 | let a = _mm512_set1_epi8(1 << 0); |
18185 | let b = _mm512_set1_epi8(1 << 0 | 1 << 1); |
18186 | let r = _mm512_test_epi8_mask(a, b); |
18187 | let e: __mmask64 = |
18188 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111; |
18189 | assert_eq!(r, e); |
18190 | } |
18191 | |
18192 | #[simd_test(enable = "avx512bw" )] |
18193 | unsafe fn test_mm512_mask_test_epi8_mask() { |
18194 | let a = _mm512_set1_epi8(1 << 0); |
18195 | let b = _mm512_set1_epi8(1 << 0 | 1 << 1); |
18196 | let r = _mm512_mask_test_epi8_mask(0, a, b); |
18197 | assert_eq!(r, 0); |
18198 | let r = _mm512_mask_test_epi8_mask( |
18199 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
18200 | a, |
18201 | b, |
18202 | ); |
18203 | let e: __mmask64 = |
18204 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111; |
18205 | assert_eq!(r, e); |
18206 | } |
18207 | |
18208 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18209 | unsafe fn test_mm256_test_epi8_mask() { |
18210 | let a = _mm256_set1_epi8(1 << 0); |
18211 | let b = _mm256_set1_epi8(1 << 0 | 1 << 1); |
18212 | let r = _mm256_test_epi8_mask(a, b); |
18213 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
18214 | assert_eq!(r, e); |
18215 | } |
18216 | |
18217 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18218 | unsafe fn test_mm256_mask_test_epi8_mask() { |
18219 | let a = _mm256_set1_epi8(1 << 0); |
18220 | let b = _mm256_set1_epi8(1 << 0 | 1 << 1); |
18221 | let r = _mm256_mask_test_epi8_mask(0, a, b); |
18222 | assert_eq!(r, 0); |
18223 | let r = _mm256_mask_test_epi8_mask(0b11111111_11111111_11111111_11111111, a, b); |
18224 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
18225 | assert_eq!(r, e); |
18226 | } |
18227 | |
18228 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18229 | unsafe fn test_mm_test_epi8_mask() { |
18230 | let a = _mm_set1_epi8(1 << 0); |
18231 | let b = _mm_set1_epi8(1 << 0 | 1 << 1); |
18232 | let r = _mm_test_epi8_mask(a, b); |
18233 | let e: __mmask16 = 0b11111111_11111111; |
18234 | assert_eq!(r, e); |
18235 | } |
18236 | |
18237 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18238 | unsafe fn test_mm_mask_test_epi8_mask() { |
18239 | let a = _mm_set1_epi8(1 << 0); |
18240 | let b = _mm_set1_epi8(1 << 0 | 1 << 1); |
18241 | let r = _mm_mask_test_epi8_mask(0, a, b); |
18242 | assert_eq!(r, 0); |
18243 | let r = _mm_mask_test_epi8_mask(0b11111111_11111111, a, b); |
18244 | let e: __mmask16 = 0b11111111_11111111; |
18245 | assert_eq!(r, e); |
18246 | } |
18247 | |
18248 | #[simd_test(enable = "avx512bw" )] |
18249 | unsafe fn test_mm512_testn_epi16_mask() { |
18250 | let a = _mm512_set1_epi16(1 << 0); |
18251 | let b = _mm512_set1_epi16(1 << 0 | 1 << 1); |
18252 | let r = _mm512_testn_epi16_mask(a, b); |
18253 | let e: __mmask32 = 0b00000000_00000000_00000000_00000000; |
18254 | assert_eq!(r, e); |
18255 | } |
18256 | |
18257 | #[simd_test(enable = "avx512bw" )] |
18258 | unsafe fn test_mm512_mask_testn_epi16_mask() { |
18259 | let a = _mm512_set1_epi16(1 << 0); |
18260 | let b = _mm512_set1_epi16(1 << 0 | 1 << 1); |
18261 | let r = _mm512_mask_testn_epi16_mask(0, a, b); |
18262 | assert_eq!(r, 0); |
18263 | let r = _mm512_mask_testn_epi16_mask(0b11111111_11111111_11111111_11111111, a, b); |
18264 | let e: __mmask32 = 0b00000000_00000000_00000000_00000000; |
18265 | assert_eq!(r, e); |
18266 | } |
18267 | |
18268 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18269 | unsafe fn test_mm256_testn_epi16_mask() { |
18270 | let a = _mm256_set1_epi16(1 << 0); |
18271 | let b = _mm256_set1_epi16(1 << 0 | 1 << 1); |
18272 | let r = _mm256_testn_epi16_mask(a, b); |
18273 | let e: __mmask16 = 0b00000000_00000000; |
18274 | assert_eq!(r, e); |
18275 | } |
18276 | |
18277 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18278 | unsafe fn test_mm256_mask_testn_epi16_mask() { |
18279 | let a = _mm256_set1_epi16(1 << 0); |
18280 | let b = _mm256_set1_epi16(1 << 0 | 1 << 1); |
18281 | let r = _mm256_mask_testn_epi16_mask(0, a, b); |
18282 | assert_eq!(r, 0); |
18283 | let r = _mm256_mask_testn_epi16_mask(0b11111111_11111111, a, b); |
18284 | let e: __mmask16 = 0b00000000_00000000; |
18285 | assert_eq!(r, e); |
18286 | } |
18287 | |
18288 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18289 | unsafe fn test_mm_testn_epi16_mask() { |
18290 | let a = _mm_set1_epi16(1 << 0); |
18291 | let b = _mm_set1_epi16(1 << 0 | 1 << 1); |
18292 | let r = _mm_testn_epi16_mask(a, b); |
18293 | let e: __mmask8 = 0b00000000; |
18294 | assert_eq!(r, e); |
18295 | } |
18296 | |
18297 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18298 | unsafe fn test_mm_mask_testn_epi16_mask() { |
18299 | let a = _mm_set1_epi16(1 << 0); |
18300 | let b = _mm_set1_epi16(1 << 0 | 1 << 1); |
18301 | let r = _mm_mask_testn_epi16_mask(0, a, b); |
18302 | assert_eq!(r, 0); |
18303 | let r = _mm_mask_testn_epi16_mask(0b11111111, a, b); |
18304 | let e: __mmask8 = 0b00000000; |
18305 | assert_eq!(r, e); |
18306 | } |
18307 | |
18308 | #[simd_test(enable = "avx512bw" )] |
18309 | unsafe fn test_mm512_testn_epi8_mask() { |
18310 | let a = _mm512_set1_epi8(1 << 0); |
18311 | let b = _mm512_set1_epi8(1 << 0 | 1 << 1); |
18312 | let r = _mm512_testn_epi8_mask(a, b); |
18313 | let e: __mmask64 = |
18314 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00000000; |
18315 | assert_eq!(r, e); |
18316 | } |
18317 | |
18318 | #[simd_test(enable = "avx512bw" )] |
18319 | unsafe fn test_mm512_mask_testn_epi8_mask() { |
18320 | let a = _mm512_set1_epi8(1 << 0); |
18321 | let b = _mm512_set1_epi8(1 << 0 | 1 << 1); |
18322 | let r = _mm512_mask_testn_epi8_mask(0, a, b); |
18323 | assert_eq!(r, 0); |
18324 | let r = _mm512_mask_testn_epi8_mask( |
18325 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
18326 | a, |
18327 | b, |
18328 | ); |
18329 | let e: __mmask64 = |
18330 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00000000; |
18331 | assert_eq!(r, e); |
18332 | } |
18333 | |
18334 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18335 | unsafe fn test_mm256_testn_epi8_mask() { |
18336 | let a = _mm256_set1_epi8(1 << 0); |
18337 | let b = _mm256_set1_epi8(1 << 0 | 1 << 1); |
18338 | let r = _mm256_testn_epi8_mask(a, b); |
18339 | let e: __mmask32 = 0b00000000_00000000_00000000_00000000; |
18340 | assert_eq!(r, e); |
18341 | } |
18342 | |
18343 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18344 | unsafe fn test_mm256_mask_testn_epi8_mask() { |
18345 | let a = _mm256_set1_epi8(1 << 0); |
18346 | let b = _mm256_set1_epi8(1 << 0 | 1 << 1); |
18347 | let r = _mm256_mask_testn_epi8_mask(0, a, b); |
18348 | assert_eq!(r, 0); |
18349 | let r = _mm256_mask_testn_epi8_mask(0b11111111_11111111_11111111_11111111, a, b); |
18350 | let e: __mmask32 = 0b00000000_00000000_00000000_00000000; |
18351 | assert_eq!(r, e); |
18352 | } |
18353 | |
18354 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18355 | unsafe fn test_mm_testn_epi8_mask() { |
18356 | let a = _mm_set1_epi8(1 << 0); |
18357 | let b = _mm_set1_epi8(1 << 0 | 1 << 1); |
18358 | let r = _mm_testn_epi8_mask(a, b); |
18359 | let e: __mmask16 = 0b00000000_00000000; |
18360 | assert_eq!(r, e); |
18361 | } |
18362 | |
18363 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18364 | unsafe fn test_mm_mask_testn_epi8_mask() { |
18365 | let a = _mm_set1_epi8(1 << 0); |
18366 | let b = _mm_set1_epi8(1 << 0 | 1 << 1); |
18367 | let r = _mm_mask_testn_epi8_mask(0, a, b); |
18368 | assert_eq!(r, 0); |
18369 | let r = _mm_mask_testn_epi8_mask(0b11111111_11111111, a, b); |
18370 | let e: __mmask16 = 0b00000000_00000000; |
18371 | assert_eq!(r, e); |
18372 | } |
18373 | |
18374 | #[simd_test(enable = "avx512bw" )] |
18375 | unsafe fn test_store_mask64() { |
18376 | let a: __mmask64 = |
18377 | 0b11111111_00000000_11111111_00000000_11111111_00000000_11111111_00000000; |
18378 | let mut r = 0; |
18379 | _store_mask64(&mut r, a); |
18380 | assert_eq!(r, a); |
18381 | } |
18382 | |
18383 | #[simd_test(enable = "avx512bw" )] |
18384 | unsafe fn test_store_mask32() { |
18385 | let a: __mmask32 = 0b11111111_00000000_11111111_00000000; |
18386 | let mut r = 0; |
18387 | _store_mask32(&mut r, a); |
18388 | assert_eq!(r, a); |
18389 | } |
18390 | |
18391 | #[simd_test(enable = "avx512bw" )] |
18392 | unsafe fn test_load_mask64() { |
18393 | let p: __mmask64 = |
18394 | 0b11111111_00000000_11111111_00000000_11111111_00000000_11111111_00000000; |
18395 | let r = _load_mask64(&p); |
18396 | let e: __mmask64 = |
18397 | 0b11111111_00000000_11111111_00000000_11111111_00000000_11111111_00000000; |
18398 | assert_eq!(r, e); |
18399 | } |
18400 | |
18401 | #[simd_test(enable = "avx512bw" )] |
18402 | unsafe fn test_load_mask32() { |
18403 | let p: __mmask32 = 0b11111111_00000000_11111111_00000000; |
18404 | let r = _load_mask32(&p); |
18405 | let e: __mmask32 = 0b11111111_00000000_11111111_00000000; |
18406 | assert_eq!(r, e); |
18407 | } |
18408 | |
18409 | #[simd_test(enable = "avx512bw" )] |
18410 | unsafe fn test_mm512_sad_epu8() { |
18411 | let a = _mm512_set1_epi8(2); |
18412 | let b = _mm512_set1_epi8(4); |
18413 | let r = _mm512_sad_epu8(a, b); |
18414 | let e = _mm512_set1_epi64(16); |
18415 | assert_eq_m512i(r, e); |
18416 | } |
18417 | |
18418 | #[simd_test(enable = "avx512bw" )] |
18419 | unsafe fn test_mm512_dbsad_epu8() { |
18420 | let a = _mm512_set1_epi8(2); |
18421 | let b = _mm512_set1_epi8(4); |
18422 | let r = _mm512_dbsad_epu8::<0>(a, b); |
18423 | let e = _mm512_set1_epi16(8); |
18424 | assert_eq_m512i(r, e); |
18425 | } |
18426 | |
18427 | #[simd_test(enable = "avx512bw" )] |
18428 | unsafe fn test_mm512_mask_dbsad_epu8() { |
18429 | let src = _mm512_set1_epi16(1); |
18430 | let a = _mm512_set1_epi8(2); |
18431 | let b = _mm512_set1_epi8(4); |
18432 | let r = _mm512_mask_dbsad_epu8::<0>(src, 0, a, b); |
18433 | assert_eq_m512i(r, src); |
18434 | let r = _mm512_mask_dbsad_epu8::<0>(src, 0b11111111_11111111_11111111_11111111, a, b); |
18435 | let e = _mm512_set1_epi16(8); |
18436 | assert_eq_m512i(r, e); |
18437 | } |
18438 | |
18439 | #[simd_test(enable = "avx512bw" )] |
18440 | unsafe fn test_mm512_maskz_dbsad_epu8() { |
18441 | let a = _mm512_set1_epi8(2); |
18442 | let b = _mm512_set1_epi8(4); |
18443 | let r = _mm512_maskz_dbsad_epu8::<0>(0, a, b); |
18444 | assert_eq_m512i(r, _mm512_setzero_si512()); |
18445 | let r = _mm512_maskz_dbsad_epu8::<0>(0b11111111_11111111_11111111_11111111, a, b); |
18446 | let e = _mm512_set1_epi16(8); |
18447 | assert_eq_m512i(r, e); |
18448 | } |
18449 | |
18450 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18451 | unsafe fn test_mm256_dbsad_epu8() { |
18452 | let a = _mm256_set1_epi8(2); |
18453 | let b = _mm256_set1_epi8(4); |
18454 | let r = _mm256_dbsad_epu8::<0>(a, b); |
18455 | let e = _mm256_set1_epi16(8); |
18456 | assert_eq_m256i(r, e); |
18457 | } |
18458 | |
18459 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18460 | unsafe fn test_mm256_mask_dbsad_epu8() { |
18461 | let src = _mm256_set1_epi16(1); |
18462 | let a = _mm256_set1_epi8(2); |
18463 | let b = _mm256_set1_epi8(4); |
18464 | let r = _mm256_mask_dbsad_epu8::<0>(src, 0, a, b); |
18465 | assert_eq_m256i(r, src); |
18466 | let r = _mm256_mask_dbsad_epu8::<0>(src, 0b11111111_11111111, a, b); |
18467 | let e = _mm256_set1_epi16(8); |
18468 | assert_eq_m256i(r, e); |
18469 | } |
18470 | |
18471 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18472 | unsafe fn test_mm256_maskz_dbsad_epu8() { |
18473 | let a = _mm256_set1_epi8(2); |
18474 | let b = _mm256_set1_epi8(4); |
18475 | let r = _mm256_maskz_dbsad_epu8::<0>(0, a, b); |
18476 | assert_eq_m256i(r, _mm256_setzero_si256()); |
18477 | let r = _mm256_maskz_dbsad_epu8::<0>(0b11111111_11111111, a, b); |
18478 | let e = _mm256_set1_epi16(8); |
18479 | assert_eq_m256i(r, e); |
18480 | } |
18481 | |
18482 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18483 | unsafe fn test_mm_dbsad_epu8() { |
18484 | let a = _mm_set1_epi8(2); |
18485 | let b = _mm_set1_epi8(4); |
18486 | let r = _mm_dbsad_epu8::<0>(a, b); |
18487 | let e = _mm_set1_epi16(8); |
18488 | assert_eq_m128i(r, e); |
18489 | } |
18490 | |
18491 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18492 | unsafe fn test_mm_mask_dbsad_epu8() { |
18493 | let src = _mm_set1_epi16(1); |
18494 | let a = _mm_set1_epi8(2); |
18495 | let b = _mm_set1_epi8(4); |
18496 | let r = _mm_mask_dbsad_epu8::<0>(src, 0, a, b); |
18497 | assert_eq_m128i(r, src); |
18498 | let r = _mm_mask_dbsad_epu8::<0>(src, 0b11111111, a, b); |
18499 | let e = _mm_set1_epi16(8); |
18500 | assert_eq_m128i(r, e); |
18501 | } |
18502 | |
18503 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18504 | unsafe fn test_mm_maskz_dbsad_epu8() { |
18505 | let a = _mm_set1_epi8(2); |
18506 | let b = _mm_set1_epi8(4); |
18507 | let r = _mm_maskz_dbsad_epu8::<0>(0, a, b); |
18508 | assert_eq_m128i(r, _mm_setzero_si128()); |
18509 | let r = _mm_maskz_dbsad_epu8::<0>(0b11111111, a, b); |
18510 | let e = _mm_set1_epi16(8); |
18511 | assert_eq_m128i(r, e); |
18512 | } |
18513 | |
18514 | #[simd_test(enable = "avx512bw" )] |
18515 | unsafe fn test_mm512_movepi16_mask() { |
18516 | let a = _mm512_set1_epi16(1 << 15); |
18517 | let r = _mm512_movepi16_mask(a); |
18518 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
18519 | assert_eq!(r, e); |
18520 | } |
18521 | |
18522 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18523 | unsafe fn test_mm256_movepi16_mask() { |
18524 | let a = _mm256_set1_epi16(1 << 15); |
18525 | let r = _mm256_movepi16_mask(a); |
18526 | let e: __mmask16 = 0b11111111_11111111; |
18527 | assert_eq!(r, e); |
18528 | } |
18529 | |
18530 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18531 | unsafe fn test_mm_movepi16_mask() { |
18532 | let a = _mm_set1_epi16(1 << 15); |
18533 | let r = _mm_movepi16_mask(a); |
18534 | let e: __mmask8 = 0b11111111; |
18535 | assert_eq!(r, e); |
18536 | } |
18537 | |
18538 | #[simd_test(enable = "avx512bw" )] |
18539 | unsafe fn test_mm512_movepi8_mask() { |
18540 | let a = _mm512_set1_epi8(1 << 7); |
18541 | let r = _mm512_movepi8_mask(a); |
18542 | let e: __mmask64 = |
18543 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111; |
18544 | assert_eq!(r, e); |
18545 | } |
18546 | |
18547 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18548 | unsafe fn test_mm256_movepi8_mask() { |
18549 | let a = _mm256_set1_epi8(1 << 7); |
18550 | let r = _mm256_movepi8_mask(a); |
18551 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
18552 | assert_eq!(r, e); |
18553 | } |
18554 | |
18555 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18556 | unsafe fn test_mm_movepi8_mask() { |
18557 | let a = _mm_set1_epi8(1 << 7); |
18558 | let r = _mm_movepi8_mask(a); |
18559 | let e: __mmask16 = 0b11111111_11111111; |
18560 | assert_eq!(r, e); |
18561 | } |
18562 | |
18563 | #[simd_test(enable = "avx512bw" )] |
18564 | unsafe fn test_mm512_movm_epi16() { |
18565 | let a: __mmask32 = 0b11111111_11111111_11111111_11111111; |
18566 | let r = _mm512_movm_epi16(a); |
18567 | let e = _mm512_set1_epi16( |
18568 | 1 << 15 |
18569 | | 1 << 14 |
18570 | | 1 << 13 |
18571 | | 1 << 12 |
18572 | | 1 << 11 |
18573 | | 1 << 10 |
18574 | | 1 << 9 |
18575 | | 1 << 8 |
18576 | | 1 << 7 |
18577 | | 1 << 6 |
18578 | | 1 << 5 |
18579 | | 1 << 4 |
18580 | | 1 << 3 |
18581 | | 1 << 2 |
18582 | | 1 << 1 |
18583 | | 1 << 0, |
18584 | ); |
18585 | assert_eq_m512i(r, e); |
18586 | } |
18587 | |
18588 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18589 | unsafe fn test_mm256_movm_epi16() { |
18590 | let a: __mmask16 = 0b11111111_11111111; |
18591 | let r = _mm256_movm_epi16(a); |
18592 | let e = _mm256_set1_epi16( |
18593 | 1 << 15 |
18594 | | 1 << 14 |
18595 | | 1 << 13 |
18596 | | 1 << 12 |
18597 | | 1 << 11 |
18598 | | 1 << 10 |
18599 | | 1 << 9 |
18600 | | 1 << 8 |
18601 | | 1 << 7 |
18602 | | 1 << 6 |
18603 | | 1 << 5 |
18604 | | 1 << 4 |
18605 | | 1 << 3 |
18606 | | 1 << 2 |
18607 | | 1 << 1 |
18608 | | 1 << 0, |
18609 | ); |
18610 | assert_eq_m256i(r, e); |
18611 | } |
18612 | |
18613 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18614 | unsafe fn test_mm_movm_epi16() { |
18615 | let a: __mmask8 = 0b11111111; |
18616 | let r = _mm_movm_epi16(a); |
18617 | let e = _mm_set1_epi16( |
18618 | 1 << 15 |
18619 | | 1 << 14 |
18620 | | 1 << 13 |
18621 | | 1 << 12 |
18622 | | 1 << 11 |
18623 | | 1 << 10 |
18624 | | 1 << 9 |
18625 | | 1 << 8 |
18626 | | 1 << 7 |
18627 | | 1 << 6 |
18628 | | 1 << 5 |
18629 | | 1 << 4 |
18630 | | 1 << 3 |
18631 | | 1 << 2 |
18632 | | 1 << 1 |
18633 | | 1 << 0, |
18634 | ); |
18635 | assert_eq_m128i(r, e); |
18636 | } |
18637 | |
18638 | #[simd_test(enable = "avx512bw" )] |
18639 | unsafe fn test_mm512_movm_epi8() { |
18640 | let a: __mmask64 = |
18641 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111; |
18642 | let r = _mm512_movm_epi8(a); |
18643 | let e = |
18644 | _mm512_set1_epi8(1 << 7 | 1 << 6 | 1 << 5 | 1 << 4 | 1 << 3 | 1 << 2 | 1 << 1 | 1 << 0); |
18645 | assert_eq_m512i(r, e); |
18646 | } |
18647 | |
18648 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18649 | unsafe fn test_mm256_movm_epi8() { |
18650 | let a: __mmask32 = 0b11111111_11111111_11111111_11111111; |
18651 | let r = _mm256_movm_epi8(a); |
18652 | let e = |
18653 | _mm256_set1_epi8(1 << 7 | 1 << 6 | 1 << 5 | 1 << 4 | 1 << 3 | 1 << 2 | 1 << 1 | 1 << 0); |
18654 | assert_eq_m256i(r, e); |
18655 | } |
18656 | |
18657 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18658 | unsafe fn test_mm_movm_epi8() { |
18659 | let a: __mmask16 = 0b11111111_11111111; |
18660 | let r = _mm_movm_epi8(a); |
18661 | let e = |
18662 | _mm_set1_epi8(1 << 7 | 1 << 6 | 1 << 5 | 1 << 4 | 1 << 3 | 1 << 2 | 1 << 1 | 1 << 0); |
18663 | assert_eq_m128i(r, e); |
18664 | } |
18665 | |
18666 | #[simd_test(enable = "avx512bw" )] |
18667 | unsafe fn test_kadd_mask32() { |
18668 | let a: __mmask32 = 11; |
18669 | let b: __mmask32 = 22; |
18670 | let r = _kadd_mask32(a, b); |
18671 | let e: __mmask32 = 33; |
18672 | assert_eq!(r, e); |
18673 | } |
18674 | |
18675 | #[simd_test(enable = "avx512bw" )] |
18676 | unsafe fn test_kadd_mask64() { |
18677 | let a: __mmask64 = 11; |
18678 | let b: __mmask64 = 22; |
18679 | let r = _kadd_mask64(a, b); |
18680 | let e: __mmask64 = 33; |
18681 | assert_eq!(r, e); |
18682 | } |
18683 | |
18684 | #[simd_test(enable = "avx512bw" )] |
18685 | unsafe fn test_kand_mask32() { |
18686 | let a: __mmask32 = 0b11001100_00110011_11001100_00110011; |
18687 | let b: __mmask32 = 0b11001100_00110011_11001100_00110011; |
18688 | let r = _kand_mask32(a, b); |
18689 | let e: __mmask32 = 0b11001100_00110011_11001100_00110011; |
18690 | assert_eq!(r, e); |
18691 | } |
18692 | |
18693 | #[simd_test(enable = "avx512bw" )] |
18694 | unsafe fn test_kand_mask64() { |
18695 | let a: __mmask64 = |
18696 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
18697 | let b: __mmask64 = |
18698 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
18699 | let r = _kand_mask64(a, b); |
18700 | let e: __mmask64 = |
18701 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
18702 | assert_eq!(r, e); |
18703 | } |
18704 | |
18705 | #[simd_test(enable = "avx512bw" )] |
18706 | unsafe fn test_knot_mask32() { |
18707 | let a: __mmask32 = 0b11001100_00110011_11001100_00110011; |
18708 | let r = _knot_mask32(a); |
18709 | let e: __mmask32 = 0b00110011_11001100_00110011_11001100; |
18710 | assert_eq!(r, e); |
18711 | } |
18712 | |
18713 | #[simd_test(enable = "avx512bw" )] |
18714 | unsafe fn test_knot_mask64() { |
18715 | let a: __mmask64 = |
18716 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
18717 | let r = _knot_mask64(a); |
18718 | let e: __mmask64 = |
18719 | 0b00110011_11001100_00110011_11001100_00110011_11001100_00110011_11001100; |
18720 | assert_eq!(r, e); |
18721 | } |
18722 | |
18723 | #[simd_test(enable = "avx512bw" )] |
18724 | unsafe fn test_kandn_mask32() { |
18725 | let a: __mmask32 = 0b11001100_00110011_11001100_00110011; |
18726 | let b: __mmask32 = 0b11001100_00110011_11001100_00110011; |
18727 | let r = _kandn_mask32(a, b); |
18728 | let e: __mmask32 = 0b00000000_00000000_00000000_00000000; |
18729 | assert_eq!(r, e); |
18730 | } |
18731 | |
18732 | #[simd_test(enable = "avx512bw" )] |
18733 | unsafe fn test_kandn_mask64() { |
18734 | let a: __mmask64 = |
18735 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
18736 | let b: __mmask64 = |
18737 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
18738 | let r = _kandn_mask64(a, b); |
18739 | let e: __mmask64 = |
18740 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00000000; |
18741 | assert_eq!(r, e); |
18742 | } |
18743 | |
18744 | #[simd_test(enable = "avx512bw" )] |
18745 | unsafe fn test_kor_mask32() { |
18746 | let a: __mmask32 = 0b00110011_11001100_00110011_11001100; |
18747 | let b: __mmask32 = 0b11001100_00110011_11001100_00110011; |
18748 | let r = _kor_mask32(a, b); |
18749 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
18750 | assert_eq!(r, e); |
18751 | } |
18752 | |
18753 | #[simd_test(enable = "avx512bw" )] |
18754 | unsafe fn test_kor_mask64() { |
18755 | let a: __mmask64 = |
18756 | 0b00110011_11001100_00110011_11001100_00110011_11001100_00110011_11001100; |
18757 | let b: __mmask64 = |
18758 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
18759 | let r = _kor_mask64(a, b); |
18760 | let e: __mmask64 = |
18761 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111; |
18762 | assert_eq!(r, e); |
18763 | } |
18764 | |
18765 | #[simd_test(enable = "avx512bw" )] |
18766 | unsafe fn test_kxor_mask32() { |
18767 | let a: __mmask32 = 0b00110011_11001100_00110011_11001100; |
18768 | let b: __mmask32 = 0b11001100_00110011_11001100_00110011; |
18769 | let r = _kxor_mask32(a, b); |
18770 | let e: __mmask32 = 0b11111111_11111111_11111111_11111111; |
18771 | assert_eq!(r, e); |
18772 | } |
18773 | |
18774 | #[simd_test(enable = "avx512bw" )] |
18775 | unsafe fn test_kxor_mask64() { |
18776 | let a: __mmask64 = |
18777 | 0b00110011_11001100_00110011_11001100_00110011_11001100_00110011_11001100; |
18778 | let b: __mmask64 = |
18779 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
18780 | let r = _kxor_mask64(a, b); |
18781 | let e: __mmask64 = |
18782 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111; |
18783 | assert_eq!(r, e); |
18784 | } |
18785 | |
18786 | #[simd_test(enable = "avx512bw" )] |
18787 | unsafe fn test_kxnor_mask32() { |
18788 | let a: __mmask32 = 0b00110011_11001100_00110011_11001100; |
18789 | let b: __mmask32 = 0b11001100_00110011_11001100_00110011; |
18790 | let r = _kxnor_mask32(a, b); |
18791 | let e: __mmask32 = 0b00000000_00000000_00000000_00000000; |
18792 | assert_eq!(r, e); |
18793 | } |
18794 | |
18795 | #[simd_test(enable = "avx512bw" )] |
18796 | unsafe fn test_kxnor_mask64() { |
18797 | let a: __mmask64 = |
18798 | 0b00110011_11001100_00110011_11001100_00110011_11001100_00110011_11001100; |
18799 | let b: __mmask64 = |
18800 | 0b11001100_00110011_11001100_00110011_11001100_00110011_11001100_00110011; |
18801 | let r = _kxnor_mask64(a, b); |
18802 | let e: __mmask64 = |
18803 | 0b00000000_00000000_00000000_00000000_00000000_00000000_00000000_00000000; |
18804 | assert_eq!(r, e); |
18805 | } |
18806 | |
18807 | #[simd_test(enable = "avx512bw" )] |
18808 | unsafe fn test_mm512_cvtepi16_epi8() { |
18809 | let a = _mm512_set1_epi16(2); |
18810 | let r = _mm512_cvtepi16_epi8(a); |
18811 | let e = _mm256_set1_epi8(2); |
18812 | assert_eq_m256i(r, e); |
18813 | } |
18814 | |
18815 | #[simd_test(enable = "avx512bw" )] |
18816 | unsafe fn test_mm512_mask_cvtepi16_epi8() { |
18817 | let src = _mm256_set1_epi8(1); |
18818 | let a = _mm512_set1_epi16(2); |
18819 | let r = _mm512_mask_cvtepi16_epi8(src, 0, a); |
18820 | assert_eq_m256i(r, src); |
18821 | let r = _mm512_mask_cvtepi16_epi8(src, 0b11111111_11111111_11111111_11111111, a); |
18822 | let e = _mm256_set1_epi8(2); |
18823 | assert_eq_m256i(r, e); |
18824 | } |
18825 | |
18826 | #[simd_test(enable = "avx512bw" )] |
18827 | unsafe fn test_mm512_maskz_cvtepi16_epi8() { |
18828 | let a = _mm512_set1_epi16(2); |
18829 | let r = _mm512_maskz_cvtepi16_epi8(0, a); |
18830 | assert_eq_m256i(r, _mm256_setzero_si256()); |
18831 | let r = _mm512_maskz_cvtepi16_epi8(0b11111111_11111111_11111111_11111111, a); |
18832 | let e = _mm256_set1_epi8(2); |
18833 | assert_eq_m256i(r, e); |
18834 | } |
18835 | |
18836 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18837 | unsafe fn test_mm256_cvtepi16_epi8() { |
18838 | let a = _mm256_set1_epi16(2); |
18839 | let r = _mm256_cvtepi16_epi8(a); |
18840 | let e = _mm_set1_epi8(2); |
18841 | assert_eq_m128i(r, e); |
18842 | } |
18843 | |
18844 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18845 | unsafe fn test_mm256_mask_cvtepi16_epi8() { |
18846 | let src = _mm_set1_epi8(1); |
18847 | let a = _mm256_set1_epi16(2); |
18848 | let r = _mm256_mask_cvtepi16_epi8(src, 0, a); |
18849 | assert_eq_m128i(r, src); |
18850 | let r = _mm256_mask_cvtepi16_epi8(src, 0b11111111_11111111, a); |
18851 | let e = _mm_set1_epi8(2); |
18852 | assert_eq_m128i(r, e); |
18853 | } |
18854 | |
18855 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18856 | unsafe fn test_mm256_maskz_cvtepi16_epi8() { |
18857 | let a = _mm256_set1_epi16(2); |
18858 | let r = _mm256_maskz_cvtepi16_epi8(0, a); |
18859 | assert_eq_m128i(r, _mm_setzero_si128()); |
18860 | let r = _mm256_maskz_cvtepi16_epi8(0b11111111_11111111, a); |
18861 | let e = _mm_set1_epi8(2); |
18862 | assert_eq_m128i(r, e); |
18863 | } |
18864 | |
18865 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18866 | unsafe fn test_mm_cvtepi16_epi8() { |
18867 | let a = _mm_set1_epi16(2); |
18868 | let r = _mm_cvtepi16_epi8(a); |
18869 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2); |
18870 | assert_eq_m128i(r, e); |
18871 | } |
18872 | |
18873 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18874 | unsafe fn test_mm_mask_cvtepi16_epi8() { |
18875 | let src = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
18876 | let a = _mm_set1_epi16(2); |
18877 | let r = _mm_mask_cvtepi16_epi8(src, 0, a); |
18878 | assert_eq_m128i(r, src); |
18879 | let r = _mm_mask_cvtepi16_epi8(src, 0b11111111, a); |
18880 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2); |
18881 | assert_eq_m128i(r, e); |
18882 | } |
18883 | |
18884 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18885 | unsafe fn test_mm_maskz_cvtepi16_epi8() { |
18886 | let a = _mm_set1_epi16(2); |
18887 | let r = _mm_maskz_cvtepi16_epi8(0, a); |
18888 | assert_eq_m128i(r, _mm_setzero_si128()); |
18889 | let r = _mm_maskz_cvtepi16_epi8(0b11111111, a); |
18890 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2); |
18891 | assert_eq_m128i(r, e); |
18892 | } |
18893 | |
18894 | #[simd_test(enable = "avx512bw" )] |
18895 | unsafe fn test_mm512_cvtsepi16_epi8() { |
18896 | let a = _mm512_set1_epi16(i16::MAX); |
18897 | let r = _mm512_cvtsepi16_epi8(a); |
18898 | let e = _mm256_set1_epi8(i8::MAX); |
18899 | assert_eq_m256i(r, e); |
18900 | } |
18901 | |
18902 | #[simd_test(enable = "avx512bw" )] |
18903 | unsafe fn test_mm512_mask_cvtsepi16_epi8() { |
18904 | let src = _mm256_set1_epi8(1); |
18905 | let a = _mm512_set1_epi16(i16::MAX); |
18906 | let r = _mm512_mask_cvtsepi16_epi8(src, 0, a); |
18907 | assert_eq_m256i(r, src); |
18908 | let r = _mm512_mask_cvtsepi16_epi8(src, 0b11111111_11111111_11111111_11111111, a); |
18909 | let e = _mm256_set1_epi8(i8::MAX); |
18910 | assert_eq_m256i(r, e); |
18911 | } |
18912 | |
18913 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18914 | unsafe fn test_mm256_cvtsepi16_epi8() { |
18915 | let a = _mm256_set1_epi16(i16::MAX); |
18916 | let r = _mm256_cvtsepi16_epi8(a); |
18917 | let e = _mm_set1_epi8(i8::MAX); |
18918 | assert_eq_m128i(r, e); |
18919 | } |
18920 | |
18921 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18922 | unsafe fn test_mm256_mask_cvtsepi16_epi8() { |
18923 | let src = _mm_set1_epi8(1); |
18924 | let a = _mm256_set1_epi16(i16::MAX); |
18925 | let r = _mm256_mask_cvtsepi16_epi8(src, 0, a); |
18926 | assert_eq_m128i(r, src); |
18927 | let r = _mm256_mask_cvtsepi16_epi8(src, 0b11111111_11111111, a); |
18928 | let e = _mm_set1_epi8(i8::MAX); |
18929 | assert_eq_m128i(r, e); |
18930 | } |
18931 | |
18932 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18933 | unsafe fn test_mm256_maskz_cvtsepi16_epi8() { |
18934 | let a = _mm256_set1_epi16(i16::MAX); |
18935 | let r = _mm256_maskz_cvtsepi16_epi8(0, a); |
18936 | assert_eq_m128i(r, _mm_setzero_si128()); |
18937 | let r = _mm256_maskz_cvtsepi16_epi8(0b11111111_11111111, a); |
18938 | let e = _mm_set1_epi8(i8::MAX); |
18939 | assert_eq_m128i(r, e); |
18940 | } |
18941 | |
18942 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18943 | unsafe fn test_mm_cvtsepi16_epi8() { |
18944 | let a = _mm_set1_epi16(i16::MAX); |
18945 | let r = _mm_cvtsepi16_epi8(a); |
18946 | #[rustfmt::skip] |
18947 | 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); |
18948 | assert_eq_m128i(r, e); |
18949 | } |
18950 | |
18951 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18952 | unsafe fn test_mm_mask_cvtsepi16_epi8() { |
18953 | let src = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
18954 | let a = _mm_set1_epi16(i16::MAX); |
18955 | let r = _mm_mask_cvtsepi16_epi8(src, 0, a); |
18956 | assert_eq_m128i(r, src); |
18957 | let r = _mm_mask_cvtsepi16_epi8(src, 0b11111111, a); |
18958 | #[rustfmt::skip] |
18959 | 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); |
18960 | assert_eq_m128i(r, e); |
18961 | } |
18962 | |
18963 | #[simd_test(enable = "avx512bw,avx512vl" )] |
18964 | unsafe fn test_mm_maskz_cvtsepi16_epi8() { |
18965 | let a = _mm_set1_epi16(i16::MAX); |
18966 | let r = _mm_maskz_cvtsepi16_epi8(0, a); |
18967 | assert_eq_m128i(r, _mm_setzero_si128()); |
18968 | let r = _mm_maskz_cvtsepi16_epi8(0b11111111, a); |
18969 | #[rustfmt::skip] |
18970 | 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); |
18971 | assert_eq_m128i(r, e); |
18972 | } |
18973 | |
18974 | #[simd_test(enable = "avx512bw" )] |
18975 | unsafe fn test_mm512_maskz_cvtsepi16_epi8() { |
18976 | let a = _mm512_set1_epi16(i16::MAX); |
18977 | let r = _mm512_maskz_cvtsepi16_epi8(0, a); |
18978 | assert_eq_m256i(r, _mm256_setzero_si256()); |
18979 | let r = _mm512_maskz_cvtsepi16_epi8(0b11111111_11111111_11111111_11111111, a); |
18980 | let e = _mm256_set1_epi8(i8::MAX); |
18981 | assert_eq_m256i(r, e); |
18982 | } |
18983 | |
18984 | #[simd_test(enable = "avx512bw" )] |
18985 | unsafe fn test_mm512_cvtusepi16_epi8() { |
18986 | let a = _mm512_set1_epi16(i16::MIN); |
18987 | let r = _mm512_cvtusepi16_epi8(a); |
18988 | let e = _mm256_set1_epi8(-1); |
18989 | assert_eq_m256i(r, e); |
18990 | } |
18991 | |
18992 | #[simd_test(enable = "avx512bw" )] |
18993 | unsafe fn test_mm512_mask_cvtusepi16_epi8() { |
18994 | let src = _mm256_set1_epi8(1); |
18995 | let a = _mm512_set1_epi16(i16::MIN); |
18996 | let r = _mm512_mask_cvtusepi16_epi8(src, 0, a); |
18997 | assert_eq_m256i(r, src); |
18998 | let r = _mm512_mask_cvtusepi16_epi8(src, 0b11111111_11111111_11111111_11111111, a); |
18999 | let e = _mm256_set1_epi8(-1); |
19000 | assert_eq_m256i(r, e); |
19001 | } |
19002 | |
19003 | #[simd_test(enable = "avx512bw" )] |
19004 | unsafe fn test_mm512_maskz_cvtusepi16_epi8() { |
19005 | let a = _mm512_set1_epi16(i16::MIN); |
19006 | let r = _mm512_maskz_cvtusepi16_epi8(0, a); |
19007 | assert_eq_m256i(r, _mm256_setzero_si256()); |
19008 | let r = _mm512_maskz_cvtusepi16_epi8(0b11111111_11111111_11111111_11111111, a); |
19009 | let e = _mm256_set1_epi8(-1); |
19010 | assert_eq_m256i(r, e); |
19011 | } |
19012 | |
19013 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19014 | unsafe fn test_mm256_cvtusepi16_epi8() { |
19015 | let a = _mm256_set1_epi16(i16::MIN); |
19016 | let r = _mm256_cvtusepi16_epi8(a); |
19017 | let e = _mm_set1_epi8(-1); |
19018 | assert_eq_m128i(r, e); |
19019 | } |
19020 | |
19021 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19022 | unsafe fn test_mm256_mask_cvtusepi16_epi8() { |
19023 | let src = _mm_set1_epi8(1); |
19024 | let a = _mm256_set1_epi16(i16::MIN); |
19025 | let r = _mm256_mask_cvtusepi16_epi8(src, 0, a); |
19026 | assert_eq_m128i(r, src); |
19027 | let r = _mm256_mask_cvtusepi16_epi8(src, 0b11111111_11111111, a); |
19028 | let e = _mm_set1_epi8(-1); |
19029 | assert_eq_m128i(r, e); |
19030 | } |
19031 | |
19032 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19033 | unsafe fn test_mm256_maskz_cvtusepi16_epi8() { |
19034 | let a = _mm256_set1_epi16(i16::MIN); |
19035 | let r = _mm256_maskz_cvtusepi16_epi8(0, a); |
19036 | assert_eq_m128i(r, _mm_setzero_si128()); |
19037 | let r = _mm256_maskz_cvtusepi16_epi8(0b11111111_11111111, a); |
19038 | let e = _mm_set1_epi8(-1); |
19039 | assert_eq_m128i(r, e); |
19040 | } |
19041 | |
19042 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19043 | unsafe fn test_mm_cvtusepi16_epi8() { |
19044 | let a = _mm_set1_epi16(i16::MIN); |
19045 | let r = _mm_cvtusepi16_epi8(a); |
19046 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
19047 | assert_eq_m128i(r, e); |
19048 | } |
19049 | |
19050 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19051 | unsafe fn test_mm_mask_cvtusepi16_epi8() { |
19052 | let src = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1); |
19053 | let a = _mm_set1_epi16(i16::MIN); |
19054 | let r = _mm_mask_cvtusepi16_epi8(src, 0, a); |
19055 | assert_eq_m128i(r, src); |
19056 | let r = _mm_mask_cvtusepi16_epi8(src, 0b11111111, a); |
19057 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
19058 | assert_eq_m128i(r, e); |
19059 | } |
19060 | |
19061 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19062 | unsafe fn test_mm_maskz_cvtusepi16_epi8() { |
19063 | let a = _mm_set1_epi16(i16::MIN); |
19064 | let r = _mm_maskz_cvtusepi16_epi8(0, a); |
19065 | assert_eq_m128i(r, _mm_setzero_si128()); |
19066 | let r = _mm_maskz_cvtusepi16_epi8(0b11111111, a); |
19067 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
19068 | assert_eq_m128i(r, e); |
19069 | } |
19070 | |
19071 | #[simd_test(enable = "avx512bw" )] |
19072 | unsafe fn test_mm512_cvtepi8_epi16() { |
19073 | let a = _mm256_set1_epi8(2); |
19074 | let r = _mm512_cvtepi8_epi16(a); |
19075 | let e = _mm512_set1_epi16(2); |
19076 | assert_eq_m512i(r, e); |
19077 | } |
19078 | |
19079 | #[simd_test(enable = "avx512bw" )] |
19080 | unsafe fn test_mm512_mask_cvtepi8_epi16() { |
19081 | let src = _mm512_set1_epi16(1); |
19082 | let a = _mm256_set1_epi8(2); |
19083 | let r = _mm512_mask_cvtepi8_epi16(src, 0, a); |
19084 | assert_eq_m512i(r, src); |
19085 | let r = _mm512_mask_cvtepi8_epi16(src, 0b11111111_11111111_11111111_11111111, a); |
19086 | let e = _mm512_set1_epi16(2); |
19087 | assert_eq_m512i(r, e); |
19088 | } |
19089 | |
19090 | #[simd_test(enable = "avx512bw" )] |
19091 | unsafe fn test_mm512_maskz_cvtepi8_epi16() { |
19092 | let a = _mm256_set1_epi8(2); |
19093 | let r = _mm512_maskz_cvtepi8_epi16(0, a); |
19094 | assert_eq_m512i(r, _mm512_setzero_si512()); |
19095 | let r = _mm512_maskz_cvtepi8_epi16(0b11111111_11111111_11111111_11111111, a); |
19096 | let e = _mm512_set1_epi16(2); |
19097 | assert_eq_m512i(r, e); |
19098 | } |
19099 | |
19100 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19101 | unsafe fn test_mm256_mask_cvtepi8_epi16() { |
19102 | let src = _mm256_set1_epi16(1); |
19103 | let a = _mm_set1_epi8(2); |
19104 | let r = _mm256_mask_cvtepi8_epi16(src, 0, a); |
19105 | assert_eq_m256i(r, src); |
19106 | let r = _mm256_mask_cvtepi8_epi16(src, 0b11111111_11111111, a); |
19107 | let e = _mm256_set1_epi16(2); |
19108 | assert_eq_m256i(r, e); |
19109 | } |
19110 | |
19111 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19112 | unsafe fn test_mm256_maskz_cvtepi8_epi16() { |
19113 | let a = _mm_set1_epi8(2); |
19114 | let r = _mm256_maskz_cvtepi8_epi16(0, a); |
19115 | assert_eq_m256i(r, _mm256_setzero_si256()); |
19116 | let r = _mm256_maskz_cvtepi8_epi16(0b11111111_11111111, a); |
19117 | let e = _mm256_set1_epi16(2); |
19118 | assert_eq_m256i(r, e); |
19119 | } |
19120 | |
19121 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19122 | unsafe fn test_mm_mask_cvtepi8_epi16() { |
19123 | let src = _mm_set1_epi16(1); |
19124 | let a = _mm_set1_epi8(2); |
19125 | let r = _mm_mask_cvtepi8_epi16(src, 0, a); |
19126 | assert_eq_m128i(r, src); |
19127 | let r = _mm_mask_cvtepi8_epi16(src, 0b11111111, a); |
19128 | let e = _mm_set1_epi16(2); |
19129 | assert_eq_m128i(r, e); |
19130 | } |
19131 | |
19132 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19133 | unsafe fn test_mm_maskz_cvtepi8_epi16() { |
19134 | let a = _mm_set1_epi8(2); |
19135 | let r = _mm_maskz_cvtepi8_epi16(0, a); |
19136 | assert_eq_m128i(r, _mm_setzero_si128()); |
19137 | let r = _mm_maskz_cvtepi8_epi16(0b11111111, a); |
19138 | let e = _mm_set1_epi16(2); |
19139 | assert_eq_m128i(r, e); |
19140 | } |
19141 | |
19142 | #[simd_test(enable = "avx512bw" )] |
19143 | unsafe fn test_mm512_cvtepu8_epi16() { |
19144 | let a = _mm256_set1_epi8(2); |
19145 | let r = _mm512_cvtepu8_epi16(a); |
19146 | let e = _mm512_set1_epi16(2); |
19147 | assert_eq_m512i(r, e); |
19148 | } |
19149 | |
19150 | #[simd_test(enable = "avx512bw" )] |
19151 | unsafe fn test_mm512_mask_cvtepu8_epi16() { |
19152 | let src = _mm512_set1_epi16(1); |
19153 | let a = _mm256_set1_epi8(2); |
19154 | let r = _mm512_mask_cvtepu8_epi16(src, 0, a); |
19155 | assert_eq_m512i(r, src); |
19156 | let r = _mm512_mask_cvtepu8_epi16(src, 0b11111111_11111111_11111111_11111111, a); |
19157 | let e = _mm512_set1_epi16(2); |
19158 | assert_eq_m512i(r, e); |
19159 | } |
19160 | |
19161 | #[simd_test(enable = "avx512bw" )] |
19162 | unsafe fn test_mm512_maskz_cvtepu8_epi16() { |
19163 | let a = _mm256_set1_epi8(2); |
19164 | let r = _mm512_maskz_cvtepu8_epi16(0, a); |
19165 | assert_eq_m512i(r, _mm512_setzero_si512()); |
19166 | let r = _mm512_maskz_cvtepu8_epi16(0b11111111_11111111_11111111_11111111, a); |
19167 | let e = _mm512_set1_epi16(2); |
19168 | assert_eq_m512i(r, e); |
19169 | } |
19170 | |
19171 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19172 | unsafe fn test_mm256_mask_cvtepu8_epi16() { |
19173 | let src = _mm256_set1_epi16(1); |
19174 | let a = _mm_set1_epi8(2); |
19175 | let r = _mm256_mask_cvtepu8_epi16(src, 0, a); |
19176 | assert_eq_m256i(r, src); |
19177 | let r = _mm256_mask_cvtepu8_epi16(src, 0b11111111_11111111, a); |
19178 | let e = _mm256_set1_epi16(2); |
19179 | assert_eq_m256i(r, e); |
19180 | } |
19181 | |
19182 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19183 | unsafe fn test_mm256_maskz_cvtepu8_epi16() { |
19184 | let a = _mm_set1_epi8(2); |
19185 | let r = _mm256_maskz_cvtepu8_epi16(0, a); |
19186 | assert_eq_m256i(r, _mm256_setzero_si256()); |
19187 | let r = _mm256_maskz_cvtepu8_epi16(0b11111111_11111111, a); |
19188 | let e = _mm256_set1_epi16(2); |
19189 | assert_eq_m256i(r, e); |
19190 | } |
19191 | |
19192 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19193 | unsafe fn test_mm_mask_cvtepu8_epi16() { |
19194 | let src = _mm_set1_epi16(1); |
19195 | let a = _mm_set1_epi8(2); |
19196 | let r = _mm_mask_cvtepu8_epi16(src, 0, a); |
19197 | assert_eq_m128i(r, src); |
19198 | let r = _mm_mask_cvtepu8_epi16(src, 0b11111111, a); |
19199 | let e = _mm_set1_epi16(2); |
19200 | assert_eq_m128i(r, e); |
19201 | } |
19202 | |
19203 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19204 | unsafe fn test_mm_maskz_cvtepu8_epi16() { |
19205 | let a = _mm_set1_epi8(2); |
19206 | let r = _mm_maskz_cvtepu8_epi16(0, a); |
19207 | assert_eq_m128i(r, _mm_setzero_si128()); |
19208 | let r = _mm_maskz_cvtepu8_epi16(0b11111111, a); |
19209 | let e = _mm_set1_epi16(2); |
19210 | assert_eq_m128i(r, e); |
19211 | } |
19212 | |
19213 | #[simd_test(enable = "avx512bw" )] |
19214 | unsafe fn test_mm512_bslli_epi128() { |
19215 | #[rustfmt::skip] |
19216 | let a = _mm512_set_epi8( |
19217 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19218 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19219 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19220 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19221 | ); |
19222 | let r = _mm512_bslli_epi128::<9>(a); |
19223 | #[rustfmt::skip] |
19224 | let e = _mm512_set_epi8( |
19225 | 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
19226 | 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
19227 | 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
19228 | 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, |
19229 | ); |
19230 | assert_eq_m512i(r, e); |
19231 | } |
19232 | |
19233 | #[simd_test(enable = "avx512bw" )] |
19234 | unsafe fn test_mm512_bsrli_epi128() { |
19235 | #[rustfmt::skip] |
19236 | let a = _mm512_set_epi8( |
19237 | 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, |
19238 | 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, |
19239 | 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, |
19240 | 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, |
19241 | ); |
19242 | let r = _mm512_bsrli_epi128::<3>(a); |
19243 | #[rustfmt::skip] |
19244 | let e = _mm512_set_epi8( |
19245 | 0, 0, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, |
19246 | 0, 0, 0, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, |
19247 | 0, 0, 0, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, |
19248 | 0, 0, 0, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, |
19249 | ); |
19250 | assert_eq_m512i(r, e); |
19251 | } |
19252 | |
19253 | #[simd_test(enable = "avx512bw" )] |
19254 | unsafe fn test_mm512_alignr_epi8() { |
19255 | #[rustfmt::skip] |
19256 | let a = _mm512_set_epi8( |
19257 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19258 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19259 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19260 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19261 | ); |
19262 | let b = _mm512_set1_epi8(1); |
19263 | let r = _mm512_alignr_epi8::<14>(a, b); |
19264 | #[rustfmt::skip] |
19265 | let e = _mm512_set_epi8( |
19266 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19267 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19268 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19269 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19270 | ); |
19271 | assert_eq_m512i(r, e); |
19272 | } |
19273 | |
19274 | #[simd_test(enable = "avx512bw" )] |
19275 | unsafe fn test_mm512_mask_alignr_epi8() { |
19276 | #[rustfmt::skip] |
19277 | let a = _mm512_set_epi8( |
19278 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19279 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19280 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19281 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19282 | ); |
19283 | let b = _mm512_set1_epi8(1); |
19284 | let r = _mm512_mask_alignr_epi8::<14>(a, 0, a, b); |
19285 | assert_eq_m512i(r, a); |
19286 | let r = _mm512_mask_alignr_epi8::<14>( |
19287 | a, |
19288 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
19289 | a, |
19290 | b, |
19291 | ); |
19292 | #[rustfmt::skip] |
19293 | let e = _mm512_set_epi8( |
19294 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19295 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19296 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19297 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19298 | ); |
19299 | assert_eq_m512i(r, e); |
19300 | } |
19301 | |
19302 | #[simd_test(enable = "avx512bw" )] |
19303 | unsafe fn test_mm512_maskz_alignr_epi8() { |
19304 | #[rustfmt::skip] |
19305 | let a = _mm512_set_epi8( |
19306 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19307 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19308 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19309 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19310 | ); |
19311 | let b = _mm512_set1_epi8(1); |
19312 | let r = _mm512_maskz_alignr_epi8::<14>(0, a, b); |
19313 | assert_eq_m512i(r, _mm512_setzero_si512()); |
19314 | let r = _mm512_maskz_alignr_epi8::<14>( |
19315 | 0b11111111_11111111_11111111_11111111_11111111_11111111_11111111_11111111, |
19316 | a, |
19317 | b, |
19318 | ); |
19319 | #[rustfmt::skip] |
19320 | let e = _mm512_set_epi8( |
19321 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19322 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19323 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19324 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19325 | ); |
19326 | assert_eq_m512i(r, e); |
19327 | } |
19328 | |
19329 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19330 | unsafe fn test_mm256_mask_alignr_epi8() { |
19331 | #[rustfmt::skip] |
19332 | let a = _mm256_set_epi8( |
19333 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19334 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19335 | ); |
19336 | let b = _mm256_set1_epi8(1); |
19337 | let r = _mm256_mask_alignr_epi8::<14>(a, 0, a, b); |
19338 | assert_eq_m256i(r, a); |
19339 | let r = _mm256_mask_alignr_epi8::<14>(a, 0b11111111_11111111_11111111_11111111, a, b); |
19340 | #[rustfmt::skip] |
19341 | let e = _mm256_set_epi8( |
19342 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19343 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19344 | ); |
19345 | assert_eq_m256i(r, e); |
19346 | } |
19347 | |
19348 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19349 | unsafe fn test_mm256_maskz_alignr_epi8() { |
19350 | #[rustfmt::skip] |
19351 | let a = _mm256_set_epi8( |
19352 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19353 | 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, |
19354 | ); |
19355 | let b = _mm256_set1_epi8(1); |
19356 | let r = _mm256_maskz_alignr_epi8::<14>(0, a, b); |
19357 | assert_eq_m256i(r, _mm256_setzero_si256()); |
19358 | let r = _mm256_maskz_alignr_epi8::<14>(0b11111111_11111111_11111111_11111111, a, b); |
19359 | #[rustfmt::skip] |
19360 | let e = _mm256_set_epi8( |
19361 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19362 | 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, |
19363 | ); |
19364 | assert_eq_m256i(r, e); |
19365 | } |
19366 | |
19367 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19368 | unsafe fn test_mm_mask_alignr_epi8() { |
19369 | let a = _mm_set_epi8(1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0); |
19370 | let b = _mm_set1_epi8(1); |
19371 | let r = _mm_mask_alignr_epi8::<14>(a, 0, a, b); |
19372 | assert_eq_m128i(r, a); |
19373 | let r = _mm_mask_alignr_epi8::<14>(a, 0b11111111_11111111, a, b); |
19374 | let e = _mm_set_epi8(0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1); |
19375 | assert_eq_m128i(r, e); |
19376 | } |
19377 | |
19378 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19379 | unsafe fn test_mm_maskz_alignr_epi8() { |
19380 | let a = _mm_set_epi8(1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0); |
19381 | let b = _mm_set1_epi8(1); |
19382 | let r = _mm_maskz_alignr_epi8::<14>(0, a, b); |
19383 | assert_eq_m128i(r, _mm_setzero_si128()); |
19384 | let r = _mm_maskz_alignr_epi8::<14>(0b11111111_11111111, a, b); |
19385 | let e = _mm_set_epi8(0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1); |
19386 | assert_eq_m128i(r, e); |
19387 | } |
19388 | |
19389 | #[simd_test(enable = "avx512bw" )] |
19390 | unsafe fn test_mm512_mask_cvtsepi16_storeu_epi8() { |
19391 | let a = _mm512_set1_epi16(i16::MAX); |
19392 | let mut r = _mm256_undefined_si256(); |
19393 | _mm512_mask_cvtsepi16_storeu_epi8( |
19394 | &mut r as *mut _ as *mut i8, |
19395 | 0b11111111_11111111_11111111_11111111, |
19396 | a, |
19397 | ); |
19398 | let e = _mm256_set1_epi8(i8::MAX); |
19399 | assert_eq_m256i(r, e); |
19400 | } |
19401 | |
19402 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19403 | unsafe fn test_mm256_mask_cvtsepi16_storeu_epi8() { |
19404 | let a = _mm256_set1_epi16(i16::MAX); |
19405 | let mut r = _mm_undefined_si128(); |
19406 | _mm256_mask_cvtsepi16_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111_11111111, a); |
19407 | let e = _mm_set1_epi8(i8::MAX); |
19408 | assert_eq_m128i(r, e); |
19409 | } |
19410 | |
19411 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19412 | unsafe fn test_mm_mask_cvtsepi16_storeu_epi8() { |
19413 | let a = _mm_set1_epi16(i16::MAX); |
19414 | let mut r = _mm_set1_epi8(0); |
19415 | _mm_mask_cvtsepi16_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
19416 | #[rustfmt::skip] |
19417 | let e = _mm_set_epi8( |
19418 | 0, 0, 0, 0, 0, 0, 0, 0, |
19419 | i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, i8::MAX, |
19420 | ); |
19421 | assert_eq_m128i(r, e); |
19422 | } |
19423 | |
19424 | #[simd_test(enable = "avx512bw" )] |
19425 | unsafe fn test_mm512_mask_cvtepi16_storeu_epi8() { |
19426 | let a = _mm512_set1_epi16(8); |
19427 | let mut r = _mm256_undefined_si256(); |
19428 | _mm512_mask_cvtepi16_storeu_epi8( |
19429 | &mut r as *mut _ as *mut i8, |
19430 | 0b11111111_11111111_11111111_11111111, |
19431 | a, |
19432 | ); |
19433 | let e = _mm256_set1_epi8(8); |
19434 | assert_eq_m256i(r, e); |
19435 | } |
19436 | |
19437 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19438 | unsafe fn test_mm256_mask_cvtepi16_storeu_epi8() { |
19439 | let a = _mm256_set1_epi16(8); |
19440 | let mut r = _mm_undefined_si128(); |
19441 | _mm256_mask_cvtepi16_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111_11111111, a); |
19442 | let e = _mm_set1_epi8(8); |
19443 | assert_eq_m128i(r, e); |
19444 | } |
19445 | |
19446 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19447 | unsafe fn test_mm_mask_cvtepi16_storeu_epi8() { |
19448 | let a = _mm_set1_epi16(8); |
19449 | let mut r = _mm_set1_epi8(0); |
19450 | _mm_mask_cvtepi16_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
19451 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 8, 8, 8, 8, 8, 8, 8); |
19452 | assert_eq_m128i(r, e); |
19453 | } |
19454 | |
19455 | #[simd_test(enable = "avx512bw" )] |
19456 | unsafe fn test_mm512_mask_cvtusepi16_storeu_epi8() { |
19457 | let a = _mm512_set1_epi16(i16::MAX); |
19458 | let mut r = _mm256_undefined_si256(); |
19459 | _mm512_mask_cvtusepi16_storeu_epi8( |
19460 | &mut r as *mut _ as *mut i8, |
19461 | 0b11111111_11111111_11111111_11111111, |
19462 | a, |
19463 | ); |
19464 | let e = _mm256_set1_epi8(u8::MAX as i8); |
19465 | assert_eq_m256i(r, e); |
19466 | } |
19467 | |
19468 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19469 | unsafe fn test_mm256_mask_cvtusepi16_storeu_epi8() { |
19470 | let a = _mm256_set1_epi16(i16::MAX); |
19471 | let mut r = _mm_undefined_si128(); |
19472 | _mm256_mask_cvtusepi16_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111_11111111, a); |
19473 | let e = _mm_set1_epi8(u8::MAX as i8); |
19474 | assert_eq_m128i(r, e); |
19475 | } |
19476 | |
19477 | #[simd_test(enable = "avx512bw,avx512vl" )] |
19478 | unsafe fn test_mm_mask_cvtusepi16_storeu_epi8() { |
19479 | let a = _mm_set1_epi16(i16::MAX); |
19480 | let mut r = _mm_set1_epi8(0); |
19481 | _mm_mask_cvtusepi16_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
19482 | #[rustfmt::skip] |
19483 | let e = _mm_set_epi8( |
19484 | 0, 0, 0, 0, |
19485 | 0, 0, 0, 0, |
19486 | u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, |
19487 | u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, |
19488 | ); |
19489 | assert_eq_m128i(r, e); |
19490 | } |
19491 | } |
19492 | |