1 | use crate::{ |
2 | core_arch::{simd::*, x86::*, x86_64::*}, |
3 | intrinsics::simd::*, |
4 | mem::transmute, |
5 | }; |
6 | |
7 | #[cfg (test)] |
8 | use stdarch_test::assert_instr; |
9 | |
10 | /// Convert the lower double-precision (64-bit) floating-point element in a to a 64-bit integer, and store the result in dst. |
11 | /// |
12 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtsd_i64&expand=1792) |
13 | #[inline ] |
14 | #[target_feature (enable = "avx512f" )] |
15 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
16 | #[cfg_attr (test, assert_instr(vcvtsd2si))] |
17 | pub unsafe fn _mm_cvtsd_i64(a: __m128d) -> i64 { |
18 | _mm_cvtsd_si64(a) |
19 | } |
20 | |
21 | /// Convert the lower single-precision (32-bit) floating-point element in a to a 64-bit integer, and store the result in dst. |
22 | /// |
23 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtss_i64&expand=1894) |
24 | #[inline ] |
25 | #[target_feature (enable = "avx512f" )] |
26 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
27 | #[cfg_attr (test, assert_instr(vcvtss2si))] |
28 | pub unsafe fn _mm_cvtss_i64(a: __m128) -> i64 { |
29 | _mm_cvtss_si64(a) |
30 | } |
31 | |
32 | /// Convert the lower single-precision (32-bit) floating-point element in a to an unsigned 64-bit integer, and store the result in dst. |
33 | /// |
34 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtss_u64&expand=1902) |
35 | #[inline ] |
36 | #[target_feature (enable = "avx512f" )] |
37 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
38 | #[cfg_attr (test, assert_instr(vcvtss2usi))] |
39 | pub unsafe fn _mm_cvtss_u64(a: __m128) -> u64 { |
40 | vcvtss2usi64(a:a.as_f32x4(), _MM_FROUND_CUR_DIRECTION) |
41 | } |
42 | |
43 | /// Convert the lower double-precision (64-bit) floating-point element in a to an unsigned 64-bit integer, and store the result in dst. |
44 | /// |
45 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtsd_u64&expand=1800) |
46 | #[inline ] |
47 | #[target_feature (enable = "avx512f" )] |
48 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
49 | #[cfg_attr (test, assert_instr(vcvtsd2usi))] |
50 | pub unsafe fn _mm_cvtsd_u64(a: __m128d) -> u64 { |
51 | vcvtsd2usi64(a:a.as_f64x2(), _MM_FROUND_CUR_DIRECTION) |
52 | } |
53 | |
54 | /// Convert the signed 64-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst. |
55 | /// |
56 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=mm_cvti32_ss&expand=1643) |
57 | #[inline ] |
58 | #[target_feature (enable = "avx512f" )] |
59 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
60 | #[cfg_attr (test, assert_instr(vcvtsi2ss))] |
61 | pub unsafe fn _mm_cvti64_ss(a: __m128, b: i64) -> __m128 { |
62 | let b: f32 = b as f32; |
63 | simd_insert!(a, 0, b) |
64 | } |
65 | |
66 | /// Convert the signed 64-bit integer b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst. |
67 | /// |
68 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvti64_sd&expand=1644) |
69 | #[inline ] |
70 | #[target_feature (enable = "avx512f" )] |
71 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
72 | #[cfg_attr (test, assert_instr(vcvtsi2sd))] |
73 | pub unsafe fn _mm_cvti64_sd(a: __m128d, b: i64) -> __m128d { |
74 | let b: f64 = b as f64; |
75 | simd_insert!(a, 0, b) |
76 | } |
77 | |
78 | /// Convert the unsigned 64-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst. |
79 | /// |
80 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtu64_ss&expand=2035) |
81 | #[inline ] |
82 | #[target_feature (enable = "avx512f" )] |
83 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
84 | #[cfg_attr (test, assert_instr(vcvtusi2ss))] |
85 | pub unsafe fn _mm_cvtu64_ss(a: __m128, b: u64) -> __m128 { |
86 | let b: f32 = b as f32; |
87 | simd_insert!(a, 0, b) |
88 | } |
89 | |
90 | /// Convert the unsigned 64-bit integer b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst. |
91 | /// |
92 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtu64_sd&expand=2034) |
93 | #[inline ] |
94 | #[target_feature (enable = "avx512f" )] |
95 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
96 | #[cfg_attr (test, assert_instr(vcvtusi2sd))] |
97 | pub unsafe fn _mm_cvtu64_sd(a: __m128d, b: u64) -> __m128d { |
98 | let b: f64 = b as f64; |
99 | simd_insert!(a, 0, b) |
100 | } |
101 | |
102 | /// Convert the lower double-precision (64-bit) floating-point element in a to a 64-bit integer with truncation, and store the result in dst. |
103 | /// |
104 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvttsd_i64&expand=2016) |
105 | #[inline ] |
106 | #[target_feature (enable = "avx512f" )] |
107 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
108 | #[cfg_attr (test, assert_instr(vcvtsd2si))] |
109 | pub unsafe fn _mm_cvttsd_i64(a: __m128d) -> i64 { |
110 | vcvtsd2si64(a:a.as_f64x2(), _MM_FROUND_CUR_DIRECTION) |
111 | } |
112 | |
113 | /// Convert the lower double-precision (64-bit) floating-point element in a to an unsigned 64-bit integer with truncation, and store the result in dst. |
114 | /// |
115 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvttsd_u64&expand=2021) |
116 | #[inline ] |
117 | #[target_feature (enable = "avx512f" )] |
118 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
119 | #[cfg_attr (test, assert_instr(vcvtsd2usi))] |
120 | pub unsafe fn _mm_cvttsd_u64(a: __m128d) -> u64 { |
121 | vcvtsd2usi64(a:a.as_f64x2(), _MM_FROUND_CUR_DIRECTION) |
122 | } |
123 | |
124 | /// Convert the lower single-precision (32-bit) floating-point element in a to a 64-bit integer with truncation, and store the result in dst. |
125 | /// |
126 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=#text=_mm_cvttss_i64&expand=2023) |
127 | #[inline ] |
128 | #[target_feature (enable = "avx512f" )] |
129 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
130 | #[cfg_attr (test, assert_instr(vcvtss2si))] |
131 | pub unsafe fn _mm_cvttss_i64(a: __m128) -> i64 { |
132 | vcvtss2si64(a:a.as_f32x4(), _MM_FROUND_CUR_DIRECTION) |
133 | } |
134 | |
135 | /// Convert the lower single-precision (32-bit) floating-point element in a to an unsigned 64-bit integer with truncation, and store the result in dst. |
136 | /// |
137 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvttss_u64&expand=2027) |
138 | #[inline ] |
139 | #[target_feature (enable = "avx512f" )] |
140 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
141 | #[cfg_attr (test, assert_instr(vcvtss2usi))] |
142 | pub unsafe fn _mm_cvttss_u64(a: __m128) -> u64 { |
143 | vcvtss2usi64(a:a.as_f32x4(), _MM_FROUND_CUR_DIRECTION) |
144 | } |
145 | |
146 | /// Convert the signed 64-bit integer b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst. |
147 | /// Rounding is done according to the rounding\[3:0\] parameter, which can be one of:\ |
148 | /// (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions\ |
149 | /// (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC) // round down, and suppress exceptions\ |
150 | /// (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC) // round up, and suppress exceptions\ |
151 | /// (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC) // truncate, and suppress exceptions\ |
152 | /// _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE |
153 | /// |
154 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_roundi64_sd&expand=1313) |
155 | #[inline ] |
156 | #[target_feature (enable = "avx512f" )] |
157 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
158 | #[cfg_attr (test, assert_instr(vcvtsi2sd, ROUNDING = 8))] |
159 | #[rustc_legacy_const_generics (2)] |
160 | pub unsafe fn _mm_cvt_roundi64_sd<const ROUNDING: i32>(a: __m128d, b: i64) -> __m128d { |
161 | static_assert_rounding!(ROUNDING); |
162 | let a: f64x2 = a.as_f64x2(); |
163 | let r: f64x2 = vcvtsi2sd64(a, b, ROUNDING); |
164 | transmute(src:r) |
165 | } |
166 | |
167 | /// Convert the signed 64-bit integer b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst. |
168 | /// Rounding is done according to the rounding\[3:0\] parameter, which can be one of:\ |
169 | /// (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions\ |
170 | /// (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC) // round down, and suppress exceptions\ |
171 | /// (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC) // round up, and suppress exceptions\ |
172 | /// (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC) // truncate, and suppress exceptions\ |
173 | /// _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE |
174 | /// |
175 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_roundsi64_sd&expand=1367) |
176 | #[inline ] |
177 | #[target_feature (enable = "avx512f" )] |
178 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
179 | #[cfg_attr (test, assert_instr(vcvtsi2sd, ROUNDING = 8))] |
180 | #[rustc_legacy_const_generics (2)] |
181 | pub unsafe fn _mm_cvt_roundsi64_sd<const ROUNDING: i32>(a: __m128d, b: i64) -> __m128d { |
182 | static_assert_rounding!(ROUNDING); |
183 | let a: f64x2 = a.as_f64x2(); |
184 | let r: f64x2 = vcvtsi2sd64(a, b, ROUNDING); |
185 | transmute(src:r) |
186 | } |
187 | |
188 | /// Convert the signed 64-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst. |
189 | /// Rounding is done according to the rounding\[3:0\] parameter, which can be one of:\ |
190 | /// (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions\ |
191 | /// (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC) // round down, and suppress exceptions\ |
192 | /// (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC) // round up, and suppress exceptions\ |
193 | /// (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC) // truncate, and suppress exceptions\ |
194 | /// _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE |
195 | /// |
196 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_roundi64_ss&expand=1314) |
197 | #[inline ] |
198 | #[target_feature (enable = "avx512f" )] |
199 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
200 | #[cfg_attr (test, assert_instr(vcvtsi2ss, ROUNDING = 8))] |
201 | #[rustc_legacy_const_generics (2)] |
202 | pub unsafe fn _mm_cvt_roundi64_ss<const ROUNDING: i32>(a: __m128, b: i64) -> __m128 { |
203 | static_assert_rounding!(ROUNDING); |
204 | let a: f32x4 = a.as_f32x4(); |
205 | let r: f32x4 = vcvtsi2ss64(a, b, ROUNDING); |
206 | transmute(src:r) |
207 | } |
208 | |
209 | /// Convert the unsigned 64-bit integer b to a double-precision (64-bit) floating-point element, store the result in the lower element of dst, and copy the upper element from a to the upper element of dst.\ |
210 | /// Rounding is done according to the rounding\[3:0\] parameter, which can be one of:\ |
211 | /// (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions\ |
212 | /// (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC) // round down, and suppress exceptions\ |
213 | /// (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC) // round up, and suppress exceptions\ |
214 | /// (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC) // truncate, and suppress exceptions\ |
215 | /// _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE |
216 | /// |
217 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_roundu64_sd&expand=1379) |
218 | #[inline ] |
219 | #[target_feature (enable = "avx512f" )] |
220 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
221 | #[cfg_attr (test, assert_instr(vcvtusi2sd, ROUNDING = 8))] |
222 | #[rustc_legacy_const_generics (2)] |
223 | pub unsafe fn _mm_cvt_roundu64_sd<const ROUNDING: i32>(a: __m128d, b: u64) -> __m128d { |
224 | static_assert_rounding!(ROUNDING); |
225 | let a: f64x2 = a.as_f64x2(); |
226 | let r: f64x2 = vcvtusi2sd64(a, b, ROUNDING); |
227 | transmute(src:r) |
228 | } |
229 | |
230 | /// Convert the signed 64-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst. |
231 | /// Rounding is done according to the rounding\[3:0\] parameter, which can be one of:\ |
232 | /// (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions\ |
233 | /// (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC) // round down, and suppress exceptions\ |
234 | /// (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC) // round up, and suppress exceptions\ |
235 | /// (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC) // truncate, and suppress exceptions\ |
236 | /// _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE |
237 | /// |
238 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_roundsi64_ss&expand=1368) |
239 | #[inline ] |
240 | #[target_feature (enable = "avx512f" )] |
241 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
242 | #[cfg_attr (test, assert_instr(vcvtsi2ss, ROUNDING = 8))] |
243 | #[rustc_legacy_const_generics (2)] |
244 | pub unsafe fn _mm_cvt_roundsi64_ss<const ROUNDING: i32>(a: __m128, b: i64) -> __m128 { |
245 | static_assert_rounding!(ROUNDING); |
246 | let a: f32x4 = a.as_f32x4(); |
247 | let r: f32x4 = vcvtsi2ss64(a, b, ROUNDING); |
248 | transmute(src:r) |
249 | } |
250 | |
251 | /// Convert the unsigned 64-bit integer b to a single-precision (32-bit) floating-point element, store the result in the lower element of dst, and copy the upper 3 packed elements from a to the upper elements of dst.\ |
252 | /// Rounding is done according to the rounding\[3:0\] parameter, which can be one of:\ |
253 | /// (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions\ |
254 | /// (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC) // round down, and suppress exceptions\ |
255 | /// (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC) // round up, and suppress exceptions\ |
256 | /// (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC) // truncate, and suppress exceptions\ |
257 | /// _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE |
258 | /// |
259 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_roundu64_ss&expand=1380) |
260 | #[inline ] |
261 | #[target_feature (enable = "avx512f" )] |
262 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
263 | #[cfg_attr (test, assert_instr(vcvtusi2ss, ROUNDING = 8))] |
264 | #[rustc_legacy_const_generics (2)] |
265 | pub unsafe fn _mm_cvt_roundu64_ss<const ROUNDING: i32>(a: __m128, b: u64) -> __m128 { |
266 | static_assert_rounding!(ROUNDING); |
267 | let a: f32x4 = a.as_f32x4(); |
268 | let r: f32x4 = vcvtusi2ss64(a, b, ROUNDING); |
269 | transmute(src:r) |
270 | } |
271 | |
272 | /// Convert the lower double-precision (64-bit) floating-point element in a to a 64-bit integer, and store the result in dst.\ |
273 | /// Rounding is done according to the rounding\[3:0\] parameter, which can be one of:\ |
274 | /// (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions\ |
275 | /// (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC) // round down, and suppress exceptions\ |
276 | /// (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC) // round up, and suppress exceptions\ |
277 | /// (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC) // truncate, and suppress exceptions\ |
278 | /// _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE |
279 | /// |
280 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_roundsd_si64&expand=1360) |
281 | #[inline ] |
282 | #[target_feature (enable = "avx512f" )] |
283 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
284 | #[cfg_attr (test, assert_instr(vcvtsd2si, ROUNDING = 8))] |
285 | #[rustc_legacy_const_generics (1)] |
286 | pub unsafe fn _mm_cvt_roundsd_si64<const ROUNDING: i32>(a: __m128d) -> i64 { |
287 | static_assert_rounding!(ROUNDING); |
288 | let a: f64x2 = a.as_f64x2(); |
289 | vcvtsd2si64(a, ROUNDING) |
290 | } |
291 | |
292 | /// Convert the lower double-precision (64-bit) floating-point element in a to a 64-bit integer, and store the result in dst.\ |
293 | /// Rounding is done according to the rounding\[3:0\] parameter, which can be one of:\ |
294 | /// (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions\ |
295 | /// (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC) // round down, and suppress exceptions\ |
296 | /// (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC) // round up, and suppress exceptions\ |
297 | /// (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC) // truncate, and suppress exceptions\ |
298 | /// _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE |
299 | /// |
300 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_roundsd_i64&expand=1358) |
301 | #[inline ] |
302 | #[target_feature (enable = "avx512f" )] |
303 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
304 | #[cfg_attr (test, assert_instr(vcvtsd2si, ROUNDING = 8))] |
305 | #[rustc_legacy_const_generics (1)] |
306 | pub unsafe fn _mm_cvt_roundsd_i64<const ROUNDING: i32>(a: __m128d) -> i64 { |
307 | static_assert_rounding!(ROUNDING); |
308 | let a: f64x2 = a.as_f64x2(); |
309 | vcvtsd2si64(a, ROUNDING) |
310 | } |
311 | |
312 | /// Convert the lower double-precision (64-bit) floating-point element in a to an unsigned 64-bit integer, and store the result in dst.\ |
313 | /// Rounding is done according to the rounding\[3:0\] parameter, which can be one of:\ |
314 | /// (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions\ |
315 | /// (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC) // round down, and suppress exceptions\ |
316 | /// (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC) // round up, and suppress exceptions\ |
317 | /// (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC) // truncate, and suppress exceptions\ |
318 | /// _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE |
319 | /// |
320 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_roundsd_u64&expand=1365) |
321 | #[inline ] |
322 | #[target_feature (enable = "avx512f" )] |
323 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
324 | #[cfg_attr (test, assert_instr(vcvtsd2usi, ROUNDING = 8))] |
325 | #[rustc_legacy_const_generics (1)] |
326 | pub unsafe fn _mm_cvt_roundsd_u64<const ROUNDING: i32>(a: __m128d) -> u64 { |
327 | static_assert_rounding!(ROUNDING); |
328 | let a: f64x2 = a.as_f64x2(); |
329 | vcvtsd2usi64(a, ROUNDING) |
330 | } |
331 | |
332 | /// Convert the lower single-precision (32-bit) floating-point element in a to a 64-bit integer, and store the result in dst.\ |
333 | /// Rounding is done according to the rounding\[3:0\] parameter, which can be one of:\ |
334 | /// (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions\ |
335 | /// (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC) // round down, and suppress exceptions\ |
336 | /// (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC) // round up, and suppress exceptions\ |
337 | /// (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC) // truncate, and suppress exceptions\ |
338 | /// _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE |
339 | /// |
340 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_roundss_si64&expand=1375) |
341 | #[inline ] |
342 | #[target_feature (enable = "avx512f" )] |
343 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
344 | #[cfg_attr (test, assert_instr(vcvtss2si, ROUNDING = 8))] |
345 | #[rustc_legacy_const_generics (1)] |
346 | pub unsafe fn _mm_cvt_roundss_si64<const ROUNDING: i32>(a: __m128) -> i64 { |
347 | static_assert_rounding!(ROUNDING); |
348 | let a: f32x4 = a.as_f32x4(); |
349 | vcvtss2si64(a, ROUNDING) |
350 | } |
351 | |
352 | /// Convert the lower single-precision (32-bit) floating-point element in a to a 64-bit integer, and store the result in dst.\ |
353 | /// Rounding is done according to the rounding\[3:0\] parameter, which can be one of:\ |
354 | /// (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions\ |
355 | /// (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC) // round down, and suppress exceptions\ |
356 | /// (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC) // round up, and suppress exceptions\ |
357 | /// (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC) // truncate, and suppress exceptions\ |
358 | /// _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE |
359 | /// |
360 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_roundss_i64&expand=1370) |
361 | #[inline ] |
362 | #[target_feature (enable = "avx512f" )] |
363 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
364 | #[cfg_attr (test, assert_instr(vcvtss2si, ROUNDING = 8))] |
365 | #[rustc_legacy_const_generics (1)] |
366 | pub unsafe fn _mm_cvt_roundss_i64<const ROUNDING: i32>(a: __m128) -> i64 { |
367 | static_assert_rounding!(ROUNDING); |
368 | let a: f32x4 = a.as_f32x4(); |
369 | vcvtss2si64(a, ROUNDING) |
370 | } |
371 | |
372 | /// Convert the lower single-precision (32-bit) floating-point element in a to an unsigned 64-bit integer, and store the result in dst.\ |
373 | /// Rounding is done according to the rounding\[3:0\] parameter, which can be one of:\ |
374 | /// (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions\ |
375 | /// (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC) // round down, and suppress exceptions\ |
376 | /// (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC) // round up, and suppress exceptions\ |
377 | /// (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC) // truncate, and suppress exceptions\ |
378 | /// _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE |
379 | /// |
380 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_roundss_u64&expand=1377) |
381 | #[inline ] |
382 | #[target_feature (enable = "avx512f" )] |
383 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
384 | #[cfg_attr (test, assert_instr(vcvtss2usi, ROUNDING = 8))] |
385 | #[rustc_legacy_const_generics (1)] |
386 | pub unsafe fn _mm_cvt_roundss_u64<const ROUNDING: i32>(a: __m128) -> u64 { |
387 | static_assert_rounding!(ROUNDING); |
388 | let a: f32x4 = a.as_f32x4(); |
389 | vcvtss2usi64(a, ROUNDING) |
390 | } |
391 | |
392 | /// Convert the lower double-precision (64-bit) floating-point element in a to a 64-bit integer with truncation, and store the result in dst.\ |
393 | /// Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. |
394 | /// |
395 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtt_roundsd_si64&expand=1931) |
396 | #[inline ] |
397 | #[target_feature (enable = "avx512f" )] |
398 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
399 | #[cfg_attr (test, assert_instr(vcvtsd2si, SAE = 8))] |
400 | #[rustc_legacy_const_generics (1)] |
401 | pub unsafe fn _mm_cvtt_roundsd_si64<const SAE: i32>(a: __m128d) -> i64 { |
402 | static_assert_sae!(SAE); |
403 | let a: f64x2 = a.as_f64x2(); |
404 | vcvtsd2si64(a, SAE) |
405 | } |
406 | |
407 | /// Convert the lower double-precision (64-bit) floating-point element in a to a 64-bit integer with truncation, and store the result in dst.\ |
408 | /// Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. |
409 | /// |
410 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtt_roundsd_i64&expand=1929) |
411 | #[inline ] |
412 | #[target_feature (enable = "avx512f" )] |
413 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
414 | #[cfg_attr (test, assert_instr(vcvtsd2si, SAE = 8))] |
415 | #[rustc_legacy_const_generics (1)] |
416 | pub unsafe fn _mm_cvtt_roundsd_i64<const SAE: i32>(a: __m128d) -> i64 { |
417 | static_assert_sae!(SAE); |
418 | let a: f64x2 = a.as_f64x2(); |
419 | vcvtsd2si64(a, SAE) |
420 | } |
421 | |
422 | /// Convert the lower double-precision (64-bit) floating-point element in a to an unsigned 64-bit integer with truncation, and store the result in dst.\ |
423 | /// Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. |
424 | /// |
425 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtt_roundsd_u64&expand=1933) |
426 | #[inline ] |
427 | #[target_feature (enable = "avx512f" )] |
428 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
429 | #[cfg_attr (test, assert_instr(vcvtsd2usi, SAE = 8))] |
430 | #[rustc_legacy_const_generics (1)] |
431 | pub unsafe fn _mm_cvtt_roundsd_u64<const SAE: i32>(a: __m128d) -> u64 { |
432 | static_assert_sae!(SAE); |
433 | let a: f64x2 = a.as_f64x2(); |
434 | vcvtsd2usi64(a, SAE) |
435 | } |
436 | |
437 | /// Convert the lower single-precision (32-bit) floating-point element in a to a 64-bit integer with truncation, and store the result in dst.\ |
438 | /// Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. |
439 | /// |
440 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtt_roundss_i64&expand=1935) |
441 | #[inline ] |
442 | #[target_feature (enable = "avx512f" )] |
443 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
444 | #[cfg_attr (test, assert_instr(vcvtss2si, SAE = 8))] |
445 | #[rustc_legacy_const_generics (1)] |
446 | pub unsafe fn _mm_cvtt_roundss_i64<const SAE: i32>(a: __m128) -> i64 { |
447 | static_assert_sae!(SAE); |
448 | let a: f32x4 = a.as_f32x4(); |
449 | vcvtss2si64(a, SAE) |
450 | } |
451 | |
452 | /// Convert the lower single-precision (32-bit) floating-point element in a to a 64-bit integer with truncation, and store the result in dst.\ |
453 | /// Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. |
454 | /// |
455 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtt_roundss_si64&expand=1937) |
456 | #[inline ] |
457 | #[target_feature (enable = "avx512f" )] |
458 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
459 | #[cfg_attr (test, assert_instr(vcvtss2si, SAE = 8))] |
460 | #[rustc_legacy_const_generics (1)] |
461 | pub unsafe fn _mm_cvtt_roundss_si64<const SAE: i32>(a: __m128) -> i64 { |
462 | static_assert_sae!(SAE); |
463 | let a: f32x4 = a.as_f32x4(); |
464 | vcvtss2si64(a, SAE) |
465 | } |
466 | |
467 | /// Convert the lower single-precision (32-bit) floating-point element in a to an unsigned 64-bit integer with truncation, and store the result in dst.\ |
468 | /// Exceptions can be suppressed by passing _MM_FROUND_NO_EXC in the sae parameter. |
469 | /// |
470 | /// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtt_roundss_u64&expand=1939) |
471 | #[inline ] |
472 | #[target_feature (enable = "avx512f" )] |
473 | #[unstable (feature = "stdarch_x86_avx512" , issue = "111137" )] |
474 | #[cfg_attr (test, assert_instr(vcvtss2usi, SAE = 8))] |
475 | #[rustc_legacy_const_generics (1)] |
476 | pub unsafe fn _mm_cvtt_roundss_u64<const SAE: i32>(a: __m128) -> u64 { |
477 | static_assert_sae!(SAE); |
478 | let a: f32x4 = a.as_f32x4(); |
479 | vcvtss2usi64(a, SAE) |
480 | } |
481 | |
482 | #[allow (improper_ctypes)] |
483 | extern "C" { |
484 | #[link_name = "llvm.x86.avx512.vcvtss2si64" ] |
485 | fn vcvtss2si64(a: f32x4, rounding: i32) -> i64; |
486 | #[link_name = "llvm.x86.avx512.vcvtss2usi64" ] |
487 | fn vcvtss2usi64(a: f32x4, rounding: i32) -> u64; |
488 | #[link_name = "llvm.x86.avx512.vcvtsd2si64" ] |
489 | fn vcvtsd2si64(a: f64x2, rounding: i32) -> i64; |
490 | #[link_name = "llvm.x86.avx512.vcvtsd2usi64" ] |
491 | fn vcvtsd2usi64(a: f64x2, rounding: i32) -> u64; |
492 | |
493 | #[link_name = "llvm.x86.avx512.cvtsi2ss64" ] |
494 | fn vcvtsi2ss64(a: f32x4, b: i64, rounding: i32) -> f32x4; |
495 | #[link_name = "llvm.x86.avx512.cvtsi2sd64" ] |
496 | fn vcvtsi2sd64(a: f64x2, b: i64, rounding: i32) -> f64x2; |
497 | #[link_name = "llvm.x86.avx512.cvtusi642ss" ] |
498 | fn vcvtusi2ss64(a: f32x4, b: u64, rounding: i32) -> f32x4; |
499 | #[link_name = "llvm.x86.avx512.cvtusi642sd" ] |
500 | fn vcvtusi2sd64(a: f64x2, b: u64, rounding: i32) -> f64x2; |
501 | } |
502 | |
503 | #[cfg (test)] |
504 | mod tests { |
505 | |
506 | use stdarch_test::simd_test; |
507 | |
508 | use crate::core_arch::x86::*; |
509 | use crate::core_arch::x86_64::*; |
510 | use crate::hint::black_box; |
511 | |
512 | #[simd_test(enable = "avx512f" )] |
513 | unsafe fn test_mm512_abs_epi64() { |
514 | let a = _mm512_set_epi64(0, 1, -1, i64::MAX, i64::MIN, 100, -100, -32); |
515 | let r = _mm512_abs_epi64(a); |
516 | let e = _mm512_set_epi64(0, 1, 1, i64::MAX, i64::MAX.wrapping_add(1), 100, 100, 32); |
517 | assert_eq_m512i(r, e); |
518 | } |
519 | |
520 | #[simd_test(enable = "avx512f" )] |
521 | unsafe fn test_mm512_mask_abs_epi64() { |
522 | let a = _mm512_set_epi64(0, 1, -1, i64::MAX, i64::MIN, 100, -100, -32); |
523 | let r = _mm512_mask_abs_epi64(a, 0, a); |
524 | assert_eq_m512i(r, a); |
525 | let r = _mm512_mask_abs_epi64(a, 0b11111111, a); |
526 | let e = _mm512_set_epi64(0, 1, 1, i64::MAX, i64::MIN, 100, 100, 32); |
527 | assert_eq_m512i(r, e); |
528 | } |
529 | |
530 | #[simd_test(enable = "avx512f" )] |
531 | unsafe fn test_mm512_maskz_abs_epi64() { |
532 | let a = _mm512_set_epi64(0, 1, -1, i64::MAX, i64::MIN, 100, -100, -32); |
533 | let r = _mm512_maskz_abs_epi64(0, a); |
534 | assert_eq_m512i(r, _mm512_setzero_si512()); |
535 | let r = _mm512_maskz_abs_epi64(0b11111111, a); |
536 | let e = _mm512_set_epi64(0, 1, 1, i64::MAX, i64::MIN, 100, 100, 32); |
537 | assert_eq_m512i(r, e); |
538 | } |
539 | |
540 | #[simd_test(enable = "avx512f,avx512vl" )] |
541 | unsafe fn test_mm256_abs_epi64() { |
542 | let a = _mm256_set_epi64x(i64::MAX, i64::MIN, 100, -100); |
543 | let r = _mm256_abs_epi64(a); |
544 | let e = _mm256_set_epi64x(i64::MAX, i64::MAX.wrapping_add(1), 100, 100); |
545 | assert_eq_m256i(r, e); |
546 | } |
547 | |
548 | #[simd_test(enable = "avx512f,avx512vl" )] |
549 | unsafe fn test_mm256_mask_abs_epi64() { |
550 | let a = _mm256_set_epi64x(i64::MAX, i64::MIN, 100, -100); |
551 | let r = _mm256_mask_abs_epi64(a, 0, a); |
552 | assert_eq_m256i(r, a); |
553 | let r = _mm256_mask_abs_epi64(a, 0b00001111, a); |
554 | let e = _mm256_set_epi64x(i64::MAX, i64::MAX.wrapping_add(1), 100, 100); |
555 | assert_eq_m256i(r, e); |
556 | } |
557 | |
558 | #[simd_test(enable = "avx512f,avx512vl" )] |
559 | unsafe fn test_mm256_maskz_abs_epi64() { |
560 | let a = _mm256_set_epi64x(i64::MAX, i64::MIN, 100, -100); |
561 | let r = _mm256_maskz_abs_epi64(0, a); |
562 | assert_eq_m256i(r, _mm256_setzero_si256()); |
563 | let r = _mm256_maskz_abs_epi64(0b00001111, a); |
564 | let e = _mm256_set_epi64x(i64::MAX, i64::MAX.wrapping_add(1), 100, 100); |
565 | assert_eq_m256i(r, e); |
566 | } |
567 | |
568 | #[simd_test(enable = "avx512f" )] |
569 | unsafe fn test_mm512_abs_pd() { |
570 | let a = _mm512_setr_pd(0., 1., -1., f64::MAX, f64::MIN, 100., -100., -32.); |
571 | let r = _mm512_abs_pd(a); |
572 | let e = _mm512_setr_pd(0., 1., 1., f64::MAX, f64::MAX, 100., 100., 32.); |
573 | assert_eq_m512d(r, e); |
574 | } |
575 | |
576 | #[simd_test(enable = "avx512f" )] |
577 | unsafe fn test_mm512_mask_abs_pd() { |
578 | let a = _mm512_setr_pd(0., 1., -1., f64::MAX, f64::MIN, 100., -100., -32.); |
579 | let r = _mm512_mask_abs_pd(a, 0, a); |
580 | assert_eq_m512d(r, a); |
581 | let r = _mm512_mask_abs_pd(a, 0b00001111, a); |
582 | let e = _mm512_setr_pd(0., 1., 1., f64::MAX, f64::MIN, 100., -100., -32.); |
583 | assert_eq_m512d(r, e); |
584 | } |
585 | |
586 | #[simd_test(enable = "avx512f" )] |
587 | unsafe fn test_mm512_mask_mov_epi64() { |
588 | let src = _mm512_set1_epi64(1); |
589 | let a = _mm512_set1_epi64(2); |
590 | let r = _mm512_mask_mov_epi64(src, 0, a); |
591 | assert_eq_m512i(r, src); |
592 | let r = _mm512_mask_mov_epi64(src, 0b11111111, a); |
593 | assert_eq_m512i(r, a); |
594 | } |
595 | |
596 | #[simd_test(enable = "avx512f" )] |
597 | unsafe fn test_mm512_maskz_mov_epi64() { |
598 | let a = _mm512_set1_epi64(2); |
599 | let r = _mm512_maskz_mov_epi64(0, a); |
600 | assert_eq_m512i(r, _mm512_setzero_si512()); |
601 | let r = _mm512_maskz_mov_epi64(0b11111111, a); |
602 | assert_eq_m512i(r, a); |
603 | } |
604 | |
605 | #[simd_test(enable = "avx512f,avx512vl" )] |
606 | unsafe fn test_mm256_mask_mov_epi64() { |
607 | let src = _mm256_set1_epi64x(1); |
608 | let a = _mm256_set1_epi64x(2); |
609 | let r = _mm256_mask_mov_epi64(src, 0, a); |
610 | assert_eq_m256i(r, src); |
611 | let r = _mm256_mask_mov_epi64(src, 0b00001111, a); |
612 | assert_eq_m256i(r, a); |
613 | } |
614 | |
615 | #[simd_test(enable = "avx512f,avx512vl" )] |
616 | unsafe fn test_mm256_maskz_mov_epi64() { |
617 | let a = _mm256_set1_epi64x(2); |
618 | let r = _mm256_maskz_mov_epi64(0, a); |
619 | assert_eq_m256i(r, _mm256_setzero_si256()); |
620 | let r = _mm256_maskz_mov_epi64(0b00001111, a); |
621 | assert_eq_m256i(r, a); |
622 | } |
623 | |
624 | #[simd_test(enable = "avx512f,avx512vl" )] |
625 | unsafe fn test_mm_mask_mov_epi64() { |
626 | let src = _mm_set1_epi64x(1); |
627 | let a = _mm_set1_epi64x(2); |
628 | let r = _mm_mask_mov_epi64(src, 0, a); |
629 | assert_eq_m128i(r, src); |
630 | let r = _mm_mask_mov_epi64(src, 0b00000011, a); |
631 | assert_eq_m128i(r, a); |
632 | } |
633 | |
634 | #[simd_test(enable = "avx512f,avx512vl" )] |
635 | unsafe fn test_mm_maskz_mov_epi64() { |
636 | let a = _mm_set1_epi64x(2); |
637 | let r = _mm_maskz_mov_epi64(0, a); |
638 | assert_eq_m128i(r, _mm_setzero_si128()); |
639 | let r = _mm_maskz_mov_epi64(0b00000011, a); |
640 | assert_eq_m128i(r, a); |
641 | } |
642 | |
643 | #[simd_test(enable = "avx512f" )] |
644 | unsafe fn test_mm512_mask_mov_pd() { |
645 | let src = _mm512_set1_pd(1.); |
646 | let a = _mm512_set1_pd(2.); |
647 | let r = _mm512_mask_mov_pd(src, 0, a); |
648 | assert_eq_m512d(r, src); |
649 | let r = _mm512_mask_mov_pd(src, 0b11111111, a); |
650 | assert_eq_m512d(r, a); |
651 | } |
652 | |
653 | #[simd_test(enable = "avx512f" )] |
654 | unsafe fn test_mm512_maskz_mov_pd() { |
655 | let a = _mm512_set1_pd(2.); |
656 | let r = _mm512_maskz_mov_pd(0, a); |
657 | assert_eq_m512d(r, _mm512_setzero_pd()); |
658 | let r = _mm512_maskz_mov_pd(0b11111111, a); |
659 | assert_eq_m512d(r, a); |
660 | } |
661 | |
662 | #[simd_test(enable = "avx512f,avx512vl" )] |
663 | unsafe fn test_mm256_mask_mov_pd() { |
664 | let src = _mm256_set1_pd(1.); |
665 | let a = _mm256_set1_pd(2.); |
666 | let r = _mm256_mask_mov_pd(src, 0, a); |
667 | assert_eq_m256d(r, src); |
668 | let r = _mm256_mask_mov_pd(src, 0b00001111, a); |
669 | assert_eq_m256d(r, a); |
670 | } |
671 | |
672 | #[simd_test(enable = "avx512f,avx512vl" )] |
673 | unsafe fn test_mm256_maskz_mov_pd() { |
674 | let a = _mm256_set1_pd(2.); |
675 | let r = _mm256_maskz_mov_pd(0, a); |
676 | assert_eq_m256d(r, _mm256_setzero_pd()); |
677 | let r = _mm256_maskz_mov_pd(0b00001111, a); |
678 | assert_eq_m256d(r, a); |
679 | } |
680 | |
681 | #[simd_test(enable = "avx512f,avx512vl" )] |
682 | unsafe fn test_mm_mask_mov_pd() { |
683 | let src = _mm_set1_pd(1.); |
684 | let a = _mm_set1_pd(2.); |
685 | let r = _mm_mask_mov_pd(src, 0, a); |
686 | assert_eq_m128d(r, src); |
687 | let r = _mm_mask_mov_pd(src, 0b00000011, a); |
688 | assert_eq_m128d(r, a); |
689 | } |
690 | |
691 | #[simd_test(enable = "avx512f,avx512vl" )] |
692 | unsafe fn test_mm_maskz_mov_pd() { |
693 | let a = _mm_set1_pd(2.); |
694 | let r = _mm_maskz_mov_pd(0, a); |
695 | assert_eq_m128d(r, _mm_setzero_pd()); |
696 | let r = _mm_maskz_mov_pd(0b00000011, a); |
697 | assert_eq_m128d(r, a); |
698 | } |
699 | |
700 | #[simd_test(enable = "avx512f" )] |
701 | unsafe fn test_mm512_add_epi64() { |
702 | let a = _mm512_setr_epi64(0, 1, -1, i64::MAX, i64::MIN, 100, -100, -32); |
703 | let b = _mm512_set1_epi64(1); |
704 | let r = _mm512_add_epi64(a, b); |
705 | let e = _mm512_setr_epi64(1, 2, 0, i64::MIN, i64::MIN + 1, 101, -99, -31); |
706 | assert_eq_m512i(r, e); |
707 | } |
708 | |
709 | #[simd_test(enable = "avx512f" )] |
710 | unsafe fn test_mm512_mask_add_epi64() { |
711 | let a = _mm512_setr_epi64(0, 1, -1, i64::MAX, i64::MIN, 100, -100, -32); |
712 | let b = _mm512_set1_epi64(1); |
713 | let r = _mm512_mask_add_epi64(a, 0, a, b); |
714 | assert_eq_m512i(r, a); |
715 | let r = _mm512_mask_add_epi64(a, 0b00001111, a, b); |
716 | let e = _mm512_setr_epi64(1, 2, 0, i64::MIN, i64::MIN, 100, -100, -32); |
717 | assert_eq_m512i(r, e); |
718 | } |
719 | |
720 | #[simd_test(enable = "avx512f" )] |
721 | unsafe fn test_mm512_maskz_add_epi64() { |
722 | let a = _mm512_setr_epi64(0, 1, -1, i64::MAX, i64::MIN, 100, -100, -32); |
723 | let b = _mm512_set1_epi64(1); |
724 | let r = _mm512_maskz_add_epi64(0, a, b); |
725 | assert_eq_m512i(r, _mm512_setzero_si512()); |
726 | let r = _mm512_maskz_add_epi64(0b00001111, a, b); |
727 | let e = _mm512_setr_epi64(1, 2, 0, i64::MIN, 0, 0, 0, 0); |
728 | assert_eq_m512i(r, e); |
729 | } |
730 | |
731 | #[simd_test(enable = "avx512f,avx512vl" )] |
732 | unsafe fn test_mm256_mask_add_epi64() { |
733 | let a = _mm256_set_epi64x(1, -1, i64::MAX, i64::MIN); |
734 | let b = _mm256_set1_epi64x(1); |
735 | let r = _mm256_mask_add_epi64(a, 0, a, b); |
736 | assert_eq_m256i(r, a); |
737 | let r = _mm256_mask_add_epi64(a, 0b00001111, a, b); |
738 | let e = _mm256_set_epi64x(2, 0, i64::MIN, i64::MIN + 1); |
739 | assert_eq_m256i(r, e); |
740 | } |
741 | |
742 | #[simd_test(enable = "avx512f,avx512vl" )] |
743 | unsafe fn test_mm256_maskz_add_epi64() { |
744 | let a = _mm256_set_epi64x(1, -1, i64::MAX, i64::MIN); |
745 | let b = _mm256_set1_epi64x(1); |
746 | let r = _mm256_maskz_add_epi64(0, a, b); |
747 | assert_eq_m256i(r, _mm256_setzero_si256()); |
748 | let r = _mm256_maskz_add_epi64(0b00001111, a, b); |
749 | let e = _mm256_set_epi64x(2, 0, i64::MIN, i64::MIN + 1); |
750 | assert_eq_m256i(r, e); |
751 | } |
752 | |
753 | #[simd_test(enable = "avx512f,avx512vl" )] |
754 | unsafe fn test_mm_mask_add_epi64() { |
755 | let a = _mm_set_epi64x(i64::MAX, i64::MIN); |
756 | let b = _mm_set1_epi64x(1); |
757 | let r = _mm_mask_add_epi64(a, 0, a, b); |
758 | assert_eq_m128i(r, a); |
759 | let r = _mm_mask_add_epi64(a, 0b00000011, a, b); |
760 | let e = _mm_set_epi64x(i64::MIN, i64::MIN + 1); |
761 | assert_eq_m128i(r, e); |
762 | } |
763 | |
764 | #[simd_test(enable = "avx512f,avx512vl" )] |
765 | unsafe fn test_mm_maskz_add_epi64() { |
766 | let a = _mm_set_epi64x(i64::MAX, i64::MIN); |
767 | let b = _mm_set1_epi64x(1); |
768 | let r = _mm_maskz_add_epi64(0, a, b); |
769 | assert_eq_m128i(r, _mm_setzero_si128()); |
770 | let r = _mm_maskz_add_epi64(0b00000011, a, b); |
771 | let e = _mm_set_epi64x(i64::MIN, i64::MIN + 1); |
772 | assert_eq_m128i(r, e); |
773 | } |
774 | |
775 | #[simd_test(enable = "avx512f" )] |
776 | unsafe fn test_mm512_add_pd() { |
777 | let a = _mm512_setr_pd(0., 1., -1., f64::MAX, f64::MIN, 100., -100., -32.); |
778 | let b = _mm512_set1_pd(1.); |
779 | let r = _mm512_add_pd(a, b); |
780 | let e = _mm512_setr_pd(1., 2., 0., f64::MAX, f64::MIN + 1., 101., -99., -31.); |
781 | assert_eq_m512d(r, e); |
782 | } |
783 | |
784 | #[simd_test(enable = "avx512f" )] |
785 | unsafe fn test_mm512_mask_add_pd() { |
786 | let a = _mm512_setr_pd(0., 1., -1., f64::MAX, f64::MIN, 100., -100., -32.); |
787 | let b = _mm512_set1_pd(1.); |
788 | let r = _mm512_mask_add_pd(a, 0, a, b); |
789 | assert_eq_m512d(r, a); |
790 | let r = _mm512_mask_add_pd(a, 0b00001111, a, b); |
791 | let e = _mm512_setr_pd(1., 2., 0., f64::MAX, f64::MIN, 100., -100., -32.); |
792 | assert_eq_m512d(r, e); |
793 | } |
794 | |
795 | #[simd_test(enable = "avx512f" )] |
796 | unsafe fn test_mm512_maskz_add_pd() { |
797 | let a = _mm512_setr_pd(0., 1., -1., f64::MAX, f64::MIN, 100., -100., -32.); |
798 | let b = _mm512_set1_pd(1.); |
799 | let r = _mm512_maskz_add_pd(0, a, b); |
800 | assert_eq_m512d(r, _mm512_setzero_pd()); |
801 | let r = _mm512_maskz_add_pd(0b00001111, a, b); |
802 | let e = _mm512_setr_pd(1., 2., 0., f64::MAX, 0., 0., 0., 0.); |
803 | assert_eq_m512d(r, e); |
804 | } |
805 | |
806 | #[simd_test(enable = "avx512f,avx512vl" )] |
807 | unsafe fn test_mm256_mask_add_pd() { |
808 | let a = _mm256_set_pd(1., -1., f64::MAX, f64::MIN); |
809 | let b = _mm256_set1_pd(1.); |
810 | let r = _mm256_mask_add_pd(a, 0, a, b); |
811 | assert_eq_m256d(r, a); |
812 | let r = _mm256_mask_add_pd(a, 0b00001111, a, b); |
813 | let e = _mm256_set_pd(2., 0., f64::MAX, f64::MIN + 1.); |
814 | assert_eq_m256d(r, e); |
815 | } |
816 | |
817 | #[simd_test(enable = "avx512f,avx512vl" )] |
818 | unsafe fn test_mm256_maskz_add_pd() { |
819 | let a = _mm256_set_pd(1., -1., f64::MAX, f64::MIN); |
820 | let b = _mm256_set1_pd(1.); |
821 | let r = _mm256_maskz_add_pd(0, a, b); |
822 | assert_eq_m256d(r, _mm256_setzero_pd()); |
823 | let r = _mm256_maskz_add_pd(0b00001111, a, b); |
824 | let e = _mm256_set_pd(2., 0., f64::MAX, f64::MIN + 1.); |
825 | assert_eq_m256d(r, e); |
826 | } |
827 | |
828 | #[simd_test(enable = "avx512f,avx512vl" )] |
829 | unsafe fn test_mm_mask_add_pd() { |
830 | let a = _mm_set_pd(f64::MAX, f64::MIN); |
831 | let b = _mm_set1_pd(1.); |
832 | let r = _mm_mask_add_pd(a, 0, a, b); |
833 | assert_eq_m128d(r, a); |
834 | let r = _mm_mask_add_pd(a, 0b00000011, a, b); |
835 | let e = _mm_set_pd(f64::MAX, f64::MIN + 1.); |
836 | assert_eq_m128d(r, e); |
837 | } |
838 | |
839 | #[simd_test(enable = "avx512f,avx512vl" )] |
840 | unsafe fn test_mm_maskz_add_pd() { |
841 | let a = _mm_set_pd(f64::MAX, f64::MIN); |
842 | let b = _mm_set1_pd(1.); |
843 | let r = _mm_maskz_add_pd(0, a, b); |
844 | assert_eq_m128d(r, _mm_setzero_pd()); |
845 | let r = _mm_maskz_add_pd(0b00000011, a, b); |
846 | let e = _mm_set_pd(f64::MAX, f64::MIN + 1.); |
847 | assert_eq_m128d(r, e); |
848 | } |
849 | |
850 | #[simd_test(enable = "avx512f" )] |
851 | unsafe fn test_mm512_sub_epi64() { |
852 | let a = _mm512_setr_epi64(0, 1, -1, i64::MAX, i64::MIN, 100, -100, -32); |
853 | let b = _mm512_set1_epi64(1); |
854 | let r = _mm512_sub_epi64(a, b); |
855 | let e = _mm512_setr_epi64(-1, 0, -2, i64::MAX - 1, i64::MAX, 99, -101, -33); |
856 | assert_eq_m512i(r, e); |
857 | } |
858 | |
859 | #[simd_test(enable = "avx512f" )] |
860 | unsafe fn test_mm512_mask_sub_epi64() { |
861 | let a = _mm512_setr_epi64(0, 1, -1, i64::MAX, i64::MIN, 100, -100, -32); |
862 | let b = _mm512_set1_epi64(1); |
863 | let r = _mm512_mask_sub_epi64(a, 0, a, b); |
864 | assert_eq_m512i(r, a); |
865 | let r = _mm512_mask_sub_epi64(a, 0b00001111, a, b); |
866 | let e = _mm512_setr_epi64(-1, 0, -2, i64::MAX - 1, i64::MIN, 100, -100, -32); |
867 | assert_eq_m512i(r, e); |
868 | } |
869 | |
870 | #[simd_test(enable = "avx512f" )] |
871 | unsafe fn test_mm512_maskz_sub_epi64() { |
872 | let a = _mm512_setr_epi64(0, 1, -1, i64::MAX, i64::MIN, 100, -100, -32); |
873 | let b = _mm512_set1_epi64(1); |
874 | let r = _mm512_maskz_sub_epi64(0, a, b); |
875 | assert_eq_m512i(r, _mm512_setzero_si512()); |
876 | let r = _mm512_maskz_sub_epi64(0b00001111, a, b); |
877 | let e = _mm512_setr_epi64(-1, 0, -2, i64::MAX - 1, 0, 0, 0, 0); |
878 | assert_eq_m512i(r, e); |
879 | } |
880 | |
881 | #[simd_test(enable = "avx512f,avx512vl" )] |
882 | unsafe fn test_mm256_mask_sub_epi64() { |
883 | let a = _mm256_set_epi64x(1, -1, i64::MAX, i64::MIN); |
884 | let b = _mm256_set1_epi64x(1); |
885 | let r = _mm256_mask_sub_epi64(a, 0, a, b); |
886 | assert_eq_m256i(r, a); |
887 | let r = _mm256_mask_sub_epi64(a, 0b00001111, a, b); |
888 | let e = _mm256_set_epi64x(0, -2, i64::MAX - 1, i64::MAX); |
889 | assert_eq_m256i(r, e); |
890 | } |
891 | |
892 | #[simd_test(enable = "avx512f,avx512vl" )] |
893 | unsafe fn test_mm256_maskz_sub_epi64() { |
894 | let a = _mm256_set_epi64x(1, -1, i64::MAX, i64::MIN); |
895 | let b = _mm256_set1_epi64x(1); |
896 | let r = _mm256_maskz_sub_epi64(0, a, b); |
897 | assert_eq_m256i(r, _mm256_setzero_si256()); |
898 | let r = _mm256_maskz_sub_epi64(0b00001111, a, b); |
899 | let e = _mm256_set_epi64x(0, -2, i64::MAX - 1, i64::MAX); |
900 | assert_eq_m256i(r, e); |
901 | } |
902 | |
903 | #[simd_test(enable = "avx512f,avx512vl" )] |
904 | unsafe fn test_mm_mask_sub_epi64() { |
905 | let a = _mm_set_epi64x(i64::MAX, i64::MIN); |
906 | let b = _mm_set1_epi64x(1); |
907 | let r = _mm_mask_sub_epi64(a, 0, a, b); |
908 | assert_eq_m128i(r, a); |
909 | let r = _mm_mask_sub_epi64(a, 0b00000011, a, b); |
910 | let e = _mm_set_epi64x(i64::MAX - 1, i64::MAX); |
911 | assert_eq_m128i(r, e); |
912 | } |
913 | |
914 | #[simd_test(enable = "avx512f,avx512vl" )] |
915 | unsafe fn test_mm_maskz_sub_epi64() { |
916 | let a = _mm_set_epi64x(i64::MAX, i64::MIN); |
917 | let b = _mm_set1_epi64x(1); |
918 | let r = _mm_maskz_sub_epi64(0, a, b); |
919 | assert_eq_m128i(r, _mm_setzero_si128()); |
920 | let r = _mm_maskz_sub_epi64(0b00000011, a, b); |
921 | let e = _mm_set_epi64x(i64::MAX - 1, i64::MAX); |
922 | assert_eq_m128i(r, e); |
923 | } |
924 | |
925 | #[simd_test(enable = "avx512f" )] |
926 | unsafe fn test_mm512_sub_pd() { |
927 | let a = _mm512_setr_pd(0., 1., -1., f64::MAX, f64::MIN, 100., -100., -32.); |
928 | let b = _mm512_set1_pd(1.); |
929 | let r = _mm512_sub_pd(a, b); |
930 | let e = _mm512_setr_pd(-1., 0., -2., f64::MAX - 1., f64::MIN, 99., -101., -33.); |
931 | assert_eq_m512d(r, e); |
932 | } |
933 | |
934 | #[simd_test(enable = "avx512f" )] |
935 | unsafe fn test_mm512_mask_sub_pd() { |
936 | let a = _mm512_setr_pd(0., 1., -1., f64::MAX, f64::MIN, 100., -100., -32.); |
937 | let b = _mm512_set1_pd(1.); |
938 | let r = _mm512_mask_sub_pd(a, 0, a, b); |
939 | assert_eq_m512d(r, a); |
940 | let r = _mm512_mask_sub_pd(a, 0b00001111, a, b); |
941 | let e = _mm512_setr_pd(-1., 0., -2., f64::MAX - 1., f64::MIN, 100., -100., -32.); |
942 | assert_eq_m512d(r, e); |
943 | } |
944 | |
945 | #[simd_test(enable = "avx512f" )] |
946 | unsafe fn test_mm512_maskz_sub_pd() { |
947 | let a = _mm512_setr_pd(0., 1., -1., f64::MAX, f64::MIN, 100., -100., -32.); |
948 | let b = _mm512_set1_pd(1.); |
949 | let r = _mm512_maskz_sub_pd(0, a, b); |
950 | assert_eq_m512d(r, _mm512_setzero_pd()); |
951 | let r = _mm512_maskz_sub_pd(0b00001111, a, b); |
952 | let e = _mm512_setr_pd(-1., 0., -2., f64::MAX - 1., 0., 0., 0., 0.); |
953 | assert_eq_m512d(r, e); |
954 | } |
955 | |
956 | #[simd_test(enable = "avx512f,avx512vl" )] |
957 | unsafe fn test_mm256_mask_sub_pd() { |
958 | let a = _mm256_set_pd(1., -1., f64::MAX, f64::MIN); |
959 | let b = _mm256_set1_pd(1.); |
960 | let r = _mm256_mask_sub_pd(a, 0, a, b); |
961 | assert_eq_m256d(r, a); |
962 | let r = _mm256_mask_sub_pd(a, 0b00001111, a, b); |
963 | let e = _mm256_set_pd(0., -2., f64::MAX - 1., f64::MIN); |
964 | assert_eq_m256d(r, e); |
965 | } |
966 | |
967 | #[simd_test(enable = "avx512f,avx512vl" )] |
968 | unsafe fn test_mm256_maskz_sub_pd() { |
969 | let a = _mm256_set_pd(1., -1., f64::MAX, f64::MIN); |
970 | let b = _mm256_set1_pd(1.); |
971 | let r = _mm256_maskz_sub_pd(0, a, b); |
972 | assert_eq_m256d(r, _mm256_setzero_pd()); |
973 | let r = _mm256_maskz_sub_pd(0b00001111, a, b); |
974 | let e = _mm256_set_pd(0., -2., f64::MAX - 1., f64::MIN); |
975 | assert_eq_m256d(r, e); |
976 | } |
977 | |
978 | #[simd_test(enable = "avx512f,avx512vl" )] |
979 | unsafe fn test_mm_mask_sub_pd() { |
980 | let a = _mm_set_pd(f64::MAX, f64::MIN); |
981 | let b = _mm_set1_pd(1.); |
982 | let r = _mm_mask_sub_pd(a, 0, a, b); |
983 | assert_eq_m128d(r, a); |
984 | let r = _mm_mask_sub_pd(a, 0b00000011, a, b); |
985 | let e = _mm_set_pd(f64::MAX - 1., f64::MIN); |
986 | assert_eq_m128d(r, e); |
987 | } |
988 | |
989 | #[simd_test(enable = "avx512f,avx512vl" )] |
990 | unsafe fn test_mm_maskz_sub_pd() { |
991 | let a = _mm_set_pd(f64::MAX, f64::MIN); |
992 | let b = _mm_set1_pd(1.); |
993 | let r = _mm_maskz_sub_pd(0, a, b); |
994 | assert_eq_m128d(r, _mm_setzero_pd()); |
995 | let r = _mm_maskz_sub_pd(0b00000011, a, b); |
996 | let e = _mm_set_pd(f64::MAX - 1., f64::MIN); |
997 | assert_eq_m128d(r, e); |
998 | } |
999 | |
1000 | #[simd_test(enable = "avx512f" )] |
1001 | unsafe fn test_mm512_mul_epi32() { |
1002 | let a = _mm512_set1_epi32(1); |
1003 | let b = _mm512_setr_epi32(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
1004 | let r = _mm512_mul_epi32(a, b); |
1005 | let e = _mm512_set_epi64(15, 13, 11, 9, 7, 5, 3, 1); |
1006 | assert_eq_m512i(r, e); |
1007 | } |
1008 | |
1009 | #[simd_test(enable = "avx512f" )] |
1010 | unsafe fn test_mm512_mask_mul_epi32() { |
1011 | let a = _mm512_set1_epi32(1); |
1012 | let b = _mm512_setr_epi32(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
1013 | let r = _mm512_mask_mul_epi32(a, 0, a, b); |
1014 | assert_eq_m512i(r, a); |
1015 | let r = _mm512_mask_mul_epi32(a, 0b00001111, a, b); |
1016 | #[rustfmt::skip] |
1017 | let e = _mm512_set_epi64( |
1018 | 1 | 1 << 32, 1 | 1 << 32, 1 | 1 << 32, 1 | 1 << 32, |
1019 | 7, 5, 3, 1, |
1020 | ); |
1021 | assert_eq_m512i(r, e); |
1022 | } |
1023 | |
1024 | #[simd_test(enable = "avx512f" )] |
1025 | unsafe fn test_mm512_maskz_mul_epi32() { |
1026 | let a = _mm512_set1_epi32(1); |
1027 | let b = _mm512_setr_epi32(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
1028 | let r = _mm512_maskz_mul_epi32(0, a, b); |
1029 | assert_eq_m512i(r, _mm512_setzero_si512()); |
1030 | let r = _mm512_maskz_mul_epi32(0b00001111, a, b); |
1031 | let e = _mm512_set_epi64(0, 0, 0, 0, 7, 5, 3, 1); |
1032 | assert_eq_m512i(r, e); |
1033 | } |
1034 | |
1035 | #[simd_test(enable = "avx512f,avx512vl" )] |
1036 | unsafe fn test_mm256_mask_mul_epi32() { |
1037 | let a = _mm256_set1_epi32(1); |
1038 | let b = _mm256_set_epi32(1, 2, 3, 4, 5, 6, 7, 8); |
1039 | let r = _mm256_mask_mul_epi32(a, 0, a, b); |
1040 | assert_eq_m256i(r, a); |
1041 | let r = _mm256_mask_mul_epi32(a, 0b00001111, a, b); |
1042 | let e = _mm256_set_epi64x(2, 4, 6, 8); |
1043 | assert_eq_m256i(r, e); |
1044 | } |
1045 | |
1046 | #[simd_test(enable = "avx512f,avx512vl" )] |
1047 | unsafe fn test_mm256_maskz_mul_epi32() { |
1048 | let a = _mm256_set1_epi32(1); |
1049 | let b = _mm256_set_epi32(1, 2, 3, 4, 5, 6, 7, 8); |
1050 | let r = _mm256_maskz_mul_epi32(0, a, b); |
1051 | assert_eq_m256i(r, _mm256_setzero_si256()); |
1052 | let r = _mm256_maskz_mul_epi32(0b00001111, a, b); |
1053 | let e = _mm256_set_epi64x(2, 4, 6, 8); |
1054 | assert_eq_m256i(r, e); |
1055 | } |
1056 | |
1057 | #[simd_test(enable = "avx512f,avx512vl" )] |
1058 | unsafe fn test_mm_mask_mul_epi32() { |
1059 | let a = _mm_set1_epi32(1); |
1060 | let b = _mm_set_epi32(1, 2, 3, 4); |
1061 | let r = _mm_mask_mul_epi32(a, 0, a, b); |
1062 | assert_eq_m128i(r, a); |
1063 | let r = _mm_mask_mul_epi32(a, 0b00000011, a, b); |
1064 | let e = _mm_set_epi64x(2, 4); |
1065 | assert_eq_m128i(r, e); |
1066 | } |
1067 | |
1068 | #[simd_test(enable = "avx512f,avx512vl" )] |
1069 | unsafe fn test_mm_maskz_mul_epi32() { |
1070 | let a = _mm_set1_epi32(1); |
1071 | let b = _mm_set_epi32(1, 2, 3, 4); |
1072 | let r = _mm_maskz_mul_epi32(0, a, b); |
1073 | assert_eq_m128i(r, _mm_setzero_si128()); |
1074 | let r = _mm_maskz_mul_epi32(0b00000011, a, b); |
1075 | let e = _mm_set_epi64x(2, 4); |
1076 | assert_eq_m128i(r, e); |
1077 | } |
1078 | |
1079 | #[simd_test(enable = "avx512f" )] |
1080 | unsafe fn test_mm512_mul_epu32() { |
1081 | let a = _mm512_set1_epi32(1); |
1082 | let b = _mm512_setr_epi32(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
1083 | let r = _mm512_mul_epu32(a, b); |
1084 | let e = _mm512_set_epi64(15, 13, 11, 9, 7, 5, 3, 1); |
1085 | assert_eq_m512i(r, e); |
1086 | } |
1087 | |
1088 | #[simd_test(enable = "avx512f" )] |
1089 | unsafe fn test_mm512_mask_mul_epu32() { |
1090 | let a = _mm512_set1_epi32(1); |
1091 | let b = _mm512_setr_epi32(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
1092 | let r = _mm512_mask_mul_epu32(a, 0, a, b); |
1093 | assert_eq_m512i(r, a); |
1094 | let r = _mm512_mask_mul_epu32(a, 0b00001111, a, b); |
1095 | #[rustfmt::skip] |
1096 | let e = _mm512_set_epi64( |
1097 | 1 | 1 << 32, 1 | 1 << 32, 1 | 1 << 32, 1 | 1 << 32, |
1098 | 7, 5, 3, 1, |
1099 | ); |
1100 | assert_eq_m512i(r, e); |
1101 | } |
1102 | |
1103 | #[simd_test(enable = "avx512f" )] |
1104 | unsafe fn test_mm512_maskz_mul_epu32() { |
1105 | let a = _mm512_set1_epi32(1); |
1106 | let b = _mm512_setr_epi32(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16); |
1107 | let r = _mm512_maskz_mul_epu32(0, a, b); |
1108 | assert_eq_m512i(r, _mm512_setzero_si512()); |
1109 | let r = _mm512_maskz_mul_epu32(0b00001111, a, b); |
1110 | let e = _mm512_set_epi64(0, 0, 0, 0, 7, 5, 3, 1); |
1111 | assert_eq_m512i(r, e); |
1112 | } |
1113 | |
1114 | #[simd_test(enable = "avx512f,avx512vl" )] |
1115 | unsafe fn test_mm256_mask_mul_epu32() { |
1116 | let a = _mm256_set1_epi32(1); |
1117 | let b = _mm256_set_epi32(1, 2, 3, 4, 5, 6, 7, 8); |
1118 | let r = _mm256_mask_mul_epu32(a, 0, a, b); |
1119 | assert_eq_m256i(r, a); |
1120 | let r = _mm256_mask_mul_epu32(a, 0b00001111, a, b); |
1121 | let e = _mm256_set_epi64x(2, 4, 6, 8); |
1122 | assert_eq_m256i(r, e); |
1123 | } |
1124 | |
1125 | #[simd_test(enable = "avx512f,avx512vl" )] |
1126 | unsafe fn test_mm256_maskz_mul_epu32() { |
1127 | let a = _mm256_set1_epi32(1); |
1128 | let b = _mm256_set_epi32(1, 2, 3, 4, 5, 6, 7, 8); |
1129 | let r = _mm256_maskz_mul_epu32(0, a, b); |
1130 | assert_eq_m256i(r, _mm256_setzero_si256()); |
1131 | let r = _mm256_maskz_mul_epu32(0b00001111, a, b); |
1132 | let e = _mm256_set_epi64x(2, 4, 6, 8); |
1133 | assert_eq_m256i(r, e); |
1134 | } |
1135 | |
1136 | #[simd_test(enable = "avx512f,avx512vl" )] |
1137 | unsafe fn test_mm_mask_mul_epu32() { |
1138 | let a = _mm_set1_epi32(1); |
1139 | let b = _mm_set_epi32(1, 2, 3, 4); |
1140 | let r = _mm_mask_mul_epu32(a, 0, a, b); |
1141 | assert_eq_m128i(r, a); |
1142 | let r = _mm_mask_mul_epu32(a, 0b00000011, a, b); |
1143 | let e = _mm_set_epi64x(2, 4); |
1144 | assert_eq_m128i(r, e); |
1145 | } |
1146 | |
1147 | #[simd_test(enable = "avx512f,avx512vl" )] |
1148 | unsafe fn test_mm_maskz_mul_epu32() { |
1149 | let a = _mm_set1_epi32(1); |
1150 | let b = _mm_set_epi32(1, 2, 3, 4); |
1151 | let r = _mm_maskz_mul_epu32(0, a, b); |
1152 | assert_eq_m128i(r, _mm_setzero_si128()); |
1153 | let r = _mm_maskz_mul_epu32(0b00000011, a, b); |
1154 | let e = _mm_set_epi64x(2, 4); |
1155 | assert_eq_m128i(r, e); |
1156 | } |
1157 | |
1158 | #[simd_test(enable = "avx512f" )] |
1159 | unsafe fn test_mm512_mullox_epi64() { |
1160 | let a = _mm512_setr_epi64(0, 1, i64::MAX, i64::MIN, i64::MAX, 100, -100, -32); |
1161 | let b = _mm512_set1_epi64(2); |
1162 | let r = _mm512_mullox_epi64(a, b); |
1163 | let e = _mm512_setr_epi64(0, 2, -2, 0, -2, 200, -200, -64); |
1164 | assert_eq_m512i(r, e); |
1165 | } |
1166 | |
1167 | #[simd_test(enable = "avx512f" )] |
1168 | unsafe fn test_mm512_mask_mullox_epi64() { |
1169 | let a = _mm512_setr_epi64(0, 1, i64::MAX, i64::MIN, i64::MAX, 100, -100, -32); |
1170 | let b = _mm512_set1_epi64(2); |
1171 | let r = _mm512_mask_mullox_epi64(a, 0, a, b); |
1172 | assert_eq_m512i(r, a); |
1173 | let r = _mm512_mask_mullox_epi64(a, 0b00001111, a, b); |
1174 | let e = _mm512_setr_epi64(0, 2, -2, 0, i64::MAX, 100, -100, -32); |
1175 | assert_eq_m512i(r, e); |
1176 | } |
1177 | |
1178 | #[simd_test(enable = "avx512f" )] |
1179 | unsafe fn test_mm512_mul_pd() { |
1180 | let a = _mm512_setr_pd(0., 1., f64::MAX, f64::MIN, f64::MAX, f64::MIN, -100., -32.); |
1181 | let b = _mm512_set1_pd(2.); |
1182 | let r = _mm512_mul_pd(a, b); |
1183 | #[rustfmt::skip] |
1184 | let e = _mm512_setr_pd( |
1185 | 0., 2., f64::INFINITY, f64::NEG_INFINITY, |
1186 | f64::INFINITY, f64::NEG_INFINITY, -200., -64., |
1187 | ); |
1188 | assert_eq_m512d(r, e); |
1189 | } |
1190 | |
1191 | #[simd_test(enable = "avx512f" )] |
1192 | unsafe fn test_mm512_mask_mul_pd() { |
1193 | let a = _mm512_setr_pd(0., 1., f64::MAX, f64::MIN, f64::MAX, f64::MIN, -100., -32.); |
1194 | let b = _mm512_set1_pd(2.); |
1195 | let r = _mm512_mask_mul_pd(a, 0, a, b); |
1196 | assert_eq_m512d(r, a); |
1197 | let r = _mm512_mask_mul_pd(a, 0b00001111, a, b); |
1198 | #[rustfmt::skip] |
1199 | let e = _mm512_setr_pd( |
1200 | 0., 2., f64::INFINITY, f64::NEG_INFINITY, |
1201 | f64::MAX, f64::MIN, -100., -32., |
1202 | ); |
1203 | assert_eq_m512d(r, e); |
1204 | } |
1205 | |
1206 | #[simd_test(enable = "avx512f" )] |
1207 | unsafe fn test_mm512_maskz_mul_pd() { |
1208 | let a = _mm512_setr_pd(0., 1., f64::MAX, f64::MIN, f64::MAX, f64::MIN, -100., -32.); |
1209 | let b = _mm512_set1_pd(2.); |
1210 | let r = _mm512_maskz_mul_pd(0, a, b); |
1211 | assert_eq_m512d(r, _mm512_setzero_pd()); |
1212 | let r = _mm512_maskz_mul_pd(0b00001111, a, b); |
1213 | let e = _mm512_setr_pd(0., 2., f64::INFINITY, f64::NEG_INFINITY, 0., 0., 0., 0.); |
1214 | assert_eq_m512d(r, e); |
1215 | } |
1216 | |
1217 | #[simd_test(enable = "avx512f,avx512vl" )] |
1218 | unsafe fn test_mm256_mask_mul_pd() { |
1219 | let a = _mm256_set_pd(0., 1., f64::MAX, f64::MIN); |
1220 | let b = _mm256_set1_pd(2.); |
1221 | let r = _mm256_mask_mul_pd(a, 0, a, b); |
1222 | assert_eq_m256d(r, a); |
1223 | let r = _mm256_mask_mul_pd(a, 0b00001111, a, b); |
1224 | let e = _mm256_set_pd(0., 2., f64::INFINITY, f64::NEG_INFINITY); |
1225 | assert_eq_m256d(r, e); |
1226 | } |
1227 | |
1228 | #[simd_test(enable = "avx512f,avx512vl" )] |
1229 | unsafe fn test_mm256_maskz_mul_pd() { |
1230 | let a = _mm256_set_pd(0., 1., f64::MAX, f64::MIN); |
1231 | let b = _mm256_set1_pd(2.); |
1232 | let r = _mm256_maskz_mul_pd(0, a, b); |
1233 | assert_eq_m256d(r, _mm256_setzero_pd()); |
1234 | let r = _mm256_maskz_mul_pd(0b00001111, a, b); |
1235 | let e = _mm256_set_pd(0., 2., f64::INFINITY, f64::NEG_INFINITY); |
1236 | assert_eq_m256d(r, e); |
1237 | } |
1238 | |
1239 | #[simd_test(enable = "avx512f,avx512vl" )] |
1240 | unsafe fn test_mm_mask_mul_pd() { |
1241 | let a = _mm_set_pd(f64::MAX, f64::MIN); |
1242 | let b = _mm_set1_pd(2.); |
1243 | let r = _mm_mask_mul_pd(a, 0, a, b); |
1244 | assert_eq_m128d(r, a); |
1245 | let r = _mm_mask_mul_pd(a, 0b00000011, a, b); |
1246 | let e = _mm_set_pd(f64::INFINITY, f64::NEG_INFINITY); |
1247 | assert_eq_m128d(r, e); |
1248 | } |
1249 | |
1250 | #[simd_test(enable = "avx512f,avx512vl" )] |
1251 | unsafe fn test_mm_maskz_mul_pd() { |
1252 | let a = _mm_set_pd(f64::MAX, f64::MIN); |
1253 | let b = _mm_set1_pd(2.); |
1254 | let r = _mm_maskz_mul_pd(0, a, b); |
1255 | assert_eq_m128d(r, _mm_setzero_pd()); |
1256 | let r = _mm_maskz_mul_pd(0b00000011, a, b); |
1257 | let e = _mm_set_pd(f64::INFINITY, f64::NEG_INFINITY); |
1258 | assert_eq_m128d(r, e); |
1259 | } |
1260 | |
1261 | #[simd_test(enable = "avx512f" )] |
1262 | unsafe fn test_mm512_div_pd() { |
1263 | let a = _mm512_setr_pd(0., 1., f64::MAX, f64::MIN, f64::MAX, f64::MIN, -100., -32.); |
1264 | let b = _mm512_setr_pd(2., 2., 0., 0., 0., 0., 2., 2.); |
1265 | let r = _mm512_div_pd(a, b); |
1266 | #[rustfmt::skip] |
1267 | let e = _mm512_setr_pd( |
1268 | 0., 0.5, f64::INFINITY, f64::NEG_INFINITY, |
1269 | f64::INFINITY, f64::NEG_INFINITY, -50., -16., |
1270 | ); |
1271 | assert_eq_m512d(r, e); |
1272 | } |
1273 | |
1274 | #[simd_test(enable = "avx512f" )] |
1275 | unsafe fn test_mm512_mask_div_pd() { |
1276 | let a = _mm512_setr_pd(0., 1., f64::MAX, f64::MIN, f64::MAX, f64::MIN, -100., -32.); |
1277 | let b = _mm512_setr_pd(2., 2., 0., 0., 0., 0., 2., 2.); |
1278 | let r = _mm512_mask_div_pd(a, 0, a, b); |
1279 | assert_eq_m512d(r, a); |
1280 | let r = _mm512_mask_div_pd(a, 0b00001111, a, b); |
1281 | #[rustfmt::skip] |
1282 | let e = _mm512_setr_pd( |
1283 | 0., 0.5, f64::INFINITY, f64::NEG_INFINITY, |
1284 | f64::MAX, f64::MIN, -100., -32., |
1285 | ); |
1286 | assert_eq_m512d(r, e); |
1287 | } |
1288 | |
1289 | #[simd_test(enable = "avx512f" )] |
1290 | unsafe fn test_mm512_maskz_div_pd() { |
1291 | let a = _mm512_setr_pd(0., 1., f64::MAX, f64::MIN, f64::MAX, f64::MIN, -100., -32.); |
1292 | let b = _mm512_setr_pd(2., 2., 0., 0., 0., 0., 2., 2.); |
1293 | let r = _mm512_maskz_div_pd(0, a, b); |
1294 | assert_eq_m512d(r, _mm512_setzero_pd()); |
1295 | let r = _mm512_maskz_div_pd(0b00001111, a, b); |
1296 | let e = _mm512_setr_pd(0., 0.5, f64::INFINITY, f64::NEG_INFINITY, 0., 0., 0., 0.); |
1297 | assert_eq_m512d(r, e); |
1298 | } |
1299 | |
1300 | #[simd_test(enable = "avx512f,avx512vl" )] |
1301 | unsafe fn test_mm256_mask_div_pd() { |
1302 | let a = _mm256_set_pd(0., 1., f64::MAX, f64::MIN); |
1303 | let b = _mm256_set_pd(2., 2., 0., 0.); |
1304 | let r = _mm256_mask_div_pd(a, 0, a, b); |
1305 | assert_eq_m256d(r, a); |
1306 | let r = _mm256_mask_div_pd(a, 0b00001111, a, b); |
1307 | let e = _mm256_set_pd(0., 0.5, f64::INFINITY, f64::NEG_INFINITY); |
1308 | assert_eq_m256d(r, e); |
1309 | } |
1310 | |
1311 | #[simd_test(enable = "avx512f,avx512vl" )] |
1312 | unsafe fn test_mm256_maskz_div_pd() { |
1313 | let a = _mm256_set_pd(0., 1., f64::MAX, f64::MIN); |
1314 | let b = _mm256_set_pd(2., 2., 0., 0.); |
1315 | let r = _mm256_maskz_div_pd(0, a, b); |
1316 | assert_eq_m256d(r, _mm256_setzero_pd()); |
1317 | let r = _mm256_maskz_div_pd(0b00001111, a, b); |
1318 | let e = _mm256_set_pd(0., 0.5, f64::INFINITY, f64::NEG_INFINITY); |
1319 | assert_eq_m256d(r, e); |
1320 | } |
1321 | |
1322 | #[simd_test(enable = "avx512f,avx512vl" )] |
1323 | unsafe fn test_mm_mask_div_pd() { |
1324 | let a = _mm_set_pd(f64::MAX, f64::MIN); |
1325 | let b = _mm_set_pd(0., 0.); |
1326 | let r = _mm_mask_div_pd(a, 0, a, b); |
1327 | assert_eq_m128d(r, a); |
1328 | let r = _mm_mask_div_pd(a, 0b00000011, a, b); |
1329 | let e = _mm_set_pd(f64::INFINITY, f64::NEG_INFINITY); |
1330 | assert_eq_m128d(r, e); |
1331 | } |
1332 | |
1333 | #[simd_test(enable = "avx512f,avx512vl" )] |
1334 | unsafe fn test_mm_maskz_div_pd() { |
1335 | let a = _mm_set_pd(f64::MAX, f64::MIN); |
1336 | let b = _mm_set_pd(0., 0.); |
1337 | let r = _mm_maskz_div_pd(0, a, b); |
1338 | assert_eq_m128d(r, _mm_setzero_pd()); |
1339 | let r = _mm_maskz_div_pd(0b00000011, a, b); |
1340 | let e = _mm_set_pd(f64::INFINITY, f64::NEG_INFINITY); |
1341 | assert_eq_m128d(r, e); |
1342 | } |
1343 | |
1344 | #[simd_test(enable = "avx512f" )] |
1345 | unsafe fn test_mm512_max_epi64() { |
1346 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1347 | let b = _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0); |
1348 | let r = _mm512_max_epi64(a, b); |
1349 | let e = _mm512_setr_epi64(7, 6, 5, 4, 4, 5, 6, 7); |
1350 | assert_eq_m512i(r, e); |
1351 | } |
1352 | |
1353 | #[simd_test(enable = "avx512f" )] |
1354 | unsafe fn test_mm512_mask_max_epi64() { |
1355 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1356 | let b = _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0); |
1357 | let r = _mm512_mask_max_epi64(a, 0, a, b); |
1358 | assert_eq_m512i(r, a); |
1359 | let r = _mm512_mask_max_epi64(a, 0b00001111, a, b); |
1360 | let e = _mm512_setr_epi64(7, 6, 5, 4, 4, 5, 6, 7); |
1361 | assert_eq_m512i(r, e); |
1362 | } |
1363 | |
1364 | #[simd_test(enable = "avx512f" )] |
1365 | unsafe fn test_mm512_maskz_max_epi64() { |
1366 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1367 | let b = _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0); |
1368 | let r = _mm512_maskz_max_epi64(0, a, b); |
1369 | assert_eq_m512i(r, _mm512_setzero_si512()); |
1370 | let r = _mm512_maskz_max_epi64(0b00001111, a, b); |
1371 | let e = _mm512_setr_epi64(7, 6, 5, 4, 0, 0, 0, 0); |
1372 | assert_eq_m512i(r, e); |
1373 | } |
1374 | |
1375 | #[simd_test(enable = "avx512f,avx512vl" )] |
1376 | unsafe fn test_mm256_max_epi64() { |
1377 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
1378 | let b = _mm256_set_epi64x(3, 2, 1, 0); |
1379 | let r = _mm256_max_epi64(a, b); |
1380 | let e = _mm256_set_epi64x(3, 2, 2, 3); |
1381 | assert_eq_m256i(r, e); |
1382 | } |
1383 | |
1384 | #[simd_test(enable = "avx512f,avx512vl" )] |
1385 | unsafe fn test_mm256_mask_max_epi64() { |
1386 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
1387 | let b = _mm256_set_epi64x(3, 2, 1, 0); |
1388 | let r = _mm256_mask_max_epi64(a, 0, a, b); |
1389 | assert_eq_m256i(r, a); |
1390 | let r = _mm256_mask_max_epi64(a, 0b00001111, a, b); |
1391 | let e = _mm256_set_epi64x(3, 2, 2, 3); |
1392 | assert_eq_m256i(r, e); |
1393 | } |
1394 | |
1395 | #[simd_test(enable = "avx512f,avx512vl" )] |
1396 | unsafe fn test_mm256_maskz_max_epi64() { |
1397 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
1398 | let b = _mm256_set_epi64x(3, 2, 1, 0); |
1399 | let r = _mm256_maskz_max_epi64(0, a, b); |
1400 | assert_eq_m256i(r, _mm256_setzero_si256()); |
1401 | let r = _mm256_maskz_max_epi64(0b00001111, a, b); |
1402 | let e = _mm256_set_epi64x(3, 2, 2, 3); |
1403 | assert_eq_m256i(r, e); |
1404 | } |
1405 | |
1406 | #[simd_test(enable = "avx512f,avx512vl" )] |
1407 | unsafe fn test_mm_max_epi64() { |
1408 | let a = _mm_set_epi64x(2, 3); |
1409 | let b = _mm_set_epi64x(3, 2); |
1410 | let r = _mm_max_epi64(a, b); |
1411 | let e = _mm_set_epi64x(3, 3); |
1412 | assert_eq_m128i(r, e); |
1413 | } |
1414 | |
1415 | #[simd_test(enable = "avx512f,avx512vl" )] |
1416 | unsafe fn test_mm_mask_max_epi64() { |
1417 | let a = _mm_set_epi64x(2, 3); |
1418 | let b = _mm_set_epi64x(3, 2); |
1419 | let r = _mm_mask_max_epi64(a, 0, a, b); |
1420 | assert_eq_m128i(r, a); |
1421 | let r = _mm_mask_max_epi64(a, 0b00000011, a, b); |
1422 | let e = _mm_set_epi64x(3, 3); |
1423 | assert_eq_m128i(r, e); |
1424 | } |
1425 | |
1426 | #[simd_test(enable = "avx512f,avx512vl" )] |
1427 | unsafe fn test_mm_maskz_max_epi64() { |
1428 | let a = _mm_set_epi64x(2, 3); |
1429 | let b = _mm_set_epi64x(3, 2); |
1430 | let r = _mm_maskz_max_epi64(0, a, b); |
1431 | assert_eq_m128i(r, _mm_setzero_si128()); |
1432 | let r = _mm_maskz_max_epi64(0b00000011, a, b); |
1433 | let e = _mm_set_epi64x(3, 3); |
1434 | assert_eq_m128i(r, e); |
1435 | } |
1436 | |
1437 | #[simd_test(enable = "avx512f" )] |
1438 | unsafe fn test_mm512_max_pd() { |
1439 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
1440 | let b = _mm512_setr_pd(7., 6., 5., 4., 3., 2., 1., 0.); |
1441 | let r = _mm512_max_pd(a, b); |
1442 | let e = _mm512_setr_pd(7., 6., 5., 4., 4., 5., 6., 7.); |
1443 | assert_eq_m512d(r, e); |
1444 | } |
1445 | |
1446 | #[simd_test(enable = "avx512f" )] |
1447 | unsafe fn test_mm512_mask_max_pd() { |
1448 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
1449 | let b = _mm512_setr_pd(7., 6., 5., 4., 3., 2., 1., 0.); |
1450 | let r = _mm512_mask_max_pd(a, 0, a, b); |
1451 | assert_eq_m512d(r, a); |
1452 | let r = _mm512_mask_max_pd(a, 0b00001111, a, b); |
1453 | let e = _mm512_setr_pd(7., 6., 5., 4., 4., 5., 6., 7.); |
1454 | assert_eq_m512d(r, e); |
1455 | } |
1456 | |
1457 | #[simd_test(enable = "avx512f" )] |
1458 | unsafe fn test_mm512_maskz_max_pd() { |
1459 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
1460 | let b = _mm512_setr_pd(7., 6., 5., 4., 3., 2., 1., 0.); |
1461 | let r = _mm512_maskz_max_pd(0, a, b); |
1462 | assert_eq_m512d(r, _mm512_setzero_pd()); |
1463 | let r = _mm512_maskz_max_pd(0b00001111, a, b); |
1464 | let e = _mm512_setr_pd(7., 6., 5., 4., 0., 0., 0., 0.); |
1465 | assert_eq_m512d(r, e); |
1466 | } |
1467 | |
1468 | #[simd_test(enable = "avx512f,avx512vl" )] |
1469 | unsafe fn test_mm256_mask_max_pd() { |
1470 | let a = _mm256_set_pd(0., 1., 2., 3.); |
1471 | let b = _mm256_set_pd(3., 2., 1., 0.); |
1472 | let r = _mm256_mask_max_pd(a, 0, a, b); |
1473 | assert_eq_m256d(r, a); |
1474 | let r = _mm256_mask_max_pd(a, 0b00001111, a, b); |
1475 | let e = _mm256_set_pd(3., 2., 2., 3.); |
1476 | assert_eq_m256d(r, e); |
1477 | } |
1478 | |
1479 | #[simd_test(enable = "avx512f,avx512vl" )] |
1480 | unsafe fn test_mm256_maskz_max_pd() { |
1481 | let a = _mm256_set_pd(0., 1., 2., 3.); |
1482 | let b = _mm256_set_pd(3., 2., 1., 0.); |
1483 | let r = _mm256_maskz_max_pd(0, a, b); |
1484 | assert_eq_m256d(r, _mm256_setzero_pd()); |
1485 | let r = _mm256_maskz_max_pd(0b00001111, a, b); |
1486 | let e = _mm256_set_pd(3., 2., 2., 3.); |
1487 | assert_eq_m256d(r, e); |
1488 | } |
1489 | |
1490 | #[simd_test(enable = "avx512f,avx512vl" )] |
1491 | unsafe fn test_mm_mask_max_pd() { |
1492 | let a = _mm_set_pd(2., 3.); |
1493 | let b = _mm_set_pd(3., 2.); |
1494 | let r = _mm_mask_max_pd(a, 0, a, b); |
1495 | assert_eq_m128d(r, a); |
1496 | let r = _mm_mask_max_pd(a, 0b00000011, a, b); |
1497 | let e = _mm_set_pd(3., 3.); |
1498 | assert_eq_m128d(r, e); |
1499 | } |
1500 | |
1501 | #[simd_test(enable = "avx512f,avx512vl" )] |
1502 | unsafe fn test_mm_maskz_max_pd() { |
1503 | let a = _mm_set_pd(2., 3.); |
1504 | let b = _mm_set_pd(3., 2.); |
1505 | let r = _mm_maskz_max_pd(0, a, b); |
1506 | assert_eq_m128d(r, _mm_setzero_pd()); |
1507 | let r = _mm_maskz_max_pd(0b00000011, a, b); |
1508 | let e = _mm_set_pd(3., 3.); |
1509 | assert_eq_m128d(r, e); |
1510 | } |
1511 | |
1512 | #[simd_test(enable = "avx512f" )] |
1513 | unsafe fn test_mm512_max_epu64() { |
1514 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1515 | let b = _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0); |
1516 | let r = _mm512_max_epu64(a, b); |
1517 | let e = _mm512_setr_epi64(7, 6, 5, 4, 4, 5, 6, 7); |
1518 | assert_eq_m512i(r, e); |
1519 | } |
1520 | |
1521 | #[simd_test(enable = "avx512f" )] |
1522 | unsafe fn test_mm512_mask_max_epu64() { |
1523 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1524 | let b = _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0); |
1525 | let r = _mm512_mask_max_epu64(a, 0, a, b); |
1526 | assert_eq_m512i(r, a); |
1527 | let r = _mm512_mask_max_epu64(a, 0b00001111, a, b); |
1528 | let e = _mm512_setr_epi64(7, 6, 5, 4, 4, 5, 6, 7); |
1529 | assert_eq_m512i(r, e); |
1530 | } |
1531 | |
1532 | #[simd_test(enable = "avx512f" )] |
1533 | unsafe fn test_mm512_maskz_max_epu64() { |
1534 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1535 | let b = _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0); |
1536 | let r = _mm512_maskz_max_epu64(0, a, b); |
1537 | assert_eq_m512i(r, _mm512_setzero_si512()); |
1538 | let r = _mm512_maskz_max_epu64(0b00001111, a, b); |
1539 | let e = _mm512_setr_epi64(7, 6, 5, 4, 0, 0, 0, 0); |
1540 | assert_eq_m512i(r, e); |
1541 | } |
1542 | |
1543 | #[simd_test(enable = "avx512f,avx512vl" )] |
1544 | unsafe fn test_mm256_max_epu64() { |
1545 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
1546 | let b = _mm256_set_epi64x(3, 2, 1, 0); |
1547 | let r = _mm256_max_epu64(a, b); |
1548 | let e = _mm256_set_epi64x(3, 2, 2, 3); |
1549 | assert_eq_m256i(r, e); |
1550 | } |
1551 | |
1552 | #[simd_test(enable = "avx512f,avx512vl" )] |
1553 | unsafe fn test_mm256_mask_max_epu64() { |
1554 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
1555 | let b = _mm256_set_epi64x(3, 2, 1, 0); |
1556 | let r = _mm256_mask_max_epu64(a, 0, a, b); |
1557 | assert_eq_m256i(r, a); |
1558 | let r = _mm256_mask_max_epu64(a, 0b00001111, a, b); |
1559 | let e = _mm256_set_epi64x(3, 2, 2, 3); |
1560 | assert_eq_m256i(r, e); |
1561 | } |
1562 | |
1563 | #[simd_test(enable = "avx512f,avx512vl" )] |
1564 | unsafe fn test_mm256_maskz_max_epu64() { |
1565 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
1566 | let b = _mm256_set_epi64x(3, 2, 1, 0); |
1567 | let r = _mm256_maskz_max_epu64(0, a, b); |
1568 | assert_eq_m256i(r, _mm256_setzero_si256()); |
1569 | let r = _mm256_maskz_max_epu64(0b00001111, a, b); |
1570 | let e = _mm256_set_epi64x(3, 2, 2, 3); |
1571 | assert_eq_m256i(r, e); |
1572 | } |
1573 | |
1574 | #[simd_test(enable = "avx512f,avx512vl" )] |
1575 | unsafe fn test_mm_max_epu64() { |
1576 | let a = _mm_set_epi64x(2, 3); |
1577 | let b = _mm_set_epi64x(3, 2); |
1578 | let r = _mm_max_epu64(a, b); |
1579 | let e = _mm_set_epi64x(3, 3); |
1580 | assert_eq_m128i(r, e); |
1581 | } |
1582 | |
1583 | #[simd_test(enable = "avx512f,avx512vl" )] |
1584 | unsafe fn test_mm_mask_max_epu64() { |
1585 | let a = _mm_set_epi64x(2, 3); |
1586 | let b = _mm_set_epi64x(3, 2); |
1587 | let r = _mm_mask_max_epu64(a, 0, a, b); |
1588 | assert_eq_m128i(r, a); |
1589 | let r = _mm_mask_max_epu64(a, 0b00000011, a, b); |
1590 | let e = _mm_set_epi64x(3, 3); |
1591 | assert_eq_m128i(r, e); |
1592 | } |
1593 | |
1594 | #[simd_test(enable = "avx512f,avx512vl" )] |
1595 | unsafe fn test_mm_maskz_max_epu64() { |
1596 | let a = _mm_set_epi64x(2, 3); |
1597 | let b = _mm_set_epi64x(3, 2); |
1598 | let r = _mm_maskz_max_epu64(0, a, b); |
1599 | assert_eq_m128i(r, _mm_setzero_si128()); |
1600 | let r = _mm_maskz_max_epu64(0b00000011, a, b); |
1601 | let e = _mm_set_epi64x(3, 3); |
1602 | assert_eq_m128i(r, e); |
1603 | } |
1604 | |
1605 | #[simd_test(enable = "avx512f" )] |
1606 | unsafe fn test_mm512_min_epi64() { |
1607 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1608 | let b = _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0); |
1609 | let r = _mm512_min_epi64(a, b); |
1610 | let e = _mm512_setr_epi64(0, 1, 2, 3, 3, 2, 1, 0); |
1611 | assert_eq_m512i(r, e); |
1612 | } |
1613 | |
1614 | #[simd_test(enable = "avx512f" )] |
1615 | unsafe fn test_mm512_mask_min_epi64() { |
1616 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1617 | let b = _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0); |
1618 | let r = _mm512_mask_min_epi64(a, 0, a, b); |
1619 | assert_eq_m512i(r, a); |
1620 | let r = _mm512_mask_min_epi64(a, 0b00001111, a, b); |
1621 | let e = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1622 | assert_eq_m512i(r, e); |
1623 | } |
1624 | |
1625 | #[simd_test(enable = "avx512f" )] |
1626 | unsafe fn test_mm512_maskz_min_epi64() { |
1627 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1628 | let b = _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0); |
1629 | let r = _mm512_maskz_min_epi64(0, a, b); |
1630 | assert_eq_m512i(r, _mm512_setzero_si512()); |
1631 | let r = _mm512_maskz_min_epi64(0b00001111, a, b); |
1632 | let e = _mm512_setr_epi64(0, 1, 2, 3, 0, 0, 0, 0); |
1633 | assert_eq_m512i(r, e); |
1634 | } |
1635 | |
1636 | #[simd_test(enable = "avx512f,avx512vl" )] |
1637 | unsafe fn test_mm256_min_epi64() { |
1638 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
1639 | let b = _mm256_set_epi64x(3, 2, 1, 0); |
1640 | let r = _mm256_min_epi64(a, b); |
1641 | let e = _mm256_set_epi64x(0, 1, 1, 0); |
1642 | assert_eq_m256i(r, e); |
1643 | } |
1644 | |
1645 | #[simd_test(enable = "avx512f,avx512vl" )] |
1646 | unsafe fn test_mm256_mask_min_epi64() { |
1647 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
1648 | let b = _mm256_set_epi64x(3, 2, 1, 0); |
1649 | let r = _mm256_mask_min_epi64(a, 0, a, b); |
1650 | assert_eq_m256i(r, a); |
1651 | let r = _mm256_mask_min_epi64(a, 0b00001111, a, b); |
1652 | let e = _mm256_set_epi64x(0, 1, 1, 0); |
1653 | assert_eq_m256i(r, e); |
1654 | } |
1655 | |
1656 | #[simd_test(enable = "avx512f,avx512vl" )] |
1657 | unsafe fn test_mm256_maskz_min_epi64() { |
1658 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
1659 | let b = _mm256_set_epi64x(3, 2, 1, 0); |
1660 | let r = _mm256_maskz_min_epi64(0, a, b); |
1661 | assert_eq_m256i(r, _mm256_setzero_si256()); |
1662 | let r = _mm256_maskz_min_epi64(0b00001111, a, b); |
1663 | let e = _mm256_set_epi64x(0, 1, 1, 0); |
1664 | assert_eq_m256i(r, e); |
1665 | } |
1666 | |
1667 | #[simd_test(enable = "avx512f" )] |
1668 | unsafe fn test_mm512_min_pd() { |
1669 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
1670 | let b = _mm512_setr_pd(7., 6., 5., 4., 3., 2., 1., 0.); |
1671 | let r = _mm512_min_pd(a, b); |
1672 | let e = _mm512_setr_pd(0., 1., 2., 3., 3., 2., 1., 0.); |
1673 | assert_eq_m512d(r, e); |
1674 | } |
1675 | |
1676 | #[simd_test(enable = "avx512f" )] |
1677 | unsafe fn test_mm512_mask_min_pd() { |
1678 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
1679 | let b = _mm512_setr_pd(7., 6., 5., 4., 3., 2., 1., 0.); |
1680 | let r = _mm512_mask_min_pd(a, 0, a, b); |
1681 | assert_eq_m512d(r, a); |
1682 | let r = _mm512_mask_min_pd(a, 0b00001111, a, b); |
1683 | let e = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
1684 | assert_eq_m512d(r, e); |
1685 | } |
1686 | |
1687 | #[simd_test(enable = "avx512f" )] |
1688 | unsafe fn test_mm512_maskz_min_pd() { |
1689 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
1690 | let b = _mm512_setr_pd(7., 6., 5., 4., 3., 2., 1., 0.); |
1691 | let r = _mm512_maskz_min_pd(0, a, b); |
1692 | assert_eq_m512d(r, _mm512_setzero_pd()); |
1693 | let r = _mm512_maskz_min_pd(0b00001111, a, b); |
1694 | let e = _mm512_setr_pd(0., 1., 2., 3., 0., 0., 0., 0.); |
1695 | assert_eq_m512d(r, e); |
1696 | } |
1697 | |
1698 | #[simd_test(enable = "avx512f,avx512vl" )] |
1699 | unsafe fn test_mm256_mask_min_pd() { |
1700 | let a = _mm256_set_pd(0., 1., 2., 3.); |
1701 | let b = _mm256_set_pd(3., 2., 1., 0.); |
1702 | let r = _mm256_mask_min_pd(a, 0, a, b); |
1703 | assert_eq_m256d(r, a); |
1704 | let r = _mm256_mask_min_pd(a, 0b00001111, a, b); |
1705 | let e = _mm256_set_pd(0., 1., 1., 0.); |
1706 | assert_eq_m256d(r, e); |
1707 | } |
1708 | |
1709 | #[simd_test(enable = "avx512f,avx512vl" )] |
1710 | unsafe fn test_mm256_maskz_min_pd() { |
1711 | let a = _mm256_set_pd(0., 1., 2., 3.); |
1712 | let b = _mm256_set_pd(3., 2., 1., 0.); |
1713 | let r = _mm256_maskz_min_pd(0, a, b); |
1714 | assert_eq_m256d(r, _mm256_setzero_pd()); |
1715 | let r = _mm256_maskz_min_pd(0b00001111, a, b); |
1716 | let e = _mm256_set_pd(0., 1., 1., 0.); |
1717 | assert_eq_m256d(r, e); |
1718 | } |
1719 | |
1720 | #[simd_test(enable = "avx512f,avx512vl" )] |
1721 | unsafe fn test_mm_mask_min_pd() { |
1722 | let a = _mm_set_pd(0., 1.); |
1723 | let b = _mm_set_pd(1., 0.); |
1724 | let r = _mm_mask_min_pd(a, 0, a, b); |
1725 | assert_eq_m128d(r, a); |
1726 | let r = _mm_mask_min_pd(a, 0b00000011, a, b); |
1727 | let e = _mm_set_pd(0., 0.); |
1728 | assert_eq_m128d(r, e); |
1729 | } |
1730 | |
1731 | #[simd_test(enable = "avx512f,avx512vl" )] |
1732 | unsafe fn test_mm_maskz_min_pd() { |
1733 | let a = _mm_set_pd(0., 1.); |
1734 | let b = _mm_set_pd(1., 0.); |
1735 | let r = _mm_maskz_min_pd(0, a, b); |
1736 | assert_eq_m128d(r, _mm_setzero_pd()); |
1737 | let r = _mm_maskz_min_pd(0b00000011, a, b); |
1738 | let e = _mm_set_pd(0., 0.); |
1739 | assert_eq_m128d(r, e); |
1740 | } |
1741 | |
1742 | #[simd_test(enable = "avx512f" )] |
1743 | unsafe fn test_mm512_min_epu64() { |
1744 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1745 | let b = _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0); |
1746 | let r = _mm512_min_epu64(a, b); |
1747 | let e = _mm512_setr_epi64(0, 1, 2, 3, 3, 2, 1, 0); |
1748 | assert_eq_m512i(r, e); |
1749 | } |
1750 | |
1751 | #[simd_test(enable = "avx512f" )] |
1752 | unsafe fn test_mm512_mask_min_epu64() { |
1753 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1754 | let b = _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0); |
1755 | let r = _mm512_mask_min_epu64(a, 0, a, b); |
1756 | assert_eq_m512i(r, a); |
1757 | let r = _mm512_mask_min_epu64(a, 0b00001111, a, b); |
1758 | let e = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1759 | assert_eq_m512i(r, e); |
1760 | } |
1761 | |
1762 | #[simd_test(enable = "avx512f" )] |
1763 | unsafe fn test_mm512_maskz_min_epu64() { |
1764 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
1765 | let b = _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0); |
1766 | let r = _mm512_maskz_min_epu64(0, a, b); |
1767 | assert_eq_m512i(r, _mm512_setzero_si512()); |
1768 | let r = _mm512_maskz_min_epu64(0b00001111, a, b); |
1769 | let e = _mm512_setr_epi64(0, 1, 2, 3, 0, 0, 0, 0); |
1770 | assert_eq_m512i(r, e); |
1771 | } |
1772 | |
1773 | #[simd_test(enable = "avx512f,avx512vl" )] |
1774 | unsafe fn test_mm256_min_epu64() { |
1775 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
1776 | let b = _mm256_set_epi64x(3, 2, 1, 0); |
1777 | let r = _mm256_min_epu64(a, b); |
1778 | let e = _mm256_set_epi64x(0, 1, 1, 0); |
1779 | assert_eq_m256i(r, e); |
1780 | } |
1781 | |
1782 | #[simd_test(enable = "avx512f,avx512vl" )] |
1783 | unsafe fn test_mm256_mask_min_epu64() { |
1784 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
1785 | let b = _mm256_set_epi64x(3, 2, 1, 0); |
1786 | let r = _mm256_mask_min_epu64(a, 0, a, b); |
1787 | assert_eq_m256i(r, a); |
1788 | let r = _mm256_mask_min_epu64(a, 0b00001111, a, b); |
1789 | let e = _mm256_set_epi64x(0, 1, 1, 0); |
1790 | assert_eq_m256i(r, e); |
1791 | } |
1792 | |
1793 | #[simd_test(enable = "avx512f,avx512vl" )] |
1794 | unsafe fn test_mm256_maskz_min_epu64() { |
1795 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
1796 | let b = _mm256_set_epi64x(3, 2, 1, 0); |
1797 | let r = _mm256_maskz_min_epu64(0, a, b); |
1798 | assert_eq_m256i(r, _mm256_setzero_si256()); |
1799 | let r = _mm256_maskz_min_epu64(0b00001111, a, b); |
1800 | let e = _mm256_set_epi64x(0, 1, 1, 0); |
1801 | assert_eq_m256i(r, e); |
1802 | } |
1803 | |
1804 | #[simd_test(enable = "avx512f,avx512vl" )] |
1805 | unsafe fn test_mm_min_epu64() { |
1806 | let a = _mm_set_epi64x(0, 1); |
1807 | let b = _mm_set_epi64x(1, 0); |
1808 | let r = _mm_min_epu64(a, b); |
1809 | let e = _mm_set_epi64x(0, 0); |
1810 | assert_eq_m128i(r, e); |
1811 | } |
1812 | |
1813 | #[simd_test(enable = "avx512f,avx512vl" )] |
1814 | unsafe fn test_mm_mask_min_epu64() { |
1815 | let a = _mm_set_epi64x(0, 1); |
1816 | let b = _mm_set_epi64x(1, 0); |
1817 | let r = _mm_mask_min_epu64(a, 0, a, b); |
1818 | assert_eq_m128i(r, a); |
1819 | let r = _mm_mask_min_epu64(a, 0b00000011, a, b); |
1820 | let e = _mm_set_epi64x(0, 0); |
1821 | assert_eq_m128i(r, e); |
1822 | } |
1823 | |
1824 | #[simd_test(enable = "avx512f,avx512vl" )] |
1825 | unsafe fn test_mm_maskz_min_epu64() { |
1826 | let a = _mm_set_epi64x(0, 1); |
1827 | let b = _mm_set_epi64x(1, 0); |
1828 | let r = _mm_maskz_min_epu64(0, a, b); |
1829 | assert_eq_m128i(r, _mm_setzero_si128()); |
1830 | let r = _mm_maskz_min_epu64(0b00000011, a, b); |
1831 | let e = _mm_set_epi64x(0, 0); |
1832 | assert_eq_m128i(r, e); |
1833 | } |
1834 | |
1835 | #[simd_test(enable = "avx512f" )] |
1836 | unsafe fn test_mm512_sqrt_pd() { |
1837 | let a = _mm512_setr_pd(0., 1., 4., 9., 16., 25., 36., 49.); |
1838 | let r = _mm512_sqrt_pd(a); |
1839 | let e = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
1840 | assert_eq_m512d(r, e); |
1841 | } |
1842 | |
1843 | #[simd_test(enable = "avx512f" )] |
1844 | unsafe fn test_mm512_mask_sqrt_pd() { |
1845 | let a = _mm512_setr_pd(0., 1., 4., 9., 16., 25., 36., 49.); |
1846 | let r = _mm512_mask_sqrt_pd(a, 0, a); |
1847 | assert_eq_m512d(r, a); |
1848 | let r = _mm512_mask_sqrt_pd(a, 0b00001111, a); |
1849 | let e = _mm512_setr_pd(0., 1., 2., 3., 16., 25., 36., 49.); |
1850 | assert_eq_m512d(r, e); |
1851 | } |
1852 | |
1853 | #[simd_test(enable = "avx512f" )] |
1854 | unsafe fn test_mm512_maskz_sqrt_pd() { |
1855 | let a = _mm512_setr_pd(0., 1., 4., 9., 16., 25., 36., 49.); |
1856 | let r = _mm512_maskz_sqrt_pd(0, a); |
1857 | assert_eq_m512d(r, _mm512_setzero_pd()); |
1858 | let r = _mm512_maskz_sqrt_pd(0b00001111, a); |
1859 | let e = _mm512_setr_pd(0., 1., 2., 3., 0., 0., 0., 0.); |
1860 | assert_eq_m512d(r, e); |
1861 | } |
1862 | |
1863 | #[simd_test(enable = "avx512f,avx512vl" )] |
1864 | unsafe fn test_mm256_mask_sqrt_pd() { |
1865 | let a = _mm256_set_pd(0., 1., 4., 9.); |
1866 | let r = _mm256_mask_sqrt_pd(a, 0, a); |
1867 | assert_eq_m256d(r, a); |
1868 | let r = _mm256_mask_sqrt_pd(a, 0b00001111, a); |
1869 | let e = _mm256_set_pd(0., 1., 2., 3.); |
1870 | assert_eq_m256d(r, e); |
1871 | } |
1872 | |
1873 | #[simd_test(enable = "avx512f,avx512vl" )] |
1874 | unsafe fn test_mm256_maskz_sqrt_pd() { |
1875 | let a = _mm256_set_pd(0., 1., 4., 9.); |
1876 | let r = _mm256_maskz_sqrt_pd(0, a); |
1877 | assert_eq_m256d(r, _mm256_setzero_pd()); |
1878 | let r = _mm256_maskz_sqrt_pd(0b00001111, a); |
1879 | let e = _mm256_set_pd(0., 1., 2., 3.); |
1880 | assert_eq_m256d(r, e); |
1881 | } |
1882 | |
1883 | #[simd_test(enable = "avx512f,avx512vl" )] |
1884 | unsafe fn test_mm_mask_sqrt_pd() { |
1885 | let a = _mm_set_pd(0., 1.); |
1886 | let r = _mm_mask_sqrt_pd(a, 0, a); |
1887 | assert_eq_m128d(r, a); |
1888 | let r = _mm_mask_sqrt_pd(a, 0b00000011, a); |
1889 | let e = _mm_set_pd(0., 1.); |
1890 | assert_eq_m128d(r, e); |
1891 | } |
1892 | |
1893 | #[simd_test(enable = "avx512f,avx512vl" )] |
1894 | unsafe fn test_mm_maskz_sqrt_pd() { |
1895 | let a = _mm_set_pd(0., 1.); |
1896 | let r = _mm_maskz_sqrt_pd(0, a); |
1897 | assert_eq_m128d(r, _mm_setzero_pd()); |
1898 | let r = _mm_maskz_sqrt_pd(0b00000011, a); |
1899 | let e = _mm_set_pd(0., 1.); |
1900 | assert_eq_m128d(r, e); |
1901 | } |
1902 | |
1903 | #[simd_test(enable = "avx512f" )] |
1904 | unsafe fn test_mm512_fmadd_pd() { |
1905 | let a = _mm512_setr_pd(1., 1., 1., 1., 1., 1., 1., 1.); |
1906 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
1907 | let c = _mm512_setr_pd(1., 1., 1., 1., 1., 1., 1., 1.); |
1908 | let r = _mm512_fmadd_pd(a, b, c); |
1909 | let e = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
1910 | assert_eq_m512d(r, e); |
1911 | } |
1912 | |
1913 | #[simd_test(enable = "avx512f" )] |
1914 | unsafe fn test_mm512_mask_fmadd_pd() { |
1915 | let a = _mm512_setr_pd(1., 1., 1., 1., 1., 1., 1., 1.); |
1916 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
1917 | let c = _mm512_setr_pd(1., 1., 1., 1., 1., 1., 1., 1.); |
1918 | let r = _mm512_mask_fmadd_pd(a, 0, b, c); |
1919 | assert_eq_m512d(r, a); |
1920 | let r = _mm512_mask_fmadd_pd(a, 0b00001111, b, c); |
1921 | let e = _mm512_setr_pd(1., 2., 3., 4., 1., 1., 1., 1.); |
1922 | assert_eq_m512d(r, e); |
1923 | } |
1924 | |
1925 | #[simd_test(enable = "avx512f" )] |
1926 | unsafe fn test_mm512_maskz_fmadd_pd() { |
1927 | let a = _mm512_setr_pd(1., 1., 1., 1., 1., 1., 1., 1.); |
1928 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
1929 | let c = _mm512_setr_pd(1., 1., 1., 1., 1., 1., 1., 1.); |
1930 | let r = _mm512_maskz_fmadd_pd(0, a, b, c); |
1931 | assert_eq_m512d(r, _mm512_setzero_pd()); |
1932 | let r = _mm512_maskz_fmadd_pd(0b00001111, a, b, c); |
1933 | let e = _mm512_setr_pd(1., 2., 3., 4., 0., 0., 0., 0.); |
1934 | assert_eq_m512d(r, e); |
1935 | } |
1936 | |
1937 | #[simd_test(enable = "avx512f" )] |
1938 | unsafe fn test_mm512_mask3_fmadd_pd() { |
1939 | let a = _mm512_setr_pd(1., 1., 1., 1., 1., 1., 1., 1.); |
1940 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
1941 | let c = _mm512_setr_pd(1., 1., 1., 1., 2., 2., 2., 2.); |
1942 | let r = _mm512_mask3_fmadd_pd(a, b, c, 0); |
1943 | assert_eq_m512d(r, c); |
1944 | let r = _mm512_mask3_fmadd_pd(a, b, c, 0b00001111); |
1945 | let e = _mm512_setr_pd(1., 2., 3., 4., 2., 2., 2., 2.); |
1946 | assert_eq_m512d(r, e); |
1947 | } |
1948 | |
1949 | #[simd_test(enable = "avx512f,avx512vl" )] |
1950 | unsafe fn test_mm256_mask_fmadd_pd() { |
1951 | let a = _mm256_set1_pd(1.); |
1952 | let b = _mm256_set_pd(0., 1., 2., 3.); |
1953 | let c = _mm256_set1_pd(1.); |
1954 | let r = _mm256_mask_fmadd_pd(a, 0, b, c); |
1955 | assert_eq_m256d(r, a); |
1956 | let r = _mm256_mask_fmadd_pd(a, 0b00001111, b, c); |
1957 | let e = _mm256_set_pd(1., 2., 3., 4.); |
1958 | assert_eq_m256d(r, e); |
1959 | } |
1960 | |
1961 | #[simd_test(enable = "avx512f,avx512vl" )] |
1962 | unsafe fn test_mm256_maskz_fmadd_pd() { |
1963 | let a = _mm256_set1_pd(1.); |
1964 | let b = _mm256_set_pd(0., 1., 2., 3.); |
1965 | let c = _mm256_set1_pd(1.); |
1966 | let r = _mm256_maskz_fmadd_pd(0, a, b, c); |
1967 | assert_eq_m256d(r, _mm256_setzero_pd()); |
1968 | let r = _mm256_maskz_fmadd_pd(0b00001111, a, b, c); |
1969 | let e = _mm256_set_pd(1., 2., 3., 4.); |
1970 | assert_eq_m256d(r, e); |
1971 | } |
1972 | |
1973 | #[simd_test(enable = "avx512f,avx512vl" )] |
1974 | unsafe fn test_mm256_mask3_fmadd_pd() { |
1975 | let a = _mm256_set1_pd(1.); |
1976 | let b = _mm256_set_pd(0., 1., 2., 3.); |
1977 | let c = _mm256_set1_pd(1.); |
1978 | let r = _mm256_mask3_fmadd_pd(a, b, c, 0); |
1979 | assert_eq_m256d(r, c); |
1980 | let r = _mm256_mask3_fmadd_pd(a, b, c, 0b00001111); |
1981 | let e = _mm256_set_pd(1., 2., 3., 4.); |
1982 | assert_eq_m256d(r, e); |
1983 | } |
1984 | |
1985 | #[simd_test(enable = "avx512f,avx512vl" )] |
1986 | unsafe fn test_mm_mask_fmadd_pd() { |
1987 | let a = _mm_set1_pd(1.); |
1988 | let b = _mm_set_pd(0., 1.); |
1989 | let c = _mm_set1_pd(1.); |
1990 | let r = _mm_mask_fmadd_pd(a, 0, b, c); |
1991 | assert_eq_m128d(r, a); |
1992 | let r = _mm_mask_fmadd_pd(a, 0b00000011, b, c); |
1993 | let e = _mm_set_pd(1., 2.); |
1994 | assert_eq_m128d(r, e); |
1995 | } |
1996 | |
1997 | #[simd_test(enable = "avx512f,avx512vl" )] |
1998 | unsafe fn test_mm_maskz_fmadd_pd() { |
1999 | let a = _mm_set1_pd(1.); |
2000 | let b = _mm_set_pd(0., 1.); |
2001 | let c = _mm_set1_pd(1.); |
2002 | let r = _mm_maskz_fmadd_pd(0, a, b, c); |
2003 | assert_eq_m128d(r, _mm_setzero_pd()); |
2004 | let r = _mm_maskz_fmadd_pd(0b00000011, a, b, c); |
2005 | let e = _mm_set_pd(1., 2.); |
2006 | assert_eq_m128d(r, e); |
2007 | } |
2008 | |
2009 | #[simd_test(enable = "avx512f,avx512vl" )] |
2010 | unsafe fn test_mm_mask3_fmadd_pd() { |
2011 | let a = _mm_set1_pd(1.); |
2012 | let b = _mm_set_pd(0., 1.); |
2013 | let c = _mm_set1_pd(1.); |
2014 | let r = _mm_mask3_fmadd_pd(a, b, c, 0); |
2015 | assert_eq_m128d(r, c); |
2016 | let r = _mm_mask3_fmadd_pd(a, b, c, 0b00000011); |
2017 | let e = _mm_set_pd(1., 2.); |
2018 | assert_eq_m128d(r, e); |
2019 | } |
2020 | |
2021 | #[simd_test(enable = "avx512f" )] |
2022 | unsafe fn test_mm512_fmsub_pd() { |
2023 | let a = _mm512_set1_pd(1.); |
2024 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2025 | let c = _mm512_set1_pd(1.); |
2026 | let r = _mm512_fmsub_pd(a, b, c); |
2027 | let e = _mm512_setr_pd(-1., 0., 1., 2., 3., 4., 5., 6.); |
2028 | assert_eq_m512d(r, e); |
2029 | } |
2030 | |
2031 | #[simd_test(enable = "avx512f" )] |
2032 | unsafe fn test_mm512_mask_fmsub_pd() { |
2033 | let a = _mm512_set1_pd(1.); |
2034 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2035 | let c = _mm512_set1_pd(1.); |
2036 | let r = _mm512_mask_fmsub_pd(a, 0, b, c); |
2037 | assert_eq_m512d(r, a); |
2038 | let r = _mm512_mask_fmsub_pd(a, 0b00001111, b, c); |
2039 | let e = _mm512_setr_pd(-1., 0., 1., 2., 1., 1., 1., 1.); |
2040 | assert_eq_m512d(r, e); |
2041 | } |
2042 | |
2043 | #[simd_test(enable = "avx512f" )] |
2044 | unsafe fn test_mm512_maskz_fmsub_pd() { |
2045 | let a = _mm512_set1_pd(1.); |
2046 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2047 | let c = _mm512_set1_pd(1.); |
2048 | let r = _mm512_maskz_fmsub_pd(0, a, b, c); |
2049 | assert_eq_m512d(r, _mm512_setzero_pd()); |
2050 | let r = _mm512_maskz_fmsub_pd(0b00001111, a, b, c); |
2051 | let e = _mm512_setr_pd(-1., 0., 1., 2., 0., 0., 0., 0.); |
2052 | assert_eq_m512d(r, e); |
2053 | } |
2054 | |
2055 | #[simd_test(enable = "avx512f" )] |
2056 | unsafe fn test_mm512_mask3_fmsub_pd() { |
2057 | let a = _mm512_set1_pd(1.); |
2058 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2059 | let c = _mm512_setr_pd(1., 1., 1., 1., 2., 2., 2., 2.); |
2060 | let r = _mm512_mask3_fmsub_pd(a, b, c, 0); |
2061 | assert_eq_m512d(r, c); |
2062 | let r = _mm512_mask3_fmsub_pd(a, b, c, 0b00001111); |
2063 | let e = _mm512_setr_pd(-1., 0., 1., 2., 2., 2., 2., 2.); |
2064 | assert_eq_m512d(r, e); |
2065 | } |
2066 | |
2067 | #[simd_test(enable = "avx512f,avx512vl" )] |
2068 | unsafe fn test_mm256_mask_fmsub_pd() { |
2069 | let a = _mm256_set1_pd(1.); |
2070 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2071 | let c = _mm256_set1_pd(1.); |
2072 | let r = _mm256_mask_fmsub_pd(a, 0, b, c); |
2073 | assert_eq_m256d(r, a); |
2074 | let r = _mm256_mask_fmsub_pd(a, 0b00001111, b, c); |
2075 | let e = _mm256_set_pd(-1., 0., 1., 2.); |
2076 | assert_eq_m256d(r, e); |
2077 | } |
2078 | |
2079 | #[simd_test(enable = "avx512f,avx512vl" )] |
2080 | unsafe fn test_mm256_maskz_fmsub_pd() { |
2081 | let a = _mm256_set1_pd(1.); |
2082 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2083 | let c = _mm256_set1_pd(1.); |
2084 | let r = _mm256_maskz_fmsub_pd(0, a, b, c); |
2085 | assert_eq_m256d(r, _mm256_setzero_pd()); |
2086 | let r = _mm256_maskz_fmsub_pd(0b00001111, a, b, c); |
2087 | let e = _mm256_set_pd(-1., 0., 1., 2.); |
2088 | assert_eq_m256d(r, e); |
2089 | } |
2090 | |
2091 | #[simd_test(enable = "avx512f,avx512vl" )] |
2092 | unsafe fn test_mm256_mask3_fmsub_pd() { |
2093 | let a = _mm256_set1_pd(1.); |
2094 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2095 | let c = _mm256_set1_pd(1.); |
2096 | let r = _mm256_mask3_fmsub_pd(a, b, c, 0); |
2097 | assert_eq_m256d(r, c); |
2098 | let r = _mm256_mask3_fmsub_pd(a, b, c, 0b00001111); |
2099 | let e = _mm256_set_pd(-1., 0., 1., 2.); |
2100 | assert_eq_m256d(r, e); |
2101 | } |
2102 | |
2103 | #[simd_test(enable = "avx512f,avx512vl" )] |
2104 | unsafe fn test_mm_mask_fmsub_pd() { |
2105 | let a = _mm_set1_pd(1.); |
2106 | let b = _mm_set_pd(0., 1.); |
2107 | let c = _mm_set1_pd(1.); |
2108 | let r = _mm_mask_fmsub_pd(a, 0, b, c); |
2109 | assert_eq_m128d(r, a); |
2110 | let r = _mm_mask_fmsub_pd(a, 0b00000011, b, c); |
2111 | let e = _mm_set_pd(-1., 0.); |
2112 | assert_eq_m128d(r, e); |
2113 | } |
2114 | |
2115 | #[simd_test(enable = "avx512f,avx512vl" )] |
2116 | unsafe fn test_mm_maskz_fmsub_pd() { |
2117 | let a = _mm_set1_pd(1.); |
2118 | let b = _mm_set_pd(0., 1.); |
2119 | let c = _mm_set1_pd(1.); |
2120 | let r = _mm_maskz_fmsub_pd(0, a, b, c); |
2121 | assert_eq_m128d(r, _mm_setzero_pd()); |
2122 | let r = _mm_maskz_fmsub_pd(0b00000011, a, b, c); |
2123 | let e = _mm_set_pd(-1., 0.); |
2124 | assert_eq_m128d(r, e); |
2125 | } |
2126 | |
2127 | #[simd_test(enable = "avx512f,avx512vl" )] |
2128 | unsafe fn test_mm_mask3_fmsub_pd() { |
2129 | let a = _mm_set1_pd(1.); |
2130 | let b = _mm_set_pd(0., 1.); |
2131 | let c = _mm_set1_pd(1.); |
2132 | let r = _mm_mask3_fmsub_pd(a, b, c, 0); |
2133 | assert_eq_m128d(r, c); |
2134 | let r = _mm_mask3_fmsub_pd(a, b, c, 0b00000011); |
2135 | let e = _mm_set_pd(-1., 0.); |
2136 | assert_eq_m128d(r, e); |
2137 | } |
2138 | |
2139 | #[simd_test(enable = "avx512f" )] |
2140 | unsafe fn test_mm512_fmaddsub_pd() { |
2141 | let a = _mm512_set1_pd(1.); |
2142 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2143 | let c = _mm512_set1_pd(1.); |
2144 | let r = _mm512_fmaddsub_pd(a, b, c); |
2145 | let e = _mm512_setr_pd(-1., 2., 1., 4., 3., 6., 5., 8.); |
2146 | assert_eq_m512d(r, e); |
2147 | } |
2148 | |
2149 | #[simd_test(enable = "avx512f" )] |
2150 | unsafe fn test_mm512_mask_fmaddsub_pd() { |
2151 | let a = _mm512_set1_pd(1.); |
2152 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2153 | let c = _mm512_set1_pd(1.); |
2154 | let r = _mm512_mask_fmaddsub_pd(a, 0, b, c); |
2155 | assert_eq_m512d(r, a); |
2156 | let r = _mm512_mask_fmaddsub_pd(a, 0b00001111, b, c); |
2157 | let e = _mm512_setr_pd(-1., 2., 1., 4., 1., 1., 1., 1.); |
2158 | assert_eq_m512d(r, e); |
2159 | } |
2160 | |
2161 | #[simd_test(enable = "avx512f" )] |
2162 | unsafe fn test_mm512_maskz_fmaddsub_pd() { |
2163 | let a = _mm512_set1_pd(1.); |
2164 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2165 | let c = _mm512_set1_pd(1.); |
2166 | let r = _mm512_maskz_fmaddsub_pd(0, a, b, c); |
2167 | assert_eq_m512d(r, _mm512_setzero_pd()); |
2168 | let r = _mm512_maskz_fmaddsub_pd(0b00001111, a, b, c); |
2169 | let e = _mm512_setr_pd(-1., 2., 1., 4., 0., 0., 0., 0.); |
2170 | assert_eq_m512d(r, e); |
2171 | } |
2172 | |
2173 | #[simd_test(enable = "avx512f" )] |
2174 | unsafe fn test_mm512_mask3_fmaddsub_pd() { |
2175 | let a = _mm512_set1_pd(1.); |
2176 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2177 | let c = _mm512_setr_pd(1., 1., 1., 1., 2., 2., 2., 2.); |
2178 | let r = _mm512_mask3_fmaddsub_pd(a, b, c, 0); |
2179 | assert_eq_m512d(r, c); |
2180 | let r = _mm512_mask3_fmaddsub_pd(a, b, c, 0b00001111); |
2181 | let e = _mm512_setr_pd(-1., 2., 1., 4., 2., 2., 2., 2.); |
2182 | assert_eq_m512d(r, e); |
2183 | } |
2184 | |
2185 | #[simd_test(enable = "avx512f,avx512vl" )] |
2186 | unsafe fn test_mm256_mask_fmaddsub_pd() { |
2187 | let a = _mm256_set1_pd(1.); |
2188 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2189 | let c = _mm256_set1_pd(1.); |
2190 | let r = _mm256_mask_fmaddsub_pd(a, 0, b, c); |
2191 | assert_eq_m256d(r, a); |
2192 | let r = _mm256_mask_fmaddsub_pd(a, 0b00001111, b, c); |
2193 | let e = _mm256_set_pd(1., 0., 3., 2.); |
2194 | assert_eq_m256d(r, e); |
2195 | } |
2196 | |
2197 | #[simd_test(enable = "avx512f,avx512vl" )] |
2198 | unsafe fn test_mm256_maskz_fmaddsub_pd() { |
2199 | let a = _mm256_set1_pd(1.); |
2200 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2201 | let c = _mm256_set1_pd(1.); |
2202 | let r = _mm256_maskz_fmaddsub_pd(0, a, b, c); |
2203 | assert_eq_m256d(r, _mm256_setzero_pd()); |
2204 | let r = _mm256_maskz_fmaddsub_pd(0b00001111, a, b, c); |
2205 | let e = _mm256_set_pd(1., 0., 3., 2.); |
2206 | assert_eq_m256d(r, e); |
2207 | } |
2208 | |
2209 | #[simd_test(enable = "avx512f,avx512vl" )] |
2210 | unsafe fn test_mm256_mask3_fmaddsub_pd() { |
2211 | let a = _mm256_set1_pd(1.); |
2212 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2213 | let c = _mm256_set1_pd(1.); |
2214 | let r = _mm256_mask3_fmaddsub_pd(a, b, c, 0); |
2215 | assert_eq_m256d(r, c); |
2216 | let r = _mm256_mask3_fmaddsub_pd(a, b, c, 0b00001111); |
2217 | let e = _mm256_set_pd(1., 0., 3., 2.); |
2218 | assert_eq_m256d(r, e); |
2219 | } |
2220 | |
2221 | #[simd_test(enable = "avx512f,avx512vl" )] |
2222 | unsafe fn test_mm_mask_fmaddsub_pd() { |
2223 | let a = _mm_set1_pd(1.); |
2224 | let b = _mm_set_pd(0., 1.); |
2225 | let c = _mm_set1_pd(1.); |
2226 | let r = _mm_mask_fmaddsub_pd(a, 0, b, c); |
2227 | assert_eq_m128d(r, a); |
2228 | let r = _mm_mask_fmaddsub_pd(a, 0b00000011, b, c); |
2229 | let e = _mm_set_pd(1., 0.); |
2230 | assert_eq_m128d(r, e); |
2231 | } |
2232 | |
2233 | #[simd_test(enable = "avx512f,avx512vl" )] |
2234 | unsafe fn test_mm_maskz_fmaddsub_pd() { |
2235 | let a = _mm_set1_pd(1.); |
2236 | let b = _mm_set_pd(0., 1.); |
2237 | let c = _mm_set1_pd(1.); |
2238 | let r = _mm_maskz_fmaddsub_pd(0, a, b, c); |
2239 | assert_eq_m128d(r, _mm_setzero_pd()); |
2240 | let r = _mm_maskz_fmaddsub_pd(0b00000011, a, b, c); |
2241 | let e = _mm_set_pd(1., 0.); |
2242 | assert_eq_m128d(r, e); |
2243 | } |
2244 | |
2245 | #[simd_test(enable = "avx512f,avx512vl" )] |
2246 | unsafe fn test_mm_mask3_fmaddsub_pd() { |
2247 | let a = _mm_set1_pd(1.); |
2248 | let b = _mm_set_pd(0., 1.); |
2249 | let c = _mm_set1_pd(1.); |
2250 | let r = _mm_mask3_fmaddsub_pd(a, b, c, 0); |
2251 | assert_eq_m128d(r, c); |
2252 | let r = _mm_mask3_fmaddsub_pd(a, b, c, 0b00000011); |
2253 | let e = _mm_set_pd(1., 0.); |
2254 | assert_eq_m128d(r, e); |
2255 | } |
2256 | |
2257 | #[simd_test(enable = "avx512f" )] |
2258 | unsafe fn test_mm512_fmsubadd_pd() { |
2259 | let a = _mm512_set1_pd(1.); |
2260 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2261 | let c = _mm512_set1_pd(1.); |
2262 | let r = _mm512_fmsubadd_pd(a, b, c); |
2263 | let e = _mm512_setr_pd(1., 0., 3., 2., 5., 4., 7., 6.); |
2264 | assert_eq_m512d(r, e); |
2265 | } |
2266 | |
2267 | #[simd_test(enable = "avx512f" )] |
2268 | unsafe fn test_mm512_mask_fmsubadd_pd() { |
2269 | let a = _mm512_set1_pd(1.); |
2270 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2271 | let c = _mm512_set1_pd(1.); |
2272 | let r = _mm512_mask_fmsubadd_pd(a, 0, b, c); |
2273 | assert_eq_m512d(r, a); |
2274 | let r = _mm512_mask_fmsubadd_pd(a, 0b00001111, b, c); |
2275 | let e = _mm512_setr_pd(1., 0., 3., 2., 1., 1., 1., 1.); |
2276 | assert_eq_m512d(r, e); |
2277 | } |
2278 | |
2279 | #[simd_test(enable = "avx512f" )] |
2280 | unsafe fn test_mm512_maskz_fmsubadd_pd() { |
2281 | let a = _mm512_set1_pd(1.); |
2282 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2283 | let c = _mm512_set1_pd(1.); |
2284 | let r = _mm512_maskz_fmsubadd_pd(0, a, b, c); |
2285 | assert_eq_m512d(r, _mm512_setzero_pd()); |
2286 | let r = _mm512_maskz_fmsubadd_pd(0b00001111, a, b, c); |
2287 | let e = _mm512_setr_pd(1., 0., 3., 2., 0., 0., 0., 0.); |
2288 | assert_eq_m512d(r, e); |
2289 | } |
2290 | |
2291 | #[simd_test(enable = "avx512f" )] |
2292 | unsafe fn test_mm512_mask3_fmsubadd_pd() { |
2293 | let a = _mm512_set1_pd(1.); |
2294 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2295 | let c = _mm512_setr_pd(1., 1., 1., 1., 2., 2., 2., 2.); |
2296 | let r = _mm512_mask3_fmsubadd_pd(a, b, c, 0); |
2297 | assert_eq_m512d(r, c); |
2298 | let r = _mm512_mask3_fmsubadd_pd(a, b, c, 0b00001111); |
2299 | let e = _mm512_setr_pd(1., 0., 3., 2., 2., 2., 2., 2.); |
2300 | assert_eq_m512d(r, e); |
2301 | } |
2302 | |
2303 | #[simd_test(enable = "avx512f,avx512vl" )] |
2304 | unsafe fn test_mm256_mask_fmsubadd_pd() { |
2305 | let a = _mm256_set1_pd(1.); |
2306 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2307 | let c = _mm256_set1_pd(1.); |
2308 | let r = _mm256_mask_fmsubadd_pd(a, 0, b, c); |
2309 | assert_eq_m256d(r, a); |
2310 | let r = _mm256_mask_fmsubadd_pd(a, 0b00001111, b, c); |
2311 | let e = _mm256_set_pd(-1., 2., 1., 4.); |
2312 | assert_eq_m256d(r, e); |
2313 | } |
2314 | |
2315 | #[simd_test(enable = "avx512f,avx512vl" )] |
2316 | unsafe fn test_mm256_maskz_fmsubadd_pd() { |
2317 | let a = _mm256_set1_pd(1.); |
2318 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2319 | let c = _mm256_set1_pd(1.); |
2320 | let r = _mm256_maskz_fmsubadd_pd(0, a, b, c); |
2321 | assert_eq_m256d(r, _mm256_setzero_pd()); |
2322 | let r = _mm256_maskz_fmsubadd_pd(0b00001111, a, b, c); |
2323 | let e = _mm256_set_pd(-1., 2., 1., 4.); |
2324 | assert_eq_m256d(r, e); |
2325 | } |
2326 | |
2327 | #[simd_test(enable = "avx512f,avx512vl" )] |
2328 | unsafe fn test_mm256_mask3_fmsubadd_pd() { |
2329 | let a = _mm256_set1_pd(1.); |
2330 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2331 | let c = _mm256_set1_pd(1.); |
2332 | let r = _mm256_mask3_fmsubadd_pd(a, b, c, 0); |
2333 | assert_eq_m256d(r, c); |
2334 | let r = _mm256_mask3_fmsubadd_pd(a, b, c, 0b00001111); |
2335 | let e = _mm256_set_pd(-1., 2., 1., 4.); |
2336 | assert_eq_m256d(r, e); |
2337 | } |
2338 | |
2339 | #[simd_test(enable = "avx512f,avx512vl" )] |
2340 | unsafe fn test_mm_mask_fmsubadd_pd() { |
2341 | let a = _mm_set1_pd(1.); |
2342 | let b = _mm_set_pd(0., 1.); |
2343 | let c = _mm_set1_pd(1.); |
2344 | let r = _mm_mask_fmsubadd_pd(a, 0, b, c); |
2345 | assert_eq_m128d(r, a); |
2346 | let r = _mm_mask_fmsubadd_pd(a, 0b00000011, b, c); |
2347 | let e = _mm_set_pd(-1., 2.); |
2348 | assert_eq_m128d(r, e); |
2349 | } |
2350 | |
2351 | #[simd_test(enable = "avx512f,avx512vl" )] |
2352 | unsafe fn test_mm_maskz_fmsubadd_pd() { |
2353 | let a = _mm_set1_pd(1.); |
2354 | let b = _mm_set_pd(0., 1.); |
2355 | let c = _mm_set1_pd(1.); |
2356 | let r = _mm_maskz_fmsubadd_pd(0, a, b, c); |
2357 | assert_eq_m128d(r, _mm_setzero_pd()); |
2358 | let r = _mm_maskz_fmsubadd_pd(0b00000011, a, b, c); |
2359 | let e = _mm_set_pd(-1., 2.); |
2360 | assert_eq_m128d(r, e); |
2361 | } |
2362 | |
2363 | #[simd_test(enable = "avx512f,avx512vl" )] |
2364 | unsafe fn test_mm_mask3_fmsubadd_pd() { |
2365 | let a = _mm_set1_pd(1.); |
2366 | let b = _mm_set_pd(0., 1.); |
2367 | let c = _mm_set1_pd(1.); |
2368 | let r = _mm_mask3_fmsubadd_pd(a, b, c, 0); |
2369 | assert_eq_m128d(r, c); |
2370 | let r = _mm_mask3_fmsubadd_pd(a, b, c, 0b00000011); |
2371 | let e = _mm_set_pd(-1., 2.); |
2372 | assert_eq_m128d(r, e); |
2373 | } |
2374 | |
2375 | #[simd_test(enable = "avx512f" )] |
2376 | unsafe fn test_mm512_fnmadd_pd() { |
2377 | let a = _mm512_set1_pd(1.); |
2378 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2379 | let c = _mm512_set1_pd(1.); |
2380 | let r = _mm512_fnmadd_pd(a, b, c); |
2381 | let e = _mm512_setr_pd(1., 0., -1., -2., -3., -4., -5., -6.); |
2382 | assert_eq_m512d(r, e); |
2383 | } |
2384 | |
2385 | #[simd_test(enable = "avx512f" )] |
2386 | unsafe fn test_mm512_mask_fnmadd_pd() { |
2387 | let a = _mm512_set1_pd(1.); |
2388 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2389 | let c = _mm512_set1_pd(1.); |
2390 | let r = _mm512_mask_fnmadd_pd(a, 0, b, c); |
2391 | assert_eq_m512d(r, a); |
2392 | let r = _mm512_mask_fnmadd_pd(a, 0b00001111, b, c); |
2393 | let e = _mm512_setr_pd(1., 0., -1., -2., 1., 1., 1., 1.); |
2394 | assert_eq_m512d(r, e); |
2395 | } |
2396 | |
2397 | #[simd_test(enable = "avx512f" )] |
2398 | unsafe fn test_mm512_maskz_fnmadd_pd() { |
2399 | let a = _mm512_set1_pd(1.); |
2400 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2401 | let c = _mm512_set1_pd(1.); |
2402 | let r = _mm512_maskz_fnmadd_pd(0, a, b, c); |
2403 | assert_eq_m512d(r, _mm512_setzero_pd()); |
2404 | let r = _mm512_maskz_fnmadd_pd(0b00001111, a, b, c); |
2405 | let e = _mm512_setr_pd(1., 0., -1., -2., 0., 0., 0., 0.); |
2406 | assert_eq_m512d(r, e); |
2407 | } |
2408 | |
2409 | #[simd_test(enable = "avx512f" )] |
2410 | unsafe fn test_mm512_mask3_fnmadd_pd() { |
2411 | let a = _mm512_set1_pd(1.); |
2412 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2413 | let c = _mm512_setr_pd(1., 1., 1., 1., 2., 2., 2., 2.); |
2414 | let r = _mm512_mask3_fnmadd_pd(a, b, c, 0); |
2415 | assert_eq_m512d(r, c); |
2416 | let r = _mm512_mask3_fnmadd_pd(a, b, c, 0b00001111); |
2417 | let e = _mm512_setr_pd(1., 0., -1., -2., 2., 2., 2., 2.); |
2418 | assert_eq_m512d(r, e); |
2419 | } |
2420 | |
2421 | #[simd_test(enable = "avx512f,avx512vl" )] |
2422 | unsafe fn test_mm256_mask_fnmadd_pd() { |
2423 | let a = _mm256_set1_pd(1.); |
2424 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2425 | let c = _mm256_set1_pd(1.); |
2426 | let r = _mm256_mask_fnmadd_pd(a, 0, b, c); |
2427 | assert_eq_m256d(r, a); |
2428 | let r = _mm256_mask_fnmadd_pd(a, 0b00001111, b, c); |
2429 | let e = _mm256_set_pd(1., 0., -1., -2.); |
2430 | assert_eq_m256d(r, e); |
2431 | } |
2432 | |
2433 | #[simd_test(enable = "avx512f,avx512vl" )] |
2434 | unsafe fn test_mm256_maskz_fnmadd_pd() { |
2435 | let a = _mm256_set1_pd(1.); |
2436 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2437 | let c = _mm256_set1_pd(1.); |
2438 | let r = _mm256_maskz_fnmadd_pd(0, a, b, c); |
2439 | assert_eq_m256d(r, _mm256_setzero_pd()); |
2440 | let r = _mm256_maskz_fnmadd_pd(0b00001111, a, b, c); |
2441 | let e = _mm256_set_pd(1., 0., -1., -2.); |
2442 | assert_eq_m256d(r, e); |
2443 | } |
2444 | |
2445 | #[simd_test(enable = "avx512f,avx512vl" )] |
2446 | unsafe fn test_mm256_mask3_fnmadd_pd() { |
2447 | let a = _mm256_set1_pd(1.); |
2448 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2449 | let c = _mm256_set1_pd(1.); |
2450 | let r = _mm256_mask3_fnmadd_pd(a, b, c, 0); |
2451 | assert_eq_m256d(r, c); |
2452 | let r = _mm256_mask3_fnmadd_pd(a, b, c, 0b00001111); |
2453 | let e = _mm256_set_pd(1., 0., -1., -2.); |
2454 | assert_eq_m256d(r, e); |
2455 | } |
2456 | |
2457 | #[simd_test(enable = "avx512f,avx512vl" )] |
2458 | unsafe fn test_mm_mask_fnmadd_pd() { |
2459 | let a = _mm_set1_pd(1.); |
2460 | let b = _mm_set_pd(0., 1.); |
2461 | let c = _mm_set1_pd(1.); |
2462 | let r = _mm_mask_fnmadd_pd(a, 0, b, c); |
2463 | assert_eq_m128d(r, a); |
2464 | let r = _mm_mask_fnmadd_pd(a, 0b00000011, b, c); |
2465 | let e = _mm_set_pd(1., 0.); |
2466 | assert_eq_m128d(r, e); |
2467 | } |
2468 | |
2469 | #[simd_test(enable = "avx512f,avx512vl" )] |
2470 | unsafe fn test_mm_maskz_fnmadd_pd() { |
2471 | let a = _mm_set1_pd(1.); |
2472 | let b = _mm_set_pd(0., 1.); |
2473 | let c = _mm_set1_pd(1.); |
2474 | let r = _mm_maskz_fnmadd_pd(0, a, b, c); |
2475 | assert_eq_m128d(r, _mm_setzero_pd()); |
2476 | let r = _mm_maskz_fnmadd_pd(0b00000011, a, b, c); |
2477 | let e = _mm_set_pd(1., 0.); |
2478 | assert_eq_m128d(r, e); |
2479 | } |
2480 | |
2481 | #[simd_test(enable = "avx512f,avx512vl" )] |
2482 | unsafe fn test_mm_mask3_fnmadd_pd() { |
2483 | let a = _mm_set1_pd(1.); |
2484 | let b = _mm_set_pd(0., 1.); |
2485 | let c = _mm_set1_pd(1.); |
2486 | let r = _mm_mask3_fnmadd_pd(a, b, c, 0); |
2487 | assert_eq_m128d(r, c); |
2488 | let r = _mm_mask3_fnmadd_pd(a, b, c, 0b00000011); |
2489 | let e = _mm_set_pd(1., 0.); |
2490 | assert_eq_m128d(r, e); |
2491 | } |
2492 | |
2493 | #[simd_test(enable = "avx512f" )] |
2494 | unsafe fn test_mm512_fnmsub_pd() { |
2495 | let a = _mm512_set1_pd(1.); |
2496 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2497 | let c = _mm512_set1_pd(1.); |
2498 | let r = _mm512_fnmsub_pd(a, b, c); |
2499 | let e = _mm512_setr_pd(-1., -2., -3., -4., -5., -6., -7., -8.); |
2500 | assert_eq_m512d(r, e); |
2501 | } |
2502 | |
2503 | #[simd_test(enable = "avx512f" )] |
2504 | unsafe fn test_mm512_mask_fnmsub_pd() { |
2505 | let a = _mm512_set1_pd(1.); |
2506 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2507 | let c = _mm512_set1_pd(1.); |
2508 | let r = _mm512_mask_fnmsub_pd(a, 0, b, c); |
2509 | assert_eq_m512d(r, a); |
2510 | let r = _mm512_mask_fnmsub_pd(a, 0b00001111, b, c); |
2511 | let e = _mm512_setr_pd(-1., -2., -3., -4., 1., 1., 1., 1.); |
2512 | assert_eq_m512d(r, e); |
2513 | } |
2514 | |
2515 | #[simd_test(enable = "avx512f" )] |
2516 | unsafe fn test_mm512_maskz_fnmsub_pd() { |
2517 | let a = _mm512_set1_pd(1.); |
2518 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2519 | let c = _mm512_set1_pd(1.); |
2520 | let r = _mm512_maskz_fnmsub_pd(0, a, b, c); |
2521 | assert_eq_m512d(r, _mm512_setzero_pd()); |
2522 | let r = _mm512_maskz_fnmsub_pd(0b00001111, a, b, c); |
2523 | let e = _mm512_setr_pd(-1., -2., -3., -4., 0., 0., 0., 0.); |
2524 | assert_eq_m512d(r, e); |
2525 | } |
2526 | |
2527 | #[simd_test(enable = "avx512f" )] |
2528 | unsafe fn test_mm512_mask3_fnmsub_pd() { |
2529 | let a = _mm512_set1_pd(1.); |
2530 | let b = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
2531 | let c = _mm512_setr_pd(1., 1., 1., 1., 2., 2., 2., 2.); |
2532 | let r = _mm512_mask3_fnmsub_pd(a, b, c, 0); |
2533 | assert_eq_m512d(r, c); |
2534 | let r = _mm512_mask3_fnmsub_pd(a, b, c, 0b00001111); |
2535 | let e = _mm512_setr_pd(-1., -2., -3., -4., 2., 2., 2., 2.); |
2536 | assert_eq_m512d(r, e); |
2537 | } |
2538 | |
2539 | #[simd_test(enable = "avx512f,avx512vl" )] |
2540 | unsafe fn test_mm256_mask_fnmsub_pd() { |
2541 | let a = _mm256_set1_pd(1.); |
2542 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2543 | let c = _mm256_set1_pd(1.); |
2544 | let r = _mm256_mask_fnmsub_pd(a, 0, b, c); |
2545 | assert_eq_m256d(r, a); |
2546 | let r = _mm256_mask_fnmsub_pd(a, 0b00001111, b, c); |
2547 | let e = _mm256_set_pd(-1., -2., -3., -4.); |
2548 | assert_eq_m256d(r, e); |
2549 | } |
2550 | |
2551 | #[simd_test(enable = "avx512f,avx512vl" )] |
2552 | unsafe fn test_mm256_maskz_fnmsub_pd() { |
2553 | let a = _mm256_set1_pd(1.); |
2554 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2555 | let c = _mm256_set1_pd(1.); |
2556 | let r = _mm256_maskz_fnmsub_pd(0, a, b, c); |
2557 | assert_eq_m256d(r, _mm256_setzero_pd()); |
2558 | let r = _mm256_maskz_fnmsub_pd(0b00001111, a, b, c); |
2559 | let e = _mm256_set_pd(-1., -2., -3., -4.); |
2560 | assert_eq_m256d(r, e); |
2561 | } |
2562 | |
2563 | #[simd_test(enable = "avx512f,avx512vl" )] |
2564 | unsafe fn test_mm256_mask3_fnmsub_pd() { |
2565 | let a = _mm256_set1_pd(1.); |
2566 | let b = _mm256_set_pd(0., 1., 2., 3.); |
2567 | let c = _mm256_set1_pd(1.); |
2568 | let r = _mm256_mask3_fnmsub_pd(a, b, c, 0); |
2569 | assert_eq_m256d(r, c); |
2570 | let r = _mm256_mask3_fnmsub_pd(a, b, c, 0b00001111); |
2571 | let e = _mm256_set_pd(-1., -2., -3., -4.); |
2572 | assert_eq_m256d(r, e); |
2573 | } |
2574 | |
2575 | #[simd_test(enable = "avx512f,avx512vl" )] |
2576 | unsafe fn test_mm_mask_fnmsub_pd() { |
2577 | let a = _mm_set1_pd(1.); |
2578 | let b = _mm_set_pd(0., 1.); |
2579 | let c = _mm_set1_pd(1.); |
2580 | let r = _mm_mask_fnmsub_pd(a, 0, b, c); |
2581 | assert_eq_m128d(r, a); |
2582 | let r = _mm_mask_fnmsub_pd(a, 0b00000011, b, c); |
2583 | let e = _mm_set_pd(-1., -2.); |
2584 | assert_eq_m128d(r, e); |
2585 | } |
2586 | |
2587 | #[simd_test(enable = "avx512f,avx512vl" )] |
2588 | unsafe fn test_mm_maskz_fnmsub_pd() { |
2589 | let a = _mm_set1_pd(1.); |
2590 | let b = _mm_set_pd(0., 1.); |
2591 | let c = _mm_set1_pd(1.); |
2592 | let r = _mm_maskz_fnmsub_pd(0, a, b, c); |
2593 | assert_eq_m128d(r, _mm_setzero_pd()); |
2594 | let r = _mm_maskz_fnmsub_pd(0b00000011, a, b, c); |
2595 | let e = _mm_set_pd(-1., -2.); |
2596 | assert_eq_m128d(r, e); |
2597 | } |
2598 | |
2599 | #[simd_test(enable = "avx512f,avx512vl" )] |
2600 | unsafe fn test_mm_mask3_fnmsub_pd() { |
2601 | let a = _mm_set1_pd(1.); |
2602 | let b = _mm_set_pd(0., 1.); |
2603 | let c = _mm_set1_pd(1.); |
2604 | let r = _mm_mask3_fnmsub_pd(a, b, c, 0); |
2605 | assert_eq_m128d(r, c); |
2606 | let r = _mm_mask3_fnmsub_pd(a, b, c, 0b00000011); |
2607 | let e = _mm_set_pd(-1., -2.); |
2608 | assert_eq_m128d(r, e); |
2609 | } |
2610 | |
2611 | #[simd_test(enable = "avx512f" )] |
2612 | unsafe fn test_mm512_rcp14_pd() { |
2613 | let a = _mm512_set1_pd(3.); |
2614 | let r = _mm512_rcp14_pd(a); |
2615 | let e = _mm512_set1_pd(0.3333320617675781); |
2616 | assert_eq_m512d(r, e); |
2617 | } |
2618 | |
2619 | #[simd_test(enable = "avx512f" )] |
2620 | unsafe fn test_mm512_mask_rcp14_pd() { |
2621 | let a = _mm512_set1_pd(3.); |
2622 | let r = _mm512_mask_rcp14_pd(a, 0, a); |
2623 | assert_eq_m512d(r, a); |
2624 | let r = _mm512_mask_rcp14_pd(a, 0b11110000, a); |
2625 | #[rustfmt::skip] |
2626 | let e = _mm512_setr_pd( |
2627 | 3., 3., 3., 3., |
2628 | 0.3333320617675781, 0.3333320617675781, 0.3333320617675781, 0.3333320617675781, |
2629 | ); |
2630 | assert_eq_m512d(r, e); |
2631 | } |
2632 | |
2633 | #[simd_test(enable = "avx512f" )] |
2634 | unsafe fn test_mm512_maskz_rcp14_pd() { |
2635 | let a = _mm512_set1_pd(3.); |
2636 | let r = _mm512_maskz_rcp14_pd(0, a); |
2637 | assert_eq_m512d(r, _mm512_setzero_pd()); |
2638 | let r = _mm512_maskz_rcp14_pd(0b11110000, a); |
2639 | #[rustfmt::skip] |
2640 | let e = _mm512_setr_pd( |
2641 | 0., 0., 0., 0., |
2642 | 0.3333320617675781, 0.3333320617675781, 0.3333320617675781, 0.3333320617675781, |
2643 | ); |
2644 | assert_eq_m512d(r, e); |
2645 | } |
2646 | |
2647 | #[simd_test(enable = "avx512f,avx512vl" )] |
2648 | unsafe fn test_mm256_rcp14_pd() { |
2649 | let a = _mm256_set1_pd(3.); |
2650 | let r = _mm256_rcp14_pd(a); |
2651 | let e = _mm256_set1_pd(0.3333320617675781); |
2652 | assert_eq_m256d(r, e); |
2653 | } |
2654 | |
2655 | #[simd_test(enable = "avx512f,avx512vl" )] |
2656 | unsafe fn test_mm256_mask_rcp14_pd() { |
2657 | let a = _mm256_set1_pd(3.); |
2658 | let r = _mm256_mask_rcp14_pd(a, 0, a); |
2659 | assert_eq_m256d(r, a); |
2660 | let r = _mm256_mask_rcp14_pd(a, 0b00001111, a); |
2661 | let e = _mm256_set1_pd(0.3333320617675781); |
2662 | assert_eq_m256d(r, e); |
2663 | } |
2664 | |
2665 | #[simd_test(enable = "avx512f,avx512vl" )] |
2666 | unsafe fn test_mm256_maskz_rcp14_pd() { |
2667 | let a = _mm256_set1_pd(3.); |
2668 | let r = _mm256_maskz_rcp14_pd(0, a); |
2669 | assert_eq_m256d(r, _mm256_setzero_pd()); |
2670 | let r = _mm256_maskz_rcp14_pd(0b00001111, a); |
2671 | let e = _mm256_set1_pd(0.3333320617675781); |
2672 | assert_eq_m256d(r, e); |
2673 | } |
2674 | |
2675 | #[simd_test(enable = "avx512f,avx512vl" )] |
2676 | unsafe fn test_mm_rcp14_pd() { |
2677 | let a = _mm_set1_pd(3.); |
2678 | let r = _mm_rcp14_pd(a); |
2679 | let e = _mm_set1_pd(0.3333320617675781); |
2680 | assert_eq_m128d(r, e); |
2681 | } |
2682 | |
2683 | #[simd_test(enable = "avx512f,avx512vl" )] |
2684 | unsafe fn test_mm_mask_rcp14_pd() { |
2685 | let a = _mm_set1_pd(3.); |
2686 | let r = _mm_mask_rcp14_pd(a, 0, a); |
2687 | assert_eq_m128d(r, a); |
2688 | let r = _mm_mask_rcp14_pd(a, 0b00000011, a); |
2689 | let e = _mm_set1_pd(0.3333320617675781); |
2690 | assert_eq_m128d(r, e); |
2691 | } |
2692 | |
2693 | #[simd_test(enable = "avx512f,avx512vl" )] |
2694 | unsafe fn test_mm_maskz_rcp14_pd() { |
2695 | let a = _mm_set1_pd(3.); |
2696 | let r = _mm_maskz_rcp14_pd(0, a); |
2697 | assert_eq_m128d(r, _mm_setzero_pd()); |
2698 | let r = _mm_maskz_rcp14_pd(0b00000011, a); |
2699 | let e = _mm_set1_pd(0.3333320617675781); |
2700 | assert_eq_m128d(r, e); |
2701 | } |
2702 | |
2703 | #[simd_test(enable = "avx512f" )] |
2704 | unsafe fn test_mm512_rsqrt14_pd() { |
2705 | let a = _mm512_set1_pd(3.); |
2706 | let r = _mm512_rsqrt14_pd(a); |
2707 | let e = _mm512_set1_pd(0.5773391723632813); |
2708 | assert_eq_m512d(r, e); |
2709 | } |
2710 | |
2711 | #[simd_test(enable = "avx512f" )] |
2712 | unsafe fn test_mm512_mask_rsqrt14_pd() { |
2713 | let a = _mm512_set1_pd(3.); |
2714 | let r = _mm512_mask_rsqrt14_pd(a, 0, a); |
2715 | assert_eq_m512d(r, a); |
2716 | let r = _mm512_mask_rsqrt14_pd(a, 0b11110000, a); |
2717 | #[rustfmt::skip] |
2718 | let e = _mm512_setr_pd( |
2719 | 3., 3., 3., 3., |
2720 | 0.5773391723632813, 0.5773391723632813, 0.5773391723632813, 0.5773391723632813, |
2721 | ); |
2722 | assert_eq_m512d(r, e); |
2723 | } |
2724 | |
2725 | #[simd_test(enable = "avx512f" )] |
2726 | unsafe fn test_mm512_maskz_rsqrt14_pd() { |
2727 | let a = _mm512_set1_pd(3.); |
2728 | let r = _mm512_maskz_rsqrt14_pd(0, a); |
2729 | assert_eq_m512d(r, _mm512_setzero_pd()); |
2730 | let r = _mm512_maskz_rsqrt14_pd(0b11110000, a); |
2731 | #[rustfmt::skip] |
2732 | let e = _mm512_setr_pd( |
2733 | 0., 0., 0., 0., |
2734 | 0.5773391723632813, 0.5773391723632813, 0.5773391723632813, 0.5773391723632813, |
2735 | ); |
2736 | assert_eq_m512d(r, e); |
2737 | } |
2738 | |
2739 | #[simd_test(enable = "avx512f,avx512vl" )] |
2740 | unsafe fn test_mm256_mask_rsqrt14_pd() { |
2741 | let a = _mm256_set1_pd(3.); |
2742 | let r = _mm256_mask_rsqrt14_pd(a, 0, a); |
2743 | assert_eq_m256d(r, a); |
2744 | let r = _mm256_mask_rsqrt14_pd(a, 0b00001111, a); |
2745 | let e = _mm256_set1_pd(0.5773391723632813); |
2746 | assert_eq_m256d(r, e); |
2747 | } |
2748 | |
2749 | #[simd_test(enable = "avx512f,avx512vl" )] |
2750 | unsafe fn test_mm256_maskz_rsqrt14_pd() { |
2751 | let a = _mm256_set1_pd(3.); |
2752 | let r = _mm256_maskz_rsqrt14_pd(0, a); |
2753 | assert_eq_m256d(r, _mm256_setzero_pd()); |
2754 | let r = _mm256_maskz_rsqrt14_pd(0b00001111, a); |
2755 | let e = _mm256_set1_pd(0.5773391723632813); |
2756 | assert_eq_m256d(r, e); |
2757 | } |
2758 | |
2759 | #[simd_test(enable = "avx512f,avx512vl" )] |
2760 | unsafe fn test_mm_mask_rsqrt14_pd() { |
2761 | let a = _mm_set1_pd(3.); |
2762 | let r = _mm_mask_rsqrt14_pd(a, 0, a); |
2763 | assert_eq_m128d(r, a); |
2764 | let r = _mm_mask_rsqrt14_pd(a, 0b00000011, a); |
2765 | let e = _mm_set1_pd(0.5773391723632813); |
2766 | assert_eq_m128d(r, e); |
2767 | } |
2768 | |
2769 | #[simd_test(enable = "avx512f,avx512vl" )] |
2770 | unsafe fn test_mm_maskz_rsqrt14_pd() { |
2771 | let a = _mm_set1_pd(3.); |
2772 | let r = _mm_maskz_rsqrt14_pd(0, a); |
2773 | assert_eq_m128d(r, _mm_setzero_pd()); |
2774 | let r = _mm_maskz_rsqrt14_pd(0b00000011, a); |
2775 | let e = _mm_set1_pd(0.5773391723632813); |
2776 | assert_eq_m128d(r, e); |
2777 | } |
2778 | |
2779 | #[simd_test(enable = "avx512f" )] |
2780 | unsafe fn test_mm512_getexp_pd() { |
2781 | let a = _mm512_set1_pd(3.); |
2782 | let r = _mm512_getexp_pd(a); |
2783 | let e = _mm512_set1_pd(1.); |
2784 | assert_eq_m512d(r, e); |
2785 | } |
2786 | |
2787 | #[simd_test(enable = "avx512f" )] |
2788 | unsafe fn test_mm512_mask_getexp_pd() { |
2789 | let a = _mm512_set1_pd(3.); |
2790 | let r = _mm512_mask_getexp_pd(a, 0, a); |
2791 | assert_eq_m512d(r, a); |
2792 | let r = _mm512_mask_getexp_pd(a, 0b11110000, a); |
2793 | let e = _mm512_setr_pd(3., 3., 3., 3., 1., 1., 1., 1.); |
2794 | assert_eq_m512d(r, e); |
2795 | } |
2796 | |
2797 | #[simd_test(enable = "avx512f" )] |
2798 | unsafe fn test_mm512_maskz_getexp_pd() { |
2799 | let a = _mm512_set1_pd(3.); |
2800 | let r = _mm512_maskz_getexp_pd(0, a); |
2801 | assert_eq_m512d(r, _mm512_setzero_pd()); |
2802 | let r = _mm512_maskz_getexp_pd(0b11110000, a); |
2803 | let e = _mm512_setr_pd(0., 0., 0., 0., 1., 1., 1., 1.); |
2804 | assert_eq_m512d(r, e); |
2805 | } |
2806 | |
2807 | #[simd_test(enable = "avx512f,avx512vl" )] |
2808 | unsafe fn test_mm256_getexp_pd() { |
2809 | let a = _mm256_set1_pd(3.); |
2810 | let r = _mm256_getexp_pd(a); |
2811 | let e = _mm256_set1_pd(1.); |
2812 | assert_eq_m256d(r, e); |
2813 | } |
2814 | |
2815 | #[simd_test(enable = "avx512f,avx512vl" )] |
2816 | unsafe fn test_mm256_mask_getexp_pd() { |
2817 | let a = _mm256_set1_pd(3.); |
2818 | let r = _mm256_mask_getexp_pd(a, 0, a); |
2819 | assert_eq_m256d(r, a); |
2820 | let r = _mm256_mask_getexp_pd(a, 0b00001111, a); |
2821 | let e = _mm256_set1_pd(1.); |
2822 | assert_eq_m256d(r, e); |
2823 | } |
2824 | |
2825 | #[simd_test(enable = "avx512f,avx512vl" )] |
2826 | unsafe fn test_mm256_maskz_getexp_pd() { |
2827 | let a = _mm256_set1_pd(3.); |
2828 | let r = _mm256_maskz_getexp_pd(0, a); |
2829 | assert_eq_m256d(r, _mm256_setzero_pd()); |
2830 | let r = _mm256_maskz_getexp_pd(0b00001111, a); |
2831 | let e = _mm256_set1_pd(1.); |
2832 | assert_eq_m256d(r, e); |
2833 | } |
2834 | |
2835 | #[simd_test(enable = "avx512f,avx512vl" )] |
2836 | unsafe fn test_mm_getexp_pd() { |
2837 | let a = _mm_set1_pd(3.); |
2838 | let r = _mm_getexp_pd(a); |
2839 | let e = _mm_set1_pd(1.); |
2840 | assert_eq_m128d(r, e); |
2841 | } |
2842 | |
2843 | #[simd_test(enable = "avx512f,avx512vl" )] |
2844 | unsafe fn test_mm_mask_getexp_pd() { |
2845 | let a = _mm_set1_pd(3.); |
2846 | let r = _mm_mask_getexp_pd(a, 0, a); |
2847 | assert_eq_m128d(r, a); |
2848 | let r = _mm_mask_getexp_pd(a, 0b00000011, a); |
2849 | let e = _mm_set1_pd(1.); |
2850 | assert_eq_m128d(r, e); |
2851 | } |
2852 | |
2853 | #[simd_test(enable = "avx512f,avx512vl" )] |
2854 | unsafe fn test_mm_maskz_getexp_pd() { |
2855 | let a = _mm_set1_pd(3.); |
2856 | let r = _mm_maskz_getexp_pd(0, a); |
2857 | assert_eq_m128d(r, _mm_setzero_pd()); |
2858 | let r = _mm_maskz_getexp_pd(0b00000011, a); |
2859 | let e = _mm_set1_pd(1.); |
2860 | assert_eq_m128d(r, e); |
2861 | } |
2862 | |
2863 | #[simd_test(enable = "avx512f" )] |
2864 | unsafe fn test_mm512_roundscale_pd() { |
2865 | let a = _mm512_set1_pd(1.1); |
2866 | let r = _mm512_roundscale_pd::<0b00_00_00_00>(a); |
2867 | let e = _mm512_set1_pd(1.0); |
2868 | assert_eq_m512d(r, e); |
2869 | } |
2870 | |
2871 | #[simd_test(enable = "avx512f" )] |
2872 | unsafe fn test_mm512_mask_roundscale_pd() { |
2873 | let a = _mm512_set1_pd(1.1); |
2874 | let r = _mm512_mask_roundscale_pd::<0b00_00_00_00>(a, 0, a); |
2875 | let e = _mm512_set1_pd(1.1); |
2876 | assert_eq_m512d(r, e); |
2877 | let r = _mm512_mask_roundscale_pd::<0b00_00_00_00>(a, 0b11111111, a); |
2878 | let e = _mm512_set1_pd(1.0); |
2879 | assert_eq_m512d(r, e); |
2880 | } |
2881 | |
2882 | #[simd_test(enable = "avx512f" )] |
2883 | unsafe fn test_mm512_maskz_roundscale_pd() { |
2884 | let a = _mm512_set1_pd(1.1); |
2885 | let r = _mm512_maskz_roundscale_pd::<0b00_00_00_00>(0, a); |
2886 | assert_eq_m512d(r, _mm512_setzero_pd()); |
2887 | let r = _mm512_maskz_roundscale_pd::<0b00_00_00_00>(0b11111111, a); |
2888 | let e = _mm512_set1_pd(1.0); |
2889 | assert_eq_m512d(r, e); |
2890 | } |
2891 | |
2892 | #[simd_test(enable = "avx512f,avx512vl" )] |
2893 | unsafe fn test_mm256_roundscale_pd() { |
2894 | let a = _mm256_set1_pd(1.1); |
2895 | let r = _mm256_roundscale_pd::<0b00_00_00_00>(a); |
2896 | let e = _mm256_set1_pd(1.0); |
2897 | assert_eq_m256d(r, e); |
2898 | } |
2899 | |
2900 | #[simd_test(enable = "avx512f,avx512vl" )] |
2901 | unsafe fn test_mm256_mask_roundscale_pd() { |
2902 | let a = _mm256_set1_pd(1.1); |
2903 | let r = _mm256_mask_roundscale_pd::<0b00_00_00_00>(a, 0, a); |
2904 | assert_eq_m256d(r, a); |
2905 | let r = _mm256_mask_roundscale_pd::<0b00_00_00_00>(a, 0b00001111, a); |
2906 | let e = _mm256_set1_pd(1.0); |
2907 | assert_eq_m256d(r, e); |
2908 | } |
2909 | |
2910 | #[simd_test(enable = "avx512f,avx512vl" )] |
2911 | unsafe fn test_mm256_maskz_roundscale_pd() { |
2912 | let a = _mm256_set1_pd(1.1); |
2913 | let r = _mm256_maskz_roundscale_pd::<0b00_00_00_00>(0, a); |
2914 | assert_eq_m256d(r, _mm256_setzero_pd()); |
2915 | let r = _mm256_maskz_roundscale_pd::<0b00_00_00_00>(0b00001111, a); |
2916 | let e = _mm256_set1_pd(1.0); |
2917 | assert_eq_m256d(r, e); |
2918 | } |
2919 | |
2920 | #[simd_test(enable = "avx512f,avx512vl" )] |
2921 | unsafe fn test_mm_roundscale_pd() { |
2922 | let a = _mm_set1_pd(1.1); |
2923 | let r = _mm_roundscale_pd::<0b00_00_00_00>(a); |
2924 | let e = _mm_set1_pd(1.0); |
2925 | assert_eq_m128d(r, e); |
2926 | } |
2927 | |
2928 | #[simd_test(enable = "avx512f,avx512vl" )] |
2929 | unsafe fn test_mm_mask_roundscale_pd() { |
2930 | let a = _mm_set1_pd(1.1); |
2931 | let r = _mm_mask_roundscale_pd::<0b00_00_00_00>(a, 0, a); |
2932 | let e = _mm_set1_pd(1.1); |
2933 | assert_eq_m128d(r, e); |
2934 | let r = _mm_mask_roundscale_pd::<0b00_00_00_00>(a, 0b00000011, a); |
2935 | let e = _mm_set1_pd(1.0); |
2936 | assert_eq_m128d(r, e); |
2937 | } |
2938 | |
2939 | #[simd_test(enable = "avx512f,avx512vl" )] |
2940 | unsafe fn test_mm_maskz_roundscale_pd() { |
2941 | let a = _mm_set1_pd(1.1); |
2942 | let r = _mm_maskz_roundscale_pd::<0b00_00_00_00>(0, a); |
2943 | assert_eq_m128d(r, _mm_setzero_pd()); |
2944 | let r = _mm_maskz_roundscale_pd::<0b00_00_00_00>(0b00000011, a); |
2945 | let e = _mm_set1_pd(1.0); |
2946 | assert_eq_m128d(r, e); |
2947 | } |
2948 | |
2949 | #[simd_test(enable = "avx512f" )] |
2950 | unsafe fn test_mm512_scalef_pd() { |
2951 | let a = _mm512_set1_pd(1.); |
2952 | let b = _mm512_set1_pd(3.); |
2953 | let r = _mm512_scalef_pd(a, b); |
2954 | let e = _mm512_set1_pd(8.); |
2955 | assert_eq_m512d(r, e); |
2956 | } |
2957 | |
2958 | #[simd_test(enable = "avx512f" )] |
2959 | unsafe fn test_mm512_mask_scalef_pd() { |
2960 | let a = _mm512_set1_pd(1.); |
2961 | let b = _mm512_set1_pd(3.); |
2962 | let r = _mm512_mask_scalef_pd(a, 0, a, b); |
2963 | assert_eq_m512d(r, a); |
2964 | let r = _mm512_mask_scalef_pd(a, 0b11110000, a, b); |
2965 | let e = _mm512_set_pd(8., 8., 8., 8., 1., 1., 1., 1.); |
2966 | assert_eq_m512d(r, e); |
2967 | } |
2968 | |
2969 | #[simd_test(enable = "avx512f" )] |
2970 | unsafe fn test_mm512_maskz_scalef_pd() { |
2971 | let a = _mm512_set1_pd(1.); |
2972 | let b = _mm512_set1_pd(3.); |
2973 | let r = _mm512_maskz_scalef_pd(0, a, b); |
2974 | assert_eq_m512d(r, _mm512_setzero_pd()); |
2975 | let r = _mm512_maskz_scalef_pd(0b11110000, a, b); |
2976 | let e = _mm512_set_pd(8., 8., 8., 8., 0., 0., 0., 0.); |
2977 | assert_eq_m512d(r, e); |
2978 | } |
2979 | |
2980 | #[simd_test(enable = "avx512f,avx512vl" )] |
2981 | unsafe fn test_mm256_scalef_pd() { |
2982 | let a = _mm256_set1_pd(1.); |
2983 | let b = _mm256_set1_pd(3.); |
2984 | let r = _mm256_scalef_pd(a, b); |
2985 | let e = _mm256_set1_pd(8.); |
2986 | assert_eq_m256d(r, e); |
2987 | } |
2988 | |
2989 | #[simd_test(enable = "avx512f,avx512vl" )] |
2990 | unsafe fn test_mm256_mask_scalef_pd() { |
2991 | let a = _mm256_set1_pd(1.); |
2992 | let b = _mm256_set1_pd(3.); |
2993 | let r = _mm256_mask_scalef_pd(a, 0, a, b); |
2994 | assert_eq_m256d(r, a); |
2995 | let r = _mm256_mask_scalef_pd(a, 0b00001111, a, b); |
2996 | let e = _mm256_set1_pd(8.); |
2997 | assert_eq_m256d(r, e); |
2998 | } |
2999 | |
3000 | #[simd_test(enable = "avx512f,avx512vl" )] |
3001 | unsafe fn test_mm256_maskz_scalef_pd() { |
3002 | let a = _mm256_set1_pd(1.); |
3003 | let b = _mm256_set1_pd(3.); |
3004 | let r = _mm256_maskz_scalef_pd(0, a, b); |
3005 | assert_eq_m256d(r, _mm256_setzero_pd()); |
3006 | let r = _mm256_maskz_scalef_pd(0b00001111, a, b); |
3007 | let e = _mm256_set1_pd(8.); |
3008 | assert_eq_m256d(r, e); |
3009 | } |
3010 | |
3011 | #[simd_test(enable = "avx512f,avx512vl" )] |
3012 | unsafe fn test_mm_scalef_pd() { |
3013 | let a = _mm_set1_pd(1.); |
3014 | let b = _mm_set1_pd(3.); |
3015 | let r = _mm_scalef_pd(a, b); |
3016 | let e = _mm_set1_pd(8.); |
3017 | assert_eq_m128d(r, e); |
3018 | } |
3019 | |
3020 | #[simd_test(enable = "avx512f,avx512vl" )] |
3021 | unsafe fn test_mm_mask_scalef_pd() { |
3022 | let a = _mm_set1_pd(1.); |
3023 | let b = _mm_set1_pd(3.); |
3024 | let r = _mm_mask_scalef_pd(a, 0, a, b); |
3025 | assert_eq_m128d(r, a); |
3026 | let r = _mm_mask_scalef_pd(a, 0b00000011, a, b); |
3027 | let e = _mm_set1_pd(8.); |
3028 | assert_eq_m128d(r, e); |
3029 | } |
3030 | |
3031 | #[simd_test(enable = "avx512f,avx512vl" )] |
3032 | unsafe fn test_mm_maskz_scalef_pd() { |
3033 | let a = _mm_set1_pd(1.); |
3034 | let b = _mm_set1_pd(3.); |
3035 | let r = _mm_maskz_scalef_pd(0, a, b); |
3036 | assert_eq_m128d(r, _mm_setzero_pd()); |
3037 | let r = _mm_maskz_scalef_pd(0b00000011, a, b); |
3038 | let e = _mm_set1_pd(8.); |
3039 | assert_eq_m128d(r, e); |
3040 | } |
3041 | |
3042 | #[simd_test(enable = "avx512f" )] |
3043 | unsafe fn test_mm512_fixupimm_pd() { |
3044 | let a = _mm512_set1_pd(f64::NAN); |
3045 | let b = _mm512_set1_pd(f64::MAX); |
3046 | let c = _mm512_set1_epi64(i32::MAX as i64); |
3047 | let r = _mm512_fixupimm_pd::<5>(a, b, c); |
3048 | let e = _mm512_set1_pd(0.0); |
3049 | assert_eq_m512d(r, e); |
3050 | } |
3051 | |
3052 | #[simd_test(enable = "avx512f" )] |
3053 | unsafe fn test_mm512_mask_fixupimm_pd() { |
3054 | let a = _mm512_set_pd(f64::NAN, f64::NAN, f64::NAN, f64::NAN, 1., 1., 1., 1.); |
3055 | let b = _mm512_set1_pd(f64::MAX); |
3056 | let c = _mm512_set1_epi64(i32::MAX as i64); |
3057 | let r = _mm512_mask_fixupimm_pd::<5>(a, 0b11110000, b, c); |
3058 | let e = _mm512_set_pd(0., 0., 0., 0., 1., 1., 1., 1.); |
3059 | assert_eq_m512d(r, e); |
3060 | } |
3061 | |
3062 | #[simd_test(enable = "avx512f" )] |
3063 | unsafe fn test_mm512_maskz_fixupimm_pd() { |
3064 | let a = _mm512_set_pd(f64::NAN, f64::NAN, f64::NAN, f64::NAN, 1., 1., 1., 1.); |
3065 | let b = _mm512_set1_pd(f64::MAX); |
3066 | let c = _mm512_set1_epi64(i32::MAX as i64); |
3067 | let r = _mm512_maskz_fixupimm_pd::<5>(0b11110000, a, b, c); |
3068 | let e = _mm512_set_pd(0., 0., 0., 0., 0., 0., 0., 0.); |
3069 | assert_eq_m512d(r, e); |
3070 | } |
3071 | |
3072 | #[simd_test(enable = "avx512f,avx512vl" )] |
3073 | unsafe fn test_mm256_fixupimm_pd() { |
3074 | let a = _mm256_set1_pd(f64::NAN); |
3075 | let b = _mm256_set1_pd(f64::MAX); |
3076 | let c = _mm256_set1_epi64x(i32::MAX as i64); |
3077 | let r = _mm256_fixupimm_pd::<5>(a, b, c); |
3078 | let e = _mm256_set1_pd(0.0); |
3079 | assert_eq_m256d(r, e); |
3080 | } |
3081 | |
3082 | #[simd_test(enable = "avx512f,avx512vl" )] |
3083 | unsafe fn test_mm256_mask_fixupimm_pd() { |
3084 | let a = _mm256_set1_pd(f64::NAN); |
3085 | let b = _mm256_set1_pd(f64::MAX); |
3086 | let c = _mm256_set1_epi64x(i32::MAX as i64); |
3087 | let r = _mm256_mask_fixupimm_pd::<5>(a, 0b00001111, b, c); |
3088 | let e = _mm256_set1_pd(0.0); |
3089 | assert_eq_m256d(r, e); |
3090 | } |
3091 | |
3092 | #[simd_test(enable = "avx512f,avx512vl" )] |
3093 | unsafe fn test_mm256_maskz_fixupimm_pd() { |
3094 | let a = _mm256_set1_pd(f64::NAN); |
3095 | let b = _mm256_set1_pd(f64::MAX); |
3096 | let c = _mm256_set1_epi64x(i32::MAX as i64); |
3097 | let r = _mm256_maskz_fixupimm_pd::<5>(0b00001111, a, b, c); |
3098 | let e = _mm256_set1_pd(0.0); |
3099 | assert_eq_m256d(r, e); |
3100 | } |
3101 | |
3102 | #[simd_test(enable = "avx512f,avx512vl" )] |
3103 | unsafe fn test_mm_fixupimm_pd() { |
3104 | let a = _mm_set1_pd(f64::NAN); |
3105 | let b = _mm_set1_pd(f64::MAX); |
3106 | let c = _mm_set1_epi64x(i32::MAX as i64); |
3107 | let r = _mm_fixupimm_pd::<5>(a, b, c); |
3108 | let e = _mm_set1_pd(0.0); |
3109 | assert_eq_m128d(r, e); |
3110 | } |
3111 | |
3112 | #[simd_test(enable = "avx512f,avx512vl" )] |
3113 | unsafe fn test_mm_mask_fixupimm_pd() { |
3114 | let a = _mm_set1_pd(f64::NAN); |
3115 | let b = _mm_set1_pd(f64::MAX); |
3116 | let c = _mm_set1_epi64x(i32::MAX as i64); |
3117 | let r = _mm_mask_fixupimm_pd::<5>(a, 0b00000011, b, c); |
3118 | let e = _mm_set1_pd(0.0); |
3119 | assert_eq_m128d(r, e); |
3120 | } |
3121 | |
3122 | #[simd_test(enable = "avx512f,avx512vl" )] |
3123 | unsafe fn test_mm_maskz_fixupimm_pd() { |
3124 | let a = _mm_set1_pd(f64::NAN); |
3125 | let b = _mm_set1_pd(f64::MAX); |
3126 | let c = _mm_set1_epi64x(i32::MAX as i64); |
3127 | let r = _mm_maskz_fixupimm_pd::<5>(0b00000011, a, b, c); |
3128 | let e = _mm_set1_pd(0.0); |
3129 | assert_eq_m128d(r, e); |
3130 | } |
3131 | |
3132 | #[simd_test(enable = "avx512f" )] |
3133 | unsafe fn test_mm512_ternarylogic_epi64() { |
3134 | let a = _mm512_set1_epi64(1 << 2); |
3135 | let b = _mm512_set1_epi64(1 << 1); |
3136 | let c = _mm512_set1_epi64(1 << 0); |
3137 | let r = _mm512_ternarylogic_epi64::<8>(a, b, c); |
3138 | let e = _mm512_set1_epi64(0); |
3139 | assert_eq_m512i(r, e); |
3140 | } |
3141 | |
3142 | #[simd_test(enable = "avx512f" )] |
3143 | unsafe fn test_mm512_mask_ternarylogic_epi64() { |
3144 | let src = _mm512_set1_epi64(1 << 2); |
3145 | let a = _mm512_set1_epi64(1 << 1); |
3146 | let b = _mm512_set1_epi64(1 << 0); |
3147 | let r = _mm512_mask_ternarylogic_epi64::<8>(src, 0, a, b); |
3148 | assert_eq_m512i(r, src); |
3149 | let r = _mm512_mask_ternarylogic_epi64::<8>(src, 0b11111111, a, b); |
3150 | let e = _mm512_set1_epi64(0); |
3151 | assert_eq_m512i(r, e); |
3152 | } |
3153 | |
3154 | #[simd_test(enable = "avx512f" )] |
3155 | unsafe fn test_mm512_maskz_ternarylogic_epi64() { |
3156 | let a = _mm512_set1_epi64(1 << 2); |
3157 | let b = _mm512_set1_epi64(1 << 1); |
3158 | let c = _mm512_set1_epi64(1 << 0); |
3159 | let r = _mm512_maskz_ternarylogic_epi64::<8>(0, a, b, c); |
3160 | assert_eq_m512i(r, _mm512_setzero_si512()); |
3161 | let r = _mm512_maskz_ternarylogic_epi64::<8>(0b11111111, a, b, c); |
3162 | let e = _mm512_set1_epi64(0); |
3163 | assert_eq_m512i(r, e); |
3164 | } |
3165 | |
3166 | #[simd_test(enable = "avx512f,avx512vl" )] |
3167 | unsafe fn test_mm256_ternarylogic_epi64() { |
3168 | let a = _mm256_set1_epi64x(1 << 2); |
3169 | let b = _mm256_set1_epi64x(1 << 1); |
3170 | let c = _mm256_set1_epi64x(1 << 0); |
3171 | let r = _mm256_ternarylogic_epi64::<8>(a, b, c); |
3172 | let e = _mm256_set1_epi64x(0); |
3173 | assert_eq_m256i(r, e); |
3174 | } |
3175 | |
3176 | #[simd_test(enable = "avx512f,avx512vl" )] |
3177 | unsafe fn test_mm256_mask_ternarylogic_epi64() { |
3178 | let src = _mm256_set1_epi64x(1 << 2); |
3179 | let a = _mm256_set1_epi64x(1 << 1); |
3180 | let b = _mm256_set1_epi64x(1 << 0); |
3181 | let r = _mm256_mask_ternarylogic_epi64::<8>(src, 0, a, b); |
3182 | assert_eq_m256i(r, src); |
3183 | let r = _mm256_mask_ternarylogic_epi64::<8>(src, 0b00001111, a, b); |
3184 | let e = _mm256_set1_epi64x(0); |
3185 | assert_eq_m256i(r, e); |
3186 | } |
3187 | |
3188 | #[simd_test(enable = "avx512f,avx512vl" )] |
3189 | unsafe fn test_mm256_maskz_ternarylogic_epi64() { |
3190 | let a = _mm256_set1_epi64x(1 << 2); |
3191 | let b = _mm256_set1_epi64x(1 << 1); |
3192 | let c = _mm256_set1_epi64x(1 << 0); |
3193 | let r = _mm256_maskz_ternarylogic_epi64::<9>(0, a, b, c); |
3194 | assert_eq_m256i(r, _mm256_setzero_si256()); |
3195 | let r = _mm256_maskz_ternarylogic_epi64::<8>(0b00001111, a, b, c); |
3196 | let e = _mm256_set1_epi64x(0); |
3197 | assert_eq_m256i(r, e); |
3198 | } |
3199 | |
3200 | #[simd_test(enable = "avx512f,avx512vl" )] |
3201 | unsafe fn test_mm_ternarylogic_epi64() { |
3202 | let a = _mm_set1_epi64x(1 << 2); |
3203 | let b = _mm_set1_epi64x(1 << 1); |
3204 | let c = _mm_set1_epi64x(1 << 0); |
3205 | let r = _mm_ternarylogic_epi64::<8>(a, b, c); |
3206 | let e = _mm_set1_epi64x(0); |
3207 | assert_eq_m128i(r, e); |
3208 | } |
3209 | |
3210 | #[simd_test(enable = "avx512f,avx512vl" )] |
3211 | unsafe fn test_mm_mask_ternarylogic_epi64() { |
3212 | let src = _mm_set1_epi64x(1 << 2); |
3213 | let a = _mm_set1_epi64x(1 << 1); |
3214 | let b = _mm_set1_epi64x(1 << 0); |
3215 | let r = _mm_mask_ternarylogic_epi64::<8>(src, 0, a, b); |
3216 | assert_eq_m128i(r, src); |
3217 | let r = _mm_mask_ternarylogic_epi64::<8>(src, 0b00000011, a, b); |
3218 | let e = _mm_set1_epi64x(0); |
3219 | assert_eq_m128i(r, e); |
3220 | } |
3221 | |
3222 | #[simd_test(enable = "avx512f,avx512vl" )] |
3223 | unsafe fn test_mm_maskz_ternarylogic_epi64() { |
3224 | let a = _mm_set1_epi64x(1 << 2); |
3225 | let b = _mm_set1_epi64x(1 << 1); |
3226 | let c = _mm_set1_epi64x(1 << 0); |
3227 | let r = _mm_maskz_ternarylogic_epi64::<9>(0, a, b, c); |
3228 | assert_eq_m128i(r, _mm_setzero_si128()); |
3229 | let r = _mm_maskz_ternarylogic_epi64::<8>(0b00000011, a, b, c); |
3230 | let e = _mm_set1_epi64x(0); |
3231 | assert_eq_m128i(r, e); |
3232 | } |
3233 | |
3234 | #[simd_test(enable = "avx512f" )] |
3235 | unsafe fn test_mm512_getmant_pd() { |
3236 | let a = _mm512_set1_pd(10.); |
3237 | let r = _mm512_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(a); |
3238 | let e = _mm512_set1_pd(1.25); |
3239 | assert_eq_m512d(r, e); |
3240 | } |
3241 | |
3242 | #[simd_test(enable = "avx512f" )] |
3243 | unsafe fn test_mm512_mask_getmant_pd() { |
3244 | let a = _mm512_set1_pd(10.); |
3245 | let r = _mm512_mask_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(a, 0, a); |
3246 | assert_eq_m512d(r, a); |
3247 | let r = _mm512_mask_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(a, 0b11110000, a); |
3248 | let e = _mm512_setr_pd(10., 10., 10., 10., 1.25, 1.25, 1.25, 1.25); |
3249 | assert_eq_m512d(r, e); |
3250 | } |
3251 | |
3252 | #[simd_test(enable = "avx512f" )] |
3253 | unsafe fn test_mm512_maskz_getmant_pd() { |
3254 | let a = _mm512_set1_pd(10.); |
3255 | let r = _mm512_maskz_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(0, a); |
3256 | assert_eq_m512d(r, _mm512_setzero_pd()); |
3257 | let r = _mm512_maskz_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(0b11110000, a); |
3258 | let e = _mm512_setr_pd(0., 0., 0., 0., 1.25, 1.25, 1.25, 1.25); |
3259 | assert_eq_m512d(r, e); |
3260 | } |
3261 | |
3262 | #[simd_test(enable = "avx512f,avx512vl" )] |
3263 | unsafe fn test_mm256_getmant_pd() { |
3264 | let a = _mm256_set1_pd(10.); |
3265 | let r = _mm256_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(a); |
3266 | let e = _mm256_set1_pd(1.25); |
3267 | assert_eq_m256d(r, e); |
3268 | } |
3269 | |
3270 | #[simd_test(enable = "avx512f,avx512vl" )] |
3271 | unsafe fn test_mm256_mask_getmant_pd() { |
3272 | let a = _mm256_set1_pd(10.); |
3273 | let r = _mm256_mask_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(a, 0, a); |
3274 | assert_eq_m256d(r, a); |
3275 | let r = _mm256_mask_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(a, 0b00001111, a); |
3276 | let e = _mm256_set1_pd(1.25); |
3277 | assert_eq_m256d(r, e); |
3278 | } |
3279 | |
3280 | #[simd_test(enable = "avx512f,avx512vl" )] |
3281 | unsafe fn test_mm256_maskz_getmant_pd() { |
3282 | let a = _mm256_set1_pd(10.); |
3283 | let r = _mm256_maskz_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(0, a); |
3284 | assert_eq_m256d(r, _mm256_setzero_pd()); |
3285 | let r = _mm256_maskz_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(0b00001111, a); |
3286 | let e = _mm256_set1_pd(1.25); |
3287 | assert_eq_m256d(r, e); |
3288 | } |
3289 | |
3290 | #[simd_test(enable = "avx512f,avx512vl" )] |
3291 | unsafe fn test_mm_getmant_pd() { |
3292 | let a = _mm_set1_pd(10.); |
3293 | let r = _mm_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(a); |
3294 | let e = _mm_set1_pd(1.25); |
3295 | assert_eq_m128d(r, e); |
3296 | } |
3297 | |
3298 | #[simd_test(enable = "avx512f,avx512vl" )] |
3299 | unsafe fn test_mm_mask_getmant_pd() { |
3300 | let a = _mm_set1_pd(10.); |
3301 | let r = _mm_mask_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(a, 0, a); |
3302 | assert_eq_m128d(r, a); |
3303 | let r = _mm_mask_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(a, 0b00000011, a); |
3304 | let e = _mm_set1_pd(1.25); |
3305 | assert_eq_m128d(r, e); |
3306 | } |
3307 | |
3308 | #[simd_test(enable = "avx512f,avx512vl" )] |
3309 | unsafe fn test_mm_maskz_getmant_pd() { |
3310 | let a = _mm_set1_pd(10.); |
3311 | let r = _mm_maskz_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(0, a); |
3312 | assert_eq_m128d(r, _mm_setzero_pd()); |
3313 | let r = _mm_maskz_getmant_pd::<_MM_MANT_NORM_1_2, _MM_MANT_SIGN_SRC>(0b00000011, a); |
3314 | let e = _mm_set1_pd(1.25); |
3315 | assert_eq_m128d(r, e); |
3316 | } |
3317 | |
3318 | #[simd_test(enable = "avx512f" )] |
3319 | unsafe fn test_mm512_cvtps_pd() { |
3320 | let a = _mm256_setr_ps(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3321 | let r = _mm512_cvtps_pd(a); |
3322 | let e = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3323 | assert_eq_m512d(r, e); |
3324 | } |
3325 | |
3326 | #[simd_test(enable = "avx512f" )] |
3327 | unsafe fn test_mm512_mask_cvtps_pd() { |
3328 | let a = _mm256_setr_ps(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3329 | let src = _mm512_set1_pd(0.); |
3330 | let r = _mm512_mask_cvtps_pd(src, 0, a); |
3331 | assert_eq_m512d(r, src); |
3332 | let r = _mm512_mask_cvtps_pd(src, 0b00001111, a); |
3333 | let e = _mm512_setr_pd(0., -1.5, 2., -3.5, 0., 0., 0., 0.); |
3334 | assert_eq_m512d(r, e); |
3335 | } |
3336 | |
3337 | #[simd_test(enable = "avx512f" )] |
3338 | unsafe fn test_mm512_maskz_cvtps_pd() { |
3339 | let a = _mm256_setr_ps(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3340 | let r = _mm512_maskz_cvtps_pd(0, a); |
3341 | assert_eq_m512d(r, _mm512_setzero_pd()); |
3342 | let r = _mm512_maskz_cvtps_pd(0b00001111, a); |
3343 | let e = _mm512_setr_pd(0., -1.5, 2., -3.5, 0., 0., 0., 0.); |
3344 | assert_eq_m512d(r, e); |
3345 | } |
3346 | |
3347 | #[simd_test(enable = "avx512f" )] |
3348 | unsafe fn test_mm512_cvtpslo_pd() { |
3349 | let v2 = _mm512_setr_ps( |
3350 | 0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5, 100., 100., 100., 100., 100., 100., 100., 100., |
3351 | ); |
3352 | let r = _mm512_cvtpslo_pd(v2); |
3353 | let e = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3354 | assert_eq_m512d(r, e); |
3355 | } |
3356 | |
3357 | #[simd_test(enable = "avx512f" )] |
3358 | unsafe fn test_mm512_mask_cvtpslo_pd() { |
3359 | let v2 = _mm512_setr_ps( |
3360 | 0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5, 100., 100., 100., 100., 100., 100., 100., 100., |
3361 | ); |
3362 | let src = _mm512_set1_pd(0.); |
3363 | let r = _mm512_mask_cvtpslo_pd(src, 0, v2); |
3364 | assert_eq_m512d(r, src); |
3365 | let r = _mm512_mask_cvtpslo_pd(src, 0b00001111, v2); |
3366 | let e = _mm512_setr_pd(0., -1.5, 2., -3.5, 0., 0., 0., 0.); |
3367 | assert_eq_m512d(r, e); |
3368 | } |
3369 | |
3370 | #[simd_test(enable = "avx512f" )] |
3371 | unsafe fn test_mm512_cvtpd_ps() { |
3372 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3373 | let r = _mm512_cvtpd_ps(a); |
3374 | let e = _mm256_setr_ps(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3375 | assert_eq_m256(r, e); |
3376 | } |
3377 | |
3378 | #[simd_test(enable = "avx512f" )] |
3379 | unsafe fn test_mm512_mask_cvtpd_ps() { |
3380 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3381 | let src = _mm256_set1_ps(0.); |
3382 | let r = _mm512_mask_cvtpd_ps(src, 0, a); |
3383 | assert_eq_m256(r, src); |
3384 | let r = _mm512_mask_cvtpd_ps(src, 0b00001111, a); |
3385 | let e = _mm256_setr_ps(0., -1.5, 2., -3.5, 0., 0., 0., 0.); |
3386 | assert_eq_m256(r, e); |
3387 | } |
3388 | |
3389 | #[simd_test(enable = "avx512f" )] |
3390 | unsafe fn test_mm512_maskz_cvtpd_ps() { |
3391 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3392 | let r = _mm512_maskz_cvtpd_ps(0, a); |
3393 | assert_eq_m256(r, _mm256_setzero_ps()); |
3394 | let r = _mm512_maskz_cvtpd_ps(0b00001111, a); |
3395 | let e = _mm256_setr_ps(0., -1.5, 2., -3.5, 0., 0., 0., 0.); |
3396 | assert_eq_m256(r, e); |
3397 | } |
3398 | |
3399 | #[simd_test(enable = "avx512f,avx512vl" )] |
3400 | unsafe fn test_mm256_mask_cvtpd_ps() { |
3401 | let a = _mm256_set_pd(4., -5.5, 6., -7.5); |
3402 | let src = _mm_set1_ps(0.); |
3403 | let r = _mm256_mask_cvtpd_ps(src, 0, a); |
3404 | assert_eq_m128(r, src); |
3405 | let r = _mm256_mask_cvtpd_ps(src, 0b00001111, a); |
3406 | let e = _mm_set_ps(4., -5.5, 6., -7.5); |
3407 | assert_eq_m128(r, e); |
3408 | } |
3409 | |
3410 | #[simd_test(enable = "avx512f,avx512vl" )] |
3411 | unsafe fn test_mm256_maskz_cvtpd_ps() { |
3412 | let a = _mm256_set_pd(4., -5.5, 6., -7.5); |
3413 | let r = _mm256_maskz_cvtpd_ps(0, a); |
3414 | assert_eq_m128(r, _mm_setzero_ps()); |
3415 | let r = _mm256_maskz_cvtpd_ps(0b00001111, a); |
3416 | let e = _mm_set_ps(4., -5.5, 6., -7.5); |
3417 | assert_eq_m128(r, e); |
3418 | } |
3419 | |
3420 | #[simd_test(enable = "avx512f,avx512vl" )] |
3421 | unsafe fn test_mm_mask_cvtpd_ps() { |
3422 | let a = _mm_set_pd(6., -7.5); |
3423 | let src = _mm_set1_ps(0.); |
3424 | let r = _mm_mask_cvtpd_ps(src, 0, a); |
3425 | assert_eq_m128(r, src); |
3426 | let r = _mm_mask_cvtpd_ps(src, 0b00000011, a); |
3427 | let e = _mm_set_ps(0., 0., 6., -7.5); |
3428 | assert_eq_m128(r, e); |
3429 | } |
3430 | |
3431 | #[simd_test(enable = "avx512f,avx512vl" )] |
3432 | unsafe fn test_mm_maskz_cvtpd_ps() { |
3433 | let a = _mm_set_pd(6., -7.5); |
3434 | let r = _mm_maskz_cvtpd_ps(0, a); |
3435 | assert_eq_m128(r, _mm_setzero_ps()); |
3436 | let r = _mm_maskz_cvtpd_ps(0b00000011, a); |
3437 | let e = _mm_set_ps(0., 0., 6., -7.5); |
3438 | assert_eq_m128(r, e); |
3439 | } |
3440 | |
3441 | #[simd_test(enable = "avx512f" )] |
3442 | unsafe fn test_mm512_cvtpd_epi32() { |
3443 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3444 | let r = _mm512_cvtpd_epi32(a); |
3445 | let e = _mm256_setr_epi32(0, -2, 2, -4, 4, -6, 6, -8); |
3446 | assert_eq_m256i(r, e); |
3447 | } |
3448 | |
3449 | #[simd_test(enable = "avx512f" )] |
3450 | unsafe fn test_mm512_mask_cvtpd_epi32() { |
3451 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3452 | let src = _mm256_set1_epi32(0); |
3453 | let r = _mm512_mask_cvtpd_epi32(src, 0, a); |
3454 | assert_eq_m256i(r, src); |
3455 | let r = _mm512_mask_cvtpd_epi32(src, 0b11111111, a); |
3456 | let e = _mm256_setr_epi32(0, -2, 2, -4, 4, -6, 6, -8); |
3457 | assert_eq_m256i(r, e); |
3458 | } |
3459 | |
3460 | #[simd_test(enable = "avx512f" )] |
3461 | unsafe fn test_mm512_maskz_cvtpd_epi32() { |
3462 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3463 | let r = _mm512_maskz_cvtpd_epi32(0, a); |
3464 | assert_eq_m256i(r, _mm256_setzero_si256()); |
3465 | let r = _mm512_maskz_cvtpd_epi32(0b11111111, a); |
3466 | let e = _mm256_setr_epi32(0, -2, 2, -4, 4, -6, 6, -8); |
3467 | assert_eq_m256i(r, e); |
3468 | } |
3469 | |
3470 | #[simd_test(enable = "avx512f,avx512vl" )] |
3471 | unsafe fn test_mm256_mask_cvtpd_epi32() { |
3472 | let a = _mm256_set_pd(4., -5.5, 6., -7.5); |
3473 | let src = _mm_set1_epi32(0); |
3474 | let r = _mm256_mask_cvtpd_epi32(src, 0, a); |
3475 | assert_eq_m128i(r, src); |
3476 | let r = _mm256_mask_cvtpd_epi32(src, 0b00001111, a); |
3477 | let e = _mm_set_epi32(4, -6, 6, -8); |
3478 | assert_eq_m128i(r, e); |
3479 | } |
3480 | |
3481 | #[simd_test(enable = "avx512f,avx512vl" )] |
3482 | unsafe fn test_mm256_maskz_cvtpd_epi32() { |
3483 | let a = _mm256_set_pd(4., -5.5, 6., -7.5); |
3484 | let r = _mm256_maskz_cvtpd_epi32(0, a); |
3485 | assert_eq_m128i(r, _mm_setzero_si128()); |
3486 | let r = _mm256_maskz_cvtpd_epi32(0b00001111, a); |
3487 | let e = _mm_set_epi32(4, -6, 6, -8); |
3488 | assert_eq_m128i(r, e); |
3489 | } |
3490 | |
3491 | #[simd_test(enable = "avx512f,avx512vl" )] |
3492 | unsafe fn test_mm_mask_cvtpd_epi32() { |
3493 | let a = _mm_set_pd(6., -7.5); |
3494 | let src = _mm_set1_epi32(0); |
3495 | let r = _mm_mask_cvtpd_epi32(src, 0, a); |
3496 | assert_eq_m128i(r, src); |
3497 | let r = _mm_mask_cvtpd_epi32(src, 0b00000011, a); |
3498 | let e = _mm_set_epi32(0, 0, 6, -8); |
3499 | assert_eq_m128i(r, e); |
3500 | } |
3501 | |
3502 | #[simd_test(enable = "avx512f,avx512vl" )] |
3503 | unsafe fn test_mm_maskz_cvtpd_epi32() { |
3504 | let a = _mm_set_pd(6., -7.5); |
3505 | let r = _mm_maskz_cvtpd_epi32(0, a); |
3506 | assert_eq_m128i(r, _mm_setzero_si128()); |
3507 | let r = _mm_maskz_cvtpd_epi32(0b00000011, a); |
3508 | let e = _mm_set_epi32(0, 0, 6, -8); |
3509 | assert_eq_m128i(r, e); |
3510 | } |
3511 | |
3512 | #[simd_test(enable = "avx512f" )] |
3513 | unsafe fn test_mm512_cvtpd_epu32() { |
3514 | let a = _mm512_setr_pd(0., 1.5, 2., 3.5, 4., 5.5, 6., 7.5); |
3515 | let r = _mm512_cvtpd_epu32(a); |
3516 | let e = _mm256_setr_epi32(0, 2, 2, 4, 4, 6, 6, 8); |
3517 | assert_eq_m256i(r, e); |
3518 | } |
3519 | |
3520 | #[simd_test(enable = "avx512f" )] |
3521 | unsafe fn test_mm512_mask_cvtpd_epu32() { |
3522 | let a = _mm512_setr_pd(0., 1.5, 2., 3.5, 4., 5.5, 6., 7.5); |
3523 | let src = _mm256_set1_epi32(0); |
3524 | let r = _mm512_mask_cvtpd_epu32(src, 0, a); |
3525 | assert_eq_m256i(r, src); |
3526 | let r = _mm512_mask_cvtpd_epu32(src, 0b11111111, a); |
3527 | let e = _mm256_setr_epi32(0, 2, 2, 4, 4, 6, 6, 8); |
3528 | assert_eq_m256i(r, e); |
3529 | } |
3530 | |
3531 | #[simd_test(enable = "avx512f" )] |
3532 | unsafe fn test_mm512_maskz_cvtpd_epu32() { |
3533 | let a = _mm512_setr_pd(0., 1.5, 2., 3.5, 4., 5.5, 6., 7.5); |
3534 | let r = _mm512_maskz_cvtpd_epu32(0, a); |
3535 | assert_eq_m256i(r, _mm256_setzero_si256()); |
3536 | let r = _mm512_maskz_cvtpd_epu32(0b11111111, a); |
3537 | let e = _mm256_setr_epi32(0, 2, 2, 4, 4, 6, 6, 8); |
3538 | assert_eq_m256i(r, e); |
3539 | } |
3540 | |
3541 | #[simd_test(enable = "avx512f,avx512vl" )] |
3542 | unsafe fn test_mm256_cvtpd_epu32() { |
3543 | let a = _mm256_set_pd(4., 5.5, 6., 7.5); |
3544 | let r = _mm256_cvtpd_epu32(a); |
3545 | let e = _mm_set_epi32(4, 6, 6, 8); |
3546 | assert_eq_m128i(r, e); |
3547 | } |
3548 | |
3549 | #[simd_test(enable = "avx512f,avx512vl" )] |
3550 | unsafe fn test_mm256_mask_cvtpd_epu32() { |
3551 | let a = _mm256_set_pd(4., 5.5, 6., 7.5); |
3552 | let src = _mm_set1_epi32(0); |
3553 | let r = _mm256_mask_cvtpd_epu32(src, 0, a); |
3554 | assert_eq_m128i(r, src); |
3555 | let r = _mm256_mask_cvtpd_epu32(src, 0b00001111, a); |
3556 | let e = _mm_set_epi32(4, 6, 6, 8); |
3557 | assert_eq_m128i(r, e); |
3558 | } |
3559 | |
3560 | #[simd_test(enable = "avx512f,avx512vl" )] |
3561 | unsafe fn test_mm256_maskz_cvtpd_epu32() { |
3562 | let a = _mm256_set_pd(4., 5.5, 6., 7.5); |
3563 | let r = _mm256_maskz_cvtpd_epu32(0, a); |
3564 | assert_eq_m128i(r, _mm_setzero_si128()); |
3565 | let r = _mm256_maskz_cvtpd_epu32(0b00001111, a); |
3566 | let e = _mm_set_epi32(4, 6, 6, 8); |
3567 | assert_eq_m128i(r, e); |
3568 | } |
3569 | |
3570 | #[simd_test(enable = "avx512f,avx512vl" )] |
3571 | unsafe fn test_mm_cvtpd_epu32() { |
3572 | let a = _mm_set_pd(6., 7.5); |
3573 | let r = _mm_cvtpd_epu32(a); |
3574 | let e = _mm_set_epi32(0, 0, 6, 8); |
3575 | assert_eq_m128i(r, e); |
3576 | } |
3577 | |
3578 | #[simd_test(enable = "avx512f,avx512vl" )] |
3579 | unsafe fn test_mm_mask_cvtpd_epu32() { |
3580 | let a = _mm_set_pd(6., 7.5); |
3581 | let src = _mm_set1_epi32(0); |
3582 | let r = _mm_mask_cvtpd_epu32(src, 0, a); |
3583 | assert_eq_m128i(r, src); |
3584 | let r = _mm_mask_cvtpd_epu32(src, 0b00000011, a); |
3585 | let e = _mm_set_epi32(0, 0, 6, 8); |
3586 | assert_eq_m128i(r, e); |
3587 | } |
3588 | |
3589 | #[simd_test(enable = "avx512f,avx512vl" )] |
3590 | unsafe fn test_mm_maskz_cvtpd_epu32() { |
3591 | let a = _mm_set_pd(6., 7.5); |
3592 | let r = _mm_maskz_cvtpd_epu32(0, a); |
3593 | assert_eq_m128i(r, _mm_setzero_si128()); |
3594 | let r = _mm_maskz_cvtpd_epu32(0b00000011, a); |
3595 | let e = _mm_set_epi32(0, 0, 6, 8); |
3596 | assert_eq_m128i(r, e); |
3597 | } |
3598 | |
3599 | #[simd_test(enable = "avx512f" )] |
3600 | unsafe fn test_mm512_cvtpd_pslo() { |
3601 | let v2 = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3602 | let r = _mm512_cvtpd_pslo(v2); |
3603 | let e = _mm512_setr_ps( |
3604 | 0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5, 0., 0., 0., 0., 0., 0., 0., 0., |
3605 | ); |
3606 | assert_eq_m512(r, e); |
3607 | } |
3608 | |
3609 | #[simd_test(enable = "avx512f" )] |
3610 | unsafe fn test_mm512_mask_cvtpd_pslo() { |
3611 | let v2 = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
3612 | let src = _mm512_set1_ps(0.); |
3613 | let r = _mm512_mask_cvtpd_pslo(src, 0, v2); |
3614 | assert_eq_m512(r, src); |
3615 | let r = _mm512_mask_cvtpd_pslo(src, 0b00001111, v2); |
3616 | let e = _mm512_setr_ps( |
3617 | 0., -1.5, 2., -3.5, 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., |
3618 | ); |
3619 | assert_eq_m512(r, e); |
3620 | } |
3621 | |
3622 | #[simd_test(enable = "avx512f" )] |
3623 | unsafe fn test_mm512_cvtepi8_epi64() { |
3624 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3625 | let r = _mm512_cvtepi8_epi64(a); |
3626 | let e = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
3627 | assert_eq_m512i(r, e); |
3628 | } |
3629 | |
3630 | #[simd_test(enable = "avx512f" )] |
3631 | unsafe fn test_mm512_mask_cvtepi8_epi64() { |
3632 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3633 | let src = _mm512_set1_epi64(-1); |
3634 | let r = _mm512_mask_cvtepi8_epi64(src, 0, a); |
3635 | assert_eq_m512i(r, src); |
3636 | let r = _mm512_mask_cvtepi8_epi64(src, 0b00001111, a); |
3637 | let e = _mm512_set_epi64(-1, -1, -1, -1, 12, 13, 14, 15); |
3638 | assert_eq_m512i(r, e); |
3639 | } |
3640 | |
3641 | #[simd_test(enable = "avx512f" )] |
3642 | unsafe fn test_mm512_maskz_cvtepi8_epi64() { |
3643 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3644 | let r = _mm512_maskz_cvtepi8_epi64(0, a); |
3645 | assert_eq_m512i(r, _mm512_setzero_si512()); |
3646 | let r = _mm512_maskz_cvtepi8_epi64(0b00001111, a); |
3647 | let e = _mm512_set_epi64(0, 0, 0, 0, 12, 13, 14, 15); |
3648 | assert_eq_m512i(r, e); |
3649 | } |
3650 | |
3651 | #[simd_test(enable = "avx512f,avx512vl" )] |
3652 | unsafe fn test_mm256_mask_cvtepi8_epi64() { |
3653 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3654 | let src = _mm256_set1_epi64x(-1); |
3655 | let r = _mm256_mask_cvtepi8_epi64(src, 0, a); |
3656 | assert_eq_m256i(r, src); |
3657 | let r = _mm256_mask_cvtepi8_epi64(src, 0b00001111, a); |
3658 | let e = _mm256_set_epi64x(12, 13, 14, 15); |
3659 | assert_eq_m256i(r, e); |
3660 | } |
3661 | |
3662 | #[simd_test(enable = "avx512f,avx512vl" )] |
3663 | unsafe fn test_mm256_maskz_cvtepi8_epi64() { |
3664 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3665 | let r = _mm256_maskz_cvtepi8_epi64(0, a); |
3666 | assert_eq_m256i(r, _mm256_setzero_si256()); |
3667 | let r = _mm256_maskz_cvtepi8_epi64(0b00001111, a); |
3668 | let e = _mm256_set_epi64x(12, 13, 14, 15); |
3669 | assert_eq_m256i(r, e); |
3670 | } |
3671 | |
3672 | #[simd_test(enable = "avx512f,avx512vl" )] |
3673 | unsafe fn test_mm_mask_cvtepi8_epi64() { |
3674 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3675 | let src = _mm_set1_epi64x(-1); |
3676 | let r = _mm_mask_cvtepi8_epi64(src, 0, a); |
3677 | assert_eq_m128i(r, src); |
3678 | let r = _mm_mask_cvtepi8_epi64(src, 0b00000011, a); |
3679 | let e = _mm_set_epi64x(14, 15); |
3680 | assert_eq_m128i(r, e); |
3681 | } |
3682 | |
3683 | #[simd_test(enable = "avx512f,avx512vl" )] |
3684 | unsafe fn test_mm_maskz_cvtepi8_epi64() { |
3685 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3686 | let r = _mm_maskz_cvtepi8_epi64(0, a); |
3687 | assert_eq_m128i(r, _mm_setzero_si128()); |
3688 | let r = _mm_maskz_cvtepi8_epi64(0b00000011, a); |
3689 | let e = _mm_set_epi64x(14, 15); |
3690 | assert_eq_m128i(r, e); |
3691 | } |
3692 | |
3693 | #[simd_test(enable = "avx512f" )] |
3694 | unsafe fn test_mm512_cvtepu8_epi64() { |
3695 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3696 | let r = _mm512_cvtepu8_epi64(a); |
3697 | let e = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
3698 | assert_eq_m512i(r, e); |
3699 | } |
3700 | |
3701 | #[simd_test(enable = "avx512f" )] |
3702 | unsafe fn test_mm512_mask_cvtepu8_epi64() { |
3703 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3704 | let src = _mm512_set1_epi64(-1); |
3705 | let r = _mm512_mask_cvtepu8_epi64(src, 0, a); |
3706 | assert_eq_m512i(r, src); |
3707 | let r = _mm512_mask_cvtepu8_epi64(src, 0b00001111, a); |
3708 | let e = _mm512_set_epi64(-1, -1, -1, -1, 12, 13, 14, 15); |
3709 | assert_eq_m512i(r, e); |
3710 | } |
3711 | |
3712 | #[simd_test(enable = "avx512f" )] |
3713 | unsafe fn test_mm512_maskz_cvtepu8_epi64() { |
3714 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3715 | let r = _mm512_maskz_cvtepu8_epi64(0, a); |
3716 | assert_eq_m512i(r, _mm512_setzero_si512()); |
3717 | let r = _mm512_maskz_cvtepu8_epi64(0b00001111, a); |
3718 | let e = _mm512_set_epi64(0, 0, 0, 0, 12, 13, 14, 15); |
3719 | assert_eq_m512i(r, e); |
3720 | } |
3721 | |
3722 | #[simd_test(enable = "avx512f,avx512vl" )] |
3723 | unsafe fn test_mm256_mask_cvtepu8_epi64() { |
3724 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3725 | let src = _mm256_set1_epi64x(-1); |
3726 | let r = _mm256_mask_cvtepu8_epi64(src, 0, a); |
3727 | assert_eq_m256i(r, src); |
3728 | let r = _mm256_mask_cvtepu8_epi64(src, 0b00001111, a); |
3729 | let e = _mm256_set_epi64x(12, 13, 14, 15); |
3730 | assert_eq_m256i(r, e); |
3731 | } |
3732 | |
3733 | #[simd_test(enable = "avx512f,avx512vl" )] |
3734 | unsafe fn test_mm256_maskz_cvtepu8_epi64() { |
3735 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3736 | let r = _mm256_maskz_cvtepu8_epi64(0, a); |
3737 | assert_eq_m256i(r, _mm256_setzero_si256()); |
3738 | let r = _mm256_maskz_cvtepu8_epi64(0b00001111, a); |
3739 | let e = _mm256_set_epi64x(12, 13, 14, 15); |
3740 | assert_eq_m256i(r, e); |
3741 | } |
3742 | |
3743 | #[simd_test(enable = "avx512f,avx512vl" )] |
3744 | unsafe fn test_mm_mask_cvtepu8_epi64() { |
3745 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3746 | let src = _mm_set1_epi64x(-1); |
3747 | let r = _mm_mask_cvtepu8_epi64(src, 0, a); |
3748 | assert_eq_m128i(r, src); |
3749 | let r = _mm_mask_cvtepu8_epi64(src, 0b00000011, a); |
3750 | let e = _mm_set_epi64x(14, 15); |
3751 | assert_eq_m128i(r, e); |
3752 | } |
3753 | |
3754 | #[simd_test(enable = "avx512f,avx512vl" )] |
3755 | unsafe fn test_mm_maskz_cvtepu8_epi64() { |
3756 | let a = _mm_set_epi8(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
3757 | let r = _mm_maskz_cvtepu8_epi64(0, a); |
3758 | assert_eq_m128i(r, _mm_setzero_si128()); |
3759 | let r = _mm_maskz_cvtepu8_epi64(0b00000011, a); |
3760 | let e = _mm_set_epi64x(14, 15); |
3761 | assert_eq_m128i(r, e); |
3762 | } |
3763 | |
3764 | #[simd_test(enable = "avx512f" )] |
3765 | unsafe fn test_mm512_cvtepi16_epi64() { |
3766 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3767 | let r = _mm512_cvtepi16_epi64(a); |
3768 | let e = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
3769 | assert_eq_m512i(r, e); |
3770 | } |
3771 | |
3772 | #[simd_test(enable = "avx512f" )] |
3773 | unsafe fn test_mm512_mask_cvtepi16_epi64() { |
3774 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3775 | let src = _mm512_set1_epi64(-1); |
3776 | let r = _mm512_mask_cvtepi16_epi64(src, 0, a); |
3777 | assert_eq_m512i(r, src); |
3778 | let r = _mm512_mask_cvtepi16_epi64(src, 0b00001111, a); |
3779 | let e = _mm512_set_epi64(-1, -1, -1, -1, 12, 13, 14, 15); |
3780 | assert_eq_m512i(r, e); |
3781 | } |
3782 | |
3783 | #[simd_test(enable = "avx512f" )] |
3784 | unsafe fn test_mm512_maskz_cvtepi16_epi64() { |
3785 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3786 | let r = _mm512_maskz_cvtepi16_epi64(0, a); |
3787 | assert_eq_m512i(r, _mm512_setzero_si512()); |
3788 | let r = _mm512_maskz_cvtepi16_epi64(0b00001111, a); |
3789 | let e = _mm512_set_epi64(0, 0, 0, 0, 12, 13, 14, 15); |
3790 | assert_eq_m512i(r, e); |
3791 | } |
3792 | |
3793 | #[simd_test(enable = "avx512f,avx512vl" )] |
3794 | unsafe fn test_mm256_mask_cvtepi16_epi64() { |
3795 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3796 | let src = _mm256_set1_epi64x(-1); |
3797 | let r = _mm256_mask_cvtepi16_epi64(src, 0, a); |
3798 | assert_eq_m256i(r, src); |
3799 | let r = _mm256_mask_cvtepi16_epi64(src, 0b00001111, a); |
3800 | let e = _mm256_set_epi64x(12, 13, 14, 15); |
3801 | assert_eq_m256i(r, e); |
3802 | } |
3803 | |
3804 | #[simd_test(enable = "avx512f,avx512vl" )] |
3805 | unsafe fn test_mm256_maskz_cvtepi16_epi64() { |
3806 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3807 | let r = _mm256_maskz_cvtepi16_epi64(0, a); |
3808 | assert_eq_m256i(r, _mm256_setzero_si256()); |
3809 | let r = _mm256_maskz_cvtepi16_epi64(0b00001111, a); |
3810 | let e = _mm256_set_epi64x(12, 13, 14, 15); |
3811 | assert_eq_m256i(r, e); |
3812 | } |
3813 | |
3814 | #[simd_test(enable = "avx512f,avx512vl" )] |
3815 | unsafe fn test_mm_mask_cvtepi16_epi64() { |
3816 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3817 | let src = _mm_set1_epi64x(-1); |
3818 | let r = _mm_mask_cvtepi16_epi64(src, 0, a); |
3819 | assert_eq_m128i(r, src); |
3820 | let r = _mm_mask_cvtepi16_epi64(src, 0b00000011, a); |
3821 | let e = _mm_set_epi64x(14, 15); |
3822 | assert_eq_m128i(r, e); |
3823 | } |
3824 | |
3825 | #[simd_test(enable = "avx512f,avx512vl" )] |
3826 | unsafe fn test_mm_maskz_cvtepi16_epi64() { |
3827 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3828 | let r = _mm_maskz_cvtepi16_epi64(0, a); |
3829 | assert_eq_m128i(r, _mm_setzero_si128()); |
3830 | let r = _mm_maskz_cvtepi16_epi64(0b00000011, a); |
3831 | let e = _mm_set_epi64x(14, 15); |
3832 | assert_eq_m128i(r, e); |
3833 | } |
3834 | |
3835 | #[simd_test(enable = "avx512f" )] |
3836 | unsafe fn test_mm512_cvtepu16_epi64() { |
3837 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3838 | let r = _mm512_cvtepu16_epi64(a); |
3839 | let e = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
3840 | assert_eq_m512i(r, e); |
3841 | } |
3842 | |
3843 | #[simd_test(enable = "avx512f" )] |
3844 | unsafe fn test_mm512_mask_cvtepu16_epi64() { |
3845 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3846 | let src = _mm512_set1_epi64(-1); |
3847 | let r = _mm512_mask_cvtepu16_epi64(src, 0, a); |
3848 | assert_eq_m512i(r, src); |
3849 | let r = _mm512_mask_cvtepu16_epi64(src, 0b00001111, a); |
3850 | let e = _mm512_set_epi64(-1, -1, -1, -1, 12, 13, 14, 15); |
3851 | assert_eq_m512i(r, e); |
3852 | } |
3853 | |
3854 | #[simd_test(enable = "avx512f" )] |
3855 | unsafe fn test_mm512_maskz_cvtepu16_epi64() { |
3856 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3857 | let r = _mm512_maskz_cvtepu16_epi64(0, a); |
3858 | assert_eq_m512i(r, _mm512_setzero_si512()); |
3859 | let r = _mm512_maskz_cvtepu16_epi64(0b00001111, a); |
3860 | let e = _mm512_set_epi64(0, 0, 0, 0, 12, 13, 14, 15); |
3861 | assert_eq_m512i(r, e); |
3862 | } |
3863 | |
3864 | #[simd_test(enable = "avx512f,avx512vl" )] |
3865 | unsafe fn test_mm256_mask_cvtepu16_epi64() { |
3866 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3867 | let src = _mm256_set1_epi64x(-1); |
3868 | let r = _mm256_mask_cvtepu16_epi64(src, 0, a); |
3869 | assert_eq_m256i(r, src); |
3870 | let r = _mm256_mask_cvtepu16_epi64(src, 0b00001111, a); |
3871 | let e = _mm256_set_epi64x(12, 13, 14, 15); |
3872 | assert_eq_m256i(r, e); |
3873 | } |
3874 | |
3875 | #[simd_test(enable = "avx512f,avx512vl" )] |
3876 | unsafe fn test_mm256_maskz_cvtepu16_epi64() { |
3877 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3878 | let r = _mm256_maskz_cvtepu16_epi64(0, a); |
3879 | assert_eq_m256i(r, _mm256_setzero_si256()); |
3880 | let r = _mm256_maskz_cvtepu16_epi64(0b00001111, a); |
3881 | let e = _mm256_set_epi64x(12, 13, 14, 15); |
3882 | assert_eq_m256i(r, e); |
3883 | } |
3884 | |
3885 | #[simd_test(enable = "avx512f,avx512vl" )] |
3886 | unsafe fn test_mm_mask_cvtepu16_epi64() { |
3887 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3888 | let src = _mm_set1_epi64x(-1); |
3889 | let r = _mm_mask_cvtepu16_epi64(src, 0, a); |
3890 | assert_eq_m128i(r, src); |
3891 | let r = _mm_mask_cvtepu16_epi64(src, 0b00000011, a); |
3892 | let e = _mm_set_epi64x(14, 15); |
3893 | assert_eq_m128i(r, e); |
3894 | } |
3895 | |
3896 | #[simd_test(enable = "avx512f,avx512vl" )] |
3897 | unsafe fn test_mm_maskz_cvtepu16_epi64() { |
3898 | let a = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
3899 | let r = _mm_maskz_cvtepu16_epi64(0, a); |
3900 | assert_eq_m128i(r, _mm_setzero_si128()); |
3901 | let r = _mm_maskz_cvtepu16_epi64(0b00000011, a); |
3902 | let e = _mm_set_epi64x(14, 15); |
3903 | assert_eq_m128i(r, e); |
3904 | } |
3905 | |
3906 | #[simd_test(enable = "avx512f" )] |
3907 | unsafe fn test_mm512_cvtepi32_epi64() { |
3908 | let a = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
3909 | let r = _mm512_cvtepi32_epi64(a); |
3910 | let e = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
3911 | assert_eq_m512i(r, e); |
3912 | } |
3913 | |
3914 | #[simd_test(enable = "avx512f" )] |
3915 | unsafe fn test_mm512_mask_cvtepi32_epi64() { |
3916 | let a = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
3917 | let src = _mm512_set1_epi64(-1); |
3918 | let r = _mm512_mask_cvtepi32_epi64(src, 0, a); |
3919 | assert_eq_m512i(r, src); |
3920 | let r = _mm512_mask_cvtepi32_epi64(src, 0b00001111, a); |
3921 | let e = _mm512_set_epi64(-1, -1, -1, -1, 12, 13, 14, 15); |
3922 | assert_eq_m512i(r, e); |
3923 | } |
3924 | |
3925 | #[simd_test(enable = "avx512f" )] |
3926 | unsafe fn test_mm512_maskz_cvtepi32_epi64() { |
3927 | let a = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
3928 | let r = _mm512_maskz_cvtepi32_epi64(0, a); |
3929 | assert_eq_m512i(r, _mm512_setzero_si512()); |
3930 | let r = _mm512_maskz_cvtepi32_epi64(0b00001111, a); |
3931 | let e = _mm512_set_epi64(0, 0, 0, 0, 12, 13, 14, 15); |
3932 | assert_eq_m512i(r, e); |
3933 | } |
3934 | |
3935 | #[simd_test(enable = "avx512f,avx512vl" )] |
3936 | unsafe fn test_mm256_mask_cvtepi32_epi64() { |
3937 | let a = _mm_set_epi32(8, 9, 10, 11); |
3938 | let src = _mm256_set1_epi64x(-1); |
3939 | let r = _mm256_mask_cvtepi32_epi64(src, 0, a); |
3940 | assert_eq_m256i(r, src); |
3941 | let r = _mm256_mask_cvtepi32_epi64(src, 0b00001111, a); |
3942 | let e = _mm256_set_epi64x(8, 9, 10, 11); |
3943 | assert_eq_m256i(r, e); |
3944 | } |
3945 | |
3946 | #[simd_test(enable = "avx512f,avx512vl" )] |
3947 | unsafe fn test_mm256_maskz_cvtepi32_epi64() { |
3948 | let a = _mm_set_epi32(8, 9, 10, 11); |
3949 | let r = _mm256_maskz_cvtepi32_epi64(0, a); |
3950 | assert_eq_m256i(r, _mm256_setzero_si256()); |
3951 | let r = _mm256_maskz_cvtepi32_epi64(0b00001111, a); |
3952 | let e = _mm256_set_epi64x(8, 9, 10, 11); |
3953 | assert_eq_m256i(r, e); |
3954 | } |
3955 | |
3956 | #[simd_test(enable = "avx512f,avx512vl" )] |
3957 | unsafe fn test_mm_mask_cvtepi32_epi64() { |
3958 | let a = _mm_set_epi32(8, 9, 10, 11); |
3959 | let src = _mm_set1_epi64x(0); |
3960 | let r = _mm_mask_cvtepi32_epi64(src, 0, a); |
3961 | assert_eq_m128i(r, src); |
3962 | let r = _mm_mask_cvtepi32_epi64(src, 0b00000011, a); |
3963 | let e = _mm_set_epi64x(10, 11); |
3964 | assert_eq_m128i(r, e); |
3965 | } |
3966 | |
3967 | #[simd_test(enable = "avx512f,avx512vl" )] |
3968 | unsafe fn test_mm_maskz_cvtepi32_epi64() { |
3969 | let a = _mm_set_epi32(8, 9, 10, 11); |
3970 | let r = _mm_maskz_cvtepi32_epi64(0, a); |
3971 | assert_eq_m128i(r, _mm_setzero_si128()); |
3972 | let r = _mm_maskz_cvtepi32_epi64(0b00000011, a); |
3973 | let e = _mm_set_epi64x(10, 11); |
3974 | assert_eq_m128i(r, e); |
3975 | } |
3976 | |
3977 | #[simd_test(enable = "avx512f" )] |
3978 | unsafe fn test_mm512_cvtepu32_epi64() { |
3979 | let a = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
3980 | let r = _mm512_cvtepu32_epi64(a); |
3981 | let e = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
3982 | assert_eq_m512i(r, e); |
3983 | } |
3984 | |
3985 | #[simd_test(enable = "avx512f" )] |
3986 | unsafe fn test_mm512_mask_cvtepu32_epi64() { |
3987 | let a = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
3988 | let src = _mm512_set1_epi64(-1); |
3989 | let r = _mm512_mask_cvtepu32_epi64(src, 0, a); |
3990 | assert_eq_m512i(r, src); |
3991 | let r = _mm512_mask_cvtepu32_epi64(src, 0b00001111, a); |
3992 | let e = _mm512_set_epi64(-1, -1, -1, -1, 12, 13, 14, 15); |
3993 | assert_eq_m512i(r, e); |
3994 | } |
3995 | |
3996 | #[simd_test(enable = "avx512f" )] |
3997 | unsafe fn test_mm512_maskz_cvtepu32_epi64() { |
3998 | let a = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
3999 | let r = _mm512_maskz_cvtepu32_epi64(0, a); |
4000 | assert_eq_m512i(r, _mm512_setzero_si512()); |
4001 | let r = _mm512_maskz_cvtepu32_epi64(0b00001111, a); |
4002 | let e = _mm512_set_epi64(0, 0, 0, 0, 12, 13, 14, 15); |
4003 | assert_eq_m512i(r, e); |
4004 | } |
4005 | |
4006 | #[simd_test(enable = "avx512f,avx512vl" )] |
4007 | unsafe fn test_mm256_mask_cvtepu32_epi64() { |
4008 | let a = _mm_set_epi32(12, 13, 14, 15); |
4009 | let src = _mm256_set1_epi64x(-1); |
4010 | let r = _mm256_mask_cvtepu32_epi64(src, 0, a); |
4011 | assert_eq_m256i(r, src); |
4012 | let r = _mm256_mask_cvtepu32_epi64(src, 0b00001111, a); |
4013 | let e = _mm256_set_epi64x(12, 13, 14, 15); |
4014 | assert_eq_m256i(r, e); |
4015 | } |
4016 | |
4017 | #[simd_test(enable = "avx512f,avx512vl" )] |
4018 | unsafe fn test_mm256_maskz_cvtepu32_epi64() { |
4019 | let a = _mm_set_epi32(12, 13, 14, 15); |
4020 | let r = _mm256_maskz_cvtepu32_epi64(0, a); |
4021 | assert_eq_m256i(r, _mm256_setzero_si256()); |
4022 | let r = _mm256_maskz_cvtepu32_epi64(0b00001111, a); |
4023 | let e = _mm256_set_epi64x(12, 13, 14, 15); |
4024 | assert_eq_m256i(r, e); |
4025 | } |
4026 | |
4027 | #[simd_test(enable = "avx512f,avx512vl" )] |
4028 | unsafe fn test_mm_mask_cvtepu32_epi64() { |
4029 | let a = _mm_set_epi32(12, 13, 14, 15); |
4030 | let src = _mm_set1_epi64x(-1); |
4031 | let r = _mm_mask_cvtepu32_epi64(src, 0, a); |
4032 | assert_eq_m128i(r, src); |
4033 | let r = _mm_mask_cvtepu32_epi64(src, 0b00000011, a); |
4034 | let e = _mm_set_epi64x(14, 15); |
4035 | assert_eq_m128i(r, e); |
4036 | } |
4037 | |
4038 | #[simd_test(enable = "avx512f,avx512vl" )] |
4039 | unsafe fn test_mm_maskz_cvtepu32_epi64() { |
4040 | let a = _mm_set_epi32(12, 13, 14, 15); |
4041 | let r = _mm_maskz_cvtepu32_epi64(0, a); |
4042 | assert_eq_m128i(r, _mm_setzero_si128()); |
4043 | let r = _mm_maskz_cvtepu32_epi64(0b00000011, a); |
4044 | let e = _mm_set_epi64x(14, 15); |
4045 | assert_eq_m128i(r, e); |
4046 | } |
4047 | |
4048 | #[simd_test(enable = "avx512f" )] |
4049 | unsafe fn test_mm512_cvtepi32_pd() { |
4050 | let a = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
4051 | let r = _mm512_cvtepi32_pd(a); |
4052 | let e = _mm512_set_pd(8., 9., 10., 11., 12., 13., 14., 15.); |
4053 | assert_eq_m512d(r, e); |
4054 | } |
4055 | |
4056 | #[simd_test(enable = "avx512f" )] |
4057 | unsafe fn test_mm512_mask_cvtepi32_pd() { |
4058 | let a = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
4059 | let src = _mm512_set1_pd(-1.); |
4060 | let r = _mm512_mask_cvtepi32_pd(src, 0, a); |
4061 | assert_eq_m512d(r, src); |
4062 | let r = _mm512_mask_cvtepi32_pd(src, 0b00001111, a); |
4063 | let e = _mm512_set_pd(-1., -1., -1., -1., 12., 13., 14., 15.); |
4064 | assert_eq_m512d(r, e); |
4065 | } |
4066 | |
4067 | #[simd_test(enable = "avx512f" )] |
4068 | unsafe fn test_mm512_maskz_cvtepi32_pd() { |
4069 | let a = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
4070 | let r = _mm512_maskz_cvtepi32_pd(0, a); |
4071 | assert_eq_m512d(r, _mm512_setzero_pd()); |
4072 | let r = _mm512_maskz_cvtepi32_pd(0b00001111, a); |
4073 | let e = _mm512_set_pd(0., 0., 0., 0., 12., 13., 14., 15.); |
4074 | assert_eq_m512d(r, e); |
4075 | } |
4076 | |
4077 | #[simd_test(enable = "avx512f,avx512vl" )] |
4078 | unsafe fn test_mm256_mask_cvtepi32_pd() { |
4079 | let a = _mm_set_epi32(12, 13, 14, 15); |
4080 | let src = _mm256_set1_pd(-1.); |
4081 | let r = _mm256_mask_cvtepi32_pd(src, 0, a); |
4082 | assert_eq_m256d(r, src); |
4083 | let r = _mm256_mask_cvtepi32_pd(src, 0b00001111, a); |
4084 | let e = _mm256_set_pd(12., 13., 14., 15.); |
4085 | assert_eq_m256d(r, e); |
4086 | } |
4087 | |
4088 | #[simd_test(enable = "avx512f,avx512vl" )] |
4089 | unsafe fn test_mm256_maskz_cvtepi32_pd() { |
4090 | let a = _mm_set_epi32(12, 13, 14, 15); |
4091 | let r = _mm256_maskz_cvtepi32_pd(0, a); |
4092 | assert_eq_m256d(r, _mm256_setzero_pd()); |
4093 | let r = _mm256_maskz_cvtepi32_pd(0b00001111, a); |
4094 | let e = _mm256_set_pd(12., 13., 14., 15.); |
4095 | assert_eq_m256d(r, e); |
4096 | } |
4097 | |
4098 | #[simd_test(enable = "avx512f,avx512vl" )] |
4099 | unsafe fn test_mm_mask_cvtepi32_pd() { |
4100 | let a = _mm_set_epi32(12, 13, 14, 15); |
4101 | let src = _mm_set1_pd(-1.); |
4102 | let r = _mm_mask_cvtepi32_pd(src, 0, a); |
4103 | assert_eq_m128d(r, src); |
4104 | let r = _mm_mask_cvtepi32_pd(src, 0b00000011, a); |
4105 | let e = _mm_set_pd(14., 15.); |
4106 | assert_eq_m128d(r, e); |
4107 | } |
4108 | |
4109 | #[simd_test(enable = "avx512f,avx512vl" )] |
4110 | unsafe fn test_mm_maskz_cvtepi32_pd() { |
4111 | let a = _mm_set_epi32(12, 13, 14, 15); |
4112 | let r = _mm_maskz_cvtepi32_pd(0, a); |
4113 | assert_eq_m128d(r, _mm_setzero_pd()); |
4114 | let r = _mm_maskz_cvtepi32_pd(0b00000011, a); |
4115 | let e = _mm_set_pd(14., 15.); |
4116 | assert_eq_m128d(r, e); |
4117 | } |
4118 | |
4119 | #[simd_test(enable = "avx512f" )] |
4120 | unsafe fn test_mm512_cvtepu32_pd() { |
4121 | let a = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
4122 | let r = _mm512_cvtepu32_pd(a); |
4123 | let e = _mm512_set_pd(8., 9., 10., 11., 12., 13., 14., 15.); |
4124 | assert_eq_m512d(r, e); |
4125 | } |
4126 | |
4127 | #[simd_test(enable = "avx512f" )] |
4128 | unsafe fn test_mm512_mask_cvtepu32_pd() { |
4129 | let a = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
4130 | let src = _mm512_set1_pd(-1.); |
4131 | let r = _mm512_mask_cvtepu32_pd(src, 0, a); |
4132 | assert_eq_m512d(r, src); |
4133 | let r = _mm512_mask_cvtepu32_pd(src, 0b00001111, a); |
4134 | let e = _mm512_set_pd(-1., -1., -1., -1., 12., 13., 14., 15.); |
4135 | assert_eq_m512d(r, e); |
4136 | } |
4137 | |
4138 | #[simd_test(enable = "avx512f" )] |
4139 | unsafe fn test_mm512_maskz_cvtepu32_pd() { |
4140 | let a = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
4141 | let r = _mm512_maskz_cvtepu32_pd(0, a); |
4142 | assert_eq_m512d(r, _mm512_setzero_pd()); |
4143 | let r = _mm512_maskz_cvtepu32_pd(0b00001111, a); |
4144 | let e = _mm512_set_pd(0., 0., 0., 0., 12., 13., 14., 15.); |
4145 | assert_eq_m512d(r, e); |
4146 | } |
4147 | |
4148 | #[simd_test(enable = "avx512f,avx512vl" )] |
4149 | unsafe fn test_mm256_cvtepu32_pd() { |
4150 | let a = _mm_set_epi32(12, 13, 14, 15); |
4151 | let r = _mm256_cvtepu32_pd(a); |
4152 | let e = _mm256_set_pd(12., 13., 14., 15.); |
4153 | assert_eq_m256d(r, e); |
4154 | } |
4155 | |
4156 | #[simd_test(enable = "avx512f,avx512vl" )] |
4157 | unsafe fn test_mm256_mask_cvtepu32_pd() { |
4158 | let a = _mm_set_epi32(12, 13, 14, 15); |
4159 | let src = _mm256_set1_pd(-1.); |
4160 | let r = _mm256_mask_cvtepu32_pd(src, 0, a); |
4161 | assert_eq_m256d(r, src); |
4162 | let r = _mm256_mask_cvtepu32_pd(src, 0b00001111, a); |
4163 | let e = _mm256_set_pd(12., 13., 14., 15.); |
4164 | assert_eq_m256d(r, e); |
4165 | } |
4166 | |
4167 | #[simd_test(enable = "avx512f,avx512vl" )] |
4168 | unsafe fn test_mm256_maskz_cvtepu32_pd() { |
4169 | let a = _mm_set_epi32(12, 13, 14, 15); |
4170 | let r = _mm256_maskz_cvtepu32_pd(0, a); |
4171 | assert_eq_m256d(r, _mm256_setzero_pd()); |
4172 | let r = _mm256_maskz_cvtepu32_pd(0b00001111, a); |
4173 | let e = _mm256_set_pd(12., 13., 14., 15.); |
4174 | assert_eq_m256d(r, e); |
4175 | } |
4176 | |
4177 | #[simd_test(enable = "avx512f,avx512vl" )] |
4178 | unsafe fn test_mm_cvtepu32_pd() { |
4179 | let a = _mm_set_epi32(12, 13, 14, 15); |
4180 | let r = _mm_cvtepu32_pd(a); |
4181 | let e = _mm_set_pd(14., 15.); |
4182 | assert_eq_m128d(r, e); |
4183 | } |
4184 | |
4185 | #[simd_test(enable = "avx512f,avx512vl" )] |
4186 | unsafe fn test_mm_mask_cvtepu32_pd() { |
4187 | let a = _mm_set_epi32(12, 13, 14, 15); |
4188 | let src = _mm_set1_pd(-1.); |
4189 | let r = _mm_mask_cvtepu32_pd(src, 0, a); |
4190 | assert_eq_m128d(r, src); |
4191 | let r = _mm_mask_cvtepu32_pd(src, 0b00000011, a); |
4192 | let e = _mm_set_pd(14., 15.); |
4193 | assert_eq_m128d(r, e); |
4194 | } |
4195 | |
4196 | #[simd_test(enable = "avx512f,avx512vl" )] |
4197 | unsafe fn test_mm_maskz_cvtepu32_pd() { |
4198 | let a = _mm_set_epi32(12, 13, 14, 15); |
4199 | let r = _mm_maskz_cvtepu32_pd(0, a); |
4200 | assert_eq_m128d(r, _mm_setzero_pd()); |
4201 | let r = _mm_maskz_cvtepu32_pd(0b00000011, a); |
4202 | let e = _mm_set_pd(14., 15.); |
4203 | assert_eq_m128d(r, e); |
4204 | } |
4205 | |
4206 | #[simd_test(enable = "avx512f" )] |
4207 | unsafe fn test_mm512_cvtepi32lo_pd() { |
4208 | let a = _mm512_set_epi32(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
4209 | let r = _mm512_cvtepi32lo_pd(a); |
4210 | let e = _mm512_set_pd(8., 9., 10., 11., 12., 13., 14., 15.); |
4211 | assert_eq_m512d(r, e); |
4212 | } |
4213 | |
4214 | #[simd_test(enable = "avx512f" )] |
4215 | unsafe fn test_mm512_mask_cvtepi32lo_pd() { |
4216 | let a = _mm512_set_epi32(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
4217 | let src = _mm512_set1_pd(-1.); |
4218 | let r = _mm512_mask_cvtepi32lo_pd(src, 0, a); |
4219 | assert_eq_m512d(r, src); |
4220 | let r = _mm512_mask_cvtepi32lo_pd(src, 0b00001111, a); |
4221 | let e = _mm512_set_pd(-1., -1., -1., -1., 12., 13., 14., 15.); |
4222 | assert_eq_m512d(r, e); |
4223 | } |
4224 | |
4225 | #[simd_test(enable = "avx512f" )] |
4226 | unsafe fn test_mm512_cvtepu32lo_pd() { |
4227 | let a = _mm512_set_epi32(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
4228 | let r = _mm512_cvtepu32lo_pd(a); |
4229 | let e = _mm512_set_pd(8., 9., 10., 11., 12., 13., 14., 15.); |
4230 | assert_eq_m512d(r, e); |
4231 | } |
4232 | |
4233 | #[simd_test(enable = "avx512f" )] |
4234 | unsafe fn test_mm512_mask_cvtepu32lo_pd() { |
4235 | let a = _mm512_set_epi32(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); |
4236 | let src = _mm512_set1_pd(-1.); |
4237 | let r = _mm512_mask_cvtepu32lo_pd(src, 0, a); |
4238 | assert_eq_m512d(r, src); |
4239 | let r = _mm512_mask_cvtepu32lo_pd(src, 0b00001111, a); |
4240 | let e = _mm512_set_pd(-1., -1., -1., -1., 12., 13., 14., 15.); |
4241 | assert_eq_m512d(r, e); |
4242 | } |
4243 | |
4244 | #[simd_test(enable = "avx512f" )] |
4245 | unsafe fn test_mm512_cvtepi64_epi32() { |
4246 | let a = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
4247 | let r = _mm512_cvtepi64_epi32(a); |
4248 | let e = _mm256_set_epi32(8, 9, 10, 11, 12, 13, 14, 15); |
4249 | assert_eq_m256i(r, e); |
4250 | } |
4251 | |
4252 | #[simd_test(enable = "avx512f" )] |
4253 | unsafe fn test_mm512_mask_cvtepi64_epi32() { |
4254 | let a = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
4255 | let src = _mm256_set1_epi32(-1); |
4256 | let r = _mm512_mask_cvtepi64_epi32(src, 0, a); |
4257 | assert_eq_m256i(r, src); |
4258 | let r = _mm512_mask_cvtepi64_epi32(src, 0b00001111, a); |
4259 | let e = _mm256_set_epi32(-1, -1, -1, -1, 12, 13, 14, 15); |
4260 | assert_eq_m256i(r, e); |
4261 | } |
4262 | |
4263 | #[simd_test(enable = "avx512f" )] |
4264 | unsafe fn test_mm512_maskz_cvtepi64_epi32() { |
4265 | let a = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
4266 | let r = _mm512_maskz_cvtepi64_epi32(0, a); |
4267 | assert_eq_m256i(r, _mm256_setzero_si256()); |
4268 | let r = _mm512_maskz_cvtepi64_epi32(0b00001111, a); |
4269 | let e = _mm256_set_epi32(0, 0, 0, 0, 12, 13, 14, 15); |
4270 | assert_eq_m256i(r, e); |
4271 | } |
4272 | |
4273 | #[simd_test(enable = "avx512f,avx512vl" )] |
4274 | unsafe fn test_mm256_cvtepi64_epi32() { |
4275 | let a = _mm256_set_epi64x(1, 2, 3, 4); |
4276 | let r = _mm256_cvtepi64_epi32(a); |
4277 | let e = _mm_set_epi32(1, 2, 3, 4); |
4278 | assert_eq_m128i(r, e); |
4279 | } |
4280 | |
4281 | #[simd_test(enable = "avx512f,avx512vl" )] |
4282 | unsafe fn test_mm256_mask_cvtepi64_epi32() { |
4283 | let a = _mm256_set_epi64x(1, 2, 3, 4); |
4284 | let src = _mm_set1_epi32(0); |
4285 | let r = _mm256_mask_cvtepi64_epi32(src, 0, a); |
4286 | assert_eq_m128i(r, src); |
4287 | let r = _mm256_mask_cvtepi64_epi32(src, 0b00001111, a); |
4288 | let e = _mm_set_epi32(1, 2, 3, 4); |
4289 | assert_eq_m128i(r, e); |
4290 | } |
4291 | |
4292 | #[simd_test(enable = "avx512f,avx512vl" )] |
4293 | unsafe fn test_mm256_maskz_cvtepi64_epi32() { |
4294 | let a = _mm256_set_epi64x(1, 2, 3, 4); |
4295 | let r = _mm256_maskz_cvtepi64_epi32(0, a); |
4296 | assert_eq_m128i(r, _mm_setzero_si128()); |
4297 | let r = _mm256_maskz_cvtepi64_epi32(0b00001111, a); |
4298 | let e = _mm_set_epi32(1, 2, 3, 4); |
4299 | assert_eq_m128i(r, e); |
4300 | } |
4301 | |
4302 | #[simd_test(enable = "avx512f,avx512vl" )] |
4303 | unsafe fn test_mm_cvtepi64_epi32() { |
4304 | let a = _mm_set_epi64x(3, 4); |
4305 | let r = _mm_cvtepi64_epi32(a); |
4306 | let e = _mm_set_epi32(0, 0, 3, 4); |
4307 | assert_eq_m128i(r, e); |
4308 | } |
4309 | |
4310 | #[simd_test(enable = "avx512f,avx512vl" )] |
4311 | unsafe fn test_mm_mask_cvtepi64_epi32() { |
4312 | let a = _mm_set_epi64x(3, 4); |
4313 | let src = _mm_set1_epi32(0); |
4314 | let r = _mm_mask_cvtepi64_epi32(src, 0, a); |
4315 | assert_eq_m128i(r, src); |
4316 | let r = _mm_mask_cvtepi64_epi32(src, 0b00000011, a); |
4317 | let e = _mm_set_epi32(0, 0, 3, 4); |
4318 | assert_eq_m128i(r, e); |
4319 | } |
4320 | |
4321 | #[simd_test(enable = "avx512f,avx512vl" )] |
4322 | unsafe fn test_mm_maskz_cvtepi64_epi32() { |
4323 | let a = _mm_set_epi64x(3, 4); |
4324 | let r = _mm_maskz_cvtepi64_epi32(0, a); |
4325 | assert_eq_m128i(r, _mm_setzero_si128()); |
4326 | let r = _mm_maskz_cvtepi64_epi32(0b00000011, a); |
4327 | let e = _mm_set_epi32(0, 0, 3, 4); |
4328 | assert_eq_m128i(r, e); |
4329 | } |
4330 | |
4331 | #[simd_test(enable = "avx512f" )] |
4332 | unsafe fn test_mm512_cvtepi64_epi16() { |
4333 | let a = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
4334 | let r = _mm512_cvtepi64_epi16(a); |
4335 | let e = _mm_set_epi16(8, 9, 10, 11, 12, 13, 14, 15); |
4336 | assert_eq_m128i(r, e); |
4337 | } |
4338 | |
4339 | #[simd_test(enable = "avx512f" )] |
4340 | unsafe fn test_mm512_mask_cvtepi64_epi16() { |
4341 | let a = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
4342 | let src = _mm_set1_epi16(-1); |
4343 | let r = _mm512_mask_cvtepi64_epi16(src, 0, a); |
4344 | assert_eq_m128i(r, src); |
4345 | let r = _mm512_mask_cvtepi64_epi16(src, 0b00001111, a); |
4346 | let e = _mm_set_epi16(-1, -1, -1, -1, 12, 13, 14, 15); |
4347 | assert_eq_m128i(r, e); |
4348 | } |
4349 | |
4350 | #[simd_test(enable = "avx512f" )] |
4351 | unsafe fn test_mm512_maskz_cvtepi64_epi16() { |
4352 | let a = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
4353 | let r = _mm512_maskz_cvtepi64_epi16(0, a); |
4354 | assert_eq_m128i(r, _mm_setzero_si128()); |
4355 | let r = _mm512_maskz_cvtepi64_epi16(0b00001111, a); |
4356 | let e = _mm_set_epi16(0, 0, 0, 0, 12, 13, 14, 15); |
4357 | assert_eq_m128i(r, e); |
4358 | } |
4359 | |
4360 | #[simd_test(enable = "avx512f,avx512vl" )] |
4361 | unsafe fn test_mm256_cvtepi64_epi16() { |
4362 | let a = _mm256_set_epi64x(12, 13, 14, 15); |
4363 | let r = _mm256_cvtepi64_epi16(a); |
4364 | let e = _mm_set_epi16(0, 0, 0, 0, 12, 13, 14, 15); |
4365 | assert_eq_m128i(r, e); |
4366 | } |
4367 | |
4368 | #[simd_test(enable = "avx512f,avx512vl" )] |
4369 | unsafe fn test_mm256_mask_cvtepi64_epi16() { |
4370 | let a = _mm256_set_epi64x(12, 13, 14, 15); |
4371 | let src = _mm_set1_epi16(0); |
4372 | let r = _mm256_mask_cvtepi64_epi16(src, 0, a); |
4373 | assert_eq_m128i(r, src); |
4374 | let r = _mm256_mask_cvtepi64_epi16(src, 0b11111111, a); |
4375 | let e = _mm_set_epi16(0, 0, 0, 0, 12, 13, 14, 15); |
4376 | assert_eq_m128i(r, e); |
4377 | } |
4378 | |
4379 | #[simd_test(enable = "avx512f,avx512vl" )] |
4380 | unsafe fn test_mm256_maskz_cvtepi64_epi16() { |
4381 | let a = _mm256_set_epi64x(12, 13, 14, 15); |
4382 | let r = _mm256_maskz_cvtepi64_epi16(0, a); |
4383 | assert_eq_m128i(r, _mm_setzero_si128()); |
4384 | let r = _mm256_maskz_cvtepi64_epi16(0b11111111, a); |
4385 | let e = _mm_set_epi16(0, 0, 0, 0, 12, 13, 14, 15); |
4386 | assert_eq_m128i(r, e); |
4387 | } |
4388 | |
4389 | #[simd_test(enable = "avx512f,avx512vl" )] |
4390 | unsafe fn test_mm_cvtepi64_epi16() { |
4391 | let a = _mm_set_epi64x(14, 15); |
4392 | let r = _mm_cvtepi64_epi16(a); |
4393 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 14, 15); |
4394 | assert_eq_m128i(r, e); |
4395 | } |
4396 | |
4397 | #[simd_test(enable = "avx512f,avx512vl" )] |
4398 | unsafe fn test_mm_mask_cvtepi64_epi16() { |
4399 | let a = _mm_set_epi64x(14, 15); |
4400 | let src = _mm_set1_epi16(0); |
4401 | let r = _mm_mask_cvtepi64_epi16(src, 0, a); |
4402 | assert_eq_m128i(r, src); |
4403 | let r = _mm_mask_cvtepi64_epi16(src, 0b11111111, a); |
4404 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 14, 15); |
4405 | assert_eq_m128i(r, e); |
4406 | } |
4407 | |
4408 | #[simd_test(enable = "avx512f,avx512vl" )] |
4409 | unsafe fn test_mm_maskz_cvtepi64_epi16() { |
4410 | let a = _mm_set_epi64x(14, 15); |
4411 | let r = _mm_maskz_cvtepi64_epi16(0, a); |
4412 | assert_eq_m128i(r, _mm_setzero_si128()); |
4413 | let r = _mm_maskz_cvtepi64_epi16(0b11111111, a); |
4414 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 14, 15); |
4415 | assert_eq_m128i(r, e); |
4416 | } |
4417 | |
4418 | #[simd_test(enable = "avx512f" )] |
4419 | unsafe fn test_mm512_cvtepi64_epi8() { |
4420 | let a = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
4421 | let r = _mm512_cvtepi64_epi8(a); |
4422 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 8, 9, 10, 11, 12, 13, 14, 15); |
4423 | assert_eq_m128i(r, e); |
4424 | } |
4425 | |
4426 | #[simd_test(enable = "avx512f" )] |
4427 | unsafe fn test_mm512_mask_cvtepi64_epi8() { |
4428 | let a = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
4429 | let src = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
4430 | let r = _mm512_mask_cvtepi64_epi8(src, 0, a); |
4431 | assert_eq_m128i(r, src); |
4432 | let r = _mm512_mask_cvtepi64_epi8(src, 0b00001111, a); |
4433 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, 12, 13, 14, 15); |
4434 | assert_eq_m128i(r, e); |
4435 | } |
4436 | |
4437 | #[simd_test(enable = "avx512f" )] |
4438 | unsafe fn test_mm512_maskz_cvtepi64_epi8() { |
4439 | let a = _mm512_set_epi64(8, 9, 10, 11, 12, 13, 14, 15); |
4440 | let r = _mm512_maskz_cvtepi64_epi8(0, a); |
4441 | assert_eq_m128i(r, _mm_setzero_si128()); |
4442 | let r = _mm512_maskz_cvtepi64_epi8(0b00001111, a); |
4443 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 12, 13, 14, 15); |
4444 | assert_eq_m128i(r, e); |
4445 | } |
4446 | |
4447 | #[simd_test(enable = "avx512f,avx512vl" )] |
4448 | unsafe fn test_mm256_cvtepi64_epi8() { |
4449 | let a = _mm256_set_epi64x(12, 13, 14, 15); |
4450 | let r = _mm256_cvtepi64_epi8(a); |
4451 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 12, 13, 14, 15); |
4452 | assert_eq_m128i(r, e); |
4453 | } |
4454 | |
4455 | #[simd_test(enable = "avx512f,avx512vl" )] |
4456 | unsafe fn test_mm256_mask_cvtepi64_epi8() { |
4457 | let a = _mm256_set_epi64x(12, 13, 14, 15); |
4458 | let src = _mm_set1_epi8(0); |
4459 | let r = _mm256_mask_cvtepi64_epi8(src, 0, a); |
4460 | assert_eq_m128i(r, src); |
4461 | let r = _mm256_mask_cvtepi64_epi8(src, 0b00001111, a); |
4462 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 12, 13, 14, 15); |
4463 | assert_eq_m128i(r, e); |
4464 | } |
4465 | |
4466 | #[simd_test(enable = "avx512f,avx512vl" )] |
4467 | unsafe fn test_mm256_maskz_cvtepi64_epi8() { |
4468 | let a = _mm256_set_epi64x(12, 13, 14, 15); |
4469 | let r = _mm256_maskz_cvtepi64_epi8(0, a); |
4470 | assert_eq_m128i(r, _mm_setzero_si128()); |
4471 | let r = _mm256_maskz_cvtepi64_epi8(0b00001111, a); |
4472 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 12, 13, 14, 15); |
4473 | assert_eq_m128i(r, e); |
4474 | } |
4475 | |
4476 | #[simd_test(enable = "avx512f,avx512vl" )] |
4477 | unsafe fn test_mm_cvtepi64_epi8() { |
4478 | let a = _mm_set_epi64x(14, 15); |
4479 | let r = _mm_cvtepi64_epi8(a); |
4480 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 14, 15); |
4481 | assert_eq_m128i(r, e); |
4482 | } |
4483 | |
4484 | #[simd_test(enable = "avx512f,avx512vl" )] |
4485 | unsafe fn test_mm_mask_cvtepi64_epi8() { |
4486 | let a = _mm_set_epi64x(14, 15); |
4487 | let src = _mm_set1_epi8(0); |
4488 | let r = _mm_mask_cvtepi64_epi8(src, 0, a); |
4489 | assert_eq_m128i(r, src); |
4490 | let r = _mm_mask_cvtepi64_epi8(src, 0b00000011, a); |
4491 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 14, 15); |
4492 | assert_eq_m128i(r, e); |
4493 | } |
4494 | |
4495 | #[simd_test(enable = "avx512f,avx512vl" )] |
4496 | unsafe fn test_mm_maskz_cvtepi64_epi8() { |
4497 | let a = _mm_set_epi64x(14, 15); |
4498 | let r = _mm_maskz_cvtepi64_epi8(0, a); |
4499 | assert_eq_m128i(r, _mm_setzero_si128()); |
4500 | let r = _mm_maskz_cvtepi64_epi8(0b00000011, a); |
4501 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 14, 15); |
4502 | assert_eq_m128i(r, e); |
4503 | } |
4504 | |
4505 | #[simd_test(enable = "avx512f" )] |
4506 | unsafe fn test_mm512_cvtsepi64_epi32() { |
4507 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MAX); |
4508 | let r = _mm512_cvtsepi64_epi32(a); |
4509 | let e = _mm256_set_epi32(0, 1, 2, 3, 4, 5, i32::MIN, i32::MAX); |
4510 | assert_eq_m256i(r, e); |
4511 | } |
4512 | |
4513 | #[simd_test(enable = "avx512f" )] |
4514 | unsafe fn test_mm512_mask_cvtsepi64_epi32() { |
4515 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MAX); |
4516 | let src = _mm256_set1_epi32(-1); |
4517 | let r = _mm512_mask_cvtsepi64_epi32(src, 0, a); |
4518 | assert_eq_m256i(r, src); |
4519 | let r = _mm512_mask_cvtsepi64_epi32(src, 0b00001111, a); |
4520 | let e = _mm256_set_epi32(-1, -1, -1, -1, 4, 5, i32::MIN, i32::MAX); |
4521 | assert_eq_m256i(r, e); |
4522 | } |
4523 | |
4524 | #[simd_test(enable = "avx512f" )] |
4525 | unsafe fn test_mm512_maskz_cvtsepi64_epi32() { |
4526 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MAX); |
4527 | let r = _mm512_maskz_cvtsepi64_epi32(0, a); |
4528 | assert_eq_m256i(r, _mm256_setzero_si256()); |
4529 | let r = _mm512_maskz_cvtsepi64_epi32(0b00001111, a); |
4530 | let e = _mm256_set_epi32(0, 0, 0, 0, 4, 5, i32::MIN, i32::MAX); |
4531 | assert_eq_m256i(r, e); |
4532 | } |
4533 | |
4534 | #[simd_test(enable = "avx512f,avx512vl" )] |
4535 | unsafe fn test_mm256_cvtsepi64_epi32() { |
4536 | let a = _mm256_set_epi64x(4, 5, i64::MIN, i64::MAX); |
4537 | let r = _mm256_cvtsepi64_epi32(a); |
4538 | let e = _mm_set_epi32(4, 5, i32::MIN, i32::MAX); |
4539 | assert_eq_m128i(r, e); |
4540 | } |
4541 | |
4542 | #[simd_test(enable = "avx512f,avx512vl" )] |
4543 | unsafe fn test_mm256_mask_cvtsepi64_epi32() { |
4544 | let a = _mm256_set_epi64x(4, 5, i64::MIN, i64::MAX); |
4545 | let src = _mm_set1_epi32(-1); |
4546 | let r = _mm256_mask_cvtsepi64_epi32(src, 0, a); |
4547 | assert_eq_m128i(r, src); |
4548 | let r = _mm256_mask_cvtsepi64_epi32(src, 0b00001111, a); |
4549 | let e = _mm_set_epi32(4, 5, i32::MIN, i32::MAX); |
4550 | assert_eq_m128i(r, e); |
4551 | } |
4552 | |
4553 | #[simd_test(enable = "avx512f,avx512vl" )] |
4554 | unsafe fn test_mm256_maskz_cvtsepi64_epi32() { |
4555 | let a = _mm256_set_epi64x(4, 5, i64::MIN, i64::MAX); |
4556 | let r = _mm256_maskz_cvtsepi64_epi32(0, a); |
4557 | assert_eq_m128i(r, _mm_setzero_si128()); |
4558 | let r = _mm256_maskz_cvtsepi64_epi32(0b00001111, a); |
4559 | let e = _mm_set_epi32(4, 5, i32::MIN, i32::MAX); |
4560 | assert_eq_m128i(r, e); |
4561 | } |
4562 | |
4563 | #[simd_test(enable = "avx512f,avx512vl" )] |
4564 | unsafe fn test_mm_cvtsepi64_epi32() { |
4565 | let a = _mm_set_epi64x(i64::MIN, i64::MAX); |
4566 | let r = _mm_cvtsepi64_epi32(a); |
4567 | let e = _mm_set_epi32(0, 0, i32::MIN, i32::MAX); |
4568 | assert_eq_m128i(r, e); |
4569 | } |
4570 | |
4571 | #[simd_test(enable = "avx512f,avx512vl" )] |
4572 | unsafe fn test_mm_mask_cvtsepi64_epi32() { |
4573 | let a = _mm_set_epi64x(i64::MIN, i64::MAX); |
4574 | let src = _mm_set1_epi32(0); |
4575 | let r = _mm_mask_cvtsepi64_epi32(src, 0, a); |
4576 | assert_eq_m128i(r, src); |
4577 | let r = _mm_mask_cvtsepi64_epi32(src, 0b00000011, a); |
4578 | let e = _mm_set_epi32(0, 0, i32::MIN, i32::MAX); |
4579 | assert_eq_m128i(r, e); |
4580 | } |
4581 | |
4582 | #[simd_test(enable = "avx512f,avx512vl" )] |
4583 | unsafe fn test_mm_maskz_cvtsepi64_epi32() { |
4584 | let a = _mm_set_epi64x(i64::MIN, i64::MAX); |
4585 | let r = _mm_maskz_cvtsepi64_epi32(0, a); |
4586 | assert_eq_m128i(r, _mm_setzero_si128()); |
4587 | let r = _mm_maskz_cvtsepi64_epi32(0b00000011, a); |
4588 | let e = _mm_set_epi32(0, 0, i32::MIN, i32::MAX); |
4589 | assert_eq_m128i(r, e); |
4590 | } |
4591 | |
4592 | #[simd_test(enable = "avx512f" )] |
4593 | unsafe fn test_mm512_cvtsepi64_epi16() { |
4594 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MAX); |
4595 | let r = _mm512_cvtsepi64_epi16(a); |
4596 | let e = _mm_set_epi16(0, 1, 2, 3, 4, 5, i16::MIN, i16::MAX); |
4597 | assert_eq_m128i(r, e); |
4598 | } |
4599 | |
4600 | #[simd_test(enable = "avx512f" )] |
4601 | unsafe fn test_mm512_mask_cvtsepi64_epi16() { |
4602 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MAX); |
4603 | let src = _mm_set1_epi16(-1); |
4604 | let r = _mm512_mask_cvtsepi64_epi16(src, 0, a); |
4605 | assert_eq_m128i(r, src); |
4606 | let r = _mm512_mask_cvtsepi64_epi16(src, 0b00001111, a); |
4607 | let e = _mm_set_epi16(-1, -1, -1, -1, 4, 5, i16::MIN, i16::MAX); |
4608 | assert_eq_m128i(r, e); |
4609 | } |
4610 | |
4611 | #[simd_test(enable = "avx512f" )] |
4612 | unsafe fn test_mm512_maskz_cvtsepi64_epi16() { |
4613 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MAX); |
4614 | let r = _mm512_maskz_cvtsepi64_epi16(0, a); |
4615 | assert_eq_m128i(r, _mm_setzero_si128()); |
4616 | let r = _mm512_maskz_cvtsepi64_epi16(0b00001111, a); |
4617 | let e = _mm_set_epi16(0, 0, 0, 0, 4, 5, i16::MIN, i16::MAX); |
4618 | assert_eq_m128i(r, e); |
4619 | } |
4620 | |
4621 | #[simd_test(enable = "avx512f,avx512vl" )] |
4622 | unsafe fn test_mm256_cvtsepi64_epi16() { |
4623 | let a = _mm256_set_epi64x(4, 5, i64::MIN, i64::MAX); |
4624 | let r = _mm256_cvtsepi64_epi16(a); |
4625 | let e = _mm_set_epi16(0, 0, 0, 0, 4, 5, i16::MIN, i16::MAX); |
4626 | assert_eq_m128i(r, e); |
4627 | } |
4628 | |
4629 | #[simd_test(enable = "avx512f,avx512vl" )] |
4630 | unsafe fn test_mm256_mask_cvtsepi64_epi16() { |
4631 | let a = _mm256_set_epi64x(4, 5, i64::MIN, i64::MAX); |
4632 | let src = _mm_set1_epi16(0); |
4633 | let r = _mm256_mask_cvtsepi64_epi16(src, 0, a); |
4634 | assert_eq_m128i(r, src); |
4635 | let r = _mm256_mask_cvtsepi64_epi16(src, 0b00001111, a); |
4636 | let e = _mm_set_epi16(0, 0, 0, 0, 4, 5, i16::MIN, i16::MAX); |
4637 | assert_eq_m128i(r, e); |
4638 | } |
4639 | |
4640 | #[simd_test(enable = "avx512f,avx512vl" )] |
4641 | unsafe fn test_mm256_maskz_cvtsepi64_epi16() { |
4642 | let a = _mm256_set_epi64x(4, 5, i64::MIN, i64::MAX); |
4643 | let r = _mm256_maskz_cvtsepi64_epi16(0, a); |
4644 | assert_eq_m128i(r, _mm_setzero_si128()); |
4645 | let r = _mm256_maskz_cvtsepi64_epi16(0b00001111, a); |
4646 | let e = _mm_set_epi16(0, 0, 0, 0, 4, 5, i16::MIN, i16::MAX); |
4647 | assert_eq_m128i(r, e); |
4648 | } |
4649 | |
4650 | #[simd_test(enable = "avx512f,avx512vl" )] |
4651 | unsafe fn test_mm_cvtsepi64_epi16() { |
4652 | let a = _mm_set_epi64x(i64::MIN, i64::MAX); |
4653 | let r = _mm_cvtsepi64_epi16(a); |
4654 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, i16::MIN, i16::MAX); |
4655 | assert_eq_m128i(r, e); |
4656 | } |
4657 | |
4658 | #[simd_test(enable = "avx512f,avx512vl" )] |
4659 | unsafe fn test_mm_mask_cvtsepi64_epi16() { |
4660 | let a = _mm_set_epi64x(i64::MIN, i64::MAX); |
4661 | let src = _mm_set1_epi16(0); |
4662 | let r = _mm_mask_cvtsepi64_epi16(src, 0, a); |
4663 | assert_eq_m128i(r, src); |
4664 | let r = _mm_mask_cvtsepi64_epi16(src, 0b00000011, a); |
4665 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, i16::MIN, i16::MAX); |
4666 | assert_eq_m128i(r, e); |
4667 | } |
4668 | |
4669 | #[simd_test(enable = "avx512f,avx512vl" )] |
4670 | unsafe fn test_mm_maskz_cvtsepi64_epi16() { |
4671 | let a = _mm_set_epi64x(i64::MIN, i64::MAX); |
4672 | let r = _mm_maskz_cvtsepi64_epi16(0, a); |
4673 | assert_eq_m128i(r, _mm_setzero_si128()); |
4674 | let r = _mm_maskz_cvtsepi64_epi16(0b00000011, a); |
4675 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, i16::MIN, i16::MAX); |
4676 | assert_eq_m128i(r, e); |
4677 | } |
4678 | |
4679 | #[simd_test(enable = "avx512f" )] |
4680 | unsafe fn test_mm512_cvtsepi64_epi8() { |
4681 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MAX); |
4682 | let r = _mm512_cvtsepi64_epi8(a); |
4683 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 3, 4, 5, i8::MIN, i8::MAX); |
4684 | assert_eq_m128i(r, e); |
4685 | } |
4686 | |
4687 | #[simd_test(enable = "avx512f" )] |
4688 | unsafe fn test_mm512_mask_cvtsepi64_epi8() { |
4689 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MAX); |
4690 | let src = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
4691 | let r = _mm512_mask_cvtsepi64_epi8(src, 0, a); |
4692 | assert_eq_m128i(r, src); |
4693 | let r = _mm512_mask_cvtsepi64_epi8(src, 0b00001111, a); |
4694 | #[rustfmt::skip] |
4695 | let e = _mm_set_epi8( |
4696 | 0, 0, 0, 0, |
4697 | 0, 0, 0, 0, |
4698 | -1, -1, -1, -1, |
4699 | 4, 5, i8::MIN, i8::MAX, |
4700 | ); |
4701 | assert_eq_m128i(r, e); |
4702 | } |
4703 | |
4704 | #[simd_test(enable = "avx512f" )] |
4705 | unsafe fn test_mm512_maskz_cvtsepi64_epi8() { |
4706 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MAX); |
4707 | let r = _mm512_maskz_cvtsepi64_epi8(0, a); |
4708 | assert_eq_m128i(r, _mm_setzero_si128()); |
4709 | let r = _mm512_maskz_cvtsepi64_epi8(0b00001111, a); |
4710 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 5, i8::MIN, i8::MAX); |
4711 | assert_eq_m128i(r, e); |
4712 | } |
4713 | |
4714 | #[simd_test(enable = "avx512f,avx512vl" )] |
4715 | unsafe fn test_mm256_cvtsepi64_epi8() { |
4716 | let a = _mm256_set_epi64x(4, 5, i64::MIN, i64::MAX); |
4717 | let r = _mm256_cvtsepi64_epi8(a); |
4718 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 5, i8::MIN, i8::MAX); |
4719 | assert_eq_m128i(r, e); |
4720 | } |
4721 | |
4722 | #[simd_test(enable = "avx512f,avx512vl" )] |
4723 | unsafe fn test_mm256_mask_cvtsepi64_epi8() { |
4724 | let a = _mm256_set_epi64x(4, 5, i64::MIN, i64::MAX); |
4725 | let src = _mm_set1_epi8(0); |
4726 | let r = _mm256_mask_cvtsepi64_epi8(src, 0, a); |
4727 | assert_eq_m128i(r, src); |
4728 | let r = _mm256_mask_cvtsepi64_epi8(src, 0b00001111, a); |
4729 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 5, i8::MIN, i8::MAX); |
4730 | assert_eq_m128i(r, e); |
4731 | } |
4732 | |
4733 | #[simd_test(enable = "avx512f,avx512vl" )] |
4734 | unsafe fn test_mm256_maskz_cvtsepi64_epi8() { |
4735 | let a = _mm256_set_epi64x(4, 5, i64::MIN, i64::MAX); |
4736 | let r = _mm256_maskz_cvtsepi64_epi8(0, a); |
4737 | assert_eq_m128i(r, _mm_setzero_si128()); |
4738 | let r = _mm256_maskz_cvtsepi64_epi8(0b00001111, a); |
4739 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 5, i8::MIN, i8::MAX); |
4740 | assert_eq_m128i(r, e); |
4741 | } |
4742 | |
4743 | #[simd_test(enable = "avx512f,avx512vl" )] |
4744 | unsafe fn test_mm_cvtsepi64_epi8() { |
4745 | let a = _mm_set_epi64x(i64::MIN, i64::MAX); |
4746 | let r = _mm_cvtsepi64_epi8(a); |
4747 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MIN, i8::MAX); |
4748 | assert_eq_m128i(r, e); |
4749 | } |
4750 | |
4751 | #[simd_test(enable = "avx512f,avx512vl" )] |
4752 | unsafe fn test_mm_mask_cvtsepi64_epi8() { |
4753 | let a = _mm_set_epi64x(i64::MIN, i64::MAX); |
4754 | let src = _mm_set1_epi8(0); |
4755 | let r = _mm_mask_cvtsepi64_epi8(src, 0, a); |
4756 | assert_eq_m128i(r, src); |
4757 | let r = _mm_mask_cvtsepi64_epi8(src, 0b00000011, a); |
4758 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MIN, i8::MAX); |
4759 | assert_eq_m128i(r, e); |
4760 | } |
4761 | |
4762 | #[simd_test(enable = "avx512f,avx512vl" )] |
4763 | unsafe fn test_mm_maskz_cvtsepi64_epi8() { |
4764 | let a = _mm_set_epi64x(i64::MIN, i64::MAX); |
4765 | let r = _mm_maskz_cvtsepi64_epi8(0, a); |
4766 | assert_eq_m128i(r, _mm_setzero_si128()); |
4767 | let r = _mm_maskz_cvtsepi64_epi8(0b00000011, a); |
4768 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MIN, i8::MAX); |
4769 | assert_eq_m128i(r, e); |
4770 | } |
4771 | |
4772 | #[simd_test(enable = "avx512f" )] |
4773 | unsafe fn test_mm512_cvtusepi64_epi32() { |
4774 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MIN); |
4775 | let r = _mm512_cvtusepi64_epi32(a); |
4776 | let e = _mm256_set_epi32(0, 1, 2, 3, 4, 5, -1, -1); |
4777 | assert_eq_m256i(r, e); |
4778 | } |
4779 | |
4780 | #[simd_test(enable = "avx512f" )] |
4781 | unsafe fn test_mm512_mask_cvtusepi64_epi32() { |
4782 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MIN); |
4783 | let src = _mm256_set1_epi32(-1); |
4784 | let r = _mm512_mask_cvtusepi64_epi32(src, 0, a); |
4785 | assert_eq_m256i(r, src); |
4786 | let r = _mm512_mask_cvtusepi64_epi32(src, 0b00001111, a); |
4787 | let e = _mm256_set_epi32(-1, -1, -1, -1, 4, 5, -1, -1); |
4788 | assert_eq_m256i(r, e); |
4789 | } |
4790 | |
4791 | #[simd_test(enable = "avx512f" )] |
4792 | unsafe fn test_mm512_maskz_cvtusepi64_epi32() { |
4793 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MIN); |
4794 | let r = _mm512_maskz_cvtusepi64_epi32(0, a); |
4795 | assert_eq_m256i(r, _mm256_setzero_si256()); |
4796 | let r = _mm512_maskz_cvtusepi64_epi32(0b00001111, a); |
4797 | let e = _mm256_set_epi32(0, 0, 0, 0, 4, 5, -1, -1); |
4798 | assert_eq_m256i(r, e); |
4799 | } |
4800 | |
4801 | #[simd_test(enable = "avx512f,avx512vl" )] |
4802 | unsafe fn test_mm256_cvtusepi64_epi32() { |
4803 | let a = _mm256_set_epi64x(4, 5, 6, i64::MAX); |
4804 | let r = _mm256_cvtusepi64_epi32(a); |
4805 | let e = _mm_set_epi32(4, 5, 6, u32::MAX as i32); |
4806 | assert_eq_m128i(r, e); |
4807 | } |
4808 | |
4809 | #[simd_test(enable = "avx512f,avx512vl" )] |
4810 | unsafe fn test_mm256_mask_cvtusepi64_epi32() { |
4811 | let a = _mm256_set_epi64x(4, 5, 6, i64::MAX); |
4812 | let src = _mm_set1_epi32(0); |
4813 | let r = _mm256_mask_cvtusepi64_epi32(src, 0, a); |
4814 | assert_eq_m128i(r, src); |
4815 | let r = _mm256_mask_cvtusepi64_epi32(src, 0b00001111, a); |
4816 | let e = _mm_set_epi32(4, 5, 6, u32::MAX as i32); |
4817 | assert_eq_m128i(r, e); |
4818 | } |
4819 | |
4820 | #[simd_test(enable = "avx512f,avx512vl" )] |
4821 | unsafe fn test_mm256_maskz_cvtusepi64_epi32() { |
4822 | let a = _mm256_set_epi64x(4, 5, 6, i64::MAX); |
4823 | let r = _mm256_maskz_cvtusepi64_epi32(0, a); |
4824 | assert_eq_m128i(r, _mm_setzero_si128()); |
4825 | let r = _mm256_maskz_cvtusepi64_epi32(0b00001111, a); |
4826 | let e = _mm_set_epi32(4, 5, 6, u32::MAX as i32); |
4827 | assert_eq_m128i(r, e); |
4828 | } |
4829 | |
4830 | #[simd_test(enable = "avx512f,avx512vl" )] |
4831 | unsafe fn test_mm_cvtusepi64_epi32() { |
4832 | let a = _mm_set_epi64x(6, i64::MAX); |
4833 | let r = _mm_cvtusepi64_epi32(a); |
4834 | let e = _mm_set_epi32(0, 0, 6, u32::MAX as i32); |
4835 | assert_eq_m128i(r, e); |
4836 | } |
4837 | |
4838 | #[simd_test(enable = "avx512f,avx512vl" )] |
4839 | unsafe fn test_mm_mask_cvtusepi64_epi32() { |
4840 | let a = _mm_set_epi64x(6, i64::MAX); |
4841 | let src = _mm_set1_epi32(0); |
4842 | let r = _mm_mask_cvtusepi64_epi32(src, 0, a); |
4843 | assert_eq_m128i(r, src); |
4844 | let r = _mm_mask_cvtusepi64_epi32(src, 0b00000011, a); |
4845 | let e = _mm_set_epi32(0, 0, 6, u32::MAX as i32); |
4846 | assert_eq_m128i(r, e); |
4847 | } |
4848 | |
4849 | #[simd_test(enable = "avx512f,avx512vl" )] |
4850 | unsafe fn test_mm_maskz_cvtusepi64_epi32() { |
4851 | let a = _mm_set_epi64x(6, i64::MAX); |
4852 | let r = _mm_maskz_cvtusepi64_epi32(0, a); |
4853 | assert_eq_m128i(r, _mm_setzero_si128()); |
4854 | let r = _mm_maskz_cvtusepi64_epi32(0b00000011, a); |
4855 | let e = _mm_set_epi32(0, 0, 6, u32::MAX as i32); |
4856 | assert_eq_m128i(r, e); |
4857 | } |
4858 | |
4859 | #[simd_test(enable = "avx512f" )] |
4860 | unsafe fn test_mm512_cvtusepi64_epi16() { |
4861 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MIN); |
4862 | let r = _mm512_cvtusepi64_epi16(a); |
4863 | let e = _mm_set_epi16(0, 1, 2, 3, 4, 5, -1, -1); |
4864 | assert_eq_m128i(r, e); |
4865 | } |
4866 | |
4867 | #[simd_test(enable = "avx512f" )] |
4868 | unsafe fn test_mm512_mask_cvtusepi64_epi16() { |
4869 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MIN); |
4870 | let src = _mm_set1_epi16(-1); |
4871 | let r = _mm512_mask_cvtusepi64_epi16(src, 0, a); |
4872 | assert_eq_m128i(r, src); |
4873 | let r = _mm512_mask_cvtusepi64_epi16(src, 0b00001111, a); |
4874 | let e = _mm_set_epi16(-1, -1, -1, -1, 4, 5, -1, -1); |
4875 | assert_eq_m128i(r, e); |
4876 | } |
4877 | |
4878 | #[simd_test(enable = "avx512f" )] |
4879 | unsafe fn test_mm512_maskz_cvtusepi64_epi16() { |
4880 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MIN); |
4881 | let r = _mm512_maskz_cvtusepi64_epi16(0, a); |
4882 | assert_eq_m128i(r, _mm_setzero_si128()); |
4883 | let r = _mm512_maskz_cvtusepi64_epi16(0b00001111, a); |
4884 | let e = _mm_set_epi16(0, 0, 0, 0, 4, 5, -1, -1); |
4885 | assert_eq_m128i(r, e); |
4886 | } |
4887 | |
4888 | #[simd_test(enable = "avx512f,avx512vl" )] |
4889 | unsafe fn test_mm256_cvtusepi64_epi16() { |
4890 | let a = _mm256_set_epi64x(4, 5, 6, i64::MAX); |
4891 | let r = _mm256_cvtusepi64_epi16(a); |
4892 | let e = _mm_set_epi16(0, 0, 0, 0, 4, 5, 6, u16::MAX as i16); |
4893 | assert_eq_m128i(r, e); |
4894 | } |
4895 | |
4896 | #[simd_test(enable = "avx512f,avx512vl" )] |
4897 | unsafe fn test_mm256_mask_cvtusepi64_epi16() { |
4898 | let a = _mm256_set_epi64x(4, 5, 6, i64::MAX); |
4899 | let src = _mm_set1_epi16(0); |
4900 | let r = _mm256_mask_cvtusepi64_epi16(src, 0, a); |
4901 | assert_eq_m128i(r, src); |
4902 | let r = _mm256_mask_cvtusepi64_epi16(src, 0b00001111, a); |
4903 | let e = _mm_set_epi16(0, 0, 0, 0, 4, 5, 6, u16::MAX as i16); |
4904 | assert_eq_m128i(r, e); |
4905 | } |
4906 | |
4907 | #[simd_test(enable = "avx512f,avx512vl" )] |
4908 | unsafe fn test_mm256_maskz_cvtusepi64_epi16() { |
4909 | let a = _mm256_set_epi64x(4, 5, 6, i64::MAX); |
4910 | let r = _mm256_maskz_cvtusepi64_epi16(0, a); |
4911 | assert_eq_m128i(r, _mm_setzero_si128()); |
4912 | let r = _mm256_maskz_cvtusepi64_epi16(0b00001111, a); |
4913 | let e = _mm_set_epi16(0, 0, 0, 0, 4, 5, 6, u16::MAX as i16); |
4914 | assert_eq_m128i(r, e); |
4915 | } |
4916 | |
4917 | #[simd_test(enable = "avx512f,avx512vl" )] |
4918 | unsafe fn test_mm_cvtusepi64_epi16() { |
4919 | let a = _mm_set_epi64x(6, i64::MAX); |
4920 | let r = _mm_cvtusepi64_epi16(a); |
4921 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 6, u16::MAX as i16); |
4922 | assert_eq_m128i(r, e); |
4923 | } |
4924 | |
4925 | #[simd_test(enable = "avx512f,avx512vl" )] |
4926 | unsafe fn test_mm_mask_cvtusepi64_epi16() { |
4927 | let a = _mm_set_epi64x(6, i64::MAX); |
4928 | let src = _mm_set1_epi16(0); |
4929 | let r = _mm_mask_cvtusepi64_epi16(src, 0, a); |
4930 | assert_eq_m128i(r, src); |
4931 | let r = _mm_mask_cvtusepi64_epi16(src, 0b00000011, a); |
4932 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 6, u16::MAX as i16); |
4933 | assert_eq_m128i(r, e); |
4934 | } |
4935 | |
4936 | #[simd_test(enable = "avx512f,avx512vl" )] |
4937 | unsafe fn test_mm_maskz_cvtusepi64_epi16() { |
4938 | let a = _mm_set_epi64x(6, i64::MAX); |
4939 | let r = _mm_maskz_cvtusepi64_epi16(0, a); |
4940 | assert_eq_m128i(r, _mm_setzero_si128()); |
4941 | let r = _mm_maskz_cvtusepi64_epi16(0b00000011, a); |
4942 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 6, u16::MAX as i16); |
4943 | assert_eq_m128i(r, e); |
4944 | } |
4945 | |
4946 | #[simd_test(enable = "avx512f" )] |
4947 | unsafe fn test_mm512_cvtusepi64_epi8() { |
4948 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MIN); |
4949 | let r = _mm512_cvtusepi64_epi8(a); |
4950 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 3, 4, 5, -1, -1); |
4951 | assert_eq_m128i(r, e); |
4952 | } |
4953 | |
4954 | #[simd_test(enable = "avx512f" )] |
4955 | unsafe fn test_mm512_mask_cvtusepi64_epi8() { |
4956 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MIN); |
4957 | let src = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, -1, -1); |
4958 | let r = _mm512_mask_cvtusepi64_epi8(src, 0, a); |
4959 | assert_eq_m128i(r, src); |
4960 | let r = _mm512_mask_cvtusepi64_epi8(src, 0b00001111, a); |
4961 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, 4, 5, -1, -1); |
4962 | assert_eq_m128i(r, e); |
4963 | } |
4964 | |
4965 | #[simd_test(enable = "avx512f" )] |
4966 | unsafe fn test_mm512_maskz_cvtusepi64_epi8() { |
4967 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, i64::MIN, i64::MIN); |
4968 | let r = _mm512_maskz_cvtusepi64_epi8(0, a); |
4969 | assert_eq_m128i(r, _mm_setzero_si128()); |
4970 | let r = _mm512_maskz_cvtusepi64_epi8(0b00001111, a); |
4971 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 5, -1, -1); |
4972 | assert_eq_m128i(r, e); |
4973 | } |
4974 | |
4975 | #[simd_test(enable = "avx512f,avx512vl" )] |
4976 | unsafe fn test_mm256_cvtusepi64_epi8() { |
4977 | let a = _mm256_set_epi64x(4, 5, 6, i64::MAX); |
4978 | let r = _mm256_cvtusepi64_epi8(a); |
4979 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 5, 6, u8::MAX as i8); |
4980 | assert_eq_m128i(r, e); |
4981 | } |
4982 | |
4983 | #[simd_test(enable = "avx512f,avx512vl" )] |
4984 | unsafe fn test_mm256_mask_cvtusepi64_epi8() { |
4985 | let a = _mm256_set_epi64x(4, 5, 6, i64::MAX); |
4986 | let src = _mm_set1_epi8(0); |
4987 | let r = _mm256_mask_cvtusepi64_epi8(src, 0, a); |
4988 | assert_eq_m128i(r, src); |
4989 | let r = _mm256_mask_cvtusepi64_epi8(src, 0b00001111, a); |
4990 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 5, 6, u8::MAX as i8); |
4991 | assert_eq_m128i(r, e); |
4992 | } |
4993 | |
4994 | #[simd_test(enable = "avx512f,avx512vl" )] |
4995 | unsafe fn test_mm256_maskz_cvtusepi64_epi8() { |
4996 | let a = _mm256_set_epi64x(4, 5, 6, i64::MAX); |
4997 | let r = _mm256_maskz_cvtusepi64_epi8(0, a); |
4998 | assert_eq_m128i(r, _mm_setzero_si128()); |
4999 | let r = _mm256_maskz_cvtusepi64_epi8(0b00001111, a); |
5000 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 5, 6, u8::MAX as i8); |
5001 | assert_eq_m128i(r, e); |
5002 | } |
5003 | |
5004 | #[simd_test(enable = "avx512f,avx512vl" )] |
5005 | unsafe fn test_mm_cvtusepi64_epi8() { |
5006 | let a = _mm_set_epi64x(6, i64::MAX); |
5007 | let r = _mm_cvtusepi64_epi8(a); |
5008 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 6, u8::MAX as i8); |
5009 | assert_eq_m128i(r, e); |
5010 | } |
5011 | |
5012 | #[simd_test(enable = "avx512f,avx512vl" )] |
5013 | unsafe fn test_mm_mask_cvtusepi64_epi8() { |
5014 | let a = _mm_set_epi64x(6, i64::MAX); |
5015 | let src = _mm_set1_epi8(0); |
5016 | let r = _mm_mask_cvtusepi64_epi8(src, 0, a); |
5017 | assert_eq_m128i(r, src); |
5018 | let r = _mm_mask_cvtusepi64_epi8(src, 0b00000011, a); |
5019 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 6, u8::MAX as i8); |
5020 | assert_eq_m128i(r, e); |
5021 | } |
5022 | |
5023 | #[simd_test(enable = "avx512f,avx512vl" )] |
5024 | unsafe fn test_mm_maskz_cvtusepi64_epi8() { |
5025 | let a = _mm_set_epi64x(6, i64::MAX); |
5026 | let r = _mm_maskz_cvtusepi64_epi8(0, a); |
5027 | assert_eq_m128i(r, _mm_setzero_si128()); |
5028 | let r = _mm_maskz_cvtusepi64_epi8(0b00000011, a); |
5029 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 6, u8::MAX as i8); |
5030 | assert_eq_m128i(r, e); |
5031 | } |
5032 | |
5033 | #[simd_test(enable = "avx512f" )] |
5034 | unsafe fn test_mm512_cvtt_roundpd_epi32() { |
5035 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
5036 | let r = _mm512_cvtt_roundpd_epi32::<_MM_FROUND_NO_EXC>(a); |
5037 | let e = _mm256_setr_epi32(0, -1, 2, -3, 4, -5, 6, -7); |
5038 | assert_eq_m256i(r, e); |
5039 | } |
5040 | |
5041 | #[simd_test(enable = "avx512f" )] |
5042 | unsafe fn test_mm512_mask_cvtt_roundpd_epi32() { |
5043 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
5044 | let src = _mm256_set1_epi32(0); |
5045 | let r = _mm512_mask_cvtt_roundpd_epi32::<_MM_FROUND_NO_EXC>(src, 0, a); |
5046 | assert_eq_m256i(r, src); |
5047 | let r = _mm512_mask_cvtt_roundpd_epi32::<_MM_FROUND_NO_EXC>(src, 0b00001111, a); |
5048 | let e = _mm256_setr_epi32(0, -1, 2, -3, 0, 0, 0, 0); |
5049 | assert_eq_m256i(r, e); |
5050 | } |
5051 | |
5052 | #[simd_test(enable = "avx512f" )] |
5053 | unsafe fn test_mm512_maskz_cvtt_roundpd_epi32() { |
5054 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
5055 | let r = _mm512_maskz_cvtt_roundpd_epi32::<_MM_FROUND_NO_EXC>(0, a); |
5056 | assert_eq_m256i(r, _mm256_setzero_si256()); |
5057 | let r = _mm512_maskz_cvtt_roundpd_epi32::<_MM_FROUND_NO_EXC>(0b00001111, a); |
5058 | let e = _mm256_setr_epi32(0, -1, 2, -3, 0, 0, 0, 0); |
5059 | assert_eq_m256i(r, e); |
5060 | } |
5061 | |
5062 | #[simd_test(enable = "avx512f" )] |
5063 | unsafe fn test_mm512_cvtt_roundpd_epu32() { |
5064 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
5065 | let r = _mm512_cvtt_roundpd_epu32::<_MM_FROUND_NO_EXC>(a); |
5066 | let e = _mm256_setr_epi32(0, -1, 2, -1, 4, -1, 6, -1); |
5067 | assert_eq_m256i(r, e); |
5068 | } |
5069 | |
5070 | #[simd_test(enable = "avx512f" )] |
5071 | unsafe fn test_mm512_mask_cvtt_roundpd_epu32() { |
5072 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
5073 | let src = _mm256_set1_epi32(0); |
5074 | let r = _mm512_mask_cvtt_roundpd_epu32::<_MM_FROUND_NO_EXC>(src, 0, a); |
5075 | assert_eq_m256i(r, src); |
5076 | let r = _mm512_mask_cvtt_roundpd_epu32::<_MM_FROUND_NO_EXC>(src, 0b00001111, a); |
5077 | let e = _mm256_setr_epi32(0, -1, 2, -1, 0, 0, 0, 0); |
5078 | assert_eq_m256i(r, e); |
5079 | } |
5080 | |
5081 | #[simd_test(enable = "avx512f" )] |
5082 | unsafe fn test_mm512_maskz_cvtt_roundpd_epu32() { |
5083 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
5084 | let r = _mm512_maskz_cvtt_roundpd_epu32::<_MM_FROUND_NO_EXC>(0, a); |
5085 | assert_eq_m256i(r, _mm256_setzero_si256()); |
5086 | let r = _mm512_maskz_cvtt_roundpd_epu32::<_MM_FROUND_NO_EXC>(0b00001111, a); |
5087 | let e = _mm256_setr_epi32(0, -1, 2, -1, 0, 0, 0, 0); |
5088 | assert_eq_m256i(r, e); |
5089 | } |
5090 | |
5091 | #[simd_test(enable = "avx512f" )] |
5092 | unsafe fn test_mm512_cvttpd_epi32() { |
5093 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
5094 | let r = _mm512_cvttpd_epi32(a); |
5095 | let e = _mm256_setr_epi32(0, -1, 2, -3, 4, -5, 6, -7); |
5096 | assert_eq_m256i(r, e); |
5097 | } |
5098 | |
5099 | #[simd_test(enable = "avx512f" )] |
5100 | unsafe fn test_mm512_mask_cvttpd_epi32() { |
5101 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
5102 | let src = _mm256_set1_epi32(0); |
5103 | let r = _mm512_mask_cvttpd_epi32(src, 0, a); |
5104 | assert_eq_m256i(r, src); |
5105 | let r = _mm512_mask_cvttpd_epi32(src, 0b00001111, a); |
5106 | let e = _mm256_setr_epi32(0, -1, 2, -3, 0, 0, 0, 0); |
5107 | assert_eq_m256i(r, e); |
5108 | } |
5109 | |
5110 | #[simd_test(enable = "avx512f" )] |
5111 | unsafe fn test_mm512_maskz_cvttpd_epi32() { |
5112 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
5113 | let r = _mm512_maskz_cvttpd_epi32(0, a); |
5114 | assert_eq_m256i(r, _mm256_setzero_si256()); |
5115 | let r = _mm512_maskz_cvttpd_epi32(0b00001111, a); |
5116 | let e = _mm256_setr_epi32(0, -1, 2, -3, 0, 0, 0, 0); |
5117 | assert_eq_m256i(r, e); |
5118 | } |
5119 | |
5120 | #[simd_test(enable = "avx512f,avx512vl" )] |
5121 | unsafe fn test_mm256_mask_cvttpd_epi32() { |
5122 | let a = _mm256_setr_pd(4., -5.5, 6., -7.5); |
5123 | let src = _mm_set1_epi32(0); |
5124 | let r = _mm256_mask_cvttpd_epi32(src, 0, a); |
5125 | assert_eq_m128i(r, src); |
5126 | let r = _mm256_mask_cvttpd_epi32(src, 0b00001111, a); |
5127 | let e = _mm_setr_epi32(4, -5, 6, -7); |
5128 | assert_eq_m128i(r, e); |
5129 | } |
5130 | |
5131 | #[simd_test(enable = "avx512f,avx512vl" )] |
5132 | unsafe fn test_mm256_maskz_cvttpd_epi32() { |
5133 | let a = _mm256_setr_pd(4., -5.5, 6., -7.5); |
5134 | let r = _mm256_maskz_cvttpd_epi32(0, a); |
5135 | assert_eq_m128i(r, _mm_setzero_si128()); |
5136 | let r = _mm256_maskz_cvttpd_epi32(0b00001111, a); |
5137 | let e = _mm_setr_epi32(4, -5, 6, -7); |
5138 | assert_eq_m128i(r, e); |
5139 | } |
5140 | |
5141 | #[simd_test(enable = "avx512f,avx512vl" )] |
5142 | unsafe fn test_mm_mask_cvttpd_epi32() { |
5143 | let a = _mm_set_pd(6., -7.5); |
5144 | let src = _mm_set1_epi32(0); |
5145 | let r = _mm_mask_cvttpd_epi32(src, 0, a); |
5146 | assert_eq_m128i(r, src); |
5147 | let r = _mm_mask_cvttpd_epi32(src, 0b00000011, a); |
5148 | let e = _mm_set_epi32(0, 0, 6, -7); |
5149 | assert_eq_m128i(r, e); |
5150 | } |
5151 | |
5152 | #[simd_test(enable = "avx512f,avx512vl" )] |
5153 | unsafe fn test_mm_maskz_cvttpd_epi32() { |
5154 | let a = _mm_set_pd(6., -7.5); |
5155 | let r = _mm_maskz_cvttpd_epi32(0, a); |
5156 | assert_eq_m128i(r, _mm_setzero_si128()); |
5157 | let r = _mm_maskz_cvttpd_epi32(0b00000011, a); |
5158 | let e = _mm_set_epi32(0, 0, 6, -7); |
5159 | assert_eq_m128i(r, e); |
5160 | } |
5161 | |
5162 | #[simd_test(enable = "avx512f" )] |
5163 | unsafe fn test_mm512_cvttpd_epu32() { |
5164 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
5165 | let r = _mm512_cvttpd_epu32(a); |
5166 | let e = _mm256_setr_epi32(0, -1, 2, -1, 4, -1, 6, -1); |
5167 | assert_eq_m256i(r, e); |
5168 | } |
5169 | |
5170 | #[simd_test(enable = "avx512f" )] |
5171 | unsafe fn test_mm512_mask_cvttpd_epu32() { |
5172 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
5173 | let src = _mm256_set1_epi32(0); |
5174 | let r = _mm512_mask_cvttpd_epu32(src, 0, a); |
5175 | assert_eq_m256i(r, src); |
5176 | let r = _mm512_mask_cvttpd_epu32(src, 0b00001111, a); |
5177 | let e = _mm256_setr_epi32(0, -1, 2, -1, 0, 0, 0, 0); |
5178 | assert_eq_m256i(r, e); |
5179 | } |
5180 | |
5181 | #[simd_test(enable = "avx512f" )] |
5182 | unsafe fn test_mm512_maskz_cvttpd_epu32() { |
5183 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
5184 | let r = _mm512_maskz_cvttpd_epu32(0, a); |
5185 | assert_eq_m256i(r, _mm256_setzero_si256()); |
5186 | let r = _mm512_maskz_cvttpd_epu32(0b00001111, a); |
5187 | let e = _mm256_setr_epi32(0, -1, 2, -1, 0, 0, 0, 0); |
5188 | assert_eq_m256i(r, e); |
5189 | } |
5190 | |
5191 | #[simd_test(enable = "avx512f,avx512vl" )] |
5192 | unsafe fn test_mm256_cvttpd_epu32() { |
5193 | let a = _mm256_set_pd(4., 5.5, 6., 7.5); |
5194 | let r = _mm256_cvttpd_epu32(a); |
5195 | let e = _mm_set_epi32(4, 5, 6, 7); |
5196 | assert_eq_m128i(r, e); |
5197 | } |
5198 | |
5199 | #[simd_test(enable = "avx512f,avx512vl" )] |
5200 | unsafe fn test_mm256_mask_cvttpd_epu32() { |
5201 | let a = _mm256_set_pd(4., 5.5, 6., 7.5); |
5202 | let src = _mm_set1_epi32(0); |
5203 | let r = _mm256_mask_cvttpd_epu32(src, 0, a); |
5204 | assert_eq_m128i(r, src); |
5205 | let r = _mm256_mask_cvttpd_epu32(src, 0b00001111, a); |
5206 | let e = _mm_set_epi32(4, 5, 6, 7); |
5207 | assert_eq_m128i(r, e); |
5208 | } |
5209 | |
5210 | #[simd_test(enable = "avx512f,avx512vl" )] |
5211 | unsafe fn test_mm256_maskz_cvttpd_epu32() { |
5212 | let a = _mm256_set_pd(4., 5.5, 6., 7.5); |
5213 | let r = _mm256_maskz_cvttpd_epu32(0, a); |
5214 | assert_eq_m128i(r, _mm_setzero_si128()); |
5215 | let r = _mm256_maskz_cvttpd_epu32(0b00001111, a); |
5216 | let e = _mm_set_epi32(4, 5, 6, 7); |
5217 | assert_eq_m128i(r, e); |
5218 | } |
5219 | |
5220 | #[simd_test(enable = "avx512f,avx512vl" )] |
5221 | unsafe fn test_mm_cvttpd_epu32() { |
5222 | let a = _mm_set_pd(6., 7.5); |
5223 | let r = _mm_cvttpd_epu32(a); |
5224 | let e = _mm_set_epi32(0, 0, 6, 7); |
5225 | assert_eq_m128i(r, e); |
5226 | } |
5227 | |
5228 | #[simd_test(enable = "avx512f,avx512vl" )] |
5229 | unsafe fn test_mm_mask_cvttpd_epu32() { |
5230 | let a = _mm_set_pd(6., 7.5); |
5231 | let src = _mm_set1_epi32(0); |
5232 | let r = _mm_mask_cvttpd_epu32(src, 0, a); |
5233 | assert_eq_m128i(r, src); |
5234 | let r = _mm_mask_cvttpd_epu32(src, 0b00000011, a); |
5235 | let e = _mm_set_epi32(0, 0, 6, 7); |
5236 | assert_eq_m128i(r, e); |
5237 | } |
5238 | |
5239 | #[simd_test(enable = "avx512f,avx512vl" )] |
5240 | unsafe fn test_mm_maskz_cvttpd_epu32() { |
5241 | let a = _mm_set_pd(6., 7.5); |
5242 | let r = _mm_maskz_cvttpd_epu32(0, a); |
5243 | assert_eq_m128i(r, _mm_setzero_si128()); |
5244 | let r = _mm_maskz_cvttpd_epu32(0b00000011, a); |
5245 | let e = _mm_set_epi32(0, 0, 6, 7); |
5246 | assert_eq_m128i(r, e); |
5247 | } |
5248 | |
5249 | #[simd_test(enable = "avx512f" )] |
5250 | unsafe fn test_mm512_add_round_pd() { |
5251 | let a = _mm512_setr_pd(8., 9.5, 10., 11.5, 12., 13.5, 14., 0.000000000000000007); |
5252 | let b = _mm512_set1_pd(-1.); |
5253 | let r = _mm512_add_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a, b); |
5254 | let e = _mm512_setr_pd(7., 8.5, 9., 10.5, 11., 12.5, 13., -1.0); |
5255 | assert_eq_m512d(r, e); |
5256 | let r = _mm512_add_round_pd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b); |
5257 | let e = _mm512_setr_pd(7., 8.5, 9., 10.5, 11., 12.5, 13., -0.9999999999999999); |
5258 | assert_eq_m512d(r, e); |
5259 | } |
5260 | |
5261 | #[simd_test(enable = "avx512f" )] |
5262 | unsafe fn test_mm512_mask_add_round_pd() { |
5263 | let a = _mm512_setr_pd(8., 9.5, 10., 11.5, 12., 13.5, 14., 0.000000000000000007); |
5264 | let b = _mm512_set1_pd(-1.); |
5265 | let r = _mm512_mask_add_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5266 | a, 0, a, b, |
5267 | ); |
5268 | assert_eq_m512d(r, a); |
5269 | let r = _mm512_mask_add_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5270 | a, 0b11110000, a, b, |
5271 | ); |
5272 | let e = _mm512_setr_pd(8., 9.5, 10., 11.5, 11., 12.5, 13., -1.0); |
5273 | assert_eq_m512d(r, e); |
5274 | } |
5275 | |
5276 | #[simd_test(enable = "avx512f" )] |
5277 | unsafe fn test_mm512_maskz_add_round_pd() { |
5278 | let a = _mm512_setr_pd(8., 9.5, 10., 11.5, 12., 13.5, 14., 0.000000000000000007); |
5279 | let b = _mm512_set1_pd(-1.); |
5280 | let r = |
5281 | _mm512_maskz_add_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(0, a, b); |
5282 | assert_eq_m512d(r, _mm512_setzero_pd()); |
5283 | let r = _mm512_maskz_add_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5284 | 0b11110000, a, b, |
5285 | ); |
5286 | let e = _mm512_setr_pd(0., 0., 0., 0., 11., 12.5, 13., -1.0); |
5287 | assert_eq_m512d(r, e); |
5288 | } |
5289 | |
5290 | #[simd_test(enable = "avx512f" )] |
5291 | unsafe fn test_mm512_sub_round_pd() { |
5292 | let a = _mm512_setr_pd(8., 9.5, 10., 11.5, 12., 13.5, 14., 0.000000000000000007); |
5293 | let b = _mm512_set1_pd(1.); |
5294 | let r = _mm512_sub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a, b); |
5295 | let e = _mm512_setr_pd(7., 8.5, 9., 10.5, 11., 12.5, 13., -1.0); |
5296 | assert_eq_m512d(r, e); |
5297 | let r = _mm512_sub_round_pd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b); |
5298 | let e = _mm512_setr_pd(7., 8.5, 9., 10.5, 11., 12.5, 13., -0.9999999999999999); |
5299 | assert_eq_m512d(r, e); |
5300 | } |
5301 | |
5302 | #[simd_test(enable = "avx512f" )] |
5303 | unsafe fn test_mm512_mask_sub_round_pd() { |
5304 | let a = _mm512_setr_pd(8., 9.5, 10., 11.5, 12., 13.5, 14., 0.000000000000000007); |
5305 | let b = _mm512_set1_pd(1.); |
5306 | let r = _mm512_mask_sub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5307 | a, 0, a, b, |
5308 | ); |
5309 | assert_eq_m512d(r, a); |
5310 | let r = _mm512_mask_sub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5311 | a, 0b11110000, a, b, |
5312 | ); |
5313 | let e = _mm512_setr_pd(8., 9.5, 10., 11.5, 11., 12.5, 13., -1.0); |
5314 | assert_eq_m512d(r, e); |
5315 | } |
5316 | |
5317 | #[simd_test(enable = "avx512f" )] |
5318 | unsafe fn test_mm512_maskz_sub_round_pd() { |
5319 | let a = _mm512_setr_pd(8., 9.5, 10., 11.5, 12., 13.5, 14., 0.000000000000000007); |
5320 | let b = _mm512_set1_pd(1.); |
5321 | let r = |
5322 | _mm512_maskz_sub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(0, a, b); |
5323 | assert_eq_m512d(r, _mm512_setzero_pd()); |
5324 | let r = _mm512_maskz_sub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5325 | 0b11110000, a, b, |
5326 | ); |
5327 | let e = _mm512_setr_pd(0., 0., 0., 0., 11., 12.5, 13., -1.0); |
5328 | assert_eq_m512d(r, e); |
5329 | } |
5330 | |
5331 | #[simd_test(enable = "avx512f" )] |
5332 | unsafe fn test_mm512_mul_round_pd() { |
5333 | let a = _mm512_setr_pd(8., 9.5, 10., 11.5, 12., 13.5, 14., 0.); |
5334 | let b = _mm512_set1_pd(0.1); |
5335 | let r = _mm512_mul_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a, b); |
5336 | let e = _mm512_setr_pd( |
5337 | 0.8, |
5338 | 0.9500000000000001, |
5339 | 1., |
5340 | 1.1500000000000001, |
5341 | 1.2000000000000002, |
5342 | 1.35, |
5343 | 1.4000000000000001, |
5344 | 0., |
5345 | ); |
5346 | assert_eq_m512d(r, e); |
5347 | let r = _mm512_mul_round_pd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b); |
5348 | let e = _mm512_setr_pd(0.8, 0.95, 1.0, 1.15, 1.2, 1.3499999999999999, 1.4, 0.0); |
5349 | assert_eq_m512d(r, e); |
5350 | } |
5351 | |
5352 | #[simd_test(enable = "avx512f" )] |
5353 | unsafe fn test_mm512_mask_mul_round_pd() { |
5354 | let a = _mm512_setr_pd(8., 9.5, 10., 11.5, 12., 13.5, 14., 0.); |
5355 | let b = _mm512_set1_pd(0.1); |
5356 | let r = _mm512_mask_mul_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5357 | a, 0, a, b, |
5358 | ); |
5359 | assert_eq_m512d(r, a); |
5360 | let r = _mm512_mask_mul_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5361 | a, 0b11110000, a, b, |
5362 | ); |
5363 | let e = _mm512_setr_pd( |
5364 | 8., |
5365 | 9.5, |
5366 | 10., |
5367 | 11.5, |
5368 | 1.2000000000000002, |
5369 | 1.35, |
5370 | 1.4000000000000001, |
5371 | 0., |
5372 | ); |
5373 | assert_eq_m512d(r, e); |
5374 | } |
5375 | |
5376 | #[simd_test(enable = "avx512f" )] |
5377 | unsafe fn test_mm512_maskz_mul_round_pd() { |
5378 | let a = _mm512_setr_pd(8., 9.5, 10., 11.5, 12., 13.5, 14., 0.); |
5379 | let b = _mm512_set1_pd(0.1); |
5380 | let r = |
5381 | _mm512_maskz_mul_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(0, a, b); |
5382 | assert_eq_m512d(r, _mm512_setzero_pd()); |
5383 | let r = _mm512_maskz_mul_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5384 | 0b11110000, a, b, |
5385 | ); |
5386 | let e = _mm512_setr_pd( |
5387 | 0., |
5388 | 0., |
5389 | 0., |
5390 | 0., |
5391 | 1.2000000000000002, |
5392 | 1.35, |
5393 | 1.4000000000000001, |
5394 | 0., |
5395 | ); |
5396 | assert_eq_m512d(r, e); |
5397 | } |
5398 | |
5399 | #[simd_test(enable = "avx512f" )] |
5400 | unsafe fn test_mm512_div_round_pd() { |
5401 | let a = _mm512_set1_pd(1.); |
5402 | let b = _mm512_set1_pd(3.); |
5403 | let r = _mm512_div_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a, b); |
5404 | let e = _mm512_set1_pd(0.3333333333333333); |
5405 | assert_eq_m512d(r, e); |
5406 | let r = _mm512_div_round_pd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b); |
5407 | let e = _mm512_set1_pd(0.3333333333333333); |
5408 | assert_eq_m512d(r, e); |
5409 | } |
5410 | |
5411 | #[simd_test(enable = "avx512f" )] |
5412 | unsafe fn test_mm512_mask_div_round_pd() { |
5413 | let a = _mm512_set1_pd(1.); |
5414 | let b = _mm512_set1_pd(3.); |
5415 | let r = _mm512_mask_div_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5416 | a, 0, a, b, |
5417 | ); |
5418 | assert_eq_m512d(r, a); |
5419 | let r = _mm512_mask_div_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5420 | a, 0b11110000, a, b, |
5421 | ); |
5422 | let e = _mm512_setr_pd( |
5423 | 1., |
5424 | 1., |
5425 | 1., |
5426 | 1., |
5427 | 0.3333333333333333, |
5428 | 0.3333333333333333, |
5429 | 0.3333333333333333, |
5430 | 0.3333333333333333, |
5431 | ); |
5432 | assert_eq_m512d(r, e); |
5433 | } |
5434 | |
5435 | #[simd_test(enable = "avx512f" )] |
5436 | unsafe fn test_mm512_maskz_div_round_pd() { |
5437 | let a = _mm512_set1_pd(1.); |
5438 | let b = _mm512_set1_pd(3.); |
5439 | let r = |
5440 | _mm512_maskz_div_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(0, a, b); |
5441 | assert_eq_m512d(r, _mm512_setzero_pd()); |
5442 | let r = _mm512_maskz_div_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5443 | 0b11110000, a, b, |
5444 | ); |
5445 | let e = _mm512_setr_pd( |
5446 | 0., |
5447 | 0., |
5448 | 0., |
5449 | 0., |
5450 | 0.3333333333333333, |
5451 | 0.3333333333333333, |
5452 | 0.3333333333333333, |
5453 | 0.3333333333333333, |
5454 | ); |
5455 | assert_eq_m512d(r, e); |
5456 | } |
5457 | |
5458 | #[simd_test(enable = "avx512f" )] |
5459 | unsafe fn test_mm512_sqrt_round_pd() { |
5460 | let a = _mm512_set1_pd(3.); |
5461 | let r = _mm512_sqrt_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a); |
5462 | let e = _mm512_set1_pd(1.7320508075688772); |
5463 | assert_eq_m512d(r, e); |
5464 | let r = _mm512_sqrt_round_pd::<{ _MM_FROUND_TO_POS_INF | _MM_FROUND_NO_EXC }>(a); |
5465 | let e = _mm512_set1_pd(1.7320508075688774); |
5466 | assert_eq_m512d(r, e); |
5467 | } |
5468 | |
5469 | #[simd_test(enable = "avx512f" )] |
5470 | unsafe fn test_mm512_mask_sqrt_round_pd() { |
5471 | let a = _mm512_set1_pd(3.); |
5472 | let r = |
5473 | _mm512_mask_sqrt_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a, 0, a); |
5474 | assert_eq_m512d(r, a); |
5475 | let r = _mm512_mask_sqrt_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5476 | a, 0b11110000, a, |
5477 | ); |
5478 | let e = _mm512_setr_pd( |
5479 | 3., |
5480 | 3., |
5481 | 3., |
5482 | 3., |
5483 | 1.7320508075688772, |
5484 | 1.7320508075688772, |
5485 | 1.7320508075688772, |
5486 | 1.7320508075688772, |
5487 | ); |
5488 | assert_eq_m512d(r, e); |
5489 | } |
5490 | |
5491 | #[simd_test(enable = "avx512f" )] |
5492 | unsafe fn test_mm512_maskz_sqrt_round_pd() { |
5493 | let a = _mm512_set1_pd(3.); |
5494 | let r = |
5495 | _mm512_maskz_sqrt_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(0, a); |
5496 | assert_eq_m512d(r, _mm512_setzero_pd()); |
5497 | let r = _mm512_maskz_sqrt_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5498 | 0b11110000, a, |
5499 | ); |
5500 | let e = _mm512_setr_pd( |
5501 | 0., |
5502 | 0., |
5503 | 0., |
5504 | 0., |
5505 | 1.7320508075688772, |
5506 | 1.7320508075688772, |
5507 | 1.7320508075688772, |
5508 | 1.7320508075688772, |
5509 | ); |
5510 | assert_eq_m512d(r, e); |
5511 | } |
5512 | |
5513 | #[simd_test(enable = "avx512f" )] |
5514 | unsafe fn test_mm512_fmadd_round_pd() { |
5515 | let a = _mm512_set1_pd(0.000000000000000007); |
5516 | let b = _mm512_set1_pd(1.); |
5517 | let c = _mm512_set1_pd(-1.); |
5518 | let r = _mm512_fmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a, b, c); |
5519 | let e = _mm512_set1_pd(-1.); |
5520 | assert_eq_m512d(r, e); |
5521 | let r = _mm512_fmadd_round_pd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b, c); |
5522 | let e = _mm512_set1_pd(-0.9999999999999999); |
5523 | assert_eq_m512d(r, e); |
5524 | } |
5525 | |
5526 | #[simd_test(enable = "avx512f" )] |
5527 | unsafe fn test_mm512_mask_fmadd_round_pd() { |
5528 | let a = _mm512_set1_pd(0.000000000000000007); |
5529 | let b = _mm512_set1_pd(1.); |
5530 | let c = _mm512_set1_pd(-1.); |
5531 | let r = _mm512_mask_fmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5532 | a, 0, b, c, |
5533 | ); |
5534 | assert_eq_m512d(r, a); |
5535 | let r = _mm512_mask_fmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5536 | a, 0b00001111, b, c, |
5537 | ); |
5538 | let e = _mm512_setr_pd( |
5539 | -1., |
5540 | -1., |
5541 | -1., |
5542 | -1., |
5543 | 0.000000000000000007, |
5544 | 0.000000000000000007, |
5545 | 0.000000000000000007, |
5546 | 0.000000000000000007, |
5547 | ); |
5548 | assert_eq_m512d(r, e); |
5549 | } |
5550 | |
5551 | #[simd_test(enable = "avx512f" )] |
5552 | unsafe fn test_mm512_maskz_fmadd_round_pd() { |
5553 | let a = _mm512_set1_pd(0.000000000000000007); |
5554 | let b = _mm512_set1_pd(1.); |
5555 | let c = _mm512_set1_pd(-1.); |
5556 | let r = _mm512_maskz_fmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5557 | 0, a, b, c, |
5558 | ); |
5559 | assert_eq_m512d(r, _mm512_setzero_pd()); |
5560 | let r = _mm512_maskz_fmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5561 | 0b00001111, a, b, c, |
5562 | ); |
5563 | let e = _mm512_setr_pd(-1., -1., -1., -1., 0., 0., 0., 0.); |
5564 | assert_eq_m512d(r, e); |
5565 | } |
5566 | |
5567 | #[simd_test(enable = "avx512f" )] |
5568 | unsafe fn test_mm512_mask3_fmadd_round_pd() { |
5569 | let a = _mm512_set1_pd(0.000000000000000007); |
5570 | let b = _mm512_set1_pd(1.); |
5571 | let c = _mm512_set1_pd(-1.); |
5572 | let r = _mm512_mask3_fmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5573 | a, b, c, 0, |
5574 | ); |
5575 | assert_eq_m512d(r, c); |
5576 | let r = _mm512_mask3_fmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5577 | a, b, c, 0b00001111, |
5578 | ); |
5579 | let e = _mm512_setr_pd(-1., -1., -1., -1., -1., -1., -1., -1.); |
5580 | assert_eq_m512d(r, e); |
5581 | } |
5582 | |
5583 | #[simd_test(enable = "avx512f" )] |
5584 | unsafe fn test_mm512_fmsub_round_pd() { |
5585 | let a = _mm512_set1_pd(0.000000000000000007); |
5586 | let b = _mm512_set1_pd(1.); |
5587 | let c = _mm512_set1_pd(1.); |
5588 | let r = _mm512_fmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a, b, c); |
5589 | let e = _mm512_set1_pd(-1.); |
5590 | assert_eq_m512d(r, e); |
5591 | let r = _mm512_fmsub_round_pd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b, c); |
5592 | let e = _mm512_set1_pd(-0.9999999999999999); |
5593 | assert_eq_m512d(r, e); |
5594 | } |
5595 | |
5596 | #[simd_test(enable = "avx512f" )] |
5597 | unsafe fn test_mm512_mask_fmsub_round_pd() { |
5598 | let a = _mm512_set1_pd(0.000000000000000007); |
5599 | let b = _mm512_set1_pd(1.); |
5600 | let c = _mm512_set1_pd(1.); |
5601 | let r = _mm512_mask_fmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5602 | a, 0, b, c, |
5603 | ); |
5604 | assert_eq_m512d(r, a); |
5605 | let r = _mm512_mask_fmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5606 | a, 0b00001111, b, c, |
5607 | ); |
5608 | let e = _mm512_setr_pd( |
5609 | -1., |
5610 | -1., |
5611 | -1., |
5612 | -1., |
5613 | 0.000000000000000007, |
5614 | 0.000000000000000007, |
5615 | 0.000000000000000007, |
5616 | 0.000000000000000007, |
5617 | ); |
5618 | assert_eq_m512d(r, e); |
5619 | } |
5620 | |
5621 | #[simd_test(enable = "avx512f" )] |
5622 | unsafe fn test_mm512_maskz_fmsub_round_pd() { |
5623 | let a = _mm512_set1_pd(0.000000000000000007); |
5624 | let b = _mm512_set1_pd(1.); |
5625 | let c = _mm512_set1_pd(1.); |
5626 | let r = _mm512_maskz_fmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5627 | 0, a, b, c, |
5628 | ); |
5629 | assert_eq_m512d(r, _mm512_setzero_pd()); |
5630 | let r = _mm512_maskz_fmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5631 | 0b00001111, a, b, c, |
5632 | ); |
5633 | let e = _mm512_setr_pd(-1., -1., -1., -1., 0., 0., 0., 0.); |
5634 | assert_eq_m512d(r, e); |
5635 | } |
5636 | |
5637 | #[simd_test(enable = "avx512f" )] |
5638 | unsafe fn test_mm512_mask3_fmsub_round_pd() { |
5639 | let a = _mm512_set1_pd(0.000000000000000007); |
5640 | let b = _mm512_set1_pd(1.); |
5641 | let c = _mm512_set1_pd(1.); |
5642 | let r = _mm512_mask3_fmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5643 | a, b, c, 0, |
5644 | ); |
5645 | assert_eq_m512d(r, c); |
5646 | let r = _mm512_mask3_fmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5647 | a, b, c, 0b00001111, |
5648 | ); |
5649 | let e = _mm512_setr_pd(-1., -1., -1., -1., 1., 1., 1., 1.); |
5650 | assert_eq_m512d(r, e); |
5651 | } |
5652 | |
5653 | #[simd_test(enable = "avx512f" )] |
5654 | unsafe fn test_mm512_fmaddsub_round_pd() { |
5655 | let a = _mm512_set1_pd(0.000000000000000007); |
5656 | let b = _mm512_set1_pd(1.); |
5657 | let c = _mm512_set1_pd(-1.); |
5658 | let r = |
5659 | _mm512_fmaddsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a, b, c); |
5660 | let e = _mm512_setr_pd(1., -1., 1., -1., 1., -1., 1., -1.); |
5661 | assert_eq_m512d(r, e); |
5662 | let r = _mm512_fmaddsub_round_pd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b, c); |
5663 | let e = _mm512_setr_pd( |
5664 | 1., |
5665 | -0.9999999999999999, |
5666 | 1., |
5667 | -0.9999999999999999, |
5668 | 1., |
5669 | -0.9999999999999999, |
5670 | 1., |
5671 | -0.9999999999999999, |
5672 | ); |
5673 | assert_eq_m512d(r, e); |
5674 | } |
5675 | |
5676 | #[simd_test(enable = "avx512f" )] |
5677 | unsafe fn test_mm512_mask_fmaddsub_round_pd() { |
5678 | let a = _mm512_set1_pd(0.000000000000000007); |
5679 | let b = _mm512_set1_pd(1.); |
5680 | let c = _mm512_set1_pd(-1.); |
5681 | let r = _mm512_mask_fmaddsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5682 | a, 0, b, c, |
5683 | ); |
5684 | assert_eq_m512d(r, a); |
5685 | let r = _mm512_mask_fmaddsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5686 | a, 0b00001111, b, c, |
5687 | ); |
5688 | let e = _mm512_setr_pd( |
5689 | 1., |
5690 | -1., |
5691 | 1., |
5692 | -1., |
5693 | 0.000000000000000007, |
5694 | 0.000000000000000007, |
5695 | 0.000000000000000007, |
5696 | 0.000000000000000007, |
5697 | ); |
5698 | assert_eq_m512d(r, e); |
5699 | } |
5700 | |
5701 | #[simd_test(enable = "avx512f" )] |
5702 | unsafe fn test_mm512_maskz_fmaddsub_round_pd() { |
5703 | let a = _mm512_set1_pd(0.000000000000000007); |
5704 | let b = _mm512_set1_pd(1.); |
5705 | let c = _mm512_set1_pd(-1.); |
5706 | let r = _mm512_maskz_fmaddsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5707 | 0, a, b, c, |
5708 | ); |
5709 | assert_eq_m512d(r, _mm512_setzero_pd()); |
5710 | let r = _mm512_maskz_fmaddsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5711 | 0b00001111, a, b, c, |
5712 | ); |
5713 | let e = _mm512_setr_pd(1., -1., 1., -1., 0., 0., 0., 0.); |
5714 | assert_eq_m512d(r, e); |
5715 | } |
5716 | |
5717 | #[simd_test(enable = "avx512f" )] |
5718 | unsafe fn test_mm512_mask3_fmaddsub_round_pd() { |
5719 | let a = _mm512_set1_pd(0.000000000000000007); |
5720 | let b = _mm512_set1_pd(1.); |
5721 | let c = _mm512_set1_pd(-1.); |
5722 | let r = _mm512_mask3_fmaddsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5723 | a, b, c, 0, |
5724 | ); |
5725 | assert_eq_m512d(r, c); |
5726 | let r = _mm512_mask3_fmaddsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5727 | a, b, c, 0b00001111, |
5728 | ); |
5729 | let e = _mm512_setr_pd(1., -1., 1., -1., -1., -1., -1., -1.); |
5730 | assert_eq_m512d(r, e); |
5731 | } |
5732 | |
5733 | #[simd_test(enable = "avx512f" )] |
5734 | unsafe fn test_mm512_fmsubadd_round_pd() { |
5735 | let a = _mm512_set1_pd(0.000000000000000007); |
5736 | let b = _mm512_set1_pd(1.); |
5737 | let c = _mm512_set1_pd(-1.); |
5738 | let r = |
5739 | _mm512_fmsubadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a, b, c); |
5740 | let e = _mm512_setr_pd(-1., 1., -1., 1., -1., 1., -1., 1.); |
5741 | assert_eq_m512d(r, e); |
5742 | let r = _mm512_fmsubadd_round_pd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b, c); |
5743 | let e = _mm512_setr_pd( |
5744 | -0.9999999999999999, |
5745 | 1., |
5746 | -0.9999999999999999, |
5747 | 1., |
5748 | -0.9999999999999999, |
5749 | 1., |
5750 | -0.9999999999999999, |
5751 | 1., |
5752 | ); |
5753 | assert_eq_m512d(r, e); |
5754 | } |
5755 | |
5756 | #[simd_test(enable = "avx512f" )] |
5757 | unsafe fn test_mm512_mask_fmsubadd_round_pd() { |
5758 | let a = _mm512_set1_pd(0.000000000000000007); |
5759 | let b = _mm512_set1_pd(1.); |
5760 | let c = _mm512_set1_pd(-1.); |
5761 | let r = _mm512_mask_fmsubadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5762 | a, 0, b, c, |
5763 | ); |
5764 | assert_eq_m512d(r, a); |
5765 | let r = _mm512_mask_fmsubadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5766 | a, 0b00001111, b, c, |
5767 | ); |
5768 | let e = _mm512_setr_pd( |
5769 | -1., |
5770 | 1., |
5771 | -1., |
5772 | 1., |
5773 | 0.000000000000000007, |
5774 | 0.000000000000000007, |
5775 | 0.000000000000000007, |
5776 | 0.000000000000000007, |
5777 | ); |
5778 | assert_eq_m512d(r, e); |
5779 | } |
5780 | |
5781 | #[simd_test(enable = "avx512f" )] |
5782 | unsafe fn test_mm512_maskz_fmsubadd_round_pd() { |
5783 | let a = _mm512_set1_pd(0.000000000000000007); |
5784 | let b = _mm512_set1_pd(1.); |
5785 | let c = _mm512_set1_pd(-1.); |
5786 | let r = _mm512_maskz_fmsubadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5787 | 0, a, b, c, |
5788 | ); |
5789 | assert_eq_m512d(r, _mm512_setzero_pd()); |
5790 | let r = _mm512_maskz_fmsubadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5791 | 0b00001111, a, b, c, |
5792 | ); |
5793 | let e = _mm512_setr_pd(-1., 1., -1., 1., 0., 0., 0., 0.); |
5794 | assert_eq_m512d(r, e); |
5795 | } |
5796 | |
5797 | #[simd_test(enable = "avx512f" )] |
5798 | unsafe fn test_mm512_mask3_fmsubadd_round_pd() { |
5799 | let a = _mm512_set1_pd(0.000000000000000007); |
5800 | let b = _mm512_set1_pd(1.); |
5801 | let c = _mm512_set1_pd(-1.); |
5802 | let r = _mm512_mask3_fmsubadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5803 | a, b, c, 0, |
5804 | ); |
5805 | assert_eq_m512d(r, c); |
5806 | let r = _mm512_mask3_fmsubadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5807 | a, b, c, 0b00001111, |
5808 | ); |
5809 | let e = _mm512_setr_pd(-1., 1., -1., 1., -1., -1., -1., -1.); |
5810 | assert_eq_m512d(r, e); |
5811 | } |
5812 | |
5813 | #[simd_test(enable = "avx512f" )] |
5814 | unsafe fn test_mm512_fnmadd_round_pd() { |
5815 | let a = _mm512_set1_pd(0.000000000000000007); |
5816 | let b = _mm512_set1_pd(1.); |
5817 | let c = _mm512_set1_pd(1.); |
5818 | let r = |
5819 | _mm512_fnmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a, b, c); |
5820 | let e = _mm512_set1_pd(1.); |
5821 | assert_eq_m512d(r, e); |
5822 | let r = _mm512_fnmadd_round_pd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b, c); |
5823 | let e = _mm512_set1_pd(0.9999999999999999); |
5824 | assert_eq_m512d(r, e); |
5825 | } |
5826 | |
5827 | #[simd_test(enable = "avx512f" )] |
5828 | unsafe fn test_mm512_mask_fnmadd_round_pd() { |
5829 | let a = _mm512_set1_pd(0.000000000000000007); |
5830 | let b = _mm512_set1_pd(1.); |
5831 | let c = _mm512_set1_pd(1.); |
5832 | let r = _mm512_mask_fnmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5833 | a, 0, b, c, |
5834 | ); |
5835 | assert_eq_m512d(r, a); |
5836 | let r = _mm512_mask_fnmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5837 | a, 0b00001111, b, c, |
5838 | ); |
5839 | let e = _mm512_setr_pd( |
5840 | 1., |
5841 | 1., |
5842 | 1., |
5843 | 1., |
5844 | 0.000000000000000007, |
5845 | 0.000000000000000007, |
5846 | 0.000000000000000007, |
5847 | 0.000000000000000007, |
5848 | ); |
5849 | assert_eq_m512d(r, e); |
5850 | } |
5851 | |
5852 | #[simd_test(enable = "avx512f" )] |
5853 | unsafe fn test_mm512_maskz_fnmadd_round_pd() { |
5854 | let a = _mm512_set1_pd(0.000000000000000007); |
5855 | let b = _mm512_set1_pd(1.); |
5856 | let c = _mm512_set1_pd(1.); |
5857 | let r = _mm512_maskz_fnmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5858 | 0, a, b, c, |
5859 | ); |
5860 | assert_eq_m512d(r, _mm512_setzero_pd()); |
5861 | let r = _mm512_maskz_fnmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5862 | 0b00001111, a, b, c, |
5863 | ); |
5864 | let e = _mm512_setr_pd(1., 1., 1., 1., 0., 0., 0., 0.); |
5865 | assert_eq_m512d(r, e); |
5866 | } |
5867 | |
5868 | #[simd_test(enable = "avx512f" )] |
5869 | unsafe fn test_mm512_mask3_fnmadd_round_pd() { |
5870 | let a = _mm512_set1_pd(0.000000000000000007); |
5871 | let b = _mm512_set1_pd(1.); |
5872 | let c = _mm512_set1_pd(1.); |
5873 | let r = _mm512_mask3_fnmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5874 | a, b, c, 0, |
5875 | ); |
5876 | assert_eq_m512d(r, c); |
5877 | let r = _mm512_mask3_fnmadd_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5878 | a, b, c, 0b00001111, |
5879 | ); |
5880 | let e = _mm512_setr_pd(1., 1., 1., 1., 1., 1., 1., 1.); |
5881 | assert_eq_m512d(r, e); |
5882 | } |
5883 | |
5884 | #[simd_test(enable = "avx512f" )] |
5885 | unsafe fn test_mm512_fnmsub_round_pd() { |
5886 | let a = _mm512_set1_pd(0.000000000000000007); |
5887 | let b = _mm512_set1_pd(1.); |
5888 | let c = _mm512_set1_pd(-1.); |
5889 | let r = |
5890 | _mm512_fnmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a, b, c); |
5891 | let e = _mm512_set1_pd(1.); |
5892 | assert_eq_m512d(r, e); |
5893 | let r = _mm512_fnmsub_round_pd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b, c); |
5894 | let e = _mm512_set1_pd(0.9999999999999999); |
5895 | assert_eq_m512d(r, e); |
5896 | } |
5897 | |
5898 | #[simd_test(enable = "avx512f" )] |
5899 | unsafe fn test_mm512_mask_fnmsub_round_pd() { |
5900 | let a = _mm512_set1_pd(0.000000000000000007); |
5901 | let b = _mm512_set1_pd(1.); |
5902 | let c = _mm512_set1_pd(-1.); |
5903 | let r = _mm512_mask_fnmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5904 | a, 0, b, c, |
5905 | ); |
5906 | assert_eq_m512d(r, a); |
5907 | let r = _mm512_mask_fnmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5908 | a, 0b00001111, b, c, |
5909 | ); |
5910 | let e = _mm512_setr_pd( |
5911 | 1., |
5912 | 1., |
5913 | 1., |
5914 | 1., |
5915 | 0.000000000000000007, |
5916 | 0.000000000000000007, |
5917 | 0.000000000000000007, |
5918 | 0.000000000000000007, |
5919 | ); |
5920 | assert_eq_m512d(r, e); |
5921 | } |
5922 | |
5923 | #[simd_test(enable = "avx512f" )] |
5924 | unsafe fn test_mm512_maskz_fnmsub_round_pd() { |
5925 | let a = _mm512_set1_pd(0.000000000000000007); |
5926 | let b = _mm512_set1_pd(1.); |
5927 | let c = _mm512_set1_pd(-1.); |
5928 | let r = _mm512_maskz_fnmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5929 | 0, a, b, c, |
5930 | ); |
5931 | assert_eq_m512d(r, _mm512_setzero_pd()); |
5932 | let r = _mm512_maskz_fnmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5933 | 0b00001111, a, b, c, |
5934 | ); |
5935 | let e = _mm512_setr_pd(1., 1., 1., 1., 0., 0., 0., 0.); |
5936 | assert_eq_m512d(r, e); |
5937 | } |
5938 | |
5939 | #[simd_test(enable = "avx512f" )] |
5940 | unsafe fn test_mm512_mask3_fnmsub_round_pd() { |
5941 | let a = _mm512_set1_pd(0.000000000000000007); |
5942 | let b = _mm512_set1_pd(1.); |
5943 | let c = _mm512_set1_pd(-1.); |
5944 | let r = _mm512_mask3_fnmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5945 | a, b, c, 0, |
5946 | ); |
5947 | assert_eq_m512d(r, c); |
5948 | let r = _mm512_mask3_fnmsub_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
5949 | a, b, c, 0b00001111, |
5950 | ); |
5951 | let e = _mm512_setr_pd(1., 1., 1., 1., -1., -1., -1., -1.); |
5952 | assert_eq_m512d(r, e); |
5953 | } |
5954 | |
5955 | #[simd_test(enable = "avx512f" )] |
5956 | unsafe fn test_mm512_max_round_pd() { |
5957 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
5958 | let b = _mm512_setr_pd(7., 6., 5., 4., 3., 2., 1., 0.); |
5959 | let r = _mm512_max_round_pd::<_MM_FROUND_CUR_DIRECTION>(a, b); |
5960 | let e = _mm512_setr_pd(7., 6., 5., 4., 4., 5., 6., 7.); |
5961 | assert_eq_m512d(r, e); |
5962 | } |
5963 | |
5964 | #[simd_test(enable = "avx512f" )] |
5965 | unsafe fn test_mm512_mask_max_round_pd() { |
5966 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
5967 | let b = _mm512_setr_pd(7., 6., 5., 4., 3., 2., 1., 0.); |
5968 | let r = _mm512_mask_max_round_pd::<_MM_FROUND_CUR_DIRECTION>(a, 0, a, b); |
5969 | assert_eq_m512d(r, a); |
5970 | let r = _mm512_mask_max_round_pd::<_MM_FROUND_CUR_DIRECTION>(a, 0b00001111, a, b); |
5971 | let e = _mm512_setr_pd(7., 6., 5., 4., 4., 5., 6., 7.); |
5972 | assert_eq_m512d(r, e); |
5973 | } |
5974 | |
5975 | #[simd_test(enable = "avx512f" )] |
5976 | unsafe fn test_mm512_maskz_max_round_pd() { |
5977 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
5978 | let b = _mm512_setr_pd(7., 6., 5., 4., 3., 2., 1., 0.); |
5979 | let r = _mm512_maskz_max_round_pd::<_MM_FROUND_CUR_DIRECTION>(0, a, b); |
5980 | assert_eq_m512d(r, _mm512_setzero_pd()); |
5981 | let r = _mm512_maskz_max_round_pd::<_MM_FROUND_CUR_DIRECTION>(0b00001111, a, b); |
5982 | let e = _mm512_setr_pd(7., 6., 5., 4., 0., 0., 0., 0.); |
5983 | assert_eq_m512d(r, e); |
5984 | } |
5985 | |
5986 | #[simd_test(enable = "avx512f" )] |
5987 | unsafe fn test_mm512_min_round_pd() { |
5988 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
5989 | let b = _mm512_setr_pd(7., 6., 5., 4., 3., 2., 1., 0.); |
5990 | let r = _mm512_min_round_pd::<_MM_FROUND_CUR_DIRECTION>(a, b); |
5991 | let e = _mm512_setr_pd(0., 1., 2., 3., 3., 2., 1., 0.); |
5992 | assert_eq_m512d(r, e); |
5993 | } |
5994 | |
5995 | #[simd_test(enable = "avx512f" )] |
5996 | unsafe fn test_mm512_mask_min_round_pd() { |
5997 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
5998 | let b = _mm512_setr_pd(7., 6., 5., 4., 3., 2., 1., 0.); |
5999 | let r = _mm512_mask_min_round_pd::<_MM_FROUND_CUR_DIRECTION>(a, 0, a, b); |
6000 | assert_eq_m512d(r, a); |
6001 | let r = _mm512_mask_min_round_pd::<_MM_FROUND_CUR_DIRECTION>(a, 0b00001111, a, b); |
6002 | let e = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
6003 | assert_eq_m512d(r, e); |
6004 | } |
6005 | |
6006 | #[simd_test(enable = "avx512f" )] |
6007 | unsafe fn test_mm512_maskz_min_round_pd() { |
6008 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
6009 | let b = _mm512_setr_pd(7., 6., 5., 4., 3., 2., 1., 0.); |
6010 | let r = _mm512_maskz_min_round_pd::<_MM_FROUND_CUR_DIRECTION>(0, a, b); |
6011 | assert_eq_m512d(r, _mm512_setzero_pd()); |
6012 | let r = _mm512_maskz_min_round_pd::<_MM_FROUND_CUR_DIRECTION>(0b00001111, a, b); |
6013 | let e = _mm512_setr_pd(0., 1., 2., 3., 0., 0., 0., 0.); |
6014 | assert_eq_m512d(r, e); |
6015 | } |
6016 | |
6017 | #[simd_test(enable = "avx512f" )] |
6018 | unsafe fn test_mm512_getexp_round_pd() { |
6019 | let a = _mm512_set1_pd(3.); |
6020 | let r = _mm512_getexp_round_pd::<_MM_FROUND_CUR_DIRECTION>(a); |
6021 | let e = _mm512_set1_pd(1.); |
6022 | assert_eq_m512d(r, e); |
6023 | } |
6024 | |
6025 | #[simd_test(enable = "avx512f" )] |
6026 | unsafe fn test_mm512_mask_getexp_round_pd() { |
6027 | let a = _mm512_set1_pd(3.); |
6028 | let r = _mm512_mask_getexp_round_pd::<_MM_FROUND_CUR_DIRECTION>(a, 0, a); |
6029 | assert_eq_m512d(r, a); |
6030 | let r = _mm512_mask_getexp_round_pd::<_MM_FROUND_CUR_DIRECTION>(a, 0b11110000, a); |
6031 | let e = _mm512_setr_pd(3., 3., 3., 3., 1., 1., 1., 1.); |
6032 | assert_eq_m512d(r, e); |
6033 | } |
6034 | |
6035 | #[simd_test(enable = "avx512f" )] |
6036 | unsafe fn test_mm512_maskz_getexp_round_pd() { |
6037 | let a = _mm512_set1_pd(3.); |
6038 | let r = _mm512_maskz_getexp_round_pd::<_MM_FROUND_CUR_DIRECTION>(0, a); |
6039 | assert_eq_m512d(r, _mm512_setzero_pd()); |
6040 | let r = _mm512_maskz_getexp_round_pd::<_MM_FROUND_CUR_DIRECTION>(0b11110000, a); |
6041 | let e = _mm512_setr_pd(0., 0., 0., 0., 1., 1., 1., 1.); |
6042 | assert_eq_m512d(r, e); |
6043 | } |
6044 | |
6045 | #[simd_test(enable = "avx512f" )] |
6046 | unsafe fn test_mm512_roundscale_round_pd() { |
6047 | let a = _mm512_set1_pd(1.1); |
6048 | let r = _mm512_roundscale_round_pd::<0, _MM_FROUND_CUR_DIRECTION>(a); |
6049 | let e = _mm512_set1_pd(1.0); |
6050 | assert_eq_m512d(r, e); |
6051 | } |
6052 | |
6053 | #[simd_test(enable = "avx512f" )] |
6054 | unsafe fn test_mm512_mask_roundscale_round_pd() { |
6055 | let a = _mm512_set1_pd(1.1); |
6056 | let r = _mm512_mask_roundscale_round_pd::<0, _MM_FROUND_CUR_DIRECTION>(a, 0, a); |
6057 | let e = _mm512_set1_pd(1.1); |
6058 | assert_eq_m512d(r, e); |
6059 | let r = _mm512_mask_roundscale_round_pd::<0, _MM_FROUND_CUR_DIRECTION>(a, 0b11111111, a); |
6060 | let e = _mm512_set1_pd(1.0); |
6061 | assert_eq_m512d(r, e); |
6062 | } |
6063 | |
6064 | #[simd_test(enable = "avx512f" )] |
6065 | unsafe fn test_mm512_maskz_roundscale_round_pd() { |
6066 | let a = _mm512_set1_pd(1.1); |
6067 | let r = _mm512_maskz_roundscale_round_pd::<0, _MM_FROUND_CUR_DIRECTION>(0, a); |
6068 | assert_eq_m512d(r, _mm512_setzero_pd()); |
6069 | let r = _mm512_maskz_roundscale_round_pd::<0, _MM_FROUND_CUR_DIRECTION>(0b11111111, a); |
6070 | let e = _mm512_set1_pd(1.0); |
6071 | assert_eq_m512d(r, e); |
6072 | } |
6073 | |
6074 | #[simd_test(enable = "avx512f" )] |
6075 | unsafe fn test_mm512_scalef_round_pd() { |
6076 | let a = _mm512_set1_pd(1.); |
6077 | let b = _mm512_set1_pd(3.); |
6078 | let r = _mm512_scalef_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>(a, b); |
6079 | let e = _mm512_set1_pd(8.); |
6080 | assert_eq_m512d(r, e); |
6081 | } |
6082 | |
6083 | #[simd_test(enable = "avx512f" )] |
6084 | unsafe fn test_mm512_mask_scalef_round_pd() { |
6085 | let a = _mm512_set1_pd(1.); |
6086 | let b = _mm512_set1_pd(3.); |
6087 | let r = _mm512_mask_scalef_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
6088 | a, 0, a, b, |
6089 | ); |
6090 | assert_eq_m512d(r, a); |
6091 | let r = _mm512_mask_scalef_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
6092 | a, 0b11110000, a, b, |
6093 | ); |
6094 | let e = _mm512_set_pd(8., 8., 8., 8., 1., 1., 1., 1.); |
6095 | assert_eq_m512d(r, e); |
6096 | } |
6097 | |
6098 | #[simd_test(enable = "avx512f" )] |
6099 | unsafe fn test_mm512_maskz_scalef_round_pd() { |
6100 | let a = _mm512_set1_pd(1.); |
6101 | let b = _mm512_set1_pd(3.); |
6102 | let r = _mm512_maskz_scalef_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
6103 | 0, a, b, |
6104 | ); |
6105 | assert_eq_m512d(r, _mm512_setzero_pd()); |
6106 | let r = _mm512_maskz_scalef_round_pd::<{ _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC }>( |
6107 | 0b11110000, a, b, |
6108 | ); |
6109 | let e = _mm512_set_pd(8., 8., 8., 8., 0., 0., 0., 0.); |
6110 | assert_eq_m512d(r, e); |
6111 | } |
6112 | |
6113 | #[simd_test(enable = "avx512f" )] |
6114 | unsafe fn test_mm512_fixupimm_round_pd() { |
6115 | let a = _mm512_set1_pd(f64::NAN); |
6116 | let b = _mm512_set1_pd(f64::MAX); |
6117 | let c = _mm512_set1_epi64(i32::MAX as i64); |
6118 | let r = _mm512_fixupimm_round_pd::<5, _MM_FROUND_CUR_DIRECTION>(a, b, c); |
6119 | let e = _mm512_set1_pd(0.0); |
6120 | assert_eq_m512d(r, e); |
6121 | } |
6122 | |
6123 | #[simd_test(enable = "avx512f" )] |
6124 | unsafe fn test_mm512_mask_fixupimm_round_pd() { |
6125 | let a = _mm512_set_pd(f64::NAN, f64::NAN, f64::NAN, f64::NAN, 1., 1., 1., 1.); |
6126 | let b = _mm512_set1_pd(f64::MAX); |
6127 | let c = _mm512_set1_epi64(i32::MAX as i64); |
6128 | let r = _mm512_mask_fixupimm_round_pd::<5, _MM_FROUND_CUR_DIRECTION>(a, 0b11110000, b, c); |
6129 | let e = _mm512_set_pd(0., 0., 0., 0., 1., 1., 1., 1.); |
6130 | assert_eq_m512d(r, e); |
6131 | } |
6132 | |
6133 | #[simd_test(enable = "avx512f" )] |
6134 | unsafe fn test_mm512_maskz_fixupimm_round_pd() { |
6135 | let a = _mm512_set_pd(f64::NAN, f64::NAN, f64::NAN, f64::NAN, 1., 1., 1., 1.); |
6136 | let b = _mm512_set1_pd(f64::MAX); |
6137 | let c = _mm512_set1_epi64(i32::MAX as i64); |
6138 | let r = _mm512_maskz_fixupimm_round_pd::<5, _MM_FROUND_CUR_DIRECTION>(0b11110000, a, b, c); |
6139 | let e = _mm512_set_pd(0., 0., 0., 0., 0., 0., 0., 0.); |
6140 | assert_eq_m512d(r, e); |
6141 | } |
6142 | |
6143 | #[simd_test(enable = "avx512f" )] |
6144 | unsafe fn test_mm512_getmant_round_pd() { |
6145 | let a = _mm512_set1_pd(10.); |
6146 | let r = _mm512_getmant_round_pd::< |
6147 | _MM_MANT_NORM_1_2, |
6148 | _MM_MANT_SIGN_SRC, |
6149 | _MM_FROUND_CUR_DIRECTION, |
6150 | >(a); |
6151 | let e = _mm512_set1_pd(1.25); |
6152 | assert_eq_m512d(r, e); |
6153 | } |
6154 | |
6155 | #[simd_test(enable = "avx512f" )] |
6156 | unsafe fn test_mm512_mask_getmant_round_pd() { |
6157 | let a = _mm512_set1_pd(10.); |
6158 | let r = _mm512_mask_getmant_round_pd::< |
6159 | _MM_MANT_NORM_1_2, |
6160 | _MM_MANT_SIGN_SRC, |
6161 | _MM_FROUND_CUR_DIRECTION, |
6162 | >(a, 0, a); |
6163 | assert_eq_m512d(r, a); |
6164 | let r = _mm512_mask_getmant_round_pd::< |
6165 | _MM_MANT_NORM_1_2, |
6166 | _MM_MANT_SIGN_SRC, |
6167 | _MM_FROUND_CUR_DIRECTION, |
6168 | >(a, 0b11110000, a); |
6169 | let e = _mm512_setr_pd(10., 10., 10., 10., 1.25, 1.25, 1.25, 1.25); |
6170 | assert_eq_m512d(r, e); |
6171 | } |
6172 | |
6173 | #[simd_test(enable = "avx512f" )] |
6174 | unsafe fn test_mm512_maskz_getmant_round_pd() { |
6175 | let a = _mm512_set1_pd(10.); |
6176 | let r = _mm512_maskz_getmant_round_pd::< |
6177 | _MM_MANT_NORM_1_2, |
6178 | _MM_MANT_SIGN_SRC, |
6179 | _MM_FROUND_CUR_DIRECTION, |
6180 | >(0, a); |
6181 | assert_eq_m512d(r, _mm512_setzero_pd()); |
6182 | let r = _mm512_maskz_getmant_round_pd::< |
6183 | _MM_MANT_NORM_1_2, |
6184 | _MM_MANT_SIGN_SRC, |
6185 | _MM_FROUND_CUR_DIRECTION, |
6186 | >(0b11110000, a); |
6187 | let e = _mm512_setr_pd(0., 0., 0., 0., 1.25, 1.25, 1.25, 1.25); |
6188 | assert_eq_m512d(r, e); |
6189 | } |
6190 | |
6191 | #[simd_test(enable = "avx512f" )] |
6192 | unsafe fn test_mm512_cvt_roundps_pd() { |
6193 | let a = _mm256_setr_ps(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6194 | let r = _mm512_cvt_roundps_pd::<_MM_FROUND_CUR_DIRECTION>(a); |
6195 | let e = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6196 | assert_eq_m512d(r, e); |
6197 | } |
6198 | |
6199 | #[simd_test(enable = "avx512f" )] |
6200 | unsafe fn test_mm512_mask_cvt_roundps_pd() { |
6201 | let a = _mm256_setr_ps(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6202 | let src = _mm512_set1_pd(0.); |
6203 | let r = _mm512_mask_cvt_roundps_pd::<_MM_FROUND_CUR_DIRECTION>(src, 0, a); |
6204 | assert_eq_m512d(r, src); |
6205 | let r = _mm512_mask_cvt_roundps_pd::<_MM_FROUND_CUR_DIRECTION>(src, 0b00001111, a); |
6206 | let e = _mm512_setr_pd(0., -1.5, 2., -3.5, 0., 0., 0., 0.); |
6207 | assert_eq_m512d(r, e); |
6208 | } |
6209 | |
6210 | #[simd_test(enable = "avx512f" )] |
6211 | unsafe fn test_mm512_maskz_cvt_roundps_pd() { |
6212 | let a = _mm256_setr_ps(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6213 | let r = _mm512_maskz_cvt_roundps_pd::<_MM_FROUND_CUR_DIRECTION>(0, a); |
6214 | assert_eq_m512d(r, _mm512_setzero_pd()); |
6215 | let r = _mm512_maskz_cvt_roundps_pd::<_MM_FROUND_CUR_DIRECTION>(0b00001111, a); |
6216 | let e = _mm512_setr_pd(0., -1.5, 2., -3.5, 0., 0., 0., 0.); |
6217 | assert_eq_m512d(r, e); |
6218 | } |
6219 | |
6220 | #[simd_test(enable = "avx512f" )] |
6221 | unsafe fn test_mm512_cvt_roundpd_ps() { |
6222 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6223 | let r = _mm512_cvt_roundpd_ps::<_MM_FROUND_CUR_DIRECTION>(a); |
6224 | let e = _mm256_setr_ps(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6225 | assert_eq_m256(r, e); |
6226 | } |
6227 | |
6228 | #[simd_test(enable = "avx512f" )] |
6229 | unsafe fn test_mm512_mask_cvt_roundpd_ps() { |
6230 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6231 | let src = _mm256_set1_ps(0.); |
6232 | let r = _mm512_mask_cvt_roundpd_ps::<_MM_FROUND_CUR_DIRECTION>(src, 0, a); |
6233 | assert_eq_m256(r, src); |
6234 | let r = _mm512_mask_cvt_roundpd_ps::<_MM_FROUND_CUR_DIRECTION>(src, 0b00001111, a); |
6235 | let e = _mm256_setr_ps(0., -1.5, 2., -3.5, 0., 0., 0., 0.); |
6236 | assert_eq_m256(r, e); |
6237 | } |
6238 | |
6239 | #[simd_test(enable = "avx512f" )] |
6240 | unsafe fn test_mm512_maskz_cvt_roundpd_ps() { |
6241 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6242 | let r = _mm512_maskz_cvt_roundpd_ps::<_MM_FROUND_CUR_DIRECTION>(0, a); |
6243 | assert_eq_m256(r, _mm256_setzero_ps()); |
6244 | let r = _mm512_maskz_cvt_roundpd_ps::<_MM_FROUND_CUR_DIRECTION>(0b00001111, a); |
6245 | let e = _mm256_setr_ps(0., -1.5, 2., -3.5, 0., 0., 0., 0.); |
6246 | assert_eq_m256(r, e); |
6247 | } |
6248 | |
6249 | #[simd_test(enable = "avx512f" )] |
6250 | unsafe fn test_mm512_cvt_roundpd_epi32() { |
6251 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6252 | let r = _mm512_cvt_roundpd_epi32::<_MM_FROUND_CUR_DIRECTION>(a); |
6253 | let e = _mm256_setr_epi32(0, -2, 2, -4, 4, -6, 6, -8); |
6254 | assert_eq_m256i(r, e); |
6255 | } |
6256 | |
6257 | #[simd_test(enable = "avx512f" )] |
6258 | unsafe fn test_mm512_mask_cvt_roundpd_epi32() { |
6259 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6260 | let src = _mm256_set1_epi32(0); |
6261 | let r = _mm512_mask_cvt_roundpd_epi32::<_MM_FROUND_CUR_DIRECTION>(src, 0, a); |
6262 | assert_eq_m256i(r, src); |
6263 | let r = _mm512_mask_cvt_roundpd_epi32::<_MM_FROUND_CUR_DIRECTION>(src, 0b00001111, a); |
6264 | let e = _mm256_setr_epi32(0, -2, 2, -4, 0, 0, 0, 0); |
6265 | assert_eq_m256i(r, e); |
6266 | } |
6267 | |
6268 | #[simd_test(enable = "avx512f" )] |
6269 | unsafe fn test_mm512_maskz_cvt_roundpd_epi32() { |
6270 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6271 | let r = _mm512_maskz_cvt_roundpd_epi32::<_MM_FROUND_CUR_DIRECTION>(0, a); |
6272 | assert_eq_m256i(r, _mm256_setzero_si256()); |
6273 | let r = _mm512_maskz_cvt_roundpd_epi32::<_MM_FROUND_CUR_DIRECTION>(0b00001111, a); |
6274 | let e = _mm256_setr_epi32(0, -2, 2, -4, 0, 0, 0, 0); |
6275 | assert_eq_m256i(r, e); |
6276 | } |
6277 | |
6278 | #[simd_test(enable = "avx512f" )] |
6279 | unsafe fn test_mm512_cvt_roundpd_epu32() { |
6280 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6281 | let r = _mm512_cvt_roundpd_epu32::<_MM_FROUND_CUR_DIRECTION>(a); |
6282 | let e = _mm256_setr_epi32(0, -1, 2, -1, 4, -1, 6, -1); |
6283 | assert_eq_m256i(r, e); |
6284 | } |
6285 | |
6286 | #[simd_test(enable = "avx512f" )] |
6287 | unsafe fn test_mm512_mask_cvt_roundpd_epu32() { |
6288 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6289 | let src = _mm256_set1_epi32(0); |
6290 | let r = _mm512_mask_cvt_roundpd_epu32::<_MM_FROUND_CUR_DIRECTION>(src, 0, a); |
6291 | assert_eq_m256i(r, src); |
6292 | let r = _mm512_mask_cvt_roundpd_epu32::<_MM_FROUND_CUR_DIRECTION>(src, 0b00001111, a); |
6293 | let e = _mm256_setr_epi32(0, -1, 2, -1, 0, 0, 0, 0); |
6294 | assert_eq_m256i(r, e); |
6295 | } |
6296 | |
6297 | #[simd_test(enable = "avx512f" )] |
6298 | unsafe fn test_mm512_maskz_cvt_roundpd_epu32() { |
6299 | let a = _mm512_setr_pd(0., -1.5, 2., -3.5, 4., -5.5, 6., -7.5); |
6300 | let r = _mm512_maskz_cvt_roundpd_epu32::<_MM_FROUND_CUR_DIRECTION>(0, a); |
6301 | assert_eq_m256i(r, _mm256_setzero_si256()); |
6302 | let r = _mm512_maskz_cvt_roundpd_epu32::<_MM_FROUND_CUR_DIRECTION>(0b00001111, a); |
6303 | let e = _mm256_setr_epi32(0, -1, 2, -1, 0, 0, 0, 0); |
6304 | assert_eq_m256i(r, e); |
6305 | } |
6306 | |
6307 | #[simd_test(enable = "avx512f" )] |
6308 | unsafe fn test_mm512_setzero_pd() { |
6309 | assert_eq_m512d(_mm512_setzero_pd(), _mm512_set1_pd(0.)); |
6310 | } |
6311 | |
6312 | unsafe fn test_mm512_set1_epi64() { |
6313 | let r = _mm512_set_epi64(2, 2, 2, 2, 2, 2, 2, 2); |
6314 | assert_eq_m512i(r, _mm512_set1_epi64(2)); |
6315 | } |
6316 | |
6317 | #[simd_test(enable = "avx512f" )] |
6318 | unsafe fn test_mm512_set1_pd() { |
6319 | let expected = _mm512_set_pd(2., 2., 2., 2., 2., 2., 2., 2.); |
6320 | assert_eq_m512d(expected, _mm512_set1_pd(2.)); |
6321 | } |
6322 | |
6323 | unsafe fn test_mm512_set4_epi64() { |
6324 | let r = _mm512_set_epi64(4, 3, 2, 1, 4, 3, 2, 1); |
6325 | assert_eq_m512i(r, _mm512_set4_epi64(4, 3, 2, 1)); |
6326 | } |
6327 | |
6328 | #[simd_test(enable = "avx512f" )] |
6329 | unsafe fn test_mm512_set4_pd() { |
6330 | let r = _mm512_set_pd(4., 3., 2., 1., 4., 3., 2., 1.); |
6331 | assert_eq_m512d(r, _mm512_set4_pd(4., 3., 2., 1.)); |
6332 | } |
6333 | |
6334 | unsafe fn test_mm512_setr4_epi64() { |
6335 | let r = _mm512_set_epi64(4, 3, 2, 1, 4, 3, 2, 1); |
6336 | assert_eq_m512i(r, _mm512_setr4_epi64(1, 2, 3, 4)); |
6337 | } |
6338 | |
6339 | #[simd_test(enable = "avx512f" )] |
6340 | unsafe fn test_mm512_setr4_pd() { |
6341 | let r = _mm512_set_pd(4., 3., 2., 1., 4., 3., 2., 1.); |
6342 | assert_eq_m512d(r, _mm512_setr4_pd(1., 2., 3., 4.)); |
6343 | } |
6344 | |
6345 | #[simd_test(enable = "avx512f" )] |
6346 | unsafe fn test_mm512_cmplt_pd_mask() { |
6347 | #[rustfmt::skip] |
6348 | let a = _mm512_set_pd(0., 1., -1., f64::MAX, f64::NAN, f64::MIN, 100., -100.); |
6349 | let b = _mm512_set1_pd(-1.); |
6350 | let m = _mm512_cmplt_pd_mask(a, b); |
6351 | assert_eq!(m, 0b00000101); |
6352 | } |
6353 | |
6354 | #[simd_test(enable = "avx512f" )] |
6355 | unsafe fn test_mm512_mask_cmplt_pd_mask() { |
6356 | #[rustfmt::skip] |
6357 | let a = _mm512_set_pd(0., 1., -1., f64::MAX, f64::NAN, f64::MIN, 100., -100.); |
6358 | let b = _mm512_set1_pd(-1.); |
6359 | let mask = 0b01100110; |
6360 | let r = _mm512_mask_cmplt_pd_mask(mask, a, b); |
6361 | assert_eq!(r, 0b00000100); |
6362 | } |
6363 | |
6364 | #[simd_test(enable = "avx512f" )] |
6365 | unsafe fn test_mm512_cmpnlt_pd_mask() { |
6366 | #[rustfmt::skip] |
6367 | let a = _mm512_set_pd(0., 1., -1., f64::MAX, f64::NAN, f64::MIN, 100., -100.); |
6368 | let b = _mm512_set1_pd(-1.); |
6369 | assert_eq!(_mm512_cmpnlt_pd_mask(a, b), !_mm512_cmplt_pd_mask(a, b)); |
6370 | } |
6371 | |
6372 | #[simd_test(enable = "avx512f" )] |
6373 | unsafe fn test_mm512_mask_cmpnlt_pd_mask() { |
6374 | #[rustfmt::skip] |
6375 | let a = _mm512_set_pd(0., 1., -1., f64::MAX, f64::NAN, f64::MIN, 100., -100.); |
6376 | let b = _mm512_set1_pd(-1.); |
6377 | let mask = 0b01111010; |
6378 | assert_eq!(_mm512_mask_cmpnlt_pd_mask(mask, a, b), 0b01111010); |
6379 | } |
6380 | |
6381 | #[simd_test(enable = "avx512f" )] |
6382 | unsafe fn test_mm512_cmple_pd_mask() { |
6383 | #[rustfmt::skip] |
6384 | let a = _mm512_set_pd(0., 1., -1., f64::MAX, f64::NAN, f64::MIN, 100., -100.); |
6385 | let b = _mm512_set1_pd(-1.); |
6386 | assert_eq!(_mm512_cmple_pd_mask(a, b), 0b00100101); |
6387 | } |
6388 | |
6389 | #[simd_test(enable = "avx512f" )] |
6390 | unsafe fn test_mm512_mask_cmple_pd_mask() { |
6391 | #[rustfmt::skip] |
6392 | let a = _mm512_set_pd(0., 1., -1., f64::MAX, f64::NAN, f64::MIN, 100., -100.); |
6393 | let b = _mm512_set1_pd(-1.); |
6394 | let mask = 0b01111010; |
6395 | assert_eq!(_mm512_mask_cmple_pd_mask(mask, a, b), 0b00100000); |
6396 | } |
6397 | |
6398 | #[simd_test(enable = "avx512f" )] |
6399 | unsafe fn test_mm512_cmpnle_pd_mask() { |
6400 | #[rustfmt::skip] |
6401 | let a = _mm512_set_pd(0., 1., -1., f64::MAX, f64::NAN, f64::MIN, 100., -100.); |
6402 | let b = _mm512_set1_pd(-1.); |
6403 | let m = _mm512_cmpnle_pd_mask(b, a); |
6404 | assert_eq!(m, 0b00001101); |
6405 | } |
6406 | |
6407 | #[simd_test(enable = "avx512f" )] |
6408 | unsafe fn test_mm512_mask_cmpnle_pd_mask() { |
6409 | #[rustfmt::skip] |
6410 | let a = _mm512_set_pd(0., 1., -1., f64::MAX, f64::NAN, f64::MIN, 100., -100.); |
6411 | let b = _mm512_set1_pd(-1.); |
6412 | let mask = 0b01100110; |
6413 | let r = _mm512_mask_cmpnle_pd_mask(mask, b, a); |
6414 | assert_eq!(r, 0b00000100); |
6415 | } |
6416 | |
6417 | #[simd_test(enable = "avx512f" )] |
6418 | unsafe fn test_mm512_cmpeq_pd_mask() { |
6419 | let a = _mm512_set_pd(0., 1., -1., 13., f64::MAX, f64::MIN, f64::NAN, -100.); |
6420 | let b = _mm512_set_pd(0., 1., 13., 42., f64::MAX, f64::MIN, f64::NAN, -100.); |
6421 | let m = _mm512_cmpeq_pd_mask(b, a); |
6422 | assert_eq!(m, 0b11001101); |
6423 | } |
6424 | |
6425 | #[simd_test(enable = "avx512f" )] |
6426 | unsafe fn test_mm512_mask_cmpeq_pd_mask() { |
6427 | let a = _mm512_set_pd(0., 1., -1., 13., f64::MAX, f64::MIN, f64::NAN, -100.); |
6428 | let b = _mm512_set_pd(0., 1., 13., 42., f64::MAX, f64::MIN, f64::NAN, -100.); |
6429 | let mask = 0b01111010; |
6430 | let r = _mm512_mask_cmpeq_pd_mask(mask, b, a); |
6431 | assert_eq!(r, 0b01001000); |
6432 | } |
6433 | |
6434 | #[simd_test(enable = "avx512f" )] |
6435 | unsafe fn test_mm512_cmpneq_pd_mask() { |
6436 | let a = _mm512_set_pd(0., 1., -1., 13., f64::MAX, f64::MIN, f64::NAN, -100.); |
6437 | let b = _mm512_set_pd(0., 1., 13., 42., f64::MAX, f64::MIN, f64::NAN, -100.); |
6438 | let m = _mm512_cmpneq_pd_mask(b, a); |
6439 | assert_eq!(m, 0b00110010); |
6440 | } |
6441 | |
6442 | #[simd_test(enable = "avx512f" )] |
6443 | unsafe fn test_mm512_mask_cmpneq_pd_mask() { |
6444 | let a = _mm512_set_pd(0., 1., -1., 13., f64::MAX, f64::MIN, f64::NAN, -100.); |
6445 | let b = _mm512_set_pd(0., 1., 13., 42., f64::MAX, f64::MIN, f64::NAN, -100.); |
6446 | let mask = 0b01111010; |
6447 | let r = _mm512_mask_cmpneq_pd_mask(mask, b, a); |
6448 | assert_eq!(r, 0b00110010) |
6449 | } |
6450 | |
6451 | #[simd_test(enable = "avx512f" )] |
6452 | unsafe fn test_mm512_cmp_pd_mask() { |
6453 | let a = _mm512_set_pd(0., 1., -1., 13., f64::MAX, f64::MIN, 100., -100.); |
6454 | let b = _mm512_set1_pd(-1.); |
6455 | let m = _mm512_cmp_pd_mask::<_CMP_LT_OQ>(a, b); |
6456 | assert_eq!(m, 0b00000101); |
6457 | } |
6458 | |
6459 | #[simd_test(enable = "avx512f" )] |
6460 | unsafe fn test_mm512_mask_cmp_pd_mask() { |
6461 | let a = _mm512_set_pd(0., 1., -1., 13., f64::MAX, f64::MIN, 100., -100.); |
6462 | let b = _mm512_set1_pd(-1.); |
6463 | let mask = 0b01100110; |
6464 | let r = _mm512_mask_cmp_pd_mask::<_CMP_LT_OQ>(mask, a, b); |
6465 | assert_eq!(r, 0b00000100); |
6466 | } |
6467 | |
6468 | #[simd_test(enable = "avx512f,avx512vl" )] |
6469 | unsafe fn test_mm256_cmp_pd_mask() { |
6470 | let a = _mm256_set_pd(0., 1., -1., 13.); |
6471 | let b = _mm256_set1_pd(1.); |
6472 | let m = _mm256_cmp_pd_mask::<_CMP_LT_OQ>(a, b); |
6473 | assert_eq!(m, 0b00001010); |
6474 | } |
6475 | |
6476 | #[simd_test(enable = "avx512f,avx512vl" )] |
6477 | unsafe fn test_mm256_mask_cmp_pd_mask() { |
6478 | let a = _mm256_set_pd(0., 1., -1., 13.); |
6479 | let b = _mm256_set1_pd(1.); |
6480 | let mask = 0b11111111; |
6481 | let r = _mm256_mask_cmp_pd_mask::<_CMP_LT_OQ>(mask, a, b); |
6482 | assert_eq!(r, 0b00001010); |
6483 | } |
6484 | |
6485 | #[simd_test(enable = "avx512f,avx512vl" )] |
6486 | unsafe fn test_mm_cmp_pd_mask() { |
6487 | let a = _mm_set_pd(0., 1.); |
6488 | let b = _mm_set1_pd(1.); |
6489 | let m = _mm_cmp_pd_mask::<_CMP_LT_OQ>(a, b); |
6490 | assert_eq!(m, 0b00000010); |
6491 | } |
6492 | |
6493 | #[simd_test(enable = "avx512f,avx512vl" )] |
6494 | unsafe fn test_mm_mask_cmp_pd_mask() { |
6495 | let a = _mm_set_pd(0., 1.); |
6496 | let b = _mm_set1_pd(1.); |
6497 | let mask = 0b11111111; |
6498 | let r = _mm_mask_cmp_pd_mask::<_CMP_LT_OQ>(mask, a, b); |
6499 | assert_eq!(r, 0b00000010); |
6500 | } |
6501 | |
6502 | #[simd_test(enable = "avx512f" )] |
6503 | unsafe fn test_mm512_cmp_round_pd_mask() { |
6504 | #[rustfmt::skip] |
6505 | let a = _mm512_set_pd(0., 1., -1., 13., f64::MAX, f64::MIN, 100., -100.); |
6506 | let b = _mm512_set1_pd(-1.); |
6507 | let m = _mm512_cmp_round_pd_mask::<_CMP_LT_OQ, _MM_FROUND_CUR_DIRECTION>(a, b); |
6508 | assert_eq!(m, 0b00000101); |
6509 | } |
6510 | |
6511 | #[simd_test(enable = "avx512f" )] |
6512 | unsafe fn test_mm512_mask_cmp_round_pd_mask() { |
6513 | #[rustfmt::skip] |
6514 | let a = _mm512_set_pd(0., 1., -1., 13., f64::MAX, f64::MIN, 100., -100.); |
6515 | let b = _mm512_set1_pd(-1.); |
6516 | let mask = 0b01100110; |
6517 | let r = _mm512_mask_cmp_round_pd_mask::<_CMP_LT_OQ, _MM_FROUND_CUR_DIRECTION>(mask, a, b); |
6518 | assert_eq!(r, 0b00000100); |
6519 | } |
6520 | |
6521 | #[simd_test(enable = "avx512f" )] |
6522 | unsafe fn test_mm512_cmpord_pd_mask() { |
6523 | #[rustfmt::skip] |
6524 | let a = _mm512_set_pd(f64::NAN, f64::MAX, f64::NAN, f64::MIN, f64::NAN, -1., f64::NAN, 0.); |
6525 | #[rustfmt::skip] |
6526 | let b = _mm512_set_pd(f64::NAN, f64::NAN, f64::NAN, f64::NAN, f64::MIN, f64::MAX, -1., 0.); |
6527 | let m = _mm512_cmpord_pd_mask(a, b); |
6528 | assert_eq!(m, 0b00000101); |
6529 | } |
6530 | |
6531 | #[simd_test(enable = "avx512f" )] |
6532 | unsafe fn test_mm512_mask_cmpord_pd_mask() { |
6533 | #[rustfmt::skip] |
6534 | let a = _mm512_set_pd(f64::NAN, f64::MAX, f64::NAN, f64::MIN, f64::NAN, -1., f64::NAN, 0.); |
6535 | #[rustfmt::skip] |
6536 | let b = _mm512_set_pd(f64::NAN, f64::NAN, f64::NAN, f64::NAN, f64::MIN, f64::MAX, -1., 0.); |
6537 | let mask = 0b11000011; |
6538 | let m = _mm512_mask_cmpord_pd_mask(mask, a, b); |
6539 | assert_eq!(m, 0b00000001); |
6540 | } |
6541 | |
6542 | #[simd_test(enable = "avx512f" )] |
6543 | unsafe fn test_mm512_cmpunord_pd_mask() { |
6544 | #[rustfmt::skip] |
6545 | let a = _mm512_set_pd(f64::NAN, f64::MAX, f64::NAN, f64::MIN, f64::NAN, -1., f64::NAN, 0.); |
6546 | #[rustfmt::skip] |
6547 | let b = _mm512_set_pd(f64::NAN, f64::NAN, f64::NAN, f64::NAN, f64::MIN, f64::MAX, -1., 0.); |
6548 | let m = _mm512_cmpunord_pd_mask(a, b); |
6549 | |
6550 | assert_eq!(m, 0b11111010); |
6551 | } |
6552 | |
6553 | #[simd_test(enable = "avx512f" )] |
6554 | unsafe fn test_mm512_mask_cmpunord_pd_mask() { |
6555 | #[rustfmt::skip] |
6556 | let a = _mm512_set_pd(f64::NAN, f64::MAX, f64::NAN, f64::MIN, f64::NAN, -1., f64::NAN, 0.); |
6557 | #[rustfmt::skip] |
6558 | let b = _mm512_set_pd(f64::NAN, f64::NAN, f64::NAN, f64::NAN, f64::MIN, f64::MAX, -1., 0.); |
6559 | let mask = 0b00001111; |
6560 | let m = _mm512_mask_cmpunord_pd_mask(mask, a, b); |
6561 | assert_eq!(m, 0b000001010); |
6562 | } |
6563 | |
6564 | #[simd_test(enable = "avx512f" )] |
6565 | unsafe fn test_mm512_cmplt_epu64_mask() { |
6566 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6567 | let b = _mm512_set1_epi64(-1); |
6568 | let m = _mm512_cmplt_epu64_mask(a, b); |
6569 | assert_eq!(m, 0b11001111); |
6570 | } |
6571 | |
6572 | #[simd_test(enable = "avx512f" )] |
6573 | unsafe fn test_mm512_mask_cmplt_epu64_mask() { |
6574 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6575 | let b = _mm512_set1_epi64(-1); |
6576 | let mask = 0b01111010; |
6577 | let r = _mm512_mask_cmplt_epu64_mask(mask, a, b); |
6578 | assert_eq!(r, 0b01001010); |
6579 | } |
6580 | |
6581 | #[simd_test(enable = "avx512f,avx512vl" )] |
6582 | unsafe fn test_mm256_cmplt_epu64_mask() { |
6583 | let a = _mm256_set_epi64x(0, 1, 2, 100); |
6584 | let b = _mm256_set1_epi64x(2); |
6585 | let r = _mm256_cmplt_epu64_mask(a, b); |
6586 | assert_eq!(r, 0b00001100); |
6587 | } |
6588 | |
6589 | #[simd_test(enable = "avx512f,avx512vl" )] |
6590 | unsafe fn test_mm256_mask_cmplt_epu64_mask() { |
6591 | let a = _mm256_set_epi64x(0, 1, 2, 100); |
6592 | let b = _mm256_set1_epi64x(2); |
6593 | let mask = 0b11111111; |
6594 | let r = _mm256_mask_cmplt_epu64_mask(mask, a, b); |
6595 | assert_eq!(r, 0b00001100); |
6596 | } |
6597 | |
6598 | #[simd_test(enable = "avx512f,avx512vl" )] |
6599 | unsafe fn test_mm_cmplt_epu64_mask() { |
6600 | let a = _mm_set_epi64x(0, 1); |
6601 | let b = _mm_set1_epi64x(2); |
6602 | let r = _mm_cmplt_epu64_mask(a, b); |
6603 | assert_eq!(r, 0b00000011); |
6604 | } |
6605 | |
6606 | #[simd_test(enable = "avx512f,avx512vl" )] |
6607 | unsafe fn test_mm_mask_cmplt_epu64_mask() { |
6608 | let a = _mm_set_epi64x(0, 1); |
6609 | let b = _mm_set1_epi64x(2); |
6610 | let mask = 0b11111111; |
6611 | let r = _mm_mask_cmplt_epu64_mask(mask, a, b); |
6612 | assert_eq!(r, 0b00000011); |
6613 | } |
6614 | |
6615 | #[simd_test(enable = "avx512f" )] |
6616 | unsafe fn test_mm512_cmpgt_epu64_mask() { |
6617 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6618 | let b = _mm512_set1_epi64(-1); |
6619 | let m = _mm512_cmpgt_epu64_mask(b, a); |
6620 | assert_eq!(m, 0b11001111); |
6621 | } |
6622 | |
6623 | #[simd_test(enable = "avx512f" )] |
6624 | unsafe fn test_mm512_mask_cmpgt_epu64_mask() { |
6625 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6626 | let b = _mm512_set1_epi64(-1); |
6627 | let mask = 0b01111010; |
6628 | let r = _mm512_mask_cmpgt_epu64_mask(mask, b, a); |
6629 | assert_eq!(r, 0b01001010); |
6630 | } |
6631 | |
6632 | #[simd_test(enable = "avx512f,avx512vl" )] |
6633 | unsafe fn test_mm256_cmpgt_epu64_mask() { |
6634 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
6635 | let b = _mm256_set1_epi64x(1); |
6636 | let r = _mm256_cmpgt_epu64_mask(a, b); |
6637 | assert_eq!(r, 0b00000011); |
6638 | } |
6639 | |
6640 | #[simd_test(enable = "avx512f,avx512vl" )] |
6641 | unsafe fn test_mm256_mask_cmpgt_epu64_mask() { |
6642 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
6643 | let b = _mm256_set1_epi64x(1); |
6644 | let mask = 0b11111111; |
6645 | let r = _mm256_mask_cmpgt_epu64_mask(mask, a, b); |
6646 | assert_eq!(r, 0b00000011); |
6647 | } |
6648 | |
6649 | #[simd_test(enable = "avx512f,avx512vl" )] |
6650 | unsafe fn test_mm_cmpgt_epu64_mask() { |
6651 | let a = _mm_set_epi64x(1, 2); |
6652 | let b = _mm_set1_epi64x(1); |
6653 | let r = _mm_cmpgt_epu64_mask(a, b); |
6654 | assert_eq!(r, 0b00000001); |
6655 | } |
6656 | |
6657 | #[simd_test(enable = "avx512f,avx512vl" )] |
6658 | unsafe fn test_mm_mask_cmpgt_epu64_mask() { |
6659 | let a = _mm_set_epi64x(1, 2); |
6660 | let b = _mm_set1_epi64x(1); |
6661 | let mask = 0b11111111; |
6662 | let r = _mm_mask_cmpgt_epu64_mask(mask, a, b); |
6663 | assert_eq!(r, 0b00000001); |
6664 | } |
6665 | |
6666 | #[simd_test(enable = "avx512f" )] |
6667 | unsafe fn test_mm512_cmple_epu64_mask() { |
6668 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6669 | let b = _mm512_set1_epi64(-1); |
6670 | assert_eq!( |
6671 | _mm512_cmple_epu64_mask(a, b), |
6672 | !_mm512_cmpgt_epu64_mask(a, b) |
6673 | ) |
6674 | } |
6675 | |
6676 | #[simd_test(enable = "avx512f" )] |
6677 | unsafe fn test_mm512_mask_cmple_epu64_mask() { |
6678 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6679 | let b = _mm512_set1_epi64(-1); |
6680 | let mask = 0b01111010; |
6681 | assert_eq!(_mm512_mask_cmple_epu64_mask(mask, a, b), 0b01111010); |
6682 | } |
6683 | |
6684 | #[simd_test(enable = "avx512f,avx512vl" )] |
6685 | unsafe fn test_mm256_cmple_epu64_mask() { |
6686 | let a = _mm256_set_epi64x(0, 1, 2, 1); |
6687 | let b = _mm256_set1_epi64x(1); |
6688 | let r = _mm256_cmple_epu64_mask(a, b); |
6689 | assert_eq!(r, 0b00001101) |
6690 | } |
6691 | |
6692 | #[simd_test(enable = "avx512f,avx512vl" )] |
6693 | unsafe fn test_mm256_mask_cmple_epu64_mask() { |
6694 | let a = _mm256_set_epi64x(0, 1, 2, 1); |
6695 | let b = _mm256_set1_epi64x(1); |
6696 | let mask = 0b11111111; |
6697 | let r = _mm256_mask_cmple_epu64_mask(mask, a, b); |
6698 | assert_eq!(r, 0b00001101) |
6699 | } |
6700 | |
6701 | #[simd_test(enable = "avx512f,avx512vl" )] |
6702 | unsafe fn test_mm_cmple_epu64_mask() { |
6703 | let a = _mm_set_epi64x(0, 1); |
6704 | let b = _mm_set1_epi64x(1); |
6705 | let r = _mm_cmple_epu64_mask(a, b); |
6706 | assert_eq!(r, 0b00000011) |
6707 | } |
6708 | |
6709 | #[simd_test(enable = "avx512f,avx512vl" )] |
6710 | unsafe fn test_mm_mask_cmple_epu64_mask() { |
6711 | let a = _mm_set_epi64x(0, 1); |
6712 | let b = _mm_set1_epi64x(1); |
6713 | let mask = 0b11111111; |
6714 | let r = _mm_mask_cmple_epu64_mask(mask, a, b); |
6715 | assert_eq!(r, 0b00000011) |
6716 | } |
6717 | |
6718 | #[simd_test(enable = "avx512f" )] |
6719 | unsafe fn test_mm512_cmpge_epu64_mask() { |
6720 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6721 | let b = _mm512_set1_epi64(-1); |
6722 | assert_eq!( |
6723 | _mm512_cmpge_epu64_mask(a, b), |
6724 | !_mm512_cmplt_epu64_mask(a, b) |
6725 | ); |
6726 | } |
6727 | |
6728 | #[simd_test(enable = "avx512f" )] |
6729 | unsafe fn test_mm512_mask_cmpge_epu64_mask() { |
6730 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6731 | let b = _mm512_set1_epi64(-1); |
6732 | let mask = 0b11111111; |
6733 | let r = _mm512_mask_cmpge_epu64_mask(mask, a, b); |
6734 | assert_eq!(r, 0b00110000); |
6735 | } |
6736 | |
6737 | #[simd_test(enable = "avx512f,avx512vl" )] |
6738 | unsafe fn test_mm256_cmpge_epu64_mask() { |
6739 | let a = _mm256_set_epi64x(0, 1, 2, u64::MAX as i64); |
6740 | let b = _mm256_set1_epi64x(1); |
6741 | let r = _mm256_cmpge_epu64_mask(a, b); |
6742 | assert_eq!(r, 0b00000111); |
6743 | } |
6744 | |
6745 | #[simd_test(enable = "avx512f,avx512vl" )] |
6746 | unsafe fn test_mm256_mask_cmpge_epu64_mask() { |
6747 | let a = _mm256_set_epi64x(0, 1, 2, u64::MAX as i64); |
6748 | let b = _mm256_set1_epi64x(1); |
6749 | let mask = 0b11111111; |
6750 | let r = _mm256_mask_cmpge_epu64_mask(mask, a, b); |
6751 | assert_eq!(r, 0b00000111); |
6752 | } |
6753 | |
6754 | #[simd_test(enable = "avx512f,avx512vl" )] |
6755 | unsafe fn test_mm_cmpge_epu64_mask() { |
6756 | let a = _mm_set_epi64x(0, 1); |
6757 | let b = _mm_set1_epi64x(1); |
6758 | let r = _mm_cmpge_epu64_mask(a, b); |
6759 | assert_eq!(r, 0b00000001); |
6760 | } |
6761 | |
6762 | #[simd_test(enable = "avx512f,avx512vl" )] |
6763 | unsafe fn test_mm_mask_cmpge_epu64_mask() { |
6764 | let a = _mm_set_epi64x(0, 1); |
6765 | let b = _mm_set1_epi64x(1); |
6766 | let mask = 0b11111111; |
6767 | let r = _mm_mask_cmpge_epu64_mask(mask, a, b); |
6768 | assert_eq!(r, 0b00000001); |
6769 | } |
6770 | |
6771 | #[simd_test(enable = "avx512f" )] |
6772 | unsafe fn test_mm512_cmpeq_epu64_mask() { |
6773 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6774 | let b = _mm512_set_epi64(0, 1, 13, 42, i64::MAX, i64::MIN, 100, -100); |
6775 | let m = _mm512_cmpeq_epu64_mask(b, a); |
6776 | assert_eq!(m, 0b11001111); |
6777 | } |
6778 | |
6779 | #[simd_test(enable = "avx512f" )] |
6780 | unsafe fn test_mm512_mask_cmpeq_epu64_mask() { |
6781 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6782 | let b = _mm512_set_epi64(0, 1, 13, 42, i64::MAX, i64::MIN, 100, -100); |
6783 | let mask = 0b01111010; |
6784 | let r = _mm512_mask_cmpeq_epu64_mask(mask, b, a); |
6785 | assert_eq!(r, 0b01001010); |
6786 | } |
6787 | |
6788 | #[simd_test(enable = "avx512f,avx512vl" )] |
6789 | unsafe fn test_mm256_cmpeq_epu64_mask() { |
6790 | let a = _mm256_set_epi64x(0, 1, -1, u64::MAX as i64); |
6791 | let b = _mm256_set_epi64x(0, 1, 13, 42); |
6792 | let m = _mm256_cmpeq_epu64_mask(b, a); |
6793 | assert_eq!(m, 0b00001100); |
6794 | } |
6795 | |
6796 | #[simd_test(enable = "avx512f,avx512vl" )] |
6797 | unsafe fn test_mm256_mask_cmpeq_epu64_mask() { |
6798 | let a = _mm256_set_epi64x(0, 1, -1, u64::MAX as i64); |
6799 | let b = _mm256_set_epi64x(0, 1, 13, 42); |
6800 | let mask = 0b11111111; |
6801 | let r = _mm256_mask_cmpeq_epu64_mask(mask, b, a); |
6802 | assert_eq!(r, 0b00001100); |
6803 | } |
6804 | |
6805 | #[simd_test(enable = "avx512f,avx512vl" )] |
6806 | unsafe fn test_mm_cmpeq_epu64_mask() { |
6807 | let a = _mm_set_epi64x(0, 1); |
6808 | let b = _mm_set_epi64x(0, 1); |
6809 | let m = _mm_cmpeq_epu64_mask(b, a); |
6810 | assert_eq!(m, 0b00000011); |
6811 | } |
6812 | |
6813 | #[simd_test(enable = "avx512f,avx512vl" )] |
6814 | unsafe fn test_mm_mask_cmpeq_epu64_mask() { |
6815 | let a = _mm_set_epi64x(0, 1); |
6816 | let b = _mm_set_epi64x(0, 1); |
6817 | let mask = 0b11111111; |
6818 | let r = _mm_mask_cmpeq_epu64_mask(mask, b, a); |
6819 | assert_eq!(r, 0b00000011); |
6820 | } |
6821 | |
6822 | #[simd_test(enable = "avx512f" )] |
6823 | unsafe fn test_mm512_cmpneq_epu64_mask() { |
6824 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6825 | let b = _mm512_set_epi64(0, 1, 13, 42, i64::MAX, i64::MIN, 100, -100); |
6826 | let m = _mm512_cmpneq_epu64_mask(b, a); |
6827 | assert_eq!(m, !_mm512_cmpeq_epu64_mask(b, a)); |
6828 | } |
6829 | |
6830 | #[simd_test(enable = "avx512f" )] |
6831 | unsafe fn test_mm512_mask_cmpneq_epu64_mask() { |
6832 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, -100, 100); |
6833 | let b = _mm512_set_epi64(0, 1, 13, 42, i64::MAX, i64::MIN, 100, -100); |
6834 | let mask = 0b01111010; |
6835 | let r = _mm512_mask_cmpneq_epu64_mask(mask, b, a); |
6836 | assert_eq!(r, 0b00110010); |
6837 | } |
6838 | |
6839 | #[simd_test(enable = "avx512f,avx512vl" )] |
6840 | unsafe fn test_mm256_cmpneq_epu64_mask() { |
6841 | let a = _mm256_set_epi64x(0, 1, -1, u64::MAX as i64); |
6842 | let b = _mm256_set_epi64x(0, 1, 13, 42); |
6843 | let r = _mm256_cmpneq_epu64_mask(b, a); |
6844 | assert_eq!(r, 0b00000011); |
6845 | } |
6846 | |
6847 | #[simd_test(enable = "avx512f,avx512vl" )] |
6848 | unsafe fn test_mm256_mask_cmpneq_epu64_mask() { |
6849 | let a = _mm256_set_epi64x(0, 1, -1, u64::MAX as i64); |
6850 | let b = _mm256_set_epi64x(0, 1, 13, 42); |
6851 | let mask = 0b11111111; |
6852 | let r = _mm256_mask_cmpneq_epu64_mask(mask, b, a); |
6853 | assert_eq!(r, 0b00000011); |
6854 | } |
6855 | |
6856 | #[simd_test(enable = "avx512f,avx512vl" )] |
6857 | unsafe fn test_mm_cmpneq_epu64_mask() { |
6858 | let a = _mm_set_epi64x(-1, u64::MAX as i64); |
6859 | let b = _mm_set_epi64x(13, 42); |
6860 | let r = _mm_cmpneq_epu64_mask(b, a); |
6861 | assert_eq!(r, 0b00000011); |
6862 | } |
6863 | |
6864 | #[simd_test(enable = "avx512f,avx512vl" )] |
6865 | unsafe fn test_mm_mask_cmpneq_epu64_mask() { |
6866 | let a = _mm_set_epi64x(-1, u64::MAX as i64); |
6867 | let b = _mm_set_epi64x(13, 42); |
6868 | let mask = 0b11111111; |
6869 | let r = _mm_mask_cmpneq_epu64_mask(mask, b, a); |
6870 | assert_eq!(r, 0b00000011); |
6871 | } |
6872 | |
6873 | #[simd_test(enable = "avx512f" )] |
6874 | unsafe fn test_mm512_cmp_epu64_mask() { |
6875 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6876 | let b = _mm512_set1_epi64(-1); |
6877 | let m = _mm512_cmp_epu64_mask::<_MM_CMPINT_LT>(a, b); |
6878 | assert_eq!(m, 0b11001111); |
6879 | } |
6880 | |
6881 | #[simd_test(enable = "avx512f" )] |
6882 | unsafe fn test_mm512_mask_cmp_epu64_mask() { |
6883 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
6884 | let b = _mm512_set1_epi64(-1); |
6885 | let mask = 0b01111010; |
6886 | let r = _mm512_mask_cmp_epu64_mask::<_MM_CMPINT_LT>(mask, a, b); |
6887 | assert_eq!(r, 0b01001010); |
6888 | } |
6889 | |
6890 | #[simd_test(enable = "avx512f,avx512vl" )] |
6891 | unsafe fn test_mm256_cmp_epu64_mask() { |
6892 | let a = _mm256_set_epi64x(0, 1, -1, 100); |
6893 | let b = _mm256_set1_epi64x(1); |
6894 | let m = _mm256_cmp_epu64_mask::<_MM_CMPINT_LT>(a, b); |
6895 | assert_eq!(m, 0b00001000); |
6896 | } |
6897 | |
6898 | #[simd_test(enable = "avx512f,avx512vl" )] |
6899 | unsafe fn test_mm256_mask_cmp_epu64_mask() { |
6900 | let a = _mm256_set_epi64x(0, 1, -1, 100); |
6901 | let b = _mm256_set1_epi64x(1); |
6902 | let mask = 0b11111111; |
6903 | let r = _mm256_mask_cmp_epu64_mask::<_MM_CMPINT_LT>(mask, a, b); |
6904 | assert_eq!(r, 0b00001000); |
6905 | } |
6906 | |
6907 | #[simd_test(enable = "avx512f,avx512vl" )] |
6908 | unsafe fn test_mm_cmp_epu64_mask() { |
6909 | let a = _mm_set_epi64x(0, 1); |
6910 | let b = _mm_set1_epi64x(1); |
6911 | let m = _mm_cmp_epu64_mask::<_MM_CMPINT_LT>(a, b); |
6912 | assert_eq!(m, 0b00000010); |
6913 | } |
6914 | |
6915 | #[simd_test(enable = "avx512f,avx512vl" )] |
6916 | unsafe fn test_mm_mask_cmp_epu64_mask() { |
6917 | let a = _mm_set_epi64x(0, 1); |
6918 | let b = _mm_set1_epi64x(1); |
6919 | let mask = 0b11111111; |
6920 | let r = _mm_mask_cmp_epu64_mask::<_MM_CMPINT_LT>(mask, a, b); |
6921 | assert_eq!(r, 0b00000010); |
6922 | } |
6923 | |
6924 | #[simd_test(enable = "avx512f" )] |
6925 | unsafe fn test_mm512_cmplt_epi64_mask() { |
6926 | let a = _mm512_set_epi64(0, 1, -1, 13, i64::MAX, i64::MIN, 100, -100); |
6927 | let b = _mm512_set1_epi64(-1); |
6928 | let m = _mm512_cmplt_epi64_mask(a, b); |
6929 | assert_eq!(m, 0b00000101); |
6930 | } |
6931 | |
6932 | #[simd_test(enable = "avx512f" )] |
6933 | unsafe fn test_mm512_mask_cmplt_epi64_mask() { |
6934 | let a = _mm512_set_epi64(0, 1, -1, 13, i64::MAX, i64::MIN, 100, -100); |
6935 | let b = _mm512_set1_epi64(-1); |
6936 | let mask = 0b01100110; |
6937 | let r = _mm512_mask_cmplt_epi64_mask(mask, a, b); |
6938 | assert_eq!(r, 0b00000100); |
6939 | } |
6940 | |
6941 | #[simd_test(enable = "avx512f,avx512vl" )] |
6942 | unsafe fn test_mm256_cmplt_epi64_mask() { |
6943 | let a = _mm256_set_epi64x(0, 1, -1, -13); |
6944 | let b = _mm256_set1_epi64x(-1); |
6945 | let r = _mm256_cmplt_epi64_mask(a, b); |
6946 | assert_eq!(r, 0b00000001); |
6947 | } |
6948 | |
6949 | #[simd_test(enable = "avx512f,avx512vl" )] |
6950 | unsafe fn test_mm256_mask_cmplt_epi64_mask() { |
6951 | let a = _mm256_set_epi64x(0, 1, -1, -13); |
6952 | let b = _mm256_set1_epi64x(-1); |
6953 | let mask = 0b11111111; |
6954 | let r = _mm256_mask_cmplt_epi64_mask(mask, a, b); |
6955 | assert_eq!(r, 0b00000001); |
6956 | } |
6957 | |
6958 | #[simd_test(enable = "avx512f,avx512vl" )] |
6959 | unsafe fn test_mm_cmplt_epi64_mask() { |
6960 | let a = _mm_set_epi64x(-1, -13); |
6961 | let b = _mm_set1_epi64x(-1); |
6962 | let r = _mm_cmplt_epi64_mask(a, b); |
6963 | assert_eq!(r, 0b00000001); |
6964 | } |
6965 | |
6966 | #[simd_test(enable = "avx512f,avx512vl" )] |
6967 | unsafe fn test_mm_mask_cmplt_epi64_mask() { |
6968 | let a = _mm_set_epi64x(-1, -13); |
6969 | let b = _mm_set1_epi64x(-1); |
6970 | let mask = 0b11111111; |
6971 | let r = _mm_mask_cmplt_epi64_mask(mask, a, b); |
6972 | assert_eq!(r, 0b00000001); |
6973 | } |
6974 | |
6975 | #[simd_test(enable = "avx512f" )] |
6976 | unsafe fn test_mm512_cmpgt_epi64_mask() { |
6977 | let a = _mm512_set_epi64(0, 1, -1, 13, i64::MAX, i64::MIN, 100, -100); |
6978 | let b = _mm512_set1_epi64(-1); |
6979 | let m = _mm512_cmpgt_epi64_mask(b, a); |
6980 | assert_eq!(m, 0b00000101); |
6981 | } |
6982 | |
6983 | #[simd_test(enable = "avx512f" )] |
6984 | unsafe fn test_mm512_mask_cmpgt_epi64_mask() { |
6985 | let a = _mm512_set_epi64(0, 1, -1, 13, i64::MAX, i64::MIN, 100, -100); |
6986 | let b = _mm512_set1_epi64(-1); |
6987 | let mask = 0b01100110; |
6988 | let r = _mm512_mask_cmpgt_epi64_mask(mask, b, a); |
6989 | assert_eq!(r, 0b00000100); |
6990 | } |
6991 | |
6992 | #[simd_test(enable = "avx512f,avx512vl" )] |
6993 | unsafe fn test_mm256_cmpgt_epi64_mask() { |
6994 | let a = _mm256_set_epi64x(0, 1, -1, 13); |
6995 | let b = _mm256_set1_epi64x(-1); |
6996 | let r = _mm256_cmpgt_epi64_mask(a, b); |
6997 | assert_eq!(r, 0b00001101); |
6998 | } |
6999 | |
7000 | #[simd_test(enable = "avx512f,avx512vl" )] |
7001 | unsafe fn test_mm256_mask_cmpgt_epi64_mask() { |
7002 | let a = _mm256_set_epi64x(0, 1, -1, 13); |
7003 | let b = _mm256_set1_epi64x(-1); |
7004 | let mask = 0b11111111; |
7005 | let r = _mm256_mask_cmpgt_epi64_mask(mask, a, b); |
7006 | assert_eq!(r, 0b00001101); |
7007 | } |
7008 | |
7009 | #[simd_test(enable = "avx512f,avx512vl" )] |
7010 | unsafe fn test_mm_cmpgt_epi64_mask() { |
7011 | let a = _mm_set_epi64x(0, -1); |
7012 | let b = _mm_set1_epi64x(-1); |
7013 | let r = _mm_cmpgt_epi64_mask(a, b); |
7014 | assert_eq!(r, 0b00000010); |
7015 | } |
7016 | |
7017 | #[simd_test(enable = "avx512f,avx512vl" )] |
7018 | unsafe fn test_mm_mask_cmpgt_epi64_mask() { |
7019 | let a = _mm_set_epi64x(0, -1); |
7020 | let b = _mm_set1_epi64x(-1); |
7021 | let mask = 0b11111111; |
7022 | let r = _mm_mask_cmpgt_epi64_mask(mask, a, b); |
7023 | assert_eq!(r, 0b00000010); |
7024 | } |
7025 | |
7026 | #[simd_test(enable = "avx512f" )] |
7027 | unsafe fn test_mm512_cmple_epi64_mask() { |
7028 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
7029 | let b = _mm512_set1_epi64(-1); |
7030 | assert_eq!( |
7031 | _mm512_cmple_epi64_mask(a, b), |
7032 | !_mm512_cmpgt_epi64_mask(a, b) |
7033 | ) |
7034 | } |
7035 | |
7036 | #[simd_test(enable = "avx512f" )] |
7037 | unsafe fn test_mm512_mask_cmple_epi64_mask() { |
7038 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
7039 | let b = _mm512_set1_epi64(-1); |
7040 | let mask = 0b01111010; |
7041 | assert_eq!(_mm512_mask_cmple_epi64_mask(mask, a, b), 0b00110000); |
7042 | } |
7043 | |
7044 | #[simd_test(enable = "avx512f,avx512vl" )] |
7045 | unsafe fn test_mm256_cmple_epi64_mask() { |
7046 | let a = _mm256_set_epi64x(0, 1, -1, i64::MAX); |
7047 | let b = _mm256_set1_epi64x(-1); |
7048 | let r = _mm256_cmple_epi64_mask(a, b); |
7049 | assert_eq!(r, 0b00000010) |
7050 | } |
7051 | |
7052 | #[simd_test(enable = "avx512f,avx512vl" )] |
7053 | unsafe fn test_mm256_mask_cmple_epi64_mask() { |
7054 | let a = _mm256_set_epi64x(0, 1, -1, i64::MAX); |
7055 | let b = _mm256_set1_epi64x(-1); |
7056 | let mask = 0b11111111; |
7057 | let r = _mm256_mask_cmple_epi64_mask(mask, a, b); |
7058 | assert_eq!(r, 0b00000010) |
7059 | } |
7060 | |
7061 | #[simd_test(enable = "avx512f,avx512vl" )] |
7062 | unsafe fn test_mm_cmple_epi64_mask() { |
7063 | let a = _mm_set_epi64x(0, 1); |
7064 | let b = _mm_set1_epi64x(1); |
7065 | let r = _mm_cmple_epi64_mask(a, b); |
7066 | assert_eq!(r, 0b00000011) |
7067 | } |
7068 | |
7069 | #[simd_test(enable = "avx512f,avx512vl" )] |
7070 | unsafe fn test_mm_mask_cmple_epi64_mask() { |
7071 | let a = _mm_set_epi64x(0, 1); |
7072 | let b = _mm_set1_epi64x(1); |
7073 | let mask = 0b11111111; |
7074 | let r = _mm_mask_cmple_epi64_mask(mask, a, b); |
7075 | assert_eq!(r, 0b00000011) |
7076 | } |
7077 | |
7078 | #[simd_test(enable = "avx512f" )] |
7079 | unsafe fn test_mm512_cmpge_epi64_mask() { |
7080 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
7081 | let b = _mm512_set1_epi64(-1); |
7082 | assert_eq!( |
7083 | _mm512_cmpge_epi64_mask(a, b), |
7084 | !_mm512_cmplt_epi64_mask(a, b) |
7085 | ) |
7086 | } |
7087 | |
7088 | #[simd_test(enable = "avx512f" )] |
7089 | unsafe fn test_mm512_mask_cmpge_epi64_mask() { |
7090 | let a = _mm512_set_epi64(0, 1, -1, u64::MAX as i64, i64::MAX, i64::MIN, 100, -100); |
7091 | let b = _mm512_set1_epi64(-1); |
7092 | let mask = 0b11111111; |
7093 | let r = _mm512_mask_cmpge_epi64_mask(mask, a, b); |
7094 | assert_eq!(r, 0b11111010); |
7095 | } |
7096 | |
7097 | #[simd_test(enable = "avx512f,avx512vl" )] |
7098 | unsafe fn test_mm256_cmpge_epi64_mask() { |
7099 | let a = _mm256_set_epi64x(0, 1, -1, i64::MAX); |
7100 | let b = _mm256_set1_epi64x(-1); |
7101 | let r = _mm256_cmpge_epi64_mask(a, b); |
7102 | assert_eq!(r, 0b00001111); |
7103 | } |
7104 | |
7105 | #[simd_test(enable = "avx512f,avx512vl" )] |
7106 | unsafe fn test_mm256_mask_cmpge_epi64_mask() { |
7107 | let a = _mm256_set_epi64x(0, 1, -1, i64::MAX); |
7108 | let b = _mm256_set1_epi64x(-1); |
7109 | let mask = 0b11111111; |
7110 | let r = _mm256_mask_cmpge_epi64_mask(mask, a, b); |
7111 | assert_eq!(r, 0b00001111); |
7112 | } |
7113 | |
7114 | #[simd_test(enable = "avx512f,avx512vl" )] |
7115 | unsafe fn test_mm_cmpge_epi64_mask() { |
7116 | let a = _mm_set_epi64x(0, 1); |
7117 | let b = _mm_set1_epi64x(-1); |
7118 | let r = _mm_cmpge_epi64_mask(a, b); |
7119 | assert_eq!(r, 0b00000011); |
7120 | } |
7121 | |
7122 | #[simd_test(enable = "avx512f,avx512vl" )] |
7123 | unsafe fn test_mm_mask_cmpge_epi64_mask() { |
7124 | let a = _mm_set_epi64x(0, 1); |
7125 | let b = _mm_set1_epi64x(-1); |
7126 | let mask = 0b11111111; |
7127 | let r = _mm_mask_cmpge_epi64_mask(mask, a, b); |
7128 | assert_eq!(r, 0b00000011); |
7129 | } |
7130 | |
7131 | #[simd_test(enable = "avx512f" )] |
7132 | unsafe fn test_mm512_cmpeq_epi64_mask() { |
7133 | let a = _mm512_set_epi64(0, 1, -1, 13, i64::MAX, i64::MIN, 100, -100); |
7134 | let b = _mm512_set_epi64(0, 1, 13, 42, i64::MAX, i64::MIN, 100, -100); |
7135 | let m = _mm512_cmpeq_epi64_mask(b, a); |
7136 | assert_eq!(m, 0b11001111); |
7137 | } |
7138 | |
7139 | #[simd_test(enable = "avx512f" )] |
7140 | unsafe fn test_mm512_mask_cmpeq_epi64_mask() { |
7141 | let a = _mm512_set_epi64(0, 1, -1, 13, i64::MAX, i64::MIN, 100, -100); |
7142 | let b = _mm512_set_epi64(0, 1, 13, 42, i64::MAX, i64::MIN, 100, -100); |
7143 | let mask = 0b01111010; |
7144 | let r = _mm512_mask_cmpeq_epi64_mask(mask, b, a); |
7145 | assert_eq!(r, 0b01001010); |
7146 | } |
7147 | |
7148 | #[simd_test(enable = "avx512f,avx512vl" )] |
7149 | unsafe fn test_mm256_cmpeq_epi64_mask() { |
7150 | let a = _mm256_set_epi64x(0, 1, -1, 13); |
7151 | let b = _mm256_set_epi64x(0, 1, 13, 42); |
7152 | let m = _mm256_cmpeq_epi64_mask(b, a); |
7153 | assert_eq!(m, 0b00001100); |
7154 | } |
7155 | |
7156 | #[simd_test(enable = "avx512f,avx512vl" )] |
7157 | unsafe fn test_mm256_mask_cmpeq_epi64_mask() { |
7158 | let a = _mm256_set_epi64x(0, 1, -1, 13); |
7159 | let b = _mm256_set_epi64x(0, 1, 13, 42); |
7160 | let mask = 0b11111111; |
7161 | let r = _mm256_mask_cmpeq_epi64_mask(mask, b, a); |
7162 | assert_eq!(r, 0b00001100); |
7163 | } |
7164 | |
7165 | #[simd_test(enable = "avx512f,avx512vl" )] |
7166 | unsafe fn test_mm_cmpeq_epi64_mask() { |
7167 | let a = _mm_set_epi64x(0, 1); |
7168 | let b = _mm_set_epi64x(0, 1); |
7169 | let m = _mm_cmpeq_epi64_mask(b, a); |
7170 | assert_eq!(m, 0b00000011); |
7171 | } |
7172 | |
7173 | #[simd_test(enable = "avx512f,avx512vl" )] |
7174 | unsafe fn test_mm_mask_cmpeq_epi64_mask() { |
7175 | let a = _mm_set_epi64x(0, 1); |
7176 | let b = _mm_set_epi64x(0, 1); |
7177 | let mask = 0b11111111; |
7178 | let r = _mm_mask_cmpeq_epi64_mask(mask, b, a); |
7179 | assert_eq!(r, 0b00000011); |
7180 | } |
7181 | |
7182 | #[simd_test(enable = "avx512f" )] |
7183 | unsafe fn test_mm512_set_epi64() { |
7184 | let r = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
7185 | assert_eq_m512i(r, _mm512_set_epi64(7, 6, 5, 4, 3, 2, 1, 0)) |
7186 | } |
7187 | |
7188 | #[simd_test(enable = "avx512f" )] |
7189 | unsafe fn test_mm512_setr_epi64() { |
7190 | let r = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
7191 | assert_eq_m512i(r, _mm512_setr_epi64(7, 6, 5, 4, 3, 2, 1, 0)) |
7192 | } |
7193 | |
7194 | unsafe fn test_mm512_cmpneq_epi64_mask() { |
7195 | let a = _mm512_set_epi64(0, 1, -1, 13, i64::MAX, i64::MIN, 100, -100); |
7196 | let b = _mm512_set_epi64(0, 1, 13, 42, i64::MAX, i64::MIN, 100, -100); |
7197 | let m = _mm512_cmpneq_epi64_mask(b, a); |
7198 | assert_eq!(m, !_mm512_cmpeq_epi64_mask(b, a)); |
7199 | } |
7200 | |
7201 | #[simd_test(enable = "avx512f" )] |
7202 | unsafe fn test_mm512_mask_cmpneq_epi64_mask() { |
7203 | let a = _mm512_set_epi64(0, 1, -1, 13, i64::MAX, i64::MIN, -100, 100); |
7204 | let b = _mm512_set_epi64(0, 1, 13, 42, i64::MAX, i64::MIN, 100, -100); |
7205 | let mask = 0b01111010; |
7206 | let r = _mm512_mask_cmpneq_epi64_mask(mask, b, a); |
7207 | assert_eq!(r, 0b00110010) |
7208 | } |
7209 | |
7210 | #[simd_test(enable = "avx512f,avx512vl" )] |
7211 | unsafe fn test_mm256_cmpneq_epi64_mask() { |
7212 | let a = _mm256_set_epi64x(0, 1, -1, 13); |
7213 | let b = _mm256_set_epi64x(0, 1, 13, 42); |
7214 | let r = _mm256_cmpneq_epi64_mask(b, a); |
7215 | assert_eq!(r, 0b00000011) |
7216 | } |
7217 | |
7218 | #[simd_test(enable = "avx512f,avx512vl" )] |
7219 | unsafe fn test_mm256_mask_cmpneq_epi64_mask() { |
7220 | let a = _mm256_set_epi64x(0, 1, -1, 13); |
7221 | let b = _mm256_set_epi64x(0, 1, 13, 42); |
7222 | let mask = 0b11111111; |
7223 | let r = _mm256_mask_cmpneq_epi64_mask(mask, b, a); |
7224 | assert_eq!(r, 0b00000011) |
7225 | } |
7226 | |
7227 | #[simd_test(enable = "avx512f,avx512vl" )] |
7228 | unsafe fn test_mm_cmpneq_epi64_mask() { |
7229 | let a = _mm_set_epi64x(-1, 13); |
7230 | let b = _mm_set_epi64x(13, 42); |
7231 | let r = _mm_cmpneq_epi64_mask(b, a); |
7232 | assert_eq!(r, 0b00000011) |
7233 | } |
7234 | |
7235 | #[simd_test(enable = "avx512f,avx512vl" )] |
7236 | unsafe fn test_mm_mask_cmpneq_epi64_mask() { |
7237 | let a = _mm_set_epi64x(-1, 13); |
7238 | let b = _mm_set_epi64x(13, 42); |
7239 | let mask = 0b11111111; |
7240 | let r = _mm_mask_cmpneq_epi64_mask(mask, b, a); |
7241 | assert_eq!(r, 0b00000011) |
7242 | } |
7243 | |
7244 | #[simd_test(enable = "avx512f" )] |
7245 | unsafe fn test_mm512_cmp_epi64_mask() { |
7246 | let a = _mm512_set_epi64(0, 1, -1, 13, i64::MAX, i64::MIN, 100, -100); |
7247 | let b = _mm512_set1_epi64(-1); |
7248 | let m = _mm512_cmp_epi64_mask::<_MM_CMPINT_LT>(a, b); |
7249 | assert_eq!(m, 0b00000101); |
7250 | } |
7251 | |
7252 | #[simd_test(enable = "avx512f" )] |
7253 | unsafe fn test_mm512_mask_cmp_epi64_mask() { |
7254 | let a = _mm512_set_epi64(0, 1, -1, 13, i64::MAX, i64::MIN, 100, -100); |
7255 | let b = _mm512_set1_epi64(-1); |
7256 | let mask = 0b01100110; |
7257 | let r = _mm512_mask_cmp_epi64_mask::<_MM_CMPINT_LT>(mask, a, b); |
7258 | assert_eq!(r, 0b00000100); |
7259 | } |
7260 | |
7261 | #[simd_test(enable = "avx512f,avx512vl" )] |
7262 | unsafe fn test_mm256_cmp_epi64_mask() { |
7263 | let a = _mm256_set_epi64x(0, 1, -1, 13); |
7264 | let b = _mm256_set1_epi64x(1); |
7265 | let m = _mm256_cmp_epi64_mask::<_MM_CMPINT_LT>(a, b); |
7266 | assert_eq!(m, 0b00001010); |
7267 | } |
7268 | |
7269 | #[simd_test(enable = "avx512f,avx512vl" )] |
7270 | unsafe fn test_mm256_mask_cmp_epi64_mask() { |
7271 | let a = _mm256_set_epi64x(0, 1, -1, 13); |
7272 | let b = _mm256_set1_epi64x(1); |
7273 | let mask = 0b11111111; |
7274 | let r = _mm256_mask_cmp_epi64_mask::<_MM_CMPINT_LT>(mask, a, b); |
7275 | assert_eq!(r, 0b00001010); |
7276 | } |
7277 | |
7278 | #[simd_test(enable = "avx512f,avx512vl" )] |
7279 | unsafe fn test_mm_cmp_epi64_mask() { |
7280 | let a = _mm_set_epi64x(0, 1); |
7281 | let b = _mm_set1_epi64x(1); |
7282 | let m = _mm_cmp_epi64_mask::<_MM_CMPINT_LT>(a, b); |
7283 | assert_eq!(m, 0b00000010); |
7284 | } |
7285 | |
7286 | #[simd_test(enable = "avx512f,avx512vl" )] |
7287 | unsafe fn test_mm_mask_cmp_epi64_mask() { |
7288 | let a = _mm_set_epi64x(0, 1); |
7289 | let b = _mm_set1_epi64x(1); |
7290 | let mask = 0b11111111; |
7291 | let r = _mm_mask_cmp_epi64_mask::<_MM_CMPINT_LT>(mask, a, b); |
7292 | assert_eq!(r, 0b00000010); |
7293 | } |
7294 | |
7295 | #[simd_test(enable = "avx512f" )] |
7296 | unsafe fn test_mm512_i32gather_pd() { |
7297 | let arr: [f64; 128] = core::array::from_fn(|i| i as f64); |
7298 | // A multiplier of 8 is word-addressing |
7299 | let index = _mm256_setr_epi32(0, 16, 32, 48, 64, 80, 96, 112); |
7300 | let r = _mm512_i32gather_pd::<8>(index, arr.as_ptr() as *const u8); |
7301 | assert_eq_m512d(r, _mm512_setr_pd(0., 16., 32., 48., 64., 80., 96., 112.)); |
7302 | } |
7303 | |
7304 | #[simd_test(enable = "avx512f" )] |
7305 | unsafe fn test_mm512_mask_i32gather_pd() { |
7306 | let arr: [f64; 128] = core::array::from_fn(|i| i as f64); |
7307 | let src = _mm512_set1_pd(2.); |
7308 | let mask = 0b10101010; |
7309 | let index = _mm256_setr_epi32(0, 16, 32, 48, 64, 80, 96, 112); |
7310 | // A multiplier of 8 is word-addressing |
7311 | let r = _mm512_mask_i32gather_pd::<8>(src, mask, index, arr.as_ptr() as *const u8); |
7312 | assert_eq_m512d(r, _mm512_setr_pd(2., 16., 2., 48., 2., 80., 2., 112.)); |
7313 | } |
7314 | |
7315 | #[simd_test(enable = "avx512f" )] |
7316 | unsafe fn test_mm512_i64gather_pd() { |
7317 | let arr: [f64; 128] = core::array::from_fn(|i| i as f64); |
7318 | // A multiplier of 8 is word-addressing |
7319 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7320 | let r = _mm512_i64gather_pd::<8>(index, arr.as_ptr() as *const u8); |
7321 | assert_eq_m512d(r, _mm512_setr_pd(0., 16., 32., 48., 64., 80., 96., 112.)); |
7322 | } |
7323 | |
7324 | #[simd_test(enable = "avx512f" )] |
7325 | unsafe fn test_mm512_mask_i64gather_pd() { |
7326 | let arr: [f64; 128] = core::array::from_fn(|i| i as f64); |
7327 | let src = _mm512_set1_pd(2.); |
7328 | let mask = 0b10101010; |
7329 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7330 | // A multiplier of 8 is word-addressing |
7331 | let r = _mm512_mask_i64gather_pd::<8>(src, mask, index, arr.as_ptr() as *const u8); |
7332 | assert_eq_m512d(r, _mm512_setr_pd(2., 16., 2., 48., 2., 80., 2., 112.)); |
7333 | } |
7334 | |
7335 | #[simd_test(enable = "avx512f" )] |
7336 | unsafe fn test_mm512_i64gather_ps() { |
7337 | let arr: [f32; 128] = core::array::from_fn(|i| i as f32); |
7338 | // A multiplier of 4 is word-addressing |
7339 | #[rustfmt::skip] |
7340 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7341 | let r = _mm512_i64gather_ps::<4>(index, arr.as_ptr() as *const u8); |
7342 | assert_eq_m256(r, _mm256_setr_ps(0., 16., 32., 48., 64., 80., 96., 112.)); |
7343 | } |
7344 | |
7345 | #[simd_test(enable = "avx512f" )] |
7346 | unsafe fn test_mm512_mask_i64gather_ps() { |
7347 | let arr: [f32; 128] = core::array::from_fn(|i| i as f32); |
7348 | let src = _mm256_set1_ps(2.); |
7349 | let mask = 0b10101010; |
7350 | #[rustfmt::skip] |
7351 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7352 | // A multiplier of 4 is word-addressing |
7353 | let r = _mm512_mask_i64gather_ps::<4>(src, mask, index, arr.as_ptr() as *const u8); |
7354 | assert_eq_m256(r, _mm256_setr_ps(2., 16., 2., 48., 2., 80., 2., 112.)); |
7355 | } |
7356 | |
7357 | #[simd_test(enable = "avx512f" )] |
7358 | unsafe fn test_mm512_i32gather_epi64() { |
7359 | let mut arr = [0i64; 128]; |
7360 | for i in 0..128i64 { |
7361 | arr[i as usize] = i; |
7362 | } |
7363 | // A multiplier of 8 is word-addressing |
7364 | let index = _mm256_setr_epi32(0, 16, 32, 48, 64, 80, 96, 112); |
7365 | let r = _mm512_i32gather_epi64::<8>(index, arr.as_ptr() as *const u8); |
7366 | assert_eq_m512i(r, _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112)); |
7367 | } |
7368 | |
7369 | #[simd_test(enable = "avx512f" )] |
7370 | unsafe fn test_mm512_mask_i32gather_epi64() { |
7371 | let mut arr = [0i64; 128]; |
7372 | for i in 0..128i64 { |
7373 | arr[i as usize] = i; |
7374 | } |
7375 | let src = _mm512_set1_epi64(2); |
7376 | let mask = 0b10101010; |
7377 | let index = _mm256_setr_epi32(0, 16, 32, 48, 64, 80, 96, 112); |
7378 | // A multiplier of 8 is word-addressing |
7379 | let r = _mm512_mask_i32gather_epi64::<8>(src, mask, index, arr.as_ptr() as *const u8); |
7380 | assert_eq_m512i(r, _mm512_setr_epi64(2, 16, 2, 48, 2, 80, 2, 112)); |
7381 | } |
7382 | |
7383 | #[simd_test(enable = "avx512f" )] |
7384 | unsafe fn test_mm512_i64gather_epi64() { |
7385 | let mut arr = [0i64; 128]; |
7386 | for i in 0..128i64 { |
7387 | arr[i as usize] = i; |
7388 | } |
7389 | // A multiplier of 8 is word-addressing |
7390 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7391 | let r = _mm512_i64gather_epi64::<8>(index, arr.as_ptr() as *const u8); |
7392 | assert_eq_m512i(r, _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112)); |
7393 | } |
7394 | |
7395 | #[simd_test(enable = "avx512f" )] |
7396 | unsafe fn test_mm512_mask_i64gather_epi64() { |
7397 | let mut arr = [0i64; 128]; |
7398 | for i in 0..128i64 { |
7399 | arr[i as usize] = i; |
7400 | } |
7401 | let src = _mm512_set1_epi64(2); |
7402 | let mask = 0b10101010; |
7403 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7404 | // A multiplier of 8 is word-addressing |
7405 | let r = _mm512_mask_i64gather_epi64::<8>(src, mask, index, arr.as_ptr() as *const u8); |
7406 | assert_eq_m512i(r, _mm512_setr_epi64(2, 16, 2, 48, 2, 80, 2, 112)); |
7407 | } |
7408 | |
7409 | #[simd_test(enable = "avx512f" )] |
7410 | unsafe fn test_mm512_i64gather_epi32() { |
7411 | let mut arr = [0i64; 128]; |
7412 | for i in 0..128i64 { |
7413 | arr[i as usize] = i; |
7414 | } |
7415 | // A multiplier of 8 is word-addressing |
7416 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7417 | let r = _mm512_i64gather_epi32::<8>(index, arr.as_ptr() as *const u8); |
7418 | assert_eq_m256i(r, _mm256_setr_epi32(0, 16, 32, 48, 64, 80, 96, 112)); |
7419 | } |
7420 | |
7421 | #[simd_test(enable = "avx512f" )] |
7422 | unsafe fn test_mm512_mask_i64gather_epi32() { |
7423 | let mut arr = [0i64; 128]; |
7424 | for i in 0..128i64 { |
7425 | arr[i as usize] = i; |
7426 | } |
7427 | let src = _mm256_set1_epi32(2); |
7428 | let mask = 0b10101010; |
7429 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7430 | // A multiplier of 8 is word-addressing |
7431 | let r = _mm512_mask_i64gather_epi32::<8>(src, mask, index, arr.as_ptr() as *const u8); |
7432 | assert_eq_m256i(r, _mm256_setr_epi32(2, 16, 2, 48, 2, 80, 2, 112)); |
7433 | } |
7434 | |
7435 | #[simd_test(enable = "avx512f" )] |
7436 | unsafe fn test_mm512_i32scatter_pd() { |
7437 | let mut arr = [0f64; 128]; |
7438 | let index = _mm256_setr_epi32(0, 16, 32, 48, 64, 80, 96, 112); |
7439 | let src = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
7440 | // A multiplier of 8 is word-addressing |
7441 | _mm512_i32scatter_pd::<8>(arr.as_mut_ptr() as *mut u8, index, src); |
7442 | let mut expected = [0f64; 128]; |
7443 | for i in 0..8 { |
7444 | expected[i * 16] = (i + 1) as f64; |
7445 | } |
7446 | assert_eq!(&arr[..], &expected[..],); |
7447 | } |
7448 | |
7449 | #[simd_test(enable = "avx512f" )] |
7450 | unsafe fn test_mm512_mask_i32scatter_pd() { |
7451 | let mut arr = [0f64; 128]; |
7452 | let mask = 0b10101010; |
7453 | let index = _mm256_setr_epi32(0, 16, 32, 48, 64, 80, 96, 112); |
7454 | let src = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
7455 | // A multiplier of 8 is word-addressing |
7456 | _mm512_mask_i32scatter_pd::<8>(arr.as_mut_ptr() as *mut u8, mask, index, src); |
7457 | let mut expected = [0f64; 128]; |
7458 | for i in 0..4 { |
7459 | expected[i * 32 + 16] = 2. * (i + 1) as f64; |
7460 | } |
7461 | assert_eq!(&arr[..], &expected[..],); |
7462 | } |
7463 | |
7464 | #[simd_test(enable = "avx512f" )] |
7465 | unsafe fn test_mm512_i64scatter_pd() { |
7466 | let mut arr = [0f64; 128]; |
7467 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7468 | let src = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
7469 | // A multiplier of 8 is word-addressing |
7470 | _mm512_i64scatter_pd::<8>(arr.as_mut_ptr() as *mut u8, index, src); |
7471 | let mut expected = [0f64; 128]; |
7472 | for i in 0..8 { |
7473 | expected[i * 16] = (i + 1) as f64; |
7474 | } |
7475 | assert_eq!(&arr[..], &expected[..],); |
7476 | } |
7477 | |
7478 | #[simd_test(enable = "avx512f" )] |
7479 | unsafe fn test_mm512_mask_i64scatter_pd() { |
7480 | let mut arr = [0f64; 128]; |
7481 | let mask = 0b10101010; |
7482 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7483 | let src = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
7484 | // A multiplier of 8 is word-addressing |
7485 | _mm512_mask_i64scatter_pd::<8>(arr.as_mut_ptr() as *mut u8, mask, index, src); |
7486 | let mut expected = [0f64; 128]; |
7487 | for i in 0..4 { |
7488 | expected[i * 32 + 16] = 2. * (i + 1) as f64; |
7489 | } |
7490 | assert_eq!(&arr[..], &expected[..],); |
7491 | } |
7492 | |
7493 | #[simd_test(enable = "avx512f" )] |
7494 | unsafe fn test_mm512_i64scatter_ps() { |
7495 | let mut arr = [0f32; 128]; |
7496 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7497 | let src = _mm256_setr_ps(1., 2., 3., 4., 5., 6., 7., 8.); |
7498 | // A multiplier of 4 is word-addressing |
7499 | _mm512_i64scatter_ps::<4>(arr.as_mut_ptr() as *mut u8, index, src); |
7500 | let mut expected = [0f32; 128]; |
7501 | for i in 0..8 { |
7502 | expected[i * 16] = (i + 1) as f32; |
7503 | } |
7504 | assert_eq!(&arr[..], &expected[..],); |
7505 | } |
7506 | |
7507 | #[simd_test(enable = "avx512f" )] |
7508 | unsafe fn test_mm512_mask_i64scatter_ps() { |
7509 | let mut arr = [0f32; 128]; |
7510 | let mask = 0b10101010; |
7511 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7512 | let src = _mm256_setr_ps(1., 2., 3., 4., 5., 6., 7., 8.); |
7513 | // A multiplier of 4 is word-addressing |
7514 | _mm512_mask_i64scatter_ps::<4>(arr.as_mut_ptr() as *mut u8, mask, index, src); |
7515 | let mut expected = [0f32; 128]; |
7516 | for i in 0..4 { |
7517 | expected[i * 32 + 16] = 2. * (i + 1) as f32; |
7518 | } |
7519 | assert_eq!(&arr[..], &expected[..],); |
7520 | } |
7521 | |
7522 | #[simd_test(enable = "avx512f" )] |
7523 | unsafe fn test_mm512_i32scatter_epi64() { |
7524 | let mut arr = [0i64; 128]; |
7525 | let index = _mm256_setr_epi32(0, 16, 32, 48, 64, 80, 96, 112); |
7526 | let src = _mm512_setr_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
7527 | // A multiplier of 8 is word-addressing |
7528 | _mm512_i32scatter_epi64::<8>(arr.as_mut_ptr() as *mut u8, index, src); |
7529 | let mut expected = [0i64; 128]; |
7530 | for i in 0..8 { |
7531 | expected[i * 16] = (i + 1) as i64; |
7532 | } |
7533 | assert_eq!(&arr[..], &expected[..],); |
7534 | } |
7535 | |
7536 | #[simd_test(enable = "avx512f" )] |
7537 | unsafe fn test_mm512_mask_i32scatter_epi64() { |
7538 | let mut arr = [0i64; 128]; |
7539 | let mask = 0b10101010; |
7540 | let index = _mm256_setr_epi32(0, 16, 32, 48, 64, 80, 96, 112); |
7541 | let src = _mm512_setr_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
7542 | // A multiplier of 8 is word-addressing |
7543 | _mm512_mask_i32scatter_epi64::<8>(arr.as_mut_ptr() as *mut u8, mask, index, src); |
7544 | let mut expected = [0i64; 128]; |
7545 | for i in 0..4 { |
7546 | expected[i * 32 + 16] = 2 * (i + 1) as i64; |
7547 | } |
7548 | assert_eq!(&arr[..], &expected[..],); |
7549 | } |
7550 | |
7551 | #[simd_test(enable = "avx512f,avx512vl" )] |
7552 | unsafe fn test_mm256_i32scatter_epi64() { |
7553 | let mut arr = [0i64; 64]; |
7554 | let index = _mm_setr_epi32(0, 16, 32, 48); |
7555 | let src = _mm256_setr_epi64x(1, 2, 3, 4); |
7556 | // A multiplier of 8 is word-addressing |
7557 | _mm256_i32scatter_epi64::<8>(arr.as_mut_ptr() as *mut u8, index, src); |
7558 | let mut expected = [0i64; 64]; |
7559 | for i in 0..4 { |
7560 | expected[i * 16] = (i + 1) as i64; |
7561 | } |
7562 | assert_eq!(&arr[..], &expected[..],); |
7563 | } |
7564 | |
7565 | #[simd_test(enable = "avx512f" )] |
7566 | unsafe fn test_mm512_i64scatter_epi64() { |
7567 | let mut arr = [0i64; 128]; |
7568 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7569 | let src = _mm512_setr_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
7570 | // A multiplier of 8 is word-addressing |
7571 | _mm512_i64scatter_epi64::<8>(arr.as_mut_ptr() as *mut u8, index, src); |
7572 | let mut expected = [0i64; 128]; |
7573 | for i in 0..8 { |
7574 | expected[i * 16] = (i + 1) as i64; |
7575 | } |
7576 | assert_eq!(&arr[..], &expected[..],); |
7577 | } |
7578 | |
7579 | #[simd_test(enable = "avx512f" )] |
7580 | unsafe fn test_mm512_mask_i64scatter_epi64() { |
7581 | let mut arr = [0i64; 128]; |
7582 | let mask = 0b10101010; |
7583 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7584 | let src = _mm512_setr_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
7585 | // A multiplier of 8 is word-addressing |
7586 | _mm512_mask_i64scatter_epi64::<8>(arr.as_mut_ptr() as *mut u8, mask, index, src); |
7587 | let mut expected = [0i64; 128]; |
7588 | for i in 0..4 { |
7589 | expected[i * 32 + 16] = 2 * (i + 1) as i64; |
7590 | } |
7591 | assert_eq!(&arr[..], &expected[..],); |
7592 | } |
7593 | |
7594 | #[simd_test(enable = "avx512f" )] |
7595 | unsafe fn test_mm512_i64scatter_epi32() { |
7596 | let mut arr = [0i32; 128]; |
7597 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7598 | let src = _mm256_setr_epi32(1, 2, 3, 4, 5, 6, 7, 8); |
7599 | // A multiplier of 4 is word-addressing |
7600 | _mm512_i64scatter_epi32::<4>(arr.as_mut_ptr() as *mut u8, index, src); |
7601 | let mut expected = [0i32; 128]; |
7602 | for i in 0..8 { |
7603 | expected[i * 16] = (i + 1) as i32; |
7604 | } |
7605 | assert_eq!(&arr[..], &expected[..],); |
7606 | } |
7607 | |
7608 | #[simd_test(enable = "avx512f" )] |
7609 | unsafe fn test_mm512_mask_i64scatter_epi32() { |
7610 | let mut arr = [0i32; 128]; |
7611 | let mask = 0b10101010; |
7612 | let index = _mm512_setr_epi64(0, 16, 32, 48, 64, 80, 96, 112); |
7613 | let src = _mm256_setr_epi32(1, 2, 3, 4, 5, 6, 7, 8); |
7614 | // A multiplier of 4 is word-addressing |
7615 | _mm512_mask_i64scatter_epi32::<4>(arr.as_mut_ptr() as *mut u8, mask, index, src); |
7616 | let mut expected = [0i32; 128]; |
7617 | for i in 0..4 { |
7618 | expected[i * 32 + 16] = 2 * (i + 1) as i32; |
7619 | } |
7620 | assert_eq!(&arr[..], &expected[..],); |
7621 | } |
7622 | |
7623 | #[simd_test(enable = "avx512f" )] |
7624 | unsafe fn test_mm512_rol_epi64() { |
7625 | #[rustfmt::skip] |
7626 | let a = _mm512_set_epi64( |
7627 | 1 << 63, 1 << 32, 1 << 32, 1 << 32, |
7628 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
7629 | ); |
7630 | let r = _mm512_rol_epi64::<1>(a); |
7631 | #[rustfmt::skip] |
7632 | let e = _mm512_set_epi64( |
7633 | 1 << 0, 1 << 33, 1 << 33, 1 << 33, |
7634 | 1 << 33, 1 << 33, 1 << 33, 1 << 33, |
7635 | ); |
7636 | assert_eq_m512i(r, e); |
7637 | } |
7638 | |
7639 | #[simd_test(enable = "avx512f" )] |
7640 | unsafe fn test_mm512_mask_rol_epi64() { |
7641 | #[rustfmt::skip] |
7642 | let a = _mm512_set_epi64( |
7643 | 1 << 63, 1 << 32, 1 << 32, 1 << 32, |
7644 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
7645 | ); |
7646 | let r = _mm512_mask_rol_epi64::<1>(a, 0, a); |
7647 | assert_eq_m512i(r, a); |
7648 | let r = _mm512_mask_rol_epi64::<1>(a, 0b11111111, a); |
7649 | #[rustfmt::skip] |
7650 | let e = _mm512_set_epi64( |
7651 | 1 << 0, 1 << 33, 1 << 33, 1 << 33, |
7652 | 1 << 33, 1 << 33, 1 << 33, 1 << 33, |
7653 | ); |
7654 | assert_eq_m512i(r, e); |
7655 | } |
7656 | |
7657 | #[simd_test(enable = "avx512f" )] |
7658 | unsafe fn test_mm512_maskz_rol_epi64() { |
7659 | #[rustfmt::skip] |
7660 | let a = _mm512_set_epi64( |
7661 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
7662 | 1 << 32, 1 << 32, 1 << 32, 1 << 63, |
7663 | ); |
7664 | let r = _mm512_maskz_rol_epi64::<1>(0, a); |
7665 | assert_eq_m512i(r, _mm512_setzero_si512()); |
7666 | let r = _mm512_maskz_rol_epi64::<1>(0b00001111, a); |
7667 | let e = _mm512_set_epi64(0, 0, 0, 0, 1 << 33, 1 << 33, 1 << 33, 1 << 0); |
7668 | assert_eq_m512i(r, e); |
7669 | } |
7670 | |
7671 | #[simd_test(enable = "avx512f,avx512vl" )] |
7672 | unsafe fn test_mm256_rol_epi64() { |
7673 | let a = _mm256_set_epi64x(1 << 63, 1 << 32, 1 << 32, 1 << 32); |
7674 | let r = _mm256_rol_epi64::<1>(a); |
7675 | let e = _mm256_set_epi64x(1 << 0, 1 << 33, 1 << 33, 1 << 33); |
7676 | assert_eq_m256i(r, e); |
7677 | } |
7678 | |
7679 | #[simd_test(enable = "avx512f,avx512vl" )] |
7680 | unsafe fn test_mm256_mask_rol_epi64() { |
7681 | let a = _mm256_set_epi64x(1 << 63, 1 << 32, 1 << 32, 1 << 32); |
7682 | let r = _mm256_mask_rol_epi64::<1>(a, 0, a); |
7683 | assert_eq_m256i(r, a); |
7684 | let r = _mm256_mask_rol_epi64::<1>(a, 0b00001111, a); |
7685 | let e = _mm256_set_epi64x(1 << 0, 1 << 33, 1 << 33, 1 << 33); |
7686 | assert_eq_m256i(r, e); |
7687 | } |
7688 | |
7689 | #[simd_test(enable = "avx512f,avx512vl" )] |
7690 | unsafe fn test_mm256_maskz_rol_epi64() { |
7691 | let a = _mm256_set_epi64x(1 << 63, 1 << 32, 1 << 32, 1 << 32); |
7692 | let r = _mm256_maskz_rol_epi64::<1>(0, a); |
7693 | assert_eq_m256i(r, _mm256_setzero_si256()); |
7694 | let r = _mm256_maskz_rol_epi64::<1>(0b00001111, a); |
7695 | let e = _mm256_set_epi64x(1 << 0, 1 << 33, 1 << 33, 1 << 33); |
7696 | assert_eq_m256i(r, e); |
7697 | } |
7698 | |
7699 | #[simd_test(enable = "avx512f,avx512vl" )] |
7700 | unsafe fn test_mm_rol_epi64() { |
7701 | let a = _mm_set_epi64x(1 << 63, 1 << 32); |
7702 | let r = _mm_rol_epi64::<1>(a); |
7703 | let e = _mm_set_epi64x(1 << 0, 1 << 33); |
7704 | assert_eq_m128i(r, e); |
7705 | } |
7706 | |
7707 | #[simd_test(enable = "avx512f,avx512vl" )] |
7708 | unsafe fn test_mm_mask_rol_epi64() { |
7709 | let a = _mm_set_epi64x(1 << 63, 1 << 32); |
7710 | let r = _mm_mask_rol_epi64::<1>(a, 0, a); |
7711 | assert_eq_m128i(r, a); |
7712 | let r = _mm_mask_rol_epi64::<1>(a, 0b00000011, a); |
7713 | let e = _mm_set_epi64x(1 << 0, 1 << 33); |
7714 | assert_eq_m128i(r, e); |
7715 | } |
7716 | |
7717 | #[simd_test(enable = "avx512f,avx512vl" )] |
7718 | unsafe fn test_mm_maskz_rol_epi64() { |
7719 | let a = _mm_set_epi64x(1 << 63, 1 << 32); |
7720 | let r = _mm_maskz_rol_epi64::<1>(0, a); |
7721 | assert_eq_m128i(r, _mm_setzero_si128()); |
7722 | let r = _mm_maskz_rol_epi64::<1>(0b00000011, a); |
7723 | let e = _mm_set_epi64x(1 << 0, 1 << 33); |
7724 | assert_eq_m128i(r, e); |
7725 | } |
7726 | |
7727 | #[simd_test(enable = "avx512f" )] |
7728 | unsafe fn test_mm512_ror_epi64() { |
7729 | #[rustfmt::skip] |
7730 | let a = _mm512_set_epi64( |
7731 | 1 << 0, 1 << 32, 1 << 32, 1 << 32, |
7732 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
7733 | ); |
7734 | let r = _mm512_ror_epi64::<1>(a); |
7735 | #[rustfmt::skip] |
7736 | let e = _mm512_set_epi64( |
7737 | 1 << 63, 1 << 31, 1 << 31, 1 << 31, |
7738 | 1 << 31, 1 << 31, 1 << 31, 1 << 31, |
7739 | ); |
7740 | assert_eq_m512i(r, e); |
7741 | } |
7742 | |
7743 | #[simd_test(enable = "avx512f" )] |
7744 | unsafe fn test_mm512_mask_ror_epi64() { |
7745 | #[rustfmt::skip] |
7746 | let a = _mm512_set_epi64( |
7747 | 1 << 0, 1 << 32, 1 << 32, 1 << 32, |
7748 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
7749 | ); |
7750 | let r = _mm512_mask_ror_epi64::<1>(a, 0, a); |
7751 | assert_eq_m512i(r, a); |
7752 | let r = _mm512_mask_ror_epi64::<1>(a, 0b11111111, a); |
7753 | #[rustfmt::skip] |
7754 | let e = _mm512_set_epi64( |
7755 | 1 << 63, 1 << 31, 1 << 31, 1 << 31, |
7756 | 1 << 31, 1 << 31, 1 << 31, 1 << 31, |
7757 | ); |
7758 | assert_eq_m512i(r, e); |
7759 | } |
7760 | |
7761 | #[simd_test(enable = "avx512f" )] |
7762 | unsafe fn test_mm512_maskz_ror_epi64() { |
7763 | #[rustfmt::skip] |
7764 | let a = _mm512_set_epi64( |
7765 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
7766 | 1 << 32, 1 << 32, 1 << 32, 1 << 0, |
7767 | ); |
7768 | let r = _mm512_maskz_ror_epi64::<1>(0, a); |
7769 | assert_eq_m512i(r, _mm512_setzero_si512()); |
7770 | let r = _mm512_maskz_ror_epi64::<1>(0b00001111, a); |
7771 | let e = _mm512_set_epi64(0, 0, 0, 0, 1 << 31, 1 << 31, 1 << 31, 1 << 63); |
7772 | assert_eq_m512i(r, e); |
7773 | } |
7774 | |
7775 | #[simd_test(enable = "avx512f,avx512vl" )] |
7776 | unsafe fn test_mm256_ror_epi64() { |
7777 | let a = _mm256_set_epi64x(1 << 0, 1 << 32, 1 << 32, 1 << 32); |
7778 | let r = _mm256_ror_epi64::<1>(a); |
7779 | let e = _mm256_set_epi64x(1 << 63, 1 << 31, 1 << 31, 1 << 31); |
7780 | assert_eq_m256i(r, e); |
7781 | } |
7782 | |
7783 | #[simd_test(enable = "avx512f,avx512vl" )] |
7784 | unsafe fn test_mm256_mask_ror_epi64() { |
7785 | let a = _mm256_set_epi64x(1 << 0, 1 << 32, 1 << 32, 1 << 32); |
7786 | let r = _mm256_mask_ror_epi64::<1>(a, 0, a); |
7787 | assert_eq_m256i(r, a); |
7788 | let r = _mm256_mask_ror_epi64::<1>(a, 0b00001111, a); |
7789 | let e = _mm256_set_epi64x(1 << 63, 1 << 31, 1 << 31, 1 << 31); |
7790 | assert_eq_m256i(r, e); |
7791 | } |
7792 | |
7793 | #[simd_test(enable = "avx512f,avx512vl" )] |
7794 | unsafe fn test_mm256_maskz_ror_epi64() { |
7795 | let a = _mm256_set_epi64x(1 << 0, 1 << 32, 1 << 32, 1 << 32); |
7796 | let r = _mm256_maskz_ror_epi64::<1>(0, a); |
7797 | assert_eq_m256i(r, _mm256_setzero_si256()); |
7798 | let r = _mm256_maskz_ror_epi64::<1>(0b00001111, a); |
7799 | let e = _mm256_set_epi64x(1 << 63, 1 << 31, 1 << 31, 1 << 31); |
7800 | assert_eq_m256i(r, e); |
7801 | } |
7802 | |
7803 | #[simd_test(enable = "avx512f,avx512vl" )] |
7804 | unsafe fn test_mm_ror_epi64() { |
7805 | let a = _mm_set_epi64x(1 << 0, 1 << 32); |
7806 | let r = _mm_ror_epi64::<1>(a); |
7807 | let e = _mm_set_epi64x(1 << 63, 1 << 31); |
7808 | assert_eq_m128i(r, e); |
7809 | } |
7810 | |
7811 | #[simd_test(enable = "avx512f,avx512vl" )] |
7812 | unsafe fn test_mm_mask_ror_epi64() { |
7813 | let a = _mm_set_epi64x(1 << 0, 1 << 32); |
7814 | let r = _mm_mask_ror_epi64::<1>(a, 0, a); |
7815 | assert_eq_m128i(r, a); |
7816 | let r = _mm_mask_ror_epi64::<1>(a, 0b00000011, a); |
7817 | let e = _mm_set_epi64x(1 << 63, 1 << 31); |
7818 | assert_eq_m128i(r, e); |
7819 | } |
7820 | |
7821 | #[simd_test(enable = "avx512f,avx512vl" )] |
7822 | unsafe fn test_mm_maskz_ror_epi64() { |
7823 | let a = _mm_set_epi64x(1 << 0, 1 << 32); |
7824 | let r = _mm_maskz_ror_epi64::<1>(0, a); |
7825 | assert_eq_m128i(r, _mm_setzero_si128()); |
7826 | let r = _mm_maskz_ror_epi64::<1>(0b00000011, a); |
7827 | let e = _mm_set_epi64x(1 << 63, 1 << 31); |
7828 | assert_eq_m128i(r, e); |
7829 | } |
7830 | |
7831 | #[simd_test(enable = "avx512f" )] |
7832 | unsafe fn test_mm512_slli_epi64() { |
7833 | #[rustfmt::skip] |
7834 | let a = _mm512_set_epi64( |
7835 | 1 << 63, 1 << 32, 1 << 32, 1 << 32, |
7836 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
7837 | ); |
7838 | let r = _mm512_slli_epi64::<1>(a); |
7839 | #[rustfmt::skip] |
7840 | let e = _mm512_set_epi64( |
7841 | 0, 1 << 33, 1 << 33, 1 << 33, |
7842 | 1 << 33, 1 << 33, 1 << 33, 1 << 33, |
7843 | ); |
7844 | assert_eq_m512i(r, e); |
7845 | } |
7846 | |
7847 | #[simd_test(enable = "avx512f" )] |
7848 | unsafe fn test_mm512_mask_slli_epi64() { |
7849 | #[rustfmt::skip] |
7850 | let a = _mm512_set_epi64( |
7851 | 1 << 63, 1 << 32, 1 << 32, 1 << 32, |
7852 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
7853 | ); |
7854 | let r = _mm512_mask_slli_epi64::<1>(a, 0, a); |
7855 | assert_eq_m512i(r, a); |
7856 | let r = _mm512_mask_slli_epi64::<1>(a, 0b11111111, a); |
7857 | #[rustfmt::skip] |
7858 | let e = _mm512_set_epi64( |
7859 | 0, 1 << 33, 1 << 33, 1 << 33, |
7860 | 1 << 33, 1 << 33, 1 << 33, 1 << 33, |
7861 | ); |
7862 | assert_eq_m512i(r, e); |
7863 | } |
7864 | |
7865 | #[simd_test(enable = "avx512f" )] |
7866 | unsafe fn test_mm512_maskz_slli_epi64() { |
7867 | #[rustfmt::skip] |
7868 | let a = _mm512_set_epi64( |
7869 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
7870 | 1 << 32, 1 << 32, 1 << 32, 1 << 63, |
7871 | ); |
7872 | let r = _mm512_maskz_slli_epi64::<1>(0, a); |
7873 | assert_eq_m512i(r, _mm512_setzero_si512()); |
7874 | let r = _mm512_maskz_slli_epi64::<1>(0b00001111, a); |
7875 | let e = _mm512_set_epi64(0, 0, 0, 0, 1 << 33, 1 << 33, 1 << 33, 0); |
7876 | assert_eq_m512i(r, e); |
7877 | } |
7878 | |
7879 | #[simd_test(enable = "avx512f,avx512vl" )] |
7880 | unsafe fn test_mm256_mask_slli_epi64() { |
7881 | let a = _mm256_set_epi64x(1 << 63, 1 << 32, 1 << 32, 1 << 32); |
7882 | let r = _mm256_mask_slli_epi64::<1>(a, 0, a); |
7883 | assert_eq_m256i(r, a); |
7884 | let r = _mm256_mask_slli_epi64::<1>(a, 0b00001111, a); |
7885 | let e = _mm256_set_epi64x(0, 1 << 33, 1 << 33, 1 << 33); |
7886 | assert_eq_m256i(r, e); |
7887 | } |
7888 | |
7889 | #[simd_test(enable = "avx512f,avx512vl" )] |
7890 | unsafe fn test_mm256_maskz_slli_epi64() { |
7891 | let a = _mm256_set_epi64x(1 << 63, 1 << 32, 1 << 32, 1 << 32); |
7892 | let r = _mm256_maskz_slli_epi64::<1>(0, a); |
7893 | assert_eq_m256i(r, _mm256_setzero_si256()); |
7894 | let r = _mm256_maskz_slli_epi64::<1>(0b00001111, a); |
7895 | let e = _mm256_set_epi64x(0, 1 << 33, 1 << 33, 1 << 33); |
7896 | assert_eq_m256i(r, e); |
7897 | } |
7898 | |
7899 | #[simd_test(enable = "avx512f,avx512vl" )] |
7900 | unsafe fn test_mm_mask_slli_epi64() { |
7901 | let a = _mm_set_epi64x(1 << 63, 1 << 32); |
7902 | let r = _mm_mask_slli_epi64::<1>(a, 0, a); |
7903 | assert_eq_m128i(r, a); |
7904 | let r = _mm_mask_slli_epi64::<1>(a, 0b00000011, a); |
7905 | let e = _mm_set_epi64x(0, 1 << 33); |
7906 | assert_eq_m128i(r, e); |
7907 | } |
7908 | |
7909 | #[simd_test(enable = "avx512f,avx512vl" )] |
7910 | unsafe fn test_mm_maskz_slli_epi64() { |
7911 | let a = _mm_set_epi64x(1 << 63, 1 << 32); |
7912 | let r = _mm_maskz_slli_epi64::<1>(0, a); |
7913 | assert_eq_m128i(r, _mm_setzero_si128()); |
7914 | let r = _mm_maskz_slli_epi64::<1>(0b00000011, a); |
7915 | let e = _mm_set_epi64x(0, 1 << 33); |
7916 | assert_eq_m128i(r, e); |
7917 | } |
7918 | |
7919 | #[simd_test(enable = "avx512f" )] |
7920 | unsafe fn test_mm512_srli_epi64() { |
7921 | #[rustfmt::skip] |
7922 | let a = _mm512_set_epi64( |
7923 | 1 << 0, 1 << 32, 1 << 32, 1 << 32, |
7924 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
7925 | ); |
7926 | let r = _mm512_srli_epi64::<1>(a); |
7927 | #[rustfmt::skip] |
7928 | let e = _mm512_set_epi64( |
7929 | 0, 1 << 31, 1 << 31, 1 << 31, |
7930 | 1 << 31, 1 << 31, 1 << 31, 1 << 31, |
7931 | ); |
7932 | assert_eq_m512i(r, e); |
7933 | } |
7934 | |
7935 | #[simd_test(enable = "avx512f" )] |
7936 | unsafe fn test_mm512_mask_srli_epi64() { |
7937 | #[rustfmt::skip] |
7938 | let a = _mm512_set_epi64( |
7939 | 1 << 0, 1 << 32, 1 << 32, 1 << 32, |
7940 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
7941 | ); |
7942 | let r = _mm512_mask_srli_epi64::<1>(a, 0, a); |
7943 | assert_eq_m512i(r, a); |
7944 | let r = _mm512_mask_srli_epi64::<1>(a, 0b11111111, a); |
7945 | #[rustfmt::skip] |
7946 | let e = _mm512_set_epi64( |
7947 | 0, 1 << 31, 1 << 31, 1 << 31, |
7948 | 1 << 31, 1 << 31, 1 << 31, 1 << 31, |
7949 | ); |
7950 | assert_eq_m512i(r, e); |
7951 | } |
7952 | |
7953 | #[simd_test(enable = "avx512f" )] |
7954 | unsafe fn test_mm512_maskz_srli_epi64() { |
7955 | #[rustfmt::skip] |
7956 | let a = _mm512_set_epi64( |
7957 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
7958 | 1 << 32, 1 << 32, 1 << 32, 1 << 0, |
7959 | ); |
7960 | let r = _mm512_maskz_srli_epi64::<1>(0, a); |
7961 | assert_eq_m512i(r, _mm512_setzero_si512()); |
7962 | let r = _mm512_maskz_srli_epi64::<1>(0b00001111, a); |
7963 | let e = _mm512_set_epi64(0, 0, 0, 0, 1 << 31, 1 << 31, 1 << 31, 0); |
7964 | assert_eq_m512i(r, e); |
7965 | } |
7966 | |
7967 | #[simd_test(enable = "avx512f,avx512vl" )] |
7968 | unsafe fn test_mm256_mask_srli_epi64() { |
7969 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
7970 | let r = _mm256_mask_srli_epi64::<1>(a, 0, a); |
7971 | assert_eq_m256i(r, a); |
7972 | let r = _mm256_mask_srli_epi64::<1>(a, 0b00001111, a); |
7973 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
7974 | assert_eq_m256i(r, e); |
7975 | } |
7976 | |
7977 | #[simd_test(enable = "avx512f,avx512vl" )] |
7978 | unsafe fn test_mm256_maskz_srli_epi64() { |
7979 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
7980 | let r = _mm256_maskz_srli_epi64::<1>(0, a); |
7981 | assert_eq_m256i(r, _mm256_setzero_si256()); |
7982 | let r = _mm256_maskz_srli_epi64::<1>(0b00001111, a); |
7983 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
7984 | assert_eq_m256i(r, e); |
7985 | } |
7986 | |
7987 | #[simd_test(enable = "avx512f,avx512vl" )] |
7988 | unsafe fn test_mm_mask_srli_epi64() { |
7989 | let a = _mm_set_epi64x(1 << 5, 0); |
7990 | let r = _mm_mask_srli_epi64::<1>(a, 0, a); |
7991 | assert_eq_m128i(r, a); |
7992 | let r = _mm_mask_srli_epi64::<1>(a, 0b00000011, a); |
7993 | let e = _mm_set_epi64x(1 << 4, 0); |
7994 | assert_eq_m128i(r, e); |
7995 | } |
7996 | |
7997 | #[simd_test(enable = "avx512f,avx512vl" )] |
7998 | unsafe fn test_mm_maskz_srli_epi64() { |
7999 | let a = _mm_set_epi64x(1 << 5, 0); |
8000 | let r = _mm_maskz_srli_epi64::<1>(0, a); |
8001 | assert_eq_m128i(r, _mm_setzero_si128()); |
8002 | let r = _mm_maskz_srli_epi64::<1>(0b00000011, a); |
8003 | let e = _mm_set_epi64x(1 << 4, 0); |
8004 | assert_eq_m128i(r, e); |
8005 | } |
8006 | |
8007 | #[simd_test(enable = "avx512f" )] |
8008 | unsafe fn test_mm512_rolv_epi64() { |
8009 | #[rustfmt::skip] |
8010 | let a = _mm512_set_epi64( |
8011 | 1 << 32, 1 << 63, 1 << 32, 1 << 32, |
8012 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8013 | ); |
8014 | let b = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
8015 | let r = _mm512_rolv_epi64(a, b); |
8016 | #[rustfmt::skip] |
8017 | let e = _mm512_set_epi64( |
8018 | 1 << 32, 1 << 0, 1 << 34, 1 << 35, |
8019 | 1 << 36, 1 << 37, 1 << 38, 1 << 39, |
8020 | ); |
8021 | assert_eq_m512i(r, e); |
8022 | } |
8023 | |
8024 | #[simd_test(enable = "avx512f" )] |
8025 | unsafe fn test_mm512_mask_rolv_epi64() { |
8026 | #[rustfmt::skip] |
8027 | let a = _mm512_set_epi64( |
8028 | 1 << 32, 1 << 63, 1 << 32, 1 << 32, |
8029 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8030 | ); |
8031 | let b = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
8032 | let r = _mm512_mask_rolv_epi64(a, 0, a, b); |
8033 | assert_eq_m512i(r, a); |
8034 | let r = _mm512_mask_rolv_epi64(a, 0b11111111, a, b); |
8035 | #[rustfmt::skip] |
8036 | let e = _mm512_set_epi64( |
8037 | 1 << 32, 1 << 0, 1 << 34, 1 << 35, |
8038 | 1 << 36, 1 << 37, 1 << 38, 1 << 39, |
8039 | ); |
8040 | assert_eq_m512i(r, e); |
8041 | } |
8042 | |
8043 | #[simd_test(enable = "avx512f" )] |
8044 | unsafe fn test_mm512_maskz_rolv_epi64() { |
8045 | #[rustfmt::skip] |
8046 | let a = _mm512_set_epi64( |
8047 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8048 | 1 << 32, 1 << 32, 1 << 32, 1 << 62, |
8049 | ); |
8050 | let b = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 2); |
8051 | let r = _mm512_maskz_rolv_epi64(0, a, b); |
8052 | assert_eq_m512i(r, _mm512_setzero_si512()); |
8053 | let r = _mm512_maskz_rolv_epi64(0b00001111, a, b); |
8054 | let e = _mm512_set_epi64(0, 0, 0, 0, 1 << 36, 1 << 37, 1 << 38, 1 << 0); |
8055 | assert_eq_m512i(r, e); |
8056 | } |
8057 | |
8058 | #[simd_test(enable = "avx512f,avx512vl" )] |
8059 | unsafe fn test_mm256_rolv_epi64() { |
8060 | let a = _mm256_set_epi64x(1 << 32, 1 << 63, 1 << 32, 1 << 32); |
8061 | let b = _mm256_set_epi64x(0, 1, 2, 3); |
8062 | let r = _mm256_rolv_epi64(a, b); |
8063 | let e = _mm256_set_epi64x(1 << 32, 1 << 0, 1 << 34, 1 << 35); |
8064 | assert_eq_m256i(r, e); |
8065 | } |
8066 | |
8067 | #[simd_test(enable = "avx512f,avx512vl" )] |
8068 | unsafe fn test_mm256_mask_rolv_epi64() { |
8069 | let a = _mm256_set_epi64x(1 << 32, 1 << 63, 1 << 32, 1 << 32); |
8070 | let b = _mm256_set_epi64x(0, 1, 2, 3); |
8071 | let r = _mm256_mask_rolv_epi64(a, 0, a, b); |
8072 | assert_eq_m256i(r, a); |
8073 | let r = _mm256_mask_rolv_epi64(a, 0b00001111, a, b); |
8074 | let e = _mm256_set_epi64x(1 << 32, 1 << 0, 1 << 34, 1 << 35); |
8075 | assert_eq_m256i(r, e); |
8076 | } |
8077 | |
8078 | #[simd_test(enable = "avx512f,avx512vl" )] |
8079 | unsafe fn test_mm256_maskz_rolv_epi64() { |
8080 | let a = _mm256_set_epi64x(1 << 32, 1 << 63, 1 << 32, 1 << 32); |
8081 | let b = _mm256_set_epi64x(0, 1, 2, 3); |
8082 | let r = _mm256_maskz_rolv_epi64(0, a, b); |
8083 | assert_eq_m256i(r, _mm256_setzero_si256()); |
8084 | let r = _mm256_maskz_rolv_epi64(0b00001111, a, b); |
8085 | let e = _mm256_set_epi64x(1 << 32, 1 << 0, 1 << 34, 1 << 35); |
8086 | assert_eq_m256i(r, e); |
8087 | } |
8088 | |
8089 | #[simd_test(enable = "avx512f,avx512vl" )] |
8090 | unsafe fn test_mm_rolv_epi64() { |
8091 | let a = _mm_set_epi64x(1 << 32, 1 << 63); |
8092 | let b = _mm_set_epi64x(0, 1); |
8093 | let r = _mm_rolv_epi64(a, b); |
8094 | let e = _mm_set_epi64x(1 << 32, 1 << 0); |
8095 | assert_eq_m128i(r, e); |
8096 | } |
8097 | |
8098 | #[simd_test(enable = "avx512f,avx512vl" )] |
8099 | unsafe fn test_mm_mask_rolv_epi64() { |
8100 | let a = _mm_set_epi64x(1 << 32, 1 << 63); |
8101 | let b = _mm_set_epi64x(0, 1); |
8102 | let r = _mm_mask_rolv_epi64(a, 0, a, b); |
8103 | assert_eq_m128i(r, a); |
8104 | let r = _mm_mask_rolv_epi64(a, 0b00000011, a, b); |
8105 | let e = _mm_set_epi64x(1 << 32, 1 << 0); |
8106 | assert_eq_m128i(r, e); |
8107 | } |
8108 | |
8109 | #[simd_test(enable = "avx512f,avx512vl" )] |
8110 | unsafe fn test_mm_maskz_rolv_epi64() { |
8111 | let a = _mm_set_epi64x(1 << 32, 1 << 63); |
8112 | let b = _mm_set_epi64x(0, 1); |
8113 | let r = _mm_maskz_rolv_epi64(0, a, b); |
8114 | assert_eq_m128i(r, _mm_setzero_si128()); |
8115 | let r = _mm_maskz_rolv_epi64(0b00000011, a, b); |
8116 | let e = _mm_set_epi64x(1 << 32, 1 << 0); |
8117 | assert_eq_m128i(r, e); |
8118 | } |
8119 | |
8120 | #[simd_test(enable = "avx512f" )] |
8121 | unsafe fn test_mm512_rorv_epi64() { |
8122 | #[rustfmt::skip] |
8123 | let a = _mm512_set_epi64( |
8124 | 1 << 32, 1 << 0, 1 << 32, 1 << 32, |
8125 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8126 | ); |
8127 | let b = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
8128 | let r = _mm512_rorv_epi64(a, b); |
8129 | #[rustfmt::skip] |
8130 | let e = _mm512_set_epi64( |
8131 | 1 << 32, 1 << 63, 1 << 30, 1 << 29, |
8132 | 1 << 28, 1 << 27, 1 << 26, 1 << 25, |
8133 | ); |
8134 | assert_eq_m512i(r, e); |
8135 | } |
8136 | |
8137 | #[simd_test(enable = "avx512f" )] |
8138 | unsafe fn test_mm512_mask_rorv_epi64() { |
8139 | #[rustfmt::skip] |
8140 | let a = _mm512_set_epi64( |
8141 | 1 << 32, 1 << 0, 1 << 32, 1 << 32, |
8142 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8143 | ); |
8144 | let b = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
8145 | let r = _mm512_mask_rorv_epi64(a, 0, a, b); |
8146 | assert_eq_m512i(r, a); |
8147 | let r = _mm512_mask_rorv_epi64(a, 0b11111111, a, b); |
8148 | #[rustfmt::skip] |
8149 | let e = _mm512_set_epi64( |
8150 | 1 << 32, 1 << 63, 1 << 30, 1 << 29, |
8151 | 1 << 28, 1 << 27, 1 << 26, 1 << 25, |
8152 | ); |
8153 | assert_eq_m512i(r, e); |
8154 | } |
8155 | |
8156 | #[simd_test(enable = "avx512f" )] |
8157 | unsafe fn test_mm512_maskz_rorv_epi64() { |
8158 | #[rustfmt::skip] |
8159 | let a = _mm512_set_epi64( |
8160 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8161 | 1 << 32, 1 << 32, 1 << 32, 1 << 0, |
8162 | ); |
8163 | let b = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 2); |
8164 | let r = _mm512_maskz_rorv_epi64(0, a, b); |
8165 | assert_eq_m512i(r, _mm512_setzero_si512()); |
8166 | let r = _mm512_maskz_rorv_epi64(0b00001111, a, b); |
8167 | let e = _mm512_set_epi64(0, 0, 0, 0, 1 << 28, 1 << 27, 1 << 26, 1 << 62); |
8168 | assert_eq_m512i(r, e); |
8169 | } |
8170 | |
8171 | #[simd_test(enable = "avx512f,avx512vl" )] |
8172 | unsafe fn test_mm256_rorv_epi64() { |
8173 | let a = _mm256_set_epi64x(1 << 32, 1 << 0, 1 << 32, 1 << 32); |
8174 | let b = _mm256_set_epi64x(0, 1, 2, 3); |
8175 | let r = _mm256_rorv_epi64(a, b); |
8176 | let e = _mm256_set_epi64x(1 << 32, 1 << 63, 1 << 30, 1 << 29); |
8177 | assert_eq_m256i(r, e); |
8178 | } |
8179 | |
8180 | #[simd_test(enable = "avx512f,avx512vl" )] |
8181 | unsafe fn test_mm256_mask_rorv_epi64() { |
8182 | let a = _mm256_set_epi64x(1 << 32, 1 << 0, 1 << 32, 1 << 32); |
8183 | let b = _mm256_set_epi64x(0, 1, 2, 3); |
8184 | let r = _mm256_mask_rorv_epi64(a, 0, a, b); |
8185 | assert_eq_m256i(r, a); |
8186 | let r = _mm256_mask_rorv_epi64(a, 0b00001111, a, b); |
8187 | let e = _mm256_set_epi64x(1 << 32, 1 << 63, 1 << 30, 1 << 29); |
8188 | assert_eq_m256i(r, e); |
8189 | } |
8190 | |
8191 | #[simd_test(enable = "avx512f,avx512vl" )] |
8192 | unsafe fn test_mm256_maskz_rorv_epi64() { |
8193 | let a = _mm256_set_epi64x(1 << 32, 1 << 0, 1 << 32, 1 << 32); |
8194 | let b = _mm256_set_epi64x(0, 1, 2, 3); |
8195 | let r = _mm256_maskz_rorv_epi64(0, a, b); |
8196 | assert_eq_m256i(r, _mm256_setzero_si256()); |
8197 | let r = _mm256_maskz_rorv_epi64(0b00001111, a, b); |
8198 | let e = _mm256_set_epi64x(1 << 32, 1 << 63, 1 << 30, 1 << 29); |
8199 | assert_eq_m256i(r, e); |
8200 | } |
8201 | |
8202 | #[simd_test(enable = "avx512f,avx512vl" )] |
8203 | unsafe fn test_mm_rorv_epi64() { |
8204 | let a = _mm_set_epi64x(1 << 32, 1 << 0); |
8205 | let b = _mm_set_epi64x(0, 1); |
8206 | let r = _mm_rorv_epi64(a, b); |
8207 | let e = _mm_set_epi64x(1 << 32, 1 << 63); |
8208 | assert_eq_m128i(r, e); |
8209 | } |
8210 | |
8211 | #[simd_test(enable = "avx512f,avx512vl" )] |
8212 | unsafe fn test_mm_mask_rorv_epi64() { |
8213 | let a = _mm_set_epi64x(1 << 32, 1 << 0); |
8214 | let b = _mm_set_epi64x(0, 1); |
8215 | let r = _mm_mask_rorv_epi64(a, 0, a, b); |
8216 | assert_eq_m128i(r, a); |
8217 | let r = _mm_mask_rorv_epi64(a, 0b00000011, a, b); |
8218 | let e = _mm_set_epi64x(1 << 32, 1 << 63); |
8219 | assert_eq_m128i(r, e); |
8220 | } |
8221 | |
8222 | #[simd_test(enable = "avx512f,avx512vl" )] |
8223 | unsafe fn test_mm_maskz_rorv_epi64() { |
8224 | let a = _mm_set_epi64x(1 << 32, 1 << 0); |
8225 | let b = _mm_set_epi64x(0, 1); |
8226 | let r = _mm_maskz_rorv_epi64(0, a, b); |
8227 | assert_eq_m128i(r, _mm_setzero_si128()); |
8228 | let r = _mm_maskz_rorv_epi64(0b00000011, a, b); |
8229 | let e = _mm_set_epi64x(1 << 32, 1 << 63); |
8230 | assert_eq_m128i(r, e); |
8231 | } |
8232 | |
8233 | #[simd_test(enable = "avx512f" )] |
8234 | unsafe fn test_mm512_sllv_epi64() { |
8235 | #[rustfmt::skip] |
8236 | let a = _mm512_set_epi64( |
8237 | 1 << 32, 1 << 63, 1 << 32, 1 << 32, |
8238 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8239 | ); |
8240 | let count = _mm512_set_epi64(0, 2, 2, 3, 4, 5, 6, 7); |
8241 | let r = _mm512_sllv_epi64(a, count); |
8242 | #[rustfmt::skip] |
8243 | let e = _mm512_set_epi64( |
8244 | 1 << 32, 0, 1 << 34, 1 << 35, |
8245 | 1 << 36, 1 << 37, 1 << 38, 1 << 39, |
8246 | ); |
8247 | assert_eq_m512i(r, e); |
8248 | } |
8249 | |
8250 | #[simd_test(enable = "avx512f" )] |
8251 | unsafe fn test_mm512_mask_sllv_epi64() { |
8252 | #[rustfmt::skip] |
8253 | let a = _mm512_set_epi64( |
8254 | 1 << 32, 1 << 32, 1 << 63, 1 << 32, |
8255 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8256 | ); |
8257 | let count = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
8258 | let r = _mm512_mask_sllv_epi64(a, 0, a, count); |
8259 | assert_eq_m512i(r, a); |
8260 | let r = _mm512_mask_sllv_epi64(a, 0b11111111, a, count); |
8261 | #[rustfmt::skip] |
8262 | let e = _mm512_set_epi64( |
8263 | 1 << 32, 1 << 33, 0, 1 << 35, |
8264 | 1 << 36, 1 << 37, 1 << 38, 1 << 39, |
8265 | ); |
8266 | assert_eq_m512i(r, e); |
8267 | } |
8268 | |
8269 | #[simd_test(enable = "avx512f" )] |
8270 | unsafe fn test_mm512_maskz_sllv_epi64() { |
8271 | #[rustfmt::skip] |
8272 | let a = _mm512_set_epi64( |
8273 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8274 | 1 << 32, 1 << 32, 1 << 32, 1 << 63, |
8275 | ); |
8276 | let count = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 1); |
8277 | let r = _mm512_maskz_sllv_epi64(0, a, count); |
8278 | assert_eq_m512i(r, _mm512_setzero_si512()); |
8279 | let r = _mm512_maskz_sllv_epi64(0b00001111, a, count); |
8280 | let e = _mm512_set_epi64(0, 0, 0, 0, 1 << 36, 1 << 37, 1 << 38, 0); |
8281 | assert_eq_m512i(r, e); |
8282 | } |
8283 | |
8284 | #[simd_test(enable = "avx512f,avx512vl" )] |
8285 | unsafe fn test_mm256_mask_sllv_epi64() { |
8286 | let a = _mm256_set_epi64x(1 << 32, 1 << 32, 1 << 63, 1 << 32); |
8287 | let count = _mm256_set_epi64x(0, 1, 2, 3); |
8288 | let r = _mm256_mask_sllv_epi64(a, 0, a, count); |
8289 | assert_eq_m256i(r, a); |
8290 | let r = _mm256_mask_sllv_epi64(a, 0b00001111, a, count); |
8291 | let e = _mm256_set_epi64x(1 << 32, 1 << 33, 0, 1 << 35); |
8292 | assert_eq_m256i(r, e); |
8293 | } |
8294 | |
8295 | #[simd_test(enable = "avx512f,avx512vl" )] |
8296 | unsafe fn test_mm256_maskz_sllv_epi64() { |
8297 | let a = _mm256_set_epi64x(1 << 32, 1 << 32, 1 << 63, 1 << 32); |
8298 | let count = _mm256_set_epi64x(0, 1, 2, 3); |
8299 | let r = _mm256_maskz_sllv_epi64(0, a, count); |
8300 | assert_eq_m256i(r, _mm256_setzero_si256()); |
8301 | let r = _mm256_maskz_sllv_epi64(0b00001111, a, count); |
8302 | let e = _mm256_set_epi64x(1 << 32, 1 << 33, 0, 1 << 35); |
8303 | assert_eq_m256i(r, e); |
8304 | } |
8305 | |
8306 | #[simd_test(enable = "avx512f,avx512vl" )] |
8307 | unsafe fn test_mm_mask_sllv_epi64() { |
8308 | let a = _mm_set_epi64x(1 << 63, 1 << 32); |
8309 | let count = _mm_set_epi64x(2, 3); |
8310 | let r = _mm_mask_sllv_epi64(a, 0, a, count); |
8311 | assert_eq_m128i(r, a); |
8312 | let r = _mm_mask_sllv_epi64(a, 0b00000011, a, count); |
8313 | let e = _mm_set_epi64x(0, 1 << 35); |
8314 | assert_eq_m128i(r, e); |
8315 | } |
8316 | |
8317 | #[simd_test(enable = "avx512f,avx512vl" )] |
8318 | unsafe fn test_mm_maskz_sllv_epi64() { |
8319 | let a = _mm_set_epi64x(1 << 63, 1 << 32); |
8320 | let count = _mm_set_epi64x(2, 3); |
8321 | let r = _mm_maskz_sllv_epi64(0, a, count); |
8322 | assert_eq_m128i(r, _mm_setzero_si128()); |
8323 | let r = _mm_maskz_sllv_epi64(0b00000011, a, count); |
8324 | let e = _mm_set_epi64x(0, 1 << 35); |
8325 | assert_eq_m128i(r, e); |
8326 | } |
8327 | |
8328 | #[simd_test(enable = "avx512f" )] |
8329 | unsafe fn test_mm512_srlv_epi64() { |
8330 | #[rustfmt::skip] |
8331 | let a = _mm512_set_epi64( |
8332 | 1 << 32, 1 << 0, 1 << 32, 1 << 32, |
8333 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8334 | ); |
8335 | let count = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
8336 | let r = _mm512_srlv_epi64(a, count); |
8337 | #[rustfmt::skip] |
8338 | let e = _mm512_set_epi64( |
8339 | 1 << 32, 0, 1 << 30, 1 << 29, |
8340 | 1 << 28, 1 << 27, 1 << 26, 1 << 25, |
8341 | ); |
8342 | assert_eq_m512i(r, e); |
8343 | } |
8344 | |
8345 | #[simd_test(enable = "avx512f" )] |
8346 | unsafe fn test_mm512_mask_srlv_epi64() { |
8347 | #[rustfmt::skip] |
8348 | let a = _mm512_set_epi64( |
8349 | 1 << 32, 1 << 0, 1 << 32, 1 << 32, |
8350 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8351 | ); |
8352 | let count = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
8353 | let r = _mm512_mask_srlv_epi64(a, 0, a, count); |
8354 | assert_eq_m512i(r, a); |
8355 | let r = _mm512_mask_srlv_epi64(a, 0b11111111, a, count); |
8356 | #[rustfmt::skip] |
8357 | let e = _mm512_set_epi64( |
8358 | 1 << 32, 0, 1 << 30, 1 << 29, |
8359 | 1 << 28, 1 << 27, 1 << 26, 1 << 25, |
8360 | ); |
8361 | assert_eq_m512i(r, e); |
8362 | } |
8363 | |
8364 | #[simd_test(enable = "avx512f" )] |
8365 | unsafe fn test_mm512_maskz_srlv_epi64() { |
8366 | #[rustfmt::skip] |
8367 | let a = _mm512_set_epi64( |
8368 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8369 | 1 << 32, 1 << 32, 1 << 32, 1 << 0, |
8370 | ); |
8371 | let count = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
8372 | let r = _mm512_maskz_srlv_epi64(0, a, count); |
8373 | assert_eq_m512i(r, _mm512_setzero_si512()); |
8374 | let r = _mm512_maskz_srlv_epi64(0b00001111, a, count); |
8375 | let e = _mm512_set_epi64(0, 0, 0, 0, 1 << 28, 1 << 27, 1 << 26, 0); |
8376 | assert_eq_m512i(r, e); |
8377 | } |
8378 | |
8379 | #[simd_test(enable = "avx512f,avx512vl" )] |
8380 | unsafe fn test_mm256_mask_srlv_epi64() { |
8381 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8382 | let count = _mm256_set1_epi64x(1); |
8383 | let r = _mm256_mask_srlv_epi64(a, 0, a, count); |
8384 | assert_eq_m256i(r, a); |
8385 | let r = _mm256_mask_srlv_epi64(a, 0b00001111, a, count); |
8386 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8387 | assert_eq_m256i(r, e); |
8388 | } |
8389 | |
8390 | #[simd_test(enable = "avx512f,avx512vl" )] |
8391 | unsafe fn test_mm256_maskz_srlv_epi64() { |
8392 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8393 | let count = _mm256_set1_epi64x(1); |
8394 | let r = _mm256_maskz_srlv_epi64(0, a, count); |
8395 | assert_eq_m256i(r, _mm256_setzero_si256()); |
8396 | let r = _mm256_maskz_srlv_epi64(0b00001111, a, count); |
8397 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8398 | assert_eq_m256i(r, e); |
8399 | } |
8400 | |
8401 | #[simd_test(enable = "avx512f,avx512vl" )] |
8402 | unsafe fn test_mm_mask_srlv_epi64() { |
8403 | let a = _mm_set_epi64x(1 << 5, 0); |
8404 | let count = _mm_set1_epi64x(1); |
8405 | let r = _mm_mask_srlv_epi64(a, 0, a, count); |
8406 | assert_eq_m128i(r, a); |
8407 | let r = _mm_mask_srlv_epi64(a, 0b00000011, a, count); |
8408 | let e = _mm_set_epi64x(1 << 4, 0); |
8409 | assert_eq_m128i(r, e); |
8410 | } |
8411 | |
8412 | #[simd_test(enable = "avx512f,avx512vl" )] |
8413 | unsafe fn test_mm_maskz_srlv_epi64() { |
8414 | let a = _mm_set_epi64x(1 << 5, 0); |
8415 | let count = _mm_set1_epi64x(1); |
8416 | let r = _mm_maskz_srlv_epi64(0, a, count); |
8417 | assert_eq_m128i(r, _mm_setzero_si128()); |
8418 | let r = _mm_maskz_srlv_epi64(0b00000011, a, count); |
8419 | let e = _mm_set_epi64x(1 << 4, 0); |
8420 | assert_eq_m128i(r, e); |
8421 | } |
8422 | |
8423 | #[simd_test(enable = "avx512f" )] |
8424 | unsafe fn test_mm512_sll_epi64() { |
8425 | #[rustfmt::skip] |
8426 | let a = _mm512_set_epi64( |
8427 | 1 << 63, 1 << 32, 1 << 32, 1 << 32, |
8428 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8429 | ); |
8430 | let count = _mm_set_epi64x(0, 1); |
8431 | let r = _mm512_sll_epi64(a, count); |
8432 | #[rustfmt::skip] |
8433 | let e = _mm512_set_epi64( |
8434 | 0, 1 << 33, 1 << 33, 1 << 33, |
8435 | 1 << 33, 1 << 33, 1 << 33, 1 << 33, |
8436 | ); |
8437 | assert_eq_m512i(r, e); |
8438 | let count = _mm_set_epi64x(1, 0); |
8439 | let r = _mm512_sll_epi64(a, count); |
8440 | assert_eq_m512i(r, a); |
8441 | } |
8442 | |
8443 | #[simd_test(enable = "avx512f" )] |
8444 | unsafe fn test_mm512_mask_sll_epi64() { |
8445 | #[rustfmt::skip] |
8446 | let a = _mm512_set_epi64( |
8447 | 1 << 63, 1 << 32, 1 << 32, 1 << 32, |
8448 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8449 | ); |
8450 | let count = _mm_set_epi64x(0, 1); |
8451 | let r = _mm512_mask_sll_epi64(a, 0, a, count); |
8452 | assert_eq_m512i(r, a); |
8453 | let r = _mm512_mask_sll_epi64(a, 0b11111111, a, count); |
8454 | #[rustfmt::skip] |
8455 | let e = _mm512_set_epi64( |
8456 | 0, 1 << 33, 1 << 33, 1 << 33, |
8457 | 1 << 33, 1 << 33, 1 << 33, 1 << 33, |
8458 | ); |
8459 | assert_eq_m512i(r, e); |
8460 | } |
8461 | |
8462 | #[simd_test(enable = "avx512f" )] |
8463 | unsafe fn test_mm512_maskz_sll_epi64() { |
8464 | #[rustfmt::skip] |
8465 | let a = _mm512_set_epi64( |
8466 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8467 | 1 << 32, 1 << 32, 1 << 32, 1 << 63, |
8468 | ); |
8469 | let count = _mm_set_epi64x(0, 1); |
8470 | let r = _mm512_maskz_sll_epi64(0, a, count); |
8471 | assert_eq_m512i(r, _mm512_setzero_si512()); |
8472 | let r = _mm512_maskz_sll_epi64(0b00001111, a, count); |
8473 | let e = _mm512_set_epi64(0, 0, 0, 0, 1 << 33, 1 << 33, 1 << 33, 0); |
8474 | assert_eq_m512i(r, e); |
8475 | } |
8476 | |
8477 | #[simd_test(enable = "avx512f,avx512vl" )] |
8478 | unsafe fn test_mm256_mask_sll_epi64() { |
8479 | let a = _mm256_set_epi64x(1 << 63, 1 << 32, 1 << 32, 1 << 32); |
8480 | let count = _mm_set_epi64x(0, 1); |
8481 | let r = _mm256_mask_sll_epi64(a, 0, a, count); |
8482 | assert_eq_m256i(r, a); |
8483 | let r = _mm256_mask_sll_epi64(a, 0b00001111, a, count); |
8484 | let e = _mm256_set_epi64x(0, 1 << 33, 1 << 33, 1 << 33); |
8485 | assert_eq_m256i(r, e); |
8486 | } |
8487 | |
8488 | #[simd_test(enable = "avx512f,avx512vl" )] |
8489 | unsafe fn test_mm256_maskz_sll_epi64() { |
8490 | let a = _mm256_set_epi64x(1 << 63, 1 << 32, 1 << 32, 1 << 32); |
8491 | let count = _mm_set_epi64x(0, 1); |
8492 | let r = _mm256_maskz_sll_epi64(0, a, count); |
8493 | assert_eq_m256i(r, _mm256_setzero_si256()); |
8494 | let r = _mm256_maskz_sll_epi64(0b00001111, a, count); |
8495 | let e = _mm256_set_epi64x(0, 1 << 33, 1 << 33, 1 << 33); |
8496 | assert_eq_m256i(r, e); |
8497 | } |
8498 | |
8499 | #[simd_test(enable = "avx512f,avx512vl" )] |
8500 | unsafe fn test_mm_mask_sll_epi64() { |
8501 | let a = _mm_set_epi64x(1 << 63, 1 << 32); |
8502 | let count = _mm_set_epi64x(0, 1); |
8503 | let r = _mm_mask_sll_epi64(a, 0, a, count); |
8504 | assert_eq_m128i(r, a); |
8505 | let r = _mm_mask_sll_epi64(a, 0b00000011, a, count); |
8506 | let e = _mm_set_epi64x(0, 1 << 33); |
8507 | assert_eq_m128i(r, e); |
8508 | } |
8509 | |
8510 | #[simd_test(enable = "avx512f,avx512vl" )] |
8511 | unsafe fn test_mm_maskz_sll_epi64() { |
8512 | let a = _mm_set_epi64x(1 << 63, 1 << 32); |
8513 | let count = _mm_set_epi64x(0, 1); |
8514 | let r = _mm_maskz_sll_epi64(0, a, count); |
8515 | assert_eq_m128i(r, _mm_setzero_si128()); |
8516 | let r = _mm_maskz_sll_epi64(0b00000011, a, count); |
8517 | let e = _mm_set_epi64x(0, 1 << 33); |
8518 | assert_eq_m128i(r, e); |
8519 | } |
8520 | |
8521 | #[simd_test(enable = "avx512f" )] |
8522 | unsafe fn test_mm512_srl_epi64() { |
8523 | #[rustfmt::skip] |
8524 | let a = _mm512_set_epi64( |
8525 | 1 << 0, 1 << 32, 1 << 32, 1 << 32, |
8526 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8527 | ); |
8528 | let count = _mm_set_epi64x(0, 1); |
8529 | let r = _mm512_srl_epi64(a, count); |
8530 | #[rustfmt::skip] |
8531 | let e = _mm512_set_epi64( |
8532 | 0, 1 << 31, 1 << 31, 1 << 31, |
8533 | 1 << 31, 1 << 31, 1 << 31, 1 << 31, |
8534 | ); |
8535 | assert_eq_m512i(r, e); |
8536 | } |
8537 | |
8538 | #[simd_test(enable = "avx512f" )] |
8539 | unsafe fn test_mm512_mask_srl_epi64() { |
8540 | #[rustfmt::skip] |
8541 | let a = _mm512_set_epi64( |
8542 | 1 << 0, 1 << 32, 1 << 32, 1 << 32, |
8543 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8544 | ); |
8545 | let count = _mm_set_epi64x(0, 1); |
8546 | let r = _mm512_mask_srl_epi64(a, 0, a, count); |
8547 | assert_eq_m512i(r, a); |
8548 | let r = _mm512_mask_srl_epi64(a, 0b11111111, a, count); |
8549 | #[rustfmt::skip] |
8550 | let e = _mm512_set_epi64( |
8551 | 0, 1 << 31, 1 << 31, 1 << 31, |
8552 | 1 << 31, 1 << 31, 1 << 31, 1 << 31, |
8553 | ); |
8554 | assert_eq_m512i(r, e); |
8555 | } |
8556 | |
8557 | #[simd_test(enable = "avx512f" )] |
8558 | unsafe fn test_mm512_maskz_srl_epi64() { |
8559 | #[rustfmt::skip] |
8560 | let a = _mm512_set_epi64( |
8561 | 1 << 32, 1 << 32, 1 << 32, 1 << 32, |
8562 | 1 << 32, 1 << 32, 1 << 32, 1 << 0, |
8563 | ); |
8564 | let count = _mm_set_epi64x(0, 1); |
8565 | let r = _mm512_maskz_srl_epi64(0, a, count); |
8566 | assert_eq_m512i(r, _mm512_setzero_si512()); |
8567 | let r = _mm512_maskz_srl_epi64(0b00001111, a, count); |
8568 | let e = _mm512_set_epi64(0, 0, 0, 0, 1 << 31, 1 << 31, 1 << 31, 0); |
8569 | assert_eq_m512i(r, e); |
8570 | } |
8571 | |
8572 | #[simd_test(enable = "avx512f,avx512vl" )] |
8573 | unsafe fn test_mm256_mask_srl_epi64() { |
8574 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8575 | let count = _mm_set_epi64x(0, 1); |
8576 | let r = _mm256_mask_srl_epi64(a, 0, a, count); |
8577 | assert_eq_m256i(r, a); |
8578 | let r = _mm256_mask_srl_epi64(a, 0b00001111, a, count); |
8579 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8580 | assert_eq_m256i(r, e); |
8581 | } |
8582 | |
8583 | #[simd_test(enable = "avx512f,avx512vl" )] |
8584 | unsafe fn test_mm256_maskz_srl_epi64() { |
8585 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8586 | let count = _mm_set_epi64x(0, 1); |
8587 | let r = _mm256_maskz_srl_epi64(0, a, count); |
8588 | assert_eq_m256i(r, _mm256_setzero_si256()); |
8589 | let r = _mm256_maskz_srl_epi64(0b00001111, a, count); |
8590 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8591 | assert_eq_m256i(r, e); |
8592 | } |
8593 | |
8594 | #[simd_test(enable = "avx512f,avx512vl" )] |
8595 | unsafe fn test_mm_mask_srl_epi64() { |
8596 | let a = _mm_set_epi64x(1 << 5, 0); |
8597 | let count = _mm_set_epi64x(0, 1); |
8598 | let r = _mm_mask_srl_epi64(a, 0, a, count); |
8599 | assert_eq_m128i(r, a); |
8600 | let r = _mm_mask_srl_epi64(a, 0b00000011, a, count); |
8601 | let e = _mm_set_epi64x(1 << 4, 0); |
8602 | assert_eq_m128i(r, e); |
8603 | } |
8604 | |
8605 | #[simd_test(enable = "avx512f,avx512vl" )] |
8606 | unsafe fn test_mm_maskz_srl_epi64() { |
8607 | let a = _mm_set_epi64x(1 << 5, 0); |
8608 | let count = _mm_set_epi64x(0, 1); |
8609 | let r = _mm_maskz_srl_epi64(0, a, count); |
8610 | assert_eq_m128i(r, _mm_setzero_si128()); |
8611 | let r = _mm_maskz_srl_epi64(0b00000011, a, count); |
8612 | let e = _mm_set_epi64x(1 << 4, 0); |
8613 | assert_eq_m128i(r, e); |
8614 | } |
8615 | |
8616 | #[simd_test(enable = "avx512f" )] |
8617 | unsafe fn test_mm512_sra_epi64() { |
8618 | let a = _mm512_set_epi64(1, -8, 0, 0, 0, 0, 15, -16); |
8619 | let count = _mm_set_epi64x(0, 2); |
8620 | let r = _mm512_sra_epi64(a, count); |
8621 | let e = _mm512_set_epi64(0, -2, 0, 0, 0, 0, 3, -4); |
8622 | assert_eq_m512i(r, e); |
8623 | } |
8624 | |
8625 | #[simd_test(enable = "avx512f" )] |
8626 | unsafe fn test_mm512_mask_sra_epi64() { |
8627 | let a = _mm512_set_epi64(1, -8, 0, 0, 0, 0, 15, -16); |
8628 | let count = _mm_set_epi64x(0, 2); |
8629 | let r = _mm512_mask_sra_epi64(a, 0, a, count); |
8630 | assert_eq_m512i(r, a); |
8631 | let r = _mm512_mask_sra_epi64(a, 0b11111111, a, count); |
8632 | let e = _mm512_set_epi64(0, -2, 0, 0, 0, 0, 3, -4); |
8633 | assert_eq_m512i(r, e); |
8634 | } |
8635 | |
8636 | #[simd_test(enable = "avx512f" )] |
8637 | unsafe fn test_mm512_maskz_sra_epi64() { |
8638 | let a = _mm512_set_epi64(1, -8, 0, 0, 0, 0, 15, -16); |
8639 | let count = _mm_set_epi64x(0, 2); |
8640 | let r = _mm512_maskz_sra_epi64(0, a, count); |
8641 | assert_eq_m512i(r, _mm512_setzero_si512()); |
8642 | let r = _mm512_maskz_sra_epi64(0b00001111, a, count); |
8643 | let e = _mm512_set_epi64(0, 0, 0, 0, 0, 0, 3, -4); |
8644 | assert_eq_m512i(r, e); |
8645 | } |
8646 | |
8647 | #[simd_test(enable = "avx512f,avx512vl" )] |
8648 | unsafe fn test_mm256_sra_epi64() { |
8649 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8650 | let count = _mm_set_epi64x(0, 1); |
8651 | let r = _mm256_sra_epi64(a, count); |
8652 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8653 | assert_eq_m256i(r, e); |
8654 | } |
8655 | |
8656 | #[simd_test(enable = "avx512f,avx512vl" )] |
8657 | unsafe fn test_mm256_mask_sra_epi64() { |
8658 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8659 | let count = _mm_set_epi64x(0, 1); |
8660 | let r = _mm256_mask_sra_epi64(a, 0, a, count); |
8661 | assert_eq_m256i(r, a); |
8662 | let r = _mm256_mask_sra_epi64(a, 0b00001111, a, count); |
8663 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8664 | assert_eq_m256i(r, e); |
8665 | } |
8666 | |
8667 | #[simd_test(enable = "avx512f,avx512vl" )] |
8668 | unsafe fn test_mm256_maskz_sra_epi64() { |
8669 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8670 | let count = _mm_set_epi64x(0, 1); |
8671 | let r = _mm256_maskz_sra_epi64(0, a, count); |
8672 | assert_eq_m256i(r, _mm256_setzero_si256()); |
8673 | let r = _mm256_maskz_sra_epi64(0b00001111, a, count); |
8674 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8675 | assert_eq_m256i(r, e); |
8676 | } |
8677 | |
8678 | #[simd_test(enable = "avx512f,avx512vl" )] |
8679 | unsafe fn test_mm_sra_epi64() { |
8680 | let a = _mm_set_epi64x(1 << 5, 0); |
8681 | let count = _mm_set_epi64x(0, 1); |
8682 | let r = _mm_sra_epi64(a, count); |
8683 | let e = _mm_set_epi64x(1 << 4, 0); |
8684 | assert_eq_m128i(r, e); |
8685 | } |
8686 | |
8687 | #[simd_test(enable = "avx512f,avx512vl" )] |
8688 | unsafe fn test_mm_mask_sra_epi64() { |
8689 | let a = _mm_set_epi64x(1 << 5, 0); |
8690 | let count = _mm_set_epi64x(0, 1); |
8691 | let r = _mm_mask_sra_epi64(a, 0, a, count); |
8692 | assert_eq_m128i(r, a); |
8693 | let r = _mm_mask_sra_epi64(a, 0b00000011, a, count); |
8694 | let e = _mm_set_epi64x(1 << 4, 0); |
8695 | assert_eq_m128i(r, e); |
8696 | } |
8697 | |
8698 | #[simd_test(enable = "avx512f,avx512vl" )] |
8699 | unsafe fn test_mm_maskz_sra_epi64() { |
8700 | let a = _mm_set_epi64x(1 << 5, 0); |
8701 | let count = _mm_set_epi64x(0, 1); |
8702 | let r = _mm_maskz_sra_epi64(0, a, count); |
8703 | assert_eq_m128i(r, _mm_setzero_si128()); |
8704 | let r = _mm_maskz_sra_epi64(0b00000011, a, count); |
8705 | let e = _mm_set_epi64x(1 << 4, 0); |
8706 | assert_eq_m128i(r, e); |
8707 | } |
8708 | |
8709 | #[simd_test(enable = "avx512f" )] |
8710 | unsafe fn test_mm512_srav_epi64() { |
8711 | let a = _mm512_set_epi64(1, -8, 0, 0, 0, 0, 15, -16); |
8712 | let count = _mm512_set_epi64(2, 2, 0, 0, 0, 0, 2, 1); |
8713 | let r = _mm512_srav_epi64(a, count); |
8714 | let e = _mm512_set_epi64(0, -2, 0, 0, 0, 0, 3, -8); |
8715 | assert_eq_m512i(r, e); |
8716 | } |
8717 | |
8718 | #[simd_test(enable = "avx512f" )] |
8719 | unsafe fn test_mm512_mask_srav_epi64() { |
8720 | let a = _mm512_set_epi64(1, -8, 0, 0, 0, 0, 15, -16); |
8721 | let count = _mm512_set_epi64(2, 2, 0, 0, 0, 0, 2, 1); |
8722 | let r = _mm512_mask_srav_epi64(a, 0, a, count); |
8723 | assert_eq_m512i(r, a); |
8724 | let r = _mm512_mask_srav_epi64(a, 0b11111111, a, count); |
8725 | let e = _mm512_set_epi64(0, -2, 0, 0, 0, 0, 3, -8); |
8726 | assert_eq_m512i(r, e); |
8727 | } |
8728 | |
8729 | #[simd_test(enable = "avx512f" )] |
8730 | unsafe fn test_mm512_maskz_srav_epi64() { |
8731 | let a = _mm512_set_epi64(1, -8, 0, 0, 0, 0, 15, -16); |
8732 | let count = _mm512_set_epi64(2, 2, 0, 0, 0, 0, 2, 1); |
8733 | let r = _mm512_maskz_srav_epi64(0, a, count); |
8734 | assert_eq_m512i(r, _mm512_setzero_si512()); |
8735 | let r = _mm512_maskz_srav_epi64(0b00001111, a, count); |
8736 | let e = _mm512_set_epi64(0, 0, 0, 0, 0, 0, 3, -8); |
8737 | assert_eq_m512i(r, e); |
8738 | } |
8739 | |
8740 | #[simd_test(enable = "avx512f,avx512vl" )] |
8741 | unsafe fn test_mm256_srav_epi64() { |
8742 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8743 | let count = _mm256_set1_epi64x(1); |
8744 | let r = _mm256_srav_epi64(a, count); |
8745 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8746 | assert_eq_m256i(r, e); |
8747 | } |
8748 | |
8749 | #[simd_test(enable = "avx512f,avx512vl" )] |
8750 | unsafe fn test_mm256_mask_srav_epi64() { |
8751 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8752 | let count = _mm256_set1_epi64x(1); |
8753 | let r = _mm256_mask_srav_epi64(a, 0, a, count); |
8754 | assert_eq_m256i(r, a); |
8755 | let r = _mm256_mask_srav_epi64(a, 0b00001111, a, count); |
8756 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8757 | assert_eq_m256i(r, e); |
8758 | } |
8759 | |
8760 | #[simd_test(enable = "avx512f,avx512vl" )] |
8761 | unsafe fn test_mm256_maskz_srav_epi64() { |
8762 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8763 | let count = _mm256_set1_epi64x(1); |
8764 | let r = _mm256_maskz_srav_epi64(0, a, count); |
8765 | assert_eq_m256i(r, _mm256_setzero_si256()); |
8766 | let r = _mm256_maskz_srav_epi64(0b00001111, a, count); |
8767 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8768 | assert_eq_m256i(r, e); |
8769 | } |
8770 | |
8771 | #[simd_test(enable = "avx512f,avx512vl" )] |
8772 | unsafe fn test_mm_srav_epi64() { |
8773 | let a = _mm_set_epi64x(1 << 5, 0); |
8774 | let count = _mm_set1_epi64x(1); |
8775 | let r = _mm_srav_epi64(a, count); |
8776 | let e = _mm_set_epi64x(1 << 4, 0); |
8777 | assert_eq_m128i(r, e); |
8778 | } |
8779 | |
8780 | #[simd_test(enable = "avx512f,avx512vl" )] |
8781 | unsafe fn test_mm_mask_srav_epi64() { |
8782 | let a = _mm_set_epi64x(1 << 5, 0); |
8783 | let count = _mm_set1_epi64x(1); |
8784 | let r = _mm_mask_srav_epi64(a, 0, a, count); |
8785 | assert_eq_m128i(r, a); |
8786 | let r = _mm_mask_srav_epi64(a, 0b00000011, a, count); |
8787 | let e = _mm_set_epi64x(1 << 4, 0); |
8788 | assert_eq_m128i(r, e); |
8789 | } |
8790 | |
8791 | #[simd_test(enable = "avx512f,avx512vl" )] |
8792 | unsafe fn test_mm_maskz_srav_epi64() { |
8793 | let a = _mm_set_epi64x(1 << 5, 0); |
8794 | let count = _mm_set1_epi64x(1); |
8795 | let r = _mm_maskz_srav_epi64(0, a, count); |
8796 | assert_eq_m128i(r, _mm_setzero_si128()); |
8797 | let r = _mm_maskz_srav_epi64(0b00000011, a, count); |
8798 | let e = _mm_set_epi64x(1 << 4, 0); |
8799 | assert_eq_m128i(r, e); |
8800 | } |
8801 | |
8802 | #[simd_test(enable = "avx512f" )] |
8803 | unsafe fn test_mm512_srai_epi64() { |
8804 | let a = _mm512_set_epi64(1, -4, 15, 0, 0, 0, 0, -16); |
8805 | let r = _mm512_srai_epi64::<2>(a); |
8806 | let e = _mm512_set_epi64(0, -1, 3, 0, 0, 0, 0, -4); |
8807 | assert_eq_m512i(r, e); |
8808 | } |
8809 | |
8810 | #[simd_test(enable = "avx512f" )] |
8811 | unsafe fn test_mm512_mask_srai_epi64() { |
8812 | let a = _mm512_set_epi64(1, -4, 15, 0, 0, 0, 0, -16); |
8813 | let r = _mm512_mask_srai_epi64::<2>(a, 0, a); |
8814 | assert_eq_m512i(r, a); |
8815 | let r = _mm512_mask_srai_epi64::<2>(a, 0b11111111, a); |
8816 | let e = _mm512_set_epi64(0, -1, 3, 0, 0, 0, 0, -4); |
8817 | assert_eq_m512i(r, e); |
8818 | } |
8819 | |
8820 | #[simd_test(enable = "avx512f" )] |
8821 | unsafe fn test_mm512_maskz_srai_epi64() { |
8822 | let a = _mm512_set_epi64(1, -4, 15, 0, 0, 0, 0, -16); |
8823 | let r = _mm512_maskz_srai_epi64::<2>(0, a); |
8824 | assert_eq_m512i(r, _mm512_setzero_si512()); |
8825 | let r = _mm512_maskz_srai_epi64::<2>(0b00001111, a); |
8826 | let e = _mm512_set_epi64(0, 0, 0, 0, 0, 0, 0, -4); |
8827 | assert_eq_m512i(r, e); |
8828 | } |
8829 | |
8830 | #[simd_test(enable = "avx512f,avx512vl" )] |
8831 | unsafe fn test_mm256_srai_epi64() { |
8832 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8833 | let r = _mm256_srai_epi64::<1>(a); |
8834 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8835 | assert_eq_m256i(r, e); |
8836 | } |
8837 | |
8838 | #[simd_test(enable = "avx512f,avx512vl" )] |
8839 | unsafe fn test_mm256_mask_srai_epi64() { |
8840 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8841 | let r = _mm256_mask_srai_epi64::<1>(a, 0, a); |
8842 | assert_eq_m256i(r, a); |
8843 | let r = _mm256_mask_srai_epi64::<1>(a, 0b00001111, a); |
8844 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8845 | assert_eq_m256i(r, e); |
8846 | } |
8847 | |
8848 | #[simd_test(enable = "avx512f,avx512vl" )] |
8849 | unsafe fn test_mm256_maskz_srai_epi64() { |
8850 | let a = _mm256_set_epi64x(1 << 5, 0, 0, 0); |
8851 | let r = _mm256_maskz_srai_epi64::<1>(0, a); |
8852 | assert_eq_m256i(r, _mm256_setzero_si256()); |
8853 | let r = _mm256_maskz_srai_epi64::<1>(0b00001111, a); |
8854 | let e = _mm256_set_epi64x(1 << 4, 0, 0, 0); |
8855 | assert_eq_m256i(r, e); |
8856 | } |
8857 | |
8858 | #[simd_test(enable = "avx512f,avx512vl" )] |
8859 | unsafe fn test_mm_srai_epi64() { |
8860 | let a = _mm_set_epi64x(1 << 5, 0); |
8861 | let r = _mm_srai_epi64::<1>(a); |
8862 | let e = _mm_set_epi64x(1 << 4, 0); |
8863 | assert_eq_m128i(r, e); |
8864 | } |
8865 | |
8866 | #[simd_test(enable = "avx512f,avx512vl" )] |
8867 | unsafe fn test_mm_mask_srai_epi64() { |
8868 | let a = _mm_set_epi64x(1 << 5, 0); |
8869 | let r = _mm_mask_srai_epi64::<1>(a, 0, a); |
8870 | assert_eq_m128i(r, a); |
8871 | let r = _mm_mask_srai_epi64::<1>(a, 0b00000011, a); |
8872 | let e = _mm_set_epi64x(1 << 4, 0); |
8873 | assert_eq_m128i(r, e); |
8874 | } |
8875 | |
8876 | #[simd_test(enable = "avx512f,avx512vl" )] |
8877 | unsafe fn test_mm_maskz_srai_epi64() { |
8878 | let a = _mm_set_epi64x(1 << 5, 0); |
8879 | let r = _mm_maskz_srai_epi64::<1>(0, a); |
8880 | assert_eq_m128i(r, _mm_setzero_si128()); |
8881 | let r = _mm_maskz_srai_epi64::<1>(0b00000011, a); |
8882 | let e = _mm_set_epi64x(1 << 4, 0); |
8883 | assert_eq_m128i(r, e); |
8884 | } |
8885 | |
8886 | #[simd_test(enable = "avx512f" )] |
8887 | unsafe fn test_mm512_permute_pd() { |
8888 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
8889 | let r = _mm512_permute_pd::<0b11_11_11_11>(a); |
8890 | let e = _mm512_setr_pd(1., 1., 3., 3., 5., 5., 7., 7.); |
8891 | assert_eq_m512d(r, e); |
8892 | } |
8893 | |
8894 | #[simd_test(enable = "avx512f" )] |
8895 | unsafe fn test_mm512_mask_permute_pd() { |
8896 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
8897 | let r = _mm512_mask_permute_pd::<0b11_11_11_11>(a, 0, a); |
8898 | assert_eq_m512d(r, a); |
8899 | let r = _mm512_mask_permute_pd::<0b11_11_11_11>(a, 0b11111111, a); |
8900 | let e = _mm512_setr_pd(1., 1., 3., 3., 5., 5., 7., 7.); |
8901 | assert_eq_m512d(r, e); |
8902 | } |
8903 | |
8904 | #[simd_test(enable = "avx512f" )] |
8905 | unsafe fn test_mm512_maskz_permute_pd() { |
8906 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
8907 | let r = _mm512_maskz_permute_pd::<0b11_11_11_11>(0, a); |
8908 | assert_eq_m512d(r, _mm512_setzero_pd()); |
8909 | let r = _mm512_maskz_permute_pd::<0b11_11_11_11>(0b11111111, a); |
8910 | let e = _mm512_setr_pd(1., 1., 3., 3., 5., 5., 7., 7.); |
8911 | assert_eq_m512d(r, e); |
8912 | } |
8913 | |
8914 | #[simd_test(enable = "avx512f,avx512vl" )] |
8915 | unsafe fn test_mm256_mask_permute_pd() { |
8916 | let a = _mm256_set_pd(3., 2., 1., 0.); |
8917 | let r = _mm256_mask_permute_pd::<0b11_11>(a, 0, a); |
8918 | assert_eq_m256d(r, a); |
8919 | let r = _mm256_mask_permute_pd::<0b11_11>(a, 0b00001111, a); |
8920 | let e = _mm256_set_pd(3., 3., 1., 1.); |
8921 | assert_eq_m256d(r, e); |
8922 | } |
8923 | |
8924 | #[simd_test(enable = "avx512f,avx512vl" )] |
8925 | unsafe fn test_mm256_maskz_permute_pd() { |
8926 | let a = _mm256_set_pd(3., 2., 1., 0.); |
8927 | let r = _mm256_maskz_permute_pd::<0b11_11>(0, a); |
8928 | assert_eq_m256d(r, _mm256_setzero_pd()); |
8929 | let r = _mm256_maskz_permute_pd::<0b11_11>(0b00001111, a); |
8930 | let e = _mm256_set_pd(3., 3., 1., 1.); |
8931 | assert_eq_m256d(r, e); |
8932 | } |
8933 | |
8934 | #[simd_test(enable = "avx512f,avx512vl" )] |
8935 | unsafe fn test_mm_mask_permute_pd() { |
8936 | let a = _mm_set_pd(1., 0.); |
8937 | let r = _mm_mask_permute_pd::<0b11>(a, 0, a); |
8938 | assert_eq_m128d(r, a); |
8939 | let r = _mm_mask_permute_pd::<0b11>(a, 0b00000011, a); |
8940 | let e = _mm_set_pd(1., 1.); |
8941 | assert_eq_m128d(r, e); |
8942 | } |
8943 | |
8944 | #[simd_test(enable = "avx512f,avx512vl" )] |
8945 | unsafe fn test_mm_maskz_permute_pd() { |
8946 | let a = _mm_set_pd(1., 0.); |
8947 | let r = _mm_maskz_permute_pd::<0b11>(0, a); |
8948 | assert_eq_m128d(r, _mm_setzero_pd()); |
8949 | let r = _mm_maskz_permute_pd::<0b11>(0b00000011, a); |
8950 | let e = _mm_set_pd(1., 1.); |
8951 | assert_eq_m128d(r, e); |
8952 | } |
8953 | |
8954 | #[simd_test(enable = "avx512f" )] |
8955 | unsafe fn test_mm512_permutex_epi64() { |
8956 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
8957 | let r = _mm512_permutex_epi64::<0b11_11_11_11>(a); |
8958 | let e = _mm512_setr_epi64(3, 3, 3, 3, 7, 7, 7, 7); |
8959 | assert_eq_m512i(r, e); |
8960 | } |
8961 | |
8962 | #[simd_test(enable = "avx512f" )] |
8963 | unsafe fn test_mm512_mask_permutex_epi64() { |
8964 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
8965 | let r = _mm512_mask_permutex_epi64::<0b11_11_11_11>(a, 0, a); |
8966 | assert_eq_m512i(r, a); |
8967 | let r = _mm512_mask_permutex_epi64::<0b11_11_11_11>(a, 0b11111111, a); |
8968 | let e = _mm512_setr_epi64(3, 3, 3, 3, 7, 7, 7, 7); |
8969 | assert_eq_m512i(r, e); |
8970 | } |
8971 | |
8972 | #[simd_test(enable = "avx512f" )] |
8973 | unsafe fn test_mm512_maskz_permutex_epi64() { |
8974 | let a = _mm512_setr_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
8975 | let r = _mm512_maskz_permutex_epi64::<0b11_11_11_11>(0, a); |
8976 | assert_eq_m512i(r, _mm512_setzero_si512()); |
8977 | let r = _mm512_maskz_permutex_epi64::<0b11_11_11_11>(0b11111111, a); |
8978 | let e = _mm512_setr_epi64(3, 3, 3, 3, 7, 7, 7, 7); |
8979 | assert_eq_m512i(r, e); |
8980 | } |
8981 | |
8982 | #[simd_test(enable = "avx512f,avx512vl" )] |
8983 | unsafe fn test_mm256_permutex_epi64() { |
8984 | let a = _mm256_set_epi64x(3, 2, 1, 0); |
8985 | let r = _mm256_permutex_epi64::<0b11_11_11_11>(a); |
8986 | let e = _mm256_set_epi64x(3, 3, 3, 3); |
8987 | assert_eq_m256i(r, e); |
8988 | } |
8989 | |
8990 | #[simd_test(enable = "avx512f,avx512vl" )] |
8991 | unsafe fn test_mm256_mask_permutex_epi64() { |
8992 | let a = _mm256_set_epi64x(3, 2, 1, 0); |
8993 | let r = _mm256_mask_permutex_epi64::<0b11_11_11_11>(a, 0, a); |
8994 | assert_eq_m256i(r, a); |
8995 | let r = _mm256_mask_permutex_epi64::<0b11_11_11_11>(a, 0b00001111, a); |
8996 | let e = _mm256_set_epi64x(3, 3, 3, 3); |
8997 | assert_eq_m256i(r, e); |
8998 | } |
8999 | |
9000 | #[simd_test(enable = "avx512f" )] |
9001 | unsafe fn test_mm256_maskz_permutex_epi64() { |
9002 | let a = _mm256_set_epi64x(3, 2, 1, 0); |
9003 | let r = _mm256_maskz_permutex_epi64::<0b11_11_11_11>(0, a); |
9004 | assert_eq_m256i(r, _mm256_setzero_si256()); |
9005 | let r = _mm256_maskz_permutex_epi64::<0b11_11_11_11>(0b00001111, a); |
9006 | let e = _mm256_set_epi64x(3, 3, 3, 3); |
9007 | assert_eq_m256i(r, e); |
9008 | } |
9009 | |
9010 | #[simd_test(enable = "avx512f" )] |
9011 | unsafe fn test_mm512_permutex_pd() { |
9012 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9013 | let r = _mm512_permutex_pd::<0b11_11_11_11>(a); |
9014 | let e = _mm512_setr_pd(3., 3., 3., 3., 7., 7., 7., 7.); |
9015 | assert_eq_m512d(r, e); |
9016 | } |
9017 | |
9018 | #[simd_test(enable = "avx512f" )] |
9019 | unsafe fn test_mm512_mask_permutex_pd() { |
9020 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9021 | let r = _mm512_mask_permutex_pd::<0b11_11_11_11>(a, 0, a); |
9022 | assert_eq_m512d(r, a); |
9023 | let r = _mm512_mask_permutex_pd::<0b11_11_11_11>(a, 0b11111111, a); |
9024 | let e = _mm512_setr_pd(3., 3., 3., 3., 7., 7., 7., 7.); |
9025 | assert_eq_m512d(r, e); |
9026 | } |
9027 | |
9028 | #[simd_test(enable = "avx512f" )] |
9029 | unsafe fn test_mm512_maskz_permutex_pd() { |
9030 | let a = _mm512_setr_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9031 | let r = _mm512_maskz_permutex_pd::<0b11_11_11_11>(0, a); |
9032 | assert_eq_m512d(r, _mm512_setzero_pd()); |
9033 | let r = _mm512_maskz_permutex_pd::<0b11_11_11_11>(0b11111111, a); |
9034 | let e = _mm512_setr_pd(3., 3., 3., 3., 7., 7., 7., 7.); |
9035 | assert_eq_m512d(r, e); |
9036 | } |
9037 | |
9038 | #[simd_test(enable = "avx512f,avx512vl" )] |
9039 | unsafe fn test_mm256_permutex_pd() { |
9040 | let a = _mm256_set_pd(0., 1., 2., 3.); |
9041 | let r = _mm256_permutex_pd::<0b11_11_11_11>(a); |
9042 | let e = _mm256_set_pd(0., 0., 0., 0.); |
9043 | assert_eq_m256d(r, e); |
9044 | } |
9045 | |
9046 | #[simd_test(enable = "avx512f,avx512vl" )] |
9047 | unsafe fn test_mm256_mask_permutex_pd() { |
9048 | let a = _mm256_set_pd(0., 1., 2., 3.); |
9049 | let r = _mm256_mask_permutex_pd::<0b11_11_11_11>(a, 0, a); |
9050 | assert_eq_m256d(r, a); |
9051 | let r = _mm256_mask_permutex_pd::<0b11_11_11_11>(a, 0b00001111, a); |
9052 | let e = _mm256_set_pd(0., 0., 0., 0.); |
9053 | assert_eq_m256d(r, e); |
9054 | } |
9055 | |
9056 | #[simd_test(enable = "avx512f,avx512vl" )] |
9057 | unsafe fn test_mm256_maskz_permutex_pd() { |
9058 | let a = _mm256_set_pd(0., 1., 2., 3.); |
9059 | let r = _mm256_maskz_permutex_pd::<0b11_11_11_11>(0, a); |
9060 | assert_eq_m256d(r, _mm256_setzero_pd()); |
9061 | let r = _mm256_maskz_permutex_pd::<0b11_11_11_11>(0b00001111, a); |
9062 | let e = _mm256_set_pd(0., 0., 0., 0.); |
9063 | assert_eq_m256d(r, e); |
9064 | } |
9065 | |
9066 | #[simd_test(enable = "avx512f" )] |
9067 | unsafe fn test_mm512_permutevar_pd() { |
9068 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9069 | let b = _mm512_set1_epi64(0b1); |
9070 | let r = _mm512_permutevar_pd(a, b); |
9071 | let e = _mm512_set_pd(1., 1., 3., 3., 5., 5., 7., 7.); |
9072 | assert_eq_m512d(r, e); |
9073 | } |
9074 | |
9075 | #[simd_test(enable = "avx512f" )] |
9076 | unsafe fn test_mm512_mask_permutevar_pd() { |
9077 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9078 | let b = _mm512_set1_epi64(0b1); |
9079 | let r = _mm512_mask_permutevar_pd(a, 0, a, b); |
9080 | assert_eq_m512d(r, a); |
9081 | let r = _mm512_mask_permutevar_pd(a, 0b11111111, a, b); |
9082 | let e = _mm512_set_pd(1., 1., 3., 3., 5., 5., 7., 7.); |
9083 | assert_eq_m512d(r, e); |
9084 | } |
9085 | |
9086 | #[simd_test(enable = "avx512f" )] |
9087 | unsafe fn test_mm512_maskz_permutevar_pd() { |
9088 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9089 | let b = _mm512_set1_epi64(0b1); |
9090 | let r = _mm512_maskz_permutevar_pd(0, a, b); |
9091 | assert_eq_m512d(r, _mm512_setzero_pd()); |
9092 | let r = _mm512_maskz_permutevar_pd(0b00001111, a, b); |
9093 | let e = _mm512_set_pd(0., 0., 0., 0., 5., 5., 7., 7.); |
9094 | assert_eq_m512d(r, e); |
9095 | } |
9096 | |
9097 | #[simd_test(enable = "avx512f,avx512vl" )] |
9098 | unsafe fn test_mm256_mask_permutevar_pd() { |
9099 | let a = _mm256_set_pd(0., 1., 2., 3.); |
9100 | let b = _mm256_set1_epi64x(0b1); |
9101 | let r = _mm256_mask_permutevar_pd(a, 0, a, b); |
9102 | assert_eq_m256d(r, a); |
9103 | let r = _mm256_mask_permutevar_pd(a, 0b00001111, a, b); |
9104 | let e = _mm256_set_pd(1., 1., 3., 3.); |
9105 | assert_eq_m256d(r, e); |
9106 | } |
9107 | |
9108 | #[simd_test(enable = "avx512f,avx512vl" )] |
9109 | unsafe fn test_mm256_maskz_permutevar_pd() { |
9110 | let a = _mm256_set_pd(0., 1., 2., 3.); |
9111 | let b = _mm256_set1_epi64x(0b1); |
9112 | let r = _mm256_maskz_permutevar_pd(0, a, b); |
9113 | assert_eq_m256d(r, _mm256_setzero_pd()); |
9114 | let r = _mm256_maskz_permutevar_pd(0b00001111, a, b); |
9115 | let e = _mm256_set_pd(1., 1., 3., 3.); |
9116 | assert_eq_m256d(r, e); |
9117 | } |
9118 | |
9119 | #[simd_test(enable = "avx512f,avx512vl" )] |
9120 | unsafe fn test_mm_mask_permutevar_pd() { |
9121 | let a = _mm_set_pd(0., 1.); |
9122 | let b = _mm_set1_epi64x(0b1); |
9123 | let r = _mm_mask_permutevar_pd(a, 0, a, b); |
9124 | assert_eq_m128d(r, a); |
9125 | let r = _mm_mask_permutevar_pd(a, 0b00000011, a, b); |
9126 | let e = _mm_set_pd(1., 1.); |
9127 | assert_eq_m128d(r, e); |
9128 | } |
9129 | |
9130 | #[simd_test(enable = "avx512f,avx512vl" )] |
9131 | unsafe fn test_mm_maskz_permutevar_pd() { |
9132 | let a = _mm_set_pd(0., 1.); |
9133 | let b = _mm_set1_epi64x(0b1); |
9134 | let r = _mm_maskz_permutevar_pd(0, a, b); |
9135 | assert_eq_m128d(r, _mm_setzero_pd()); |
9136 | let r = _mm_maskz_permutevar_pd(0b00000011, a, b); |
9137 | let e = _mm_set_pd(1., 1.); |
9138 | assert_eq_m128d(r, e); |
9139 | } |
9140 | |
9141 | #[simd_test(enable = "avx512f" )] |
9142 | unsafe fn test_mm512_permutexvar_epi64() { |
9143 | let idx = _mm512_set1_epi64(1); |
9144 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
9145 | let r = _mm512_permutexvar_epi64(idx, a); |
9146 | let e = _mm512_set1_epi64(6); |
9147 | assert_eq_m512i(r, e); |
9148 | } |
9149 | |
9150 | #[simd_test(enable = "avx512f" )] |
9151 | unsafe fn test_mm512_mask_permutexvar_epi64() { |
9152 | let idx = _mm512_set1_epi64(1); |
9153 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
9154 | let r = _mm512_mask_permutexvar_epi64(a, 0, idx, a); |
9155 | assert_eq_m512i(r, a); |
9156 | let r = _mm512_mask_permutexvar_epi64(a, 0b11111111, idx, a); |
9157 | let e = _mm512_set1_epi64(6); |
9158 | assert_eq_m512i(r, e); |
9159 | } |
9160 | |
9161 | #[simd_test(enable = "avx512f" )] |
9162 | unsafe fn test_mm512_maskz_permutexvar_epi64() { |
9163 | let idx = _mm512_set1_epi64(1); |
9164 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
9165 | let r = _mm512_maskz_permutexvar_epi64(0, idx, a); |
9166 | assert_eq_m512i(r, _mm512_setzero_si512()); |
9167 | let r = _mm512_maskz_permutexvar_epi64(0b00001111, idx, a); |
9168 | let e = _mm512_set_epi64(0, 0, 0, 0, 6, 6, 6, 6); |
9169 | assert_eq_m512i(r, e); |
9170 | } |
9171 | |
9172 | #[simd_test(enable = "avx512f,avx512vl" )] |
9173 | unsafe fn test_mm256_permutexvar_epi64() { |
9174 | let idx = _mm256_set1_epi64x(1); |
9175 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
9176 | let r = _mm256_permutexvar_epi64(idx, a); |
9177 | let e = _mm256_set1_epi64x(2); |
9178 | assert_eq_m256i(r, e); |
9179 | } |
9180 | |
9181 | #[simd_test(enable = "avx512f,avx512vl" )] |
9182 | unsafe fn test_mm256_mask_permutexvar_epi64() { |
9183 | let idx = _mm256_set1_epi64x(1); |
9184 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
9185 | let r = _mm256_mask_permutexvar_epi64(a, 0, idx, a); |
9186 | assert_eq_m256i(r, a); |
9187 | let r = _mm256_mask_permutexvar_epi64(a, 0b00001111, idx, a); |
9188 | let e = _mm256_set1_epi64x(2); |
9189 | assert_eq_m256i(r, e); |
9190 | } |
9191 | |
9192 | #[simd_test(enable = "avx512f,avx512vl" )] |
9193 | unsafe fn test_mm256_maskz_permutexvar_epi64() { |
9194 | let idx = _mm256_set1_epi64x(1); |
9195 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
9196 | let r = _mm256_maskz_permutexvar_epi64(0, idx, a); |
9197 | assert_eq_m256i(r, _mm256_setzero_si256()); |
9198 | let r = _mm256_maskz_permutexvar_epi64(0b00001111, idx, a); |
9199 | let e = _mm256_set1_epi64x(2); |
9200 | assert_eq_m256i(r, e); |
9201 | } |
9202 | |
9203 | #[simd_test(enable = "avx512f" )] |
9204 | unsafe fn test_mm512_permutexvar_pd() { |
9205 | let idx = _mm512_set1_epi64(1); |
9206 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9207 | let r = _mm512_permutexvar_pd(idx, a); |
9208 | let e = _mm512_set1_pd(6.); |
9209 | assert_eq_m512d(r, e); |
9210 | } |
9211 | |
9212 | #[simd_test(enable = "avx512f" )] |
9213 | unsafe fn test_mm512_mask_permutexvar_pd() { |
9214 | let idx = _mm512_set1_epi64(1); |
9215 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9216 | let r = _mm512_mask_permutexvar_pd(a, 0, idx, a); |
9217 | assert_eq_m512d(r, a); |
9218 | let r = _mm512_mask_permutexvar_pd(a, 0b11111111, idx, a); |
9219 | let e = _mm512_set1_pd(6.); |
9220 | assert_eq_m512d(r, e); |
9221 | } |
9222 | |
9223 | #[simd_test(enable = "avx512f" )] |
9224 | unsafe fn test_mm512_maskz_permutexvar_pd() { |
9225 | let idx = _mm512_set1_epi64(1); |
9226 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9227 | let r = _mm512_maskz_permutexvar_pd(0, idx, a); |
9228 | assert_eq_m512d(r, _mm512_setzero_pd()); |
9229 | let r = _mm512_maskz_permutexvar_pd(0b00001111, idx, a); |
9230 | let e = _mm512_set_pd(0., 0., 0., 0., 6., 6., 6., 6.); |
9231 | assert_eq_m512d(r, e); |
9232 | } |
9233 | |
9234 | #[simd_test(enable = "avx512f,avx512vl" )] |
9235 | unsafe fn test_mm256_permutexvar_pd() { |
9236 | let idx = _mm256_set1_epi64x(1); |
9237 | let a = _mm256_set_pd(0., 1., 2., 3.); |
9238 | let r = _mm256_permutexvar_pd(idx, a); |
9239 | let e = _mm256_set1_pd(2.); |
9240 | assert_eq_m256d(r, e); |
9241 | } |
9242 | |
9243 | #[simd_test(enable = "avx512f,avx512vl" )] |
9244 | unsafe fn test_mm256_mask_permutexvar_pd() { |
9245 | let idx = _mm256_set1_epi64x(1); |
9246 | let a = _mm256_set_pd(0., 1., 2., 3.); |
9247 | let r = _mm256_mask_permutexvar_pd(a, 0, idx, a); |
9248 | assert_eq_m256d(r, a); |
9249 | let r = _mm256_mask_permutexvar_pd(a, 0b00001111, idx, a); |
9250 | let e = _mm256_set1_pd(2.); |
9251 | assert_eq_m256d(r, e); |
9252 | } |
9253 | |
9254 | #[simd_test(enable = "avx512f,avx512vl" )] |
9255 | unsafe fn test_mm256_maskz_permutexvar_pd() { |
9256 | let idx = _mm256_set1_epi64x(1); |
9257 | let a = _mm256_set_pd(0., 1., 2., 3.); |
9258 | let r = _mm256_maskz_permutexvar_pd(0, idx, a); |
9259 | assert_eq_m256d(r, _mm256_setzero_pd()); |
9260 | let r = _mm256_maskz_permutexvar_pd(0b00001111, idx, a); |
9261 | let e = _mm256_set1_pd(2.); |
9262 | assert_eq_m256d(r, e); |
9263 | } |
9264 | |
9265 | #[simd_test(enable = "avx512f" )] |
9266 | unsafe fn test_mm512_permutex2var_epi64() { |
9267 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
9268 | let idx = _mm512_set_epi64(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
9269 | let b = _mm512_set1_epi64(100); |
9270 | let r = _mm512_permutex2var_epi64(a, idx, b); |
9271 | let e = _mm512_set_epi64(6, 100, 5, 100, 4, 100, 3, 100); |
9272 | assert_eq_m512i(r, e); |
9273 | } |
9274 | |
9275 | #[simd_test(enable = "avx512f" )] |
9276 | unsafe fn test_mm512_mask_permutex2var_epi64() { |
9277 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
9278 | let idx = _mm512_set_epi64(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
9279 | let b = _mm512_set1_epi64(100); |
9280 | let r = _mm512_mask_permutex2var_epi64(a, 0, idx, b); |
9281 | assert_eq_m512i(r, a); |
9282 | let r = _mm512_mask_permutex2var_epi64(a, 0b11111111, idx, b); |
9283 | let e = _mm512_set_epi64(6, 100, 5, 100, 4, 100, 3, 100); |
9284 | assert_eq_m512i(r, e); |
9285 | } |
9286 | |
9287 | #[simd_test(enable = "avx512f" )] |
9288 | unsafe fn test_mm512_maskz_permutex2var_epi64() { |
9289 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
9290 | let idx = _mm512_set_epi64(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
9291 | let b = _mm512_set1_epi64(100); |
9292 | let r = _mm512_maskz_permutex2var_epi64(0, a, idx, b); |
9293 | assert_eq_m512i(r, _mm512_setzero_si512()); |
9294 | let r = _mm512_maskz_permutex2var_epi64(0b00001111, a, idx, b); |
9295 | let e = _mm512_set_epi64(0, 0, 0, 0, 4, 100, 3, 100); |
9296 | assert_eq_m512i(r, e); |
9297 | } |
9298 | |
9299 | #[simd_test(enable = "avx512f" )] |
9300 | unsafe fn test_mm512_mask2_permutex2var_epi64() { |
9301 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
9302 | let idx = _mm512_set_epi64(1000, 1 << 3, 2000, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
9303 | let b = _mm512_set1_epi64(100); |
9304 | let r = _mm512_mask2_permutex2var_epi64(a, idx, 0, b); |
9305 | assert_eq_m512i(r, idx); |
9306 | let r = _mm512_mask2_permutex2var_epi64(a, idx, 0b00001111, b); |
9307 | let e = _mm512_set_epi64(1000, 1 << 3, 2000, 1 << 3, 4, 100, 3, 100); |
9308 | assert_eq_m512i(r, e); |
9309 | } |
9310 | |
9311 | #[simd_test(enable = "avx512f,avx512vl" )] |
9312 | unsafe fn test_mm256_permutex2var_epi64() { |
9313 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
9314 | let idx = _mm256_set_epi64x(1, 1 << 2, 2, 1 << 2); |
9315 | let b = _mm256_set1_epi64x(100); |
9316 | let r = _mm256_permutex2var_epi64(a, idx, b); |
9317 | let e = _mm256_set_epi64x(2, 100, 1, 100); |
9318 | assert_eq_m256i(r, e); |
9319 | } |
9320 | |
9321 | #[simd_test(enable = "avx512f,avx512vl" )] |
9322 | unsafe fn test_mm256_mask_permutex2var_epi64() { |
9323 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
9324 | let idx = _mm256_set_epi64x(1, 1 << 2, 2, 1 << 2); |
9325 | let b = _mm256_set1_epi64x(100); |
9326 | let r = _mm256_mask_permutex2var_epi64(a, 0, idx, b); |
9327 | assert_eq_m256i(r, a); |
9328 | let r = _mm256_mask_permutex2var_epi64(a, 0b00001111, idx, b); |
9329 | let e = _mm256_set_epi64x(2, 100, 1, 100); |
9330 | assert_eq_m256i(r, e); |
9331 | } |
9332 | |
9333 | #[simd_test(enable = "avx512f,avx512vl" )] |
9334 | unsafe fn test_mm256_maskz_permutex2var_epi64() { |
9335 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
9336 | let idx = _mm256_set_epi64x(1, 1 << 2, 2, 1 << 2); |
9337 | let b = _mm256_set1_epi64x(100); |
9338 | let r = _mm256_maskz_permutex2var_epi64(0, a, idx, b); |
9339 | assert_eq_m256i(r, _mm256_setzero_si256()); |
9340 | let r = _mm256_maskz_permutex2var_epi64(0b00001111, a, idx, b); |
9341 | let e = _mm256_set_epi64x(2, 100, 1, 100); |
9342 | assert_eq_m256i(r, e); |
9343 | } |
9344 | |
9345 | #[simd_test(enable = "avx512f,avx512vl" )] |
9346 | unsafe fn test_mm256_mask2_permutex2var_epi64() { |
9347 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
9348 | let idx = _mm256_set_epi64x(1, 1 << 2, 2, 1 << 2); |
9349 | let b = _mm256_set1_epi64x(100); |
9350 | let r = _mm256_mask2_permutex2var_epi64(a, idx, 0, b); |
9351 | assert_eq_m256i(r, idx); |
9352 | let r = _mm256_mask2_permutex2var_epi64(a, idx, 0b00001111, b); |
9353 | let e = _mm256_set_epi64x(2, 100, 1, 100); |
9354 | assert_eq_m256i(r, e); |
9355 | } |
9356 | |
9357 | #[simd_test(enable = "avx512f,avx512vl" )] |
9358 | unsafe fn test_mm_permutex2var_epi64() { |
9359 | let a = _mm_set_epi64x(0, 1); |
9360 | let idx = _mm_set_epi64x(1, 1 << 1); |
9361 | let b = _mm_set1_epi64x(100); |
9362 | let r = _mm_permutex2var_epi64(a, idx, b); |
9363 | let e = _mm_set_epi64x(0, 100); |
9364 | assert_eq_m128i(r, e); |
9365 | } |
9366 | |
9367 | #[simd_test(enable = "avx512f,avx512vl" )] |
9368 | unsafe fn test_mm_mask_permutex2var_epi64() { |
9369 | let a = _mm_set_epi64x(0, 1); |
9370 | let idx = _mm_set_epi64x(1, 1 << 1); |
9371 | let b = _mm_set1_epi64x(100); |
9372 | let r = _mm_mask_permutex2var_epi64(a, 0, idx, b); |
9373 | assert_eq_m128i(r, a); |
9374 | let r = _mm_mask_permutex2var_epi64(a, 0b00000011, idx, b); |
9375 | let e = _mm_set_epi64x(0, 100); |
9376 | assert_eq_m128i(r, e); |
9377 | } |
9378 | |
9379 | #[simd_test(enable = "avx512f,avx512vl" )] |
9380 | unsafe fn test_mm_maskz_permutex2var_epi64() { |
9381 | let a = _mm_set_epi64x(0, 1); |
9382 | let idx = _mm_set_epi64x(1, 1 << 1); |
9383 | let b = _mm_set1_epi64x(100); |
9384 | let r = _mm_maskz_permutex2var_epi64(0, a, idx, b); |
9385 | assert_eq_m128i(r, _mm_setzero_si128()); |
9386 | let r = _mm_maskz_permutex2var_epi64(0b00000011, a, idx, b); |
9387 | let e = _mm_set_epi64x(0, 100); |
9388 | assert_eq_m128i(r, e); |
9389 | } |
9390 | |
9391 | #[simd_test(enable = "avx512f,avx512vl" )] |
9392 | unsafe fn test_mm_mask2_permutex2var_epi64() { |
9393 | let a = _mm_set_epi64x(0, 1); |
9394 | let idx = _mm_set_epi64x(1, 1 << 1); |
9395 | let b = _mm_set1_epi64x(100); |
9396 | let r = _mm_mask2_permutex2var_epi64(a, idx, 0, b); |
9397 | assert_eq_m128i(r, idx); |
9398 | let r = _mm_mask2_permutex2var_epi64(a, idx, 0b00000011, b); |
9399 | let e = _mm_set_epi64x(0, 100); |
9400 | assert_eq_m128i(r, e); |
9401 | } |
9402 | |
9403 | #[simd_test(enable = "avx512f" )] |
9404 | unsafe fn test_mm512_permutex2var_pd() { |
9405 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9406 | let idx = _mm512_set_epi64(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
9407 | let b = _mm512_set1_pd(100.); |
9408 | let r = _mm512_permutex2var_pd(a, idx, b); |
9409 | let e = _mm512_set_pd(6., 100., 5., 100., 4., 100., 3., 100.); |
9410 | assert_eq_m512d(r, e); |
9411 | } |
9412 | |
9413 | #[simd_test(enable = "avx512f" )] |
9414 | unsafe fn test_mm512_mask_permutex2var_pd() { |
9415 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9416 | let idx = _mm512_set_epi64(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
9417 | let b = _mm512_set1_pd(100.); |
9418 | let r = _mm512_mask_permutex2var_pd(a, 0, idx, b); |
9419 | assert_eq_m512d(r, a); |
9420 | let r = _mm512_mask_permutex2var_pd(a, 0b11111111, idx, b); |
9421 | let e = _mm512_set_pd(6., 100., 5., 100., 4., 100., 3., 100.); |
9422 | assert_eq_m512d(r, e); |
9423 | } |
9424 | |
9425 | #[simd_test(enable = "avx512f" )] |
9426 | unsafe fn test_mm512_maskz_permutex2var_pd() { |
9427 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9428 | let idx = _mm512_set_epi64(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
9429 | let b = _mm512_set1_pd(100.); |
9430 | let r = _mm512_maskz_permutex2var_pd(0, a, idx, b); |
9431 | assert_eq_m512d(r, _mm512_setzero_pd()); |
9432 | let r = _mm512_maskz_permutex2var_pd(0b00001111, a, idx, b); |
9433 | let e = _mm512_set_pd(0., 0., 0., 0., 4., 100., 3., 100.); |
9434 | assert_eq_m512d(r, e); |
9435 | } |
9436 | |
9437 | #[simd_test(enable = "avx512f" )] |
9438 | unsafe fn test_mm512_mask2_permutex2var_pd() { |
9439 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
9440 | let idx = _mm512_set_epi64(1, 1 << 3, 2, 1 << 3, 3, 1 << 3, 4, 1 << 3); |
9441 | let b = _mm512_set1_pd(100.); |
9442 | let r = _mm512_mask2_permutex2var_pd(a, idx, 0, b); |
9443 | assert_eq_m512d(r, _mm512_castsi512_pd(idx)); |
9444 | let r = _mm512_mask2_permutex2var_pd(a, idx, 0b11111111, b); |
9445 | let e = _mm512_set_pd(6., 100., 5., 100., 4., 100., 3., 100.); |
9446 | assert_eq_m512d(r, e); |
9447 | } |
9448 | |
9449 | #[simd_test(enable = "avx512f,avx512vl" )] |
9450 | unsafe fn test_mm256_permutex2var_pd() { |
9451 | let a = _mm256_set_pd(0., 1., 2., 3.); |
9452 | let idx = _mm256_set_epi64x(1, 1 << 2, 2, 1 << 2); |
9453 | let b = _mm256_set1_pd(100.); |
9454 | let r = _mm256_permutex2var_pd(a, idx, b); |
9455 | let e = _mm256_set_pd(2., 100., 1., 100.); |
9456 | assert_eq_m256d(r, e); |
9457 | } |
9458 | |
9459 | #[simd_test(enable = "avx512f,avx512vl" )] |
9460 | unsafe fn test_mm256_mask_permutex2var_pd() { |
9461 | let a = _mm256_set_pd(0., 1., 2., 3.); |
9462 | let idx = _mm256_set_epi64x(1, 1 << 2, 2, 1 << 2); |
9463 | let b = _mm256_set1_pd(100.); |
9464 | let r = _mm256_mask_permutex2var_pd(a, 0, idx, b); |
9465 | assert_eq_m256d(r, a); |
9466 | let r = _mm256_mask_permutex2var_pd(a, 0b00001111, idx, b); |
9467 | let e = _mm256_set_pd(2., 100., 1., 100.); |
9468 | assert_eq_m256d(r, e); |
9469 | } |
9470 | |
9471 | #[simd_test(enable = "avx512f,avx512vl" )] |
9472 | unsafe fn test_mm256_maskz_permutex2var_pd() { |
9473 | let a = _mm256_set_pd(0., 1., 2., 3.); |
9474 | let idx = _mm256_set_epi64x(1, 1 << 2, 2, 1 << 2); |
9475 | let b = _mm256_set1_pd(100.); |
9476 | let r = _mm256_maskz_permutex2var_pd(0, a, idx, b); |
9477 | assert_eq_m256d(r, _mm256_setzero_pd()); |
9478 | let r = _mm256_maskz_permutex2var_pd(0b00001111, a, idx, b); |
9479 | let e = _mm256_set_pd(2., 100., 1., 100.); |
9480 | assert_eq_m256d(r, e); |
9481 | } |
9482 | |
9483 | #[simd_test(enable = "avx512f,avx512vl" )] |
9484 | unsafe fn test_mm256_mask2_permutex2var_pd() { |
9485 | let a = _mm256_set_pd(0., 1., 2., 3.); |
9486 | let idx = _mm256_set_epi64x(1, 1 << 2, 2, 1 << 2); |
9487 | let b = _mm256_set1_pd(100.); |
9488 | let r = _mm256_mask2_permutex2var_pd(a, idx, 0, b); |
9489 | assert_eq_m256d(r, _mm256_castsi256_pd(idx)); |
9490 | let r = _mm256_mask2_permutex2var_pd(a, idx, 0b00001111, b); |
9491 | let e = _mm256_set_pd(2., 100., 1., 100.); |
9492 | assert_eq_m256d(r, e); |
9493 | } |
9494 | |
9495 | #[simd_test(enable = "avx512f,avx512vl" )] |
9496 | unsafe fn test_mm_permutex2var_pd() { |
9497 | let a = _mm_set_pd(0., 1.); |
9498 | let idx = _mm_set_epi64x(1, 1 << 1); |
9499 | let b = _mm_set1_pd(100.); |
9500 | let r = _mm_permutex2var_pd(a, idx, b); |
9501 | let e = _mm_set_pd(0., 100.); |
9502 | assert_eq_m128d(r, e); |
9503 | } |
9504 | |
9505 | #[simd_test(enable = "avx512f,avx512vl" )] |
9506 | unsafe fn test_mm_mask_permutex2var_pd() { |
9507 | let a = _mm_set_pd(0., 1.); |
9508 | let idx = _mm_set_epi64x(1, 1 << 1); |
9509 | let b = _mm_set1_pd(100.); |
9510 | let r = _mm_mask_permutex2var_pd(a, 0, idx, b); |
9511 | assert_eq_m128d(r, a); |
9512 | let r = _mm_mask_permutex2var_pd(a, 0b00000011, idx, b); |
9513 | let e = _mm_set_pd(0., 100.); |
9514 | assert_eq_m128d(r, e); |
9515 | } |
9516 | |
9517 | #[simd_test(enable = "avx512f,avx512vl" )] |
9518 | unsafe fn test_mm_maskz_permutex2var_pd() { |
9519 | let a = _mm_set_pd(0., 1.); |
9520 | let idx = _mm_set_epi64x(1, 1 << 1); |
9521 | let b = _mm_set1_pd(100.); |
9522 | let r = _mm_maskz_permutex2var_pd(0, a, idx, b); |
9523 | assert_eq_m128d(r, _mm_setzero_pd()); |
9524 | let r = _mm_maskz_permutex2var_pd(0b00000011, a, idx, b); |
9525 | let e = _mm_set_pd(0., 100.); |
9526 | assert_eq_m128d(r, e); |
9527 | } |
9528 | |
9529 | #[simd_test(enable = "avx512f,avx512vl" )] |
9530 | unsafe fn test_mm_mask2_permutex2var_pd() { |
9531 | let a = _mm_set_pd(0., 1.); |
9532 | let idx = _mm_set_epi64x(1, 1 << 1); |
9533 | let b = _mm_set1_pd(100.); |
9534 | let r = _mm_mask2_permutex2var_pd(a, idx, 0, b); |
9535 | assert_eq_m128d(r, _mm_castsi128_pd(idx)); |
9536 | let r = _mm_mask2_permutex2var_pd(a, idx, 0b00000011, b); |
9537 | let e = _mm_set_pd(0., 100.); |
9538 | assert_eq_m128d(r, e); |
9539 | } |
9540 | |
9541 | #[simd_test(enable = "avx512f,avx512vl" )] |
9542 | unsafe fn test_mm256_mask_shuffle_pd() { |
9543 | let a = _mm256_set_pd(1., 4., 5., 8.); |
9544 | let b = _mm256_set_pd(2., 3., 6., 7.); |
9545 | let r = _mm256_mask_shuffle_pd::<0b11_11_11_11>(a, 0, a, b); |
9546 | assert_eq_m256d(r, a); |
9547 | let r = _mm256_mask_shuffle_pd::<0b11_11_11_11>(a, 0b00001111, a, b); |
9548 | let e = _mm256_set_pd(2., 1., 6., 5.); |
9549 | assert_eq_m256d(r, e); |
9550 | } |
9551 | |
9552 | #[simd_test(enable = "avx512f,avx512vl" )] |
9553 | unsafe fn test_mm256_maskz_shuffle_pd() { |
9554 | let a = _mm256_set_pd(1., 4., 5., 8.); |
9555 | let b = _mm256_set_pd(2., 3., 6., 7.); |
9556 | let r = _mm256_maskz_shuffle_pd::<0b11_11_11_11>(0, a, b); |
9557 | assert_eq_m256d(r, _mm256_setzero_pd()); |
9558 | let r = _mm256_maskz_shuffle_pd::<0b11_11_11_11>(0b00001111, a, b); |
9559 | let e = _mm256_set_pd(2., 1., 6., 5.); |
9560 | assert_eq_m256d(r, e); |
9561 | } |
9562 | |
9563 | #[simd_test(enable = "avx512f,avx512vl" )] |
9564 | unsafe fn test_mm_mask_shuffle_pd() { |
9565 | let a = _mm_set_pd(1., 4.); |
9566 | let b = _mm_set_pd(2., 3.); |
9567 | let r = _mm_mask_shuffle_pd::<0b11_11_11_11>(a, 0, a, b); |
9568 | assert_eq_m128d(r, a); |
9569 | let r = _mm_mask_shuffle_pd::<0b11_11_11_11>(a, 0b00000011, a, b); |
9570 | let e = _mm_set_pd(2., 1.); |
9571 | assert_eq_m128d(r, e); |
9572 | } |
9573 | |
9574 | #[simd_test(enable = "avx512f,avx512vl" )] |
9575 | unsafe fn test_mm_maskz_shuffle_pd() { |
9576 | let a = _mm_set_pd(1., 4.); |
9577 | let b = _mm_set_pd(2., 3.); |
9578 | let r = _mm_maskz_shuffle_pd::<0b11_11_11_11>(0, a, b); |
9579 | assert_eq_m128d(r, _mm_setzero_pd()); |
9580 | let r = _mm_maskz_shuffle_pd::<0b11_11_11_11>(0b00000011, a, b); |
9581 | let e = _mm_set_pd(2., 1.); |
9582 | assert_eq_m128d(r, e); |
9583 | } |
9584 | |
9585 | #[simd_test(enable = "avx512f" )] |
9586 | unsafe fn test_mm512_shuffle_i64x2() { |
9587 | let a = _mm512_setr_epi64(1, 4, 5, 8, 9, 12, 13, 16); |
9588 | let b = _mm512_setr_epi64(2, 3, 6, 7, 10, 11, 14, 15); |
9589 | let r = _mm512_shuffle_i64x2::<0b00_00_00_00>(a, b); |
9590 | let e = _mm512_setr_epi64(1, 4, 1, 4, 2, 3, 2, 3); |
9591 | assert_eq_m512i(r, e); |
9592 | } |
9593 | |
9594 | #[simd_test(enable = "avx512f" )] |
9595 | unsafe fn test_mm512_mask_shuffle_i64x2() { |
9596 | let a = _mm512_setr_epi64(1, 4, 5, 8, 9, 12, 13, 16); |
9597 | let b = _mm512_setr_epi64(2, 3, 6, 7, 10, 11, 14, 15); |
9598 | let r = _mm512_mask_shuffle_i64x2::<0b00_00_00_00>(a, 0, a, b); |
9599 | assert_eq_m512i(r, a); |
9600 | let r = _mm512_mask_shuffle_i64x2::<0b00_00_00_00>(a, 0b11111111, a, b); |
9601 | let e = _mm512_setr_epi64(1, 4, 1, 4, 2, 3, 2, 3); |
9602 | assert_eq_m512i(r, e); |
9603 | } |
9604 | |
9605 | #[simd_test(enable = "avx512f" )] |
9606 | unsafe fn test_mm512_maskz_shuffle_i64x2() { |
9607 | let a = _mm512_setr_epi64(1, 4, 5, 8, 9, 12, 13, 16); |
9608 | let b = _mm512_setr_epi64(2, 3, 6, 7, 10, 11, 14, 15); |
9609 | let r = _mm512_maskz_shuffle_i64x2::<0b00_00_00_00>(0, a, b); |
9610 | assert_eq_m512i(r, _mm512_setzero_si512()); |
9611 | let r = _mm512_maskz_shuffle_i64x2::<0b00_00_00_00>(0b00001111, a, b); |
9612 | let e = _mm512_setr_epi64(1, 4, 1, 4, 0, 0, 0, 0); |
9613 | assert_eq_m512i(r, e); |
9614 | } |
9615 | |
9616 | #[simd_test(enable = "avx512f,avx512vl" )] |
9617 | unsafe fn test_mm256_shuffle_i64x2() { |
9618 | let a = _mm256_set_epi64x(1, 4, 5, 8); |
9619 | let b = _mm256_set_epi64x(2, 3, 6, 7); |
9620 | let r = _mm256_shuffle_i64x2::<0b00>(a, b); |
9621 | let e = _mm256_set_epi64x(6, 7, 5, 8); |
9622 | assert_eq_m256i(r, e); |
9623 | } |
9624 | |
9625 | #[simd_test(enable = "avx512f,avx512vl" )] |
9626 | unsafe fn test_mm256_mask_shuffle_i64x2() { |
9627 | let a = _mm256_set_epi64x(1, 4, 5, 8); |
9628 | let b = _mm256_set_epi64x(2, 3, 6, 7); |
9629 | let r = _mm256_mask_shuffle_i64x2::<0b00>(a, 0, a, b); |
9630 | assert_eq_m256i(r, a); |
9631 | let r = _mm256_mask_shuffle_i64x2::<0b00>(a, 0b00001111, a, b); |
9632 | let e = _mm256_set_epi64x(6, 7, 5, 8); |
9633 | assert_eq_m256i(r, e); |
9634 | } |
9635 | |
9636 | #[simd_test(enable = "avx512f,avx512vl" )] |
9637 | unsafe fn test_mm256_maskz_shuffle_i64x2() { |
9638 | let a = _mm256_set_epi64x(1, 4, 5, 8); |
9639 | let b = _mm256_set_epi64x(2, 3, 6, 7); |
9640 | let r = _mm256_maskz_shuffle_i64x2::<0b00>(0, a, b); |
9641 | assert_eq_m256i(r, _mm256_setzero_si256()); |
9642 | let r = _mm256_maskz_shuffle_i64x2::<0b00>(0b00001111, a, b); |
9643 | let e = _mm256_set_epi64x(6, 7, 5, 8); |
9644 | assert_eq_m256i(r, e); |
9645 | } |
9646 | |
9647 | #[simd_test(enable = "avx512f" )] |
9648 | unsafe fn test_mm512_shuffle_f64x2() { |
9649 | let a = _mm512_setr_pd(1., 4., 5., 8., 9., 12., 13., 16.); |
9650 | let b = _mm512_setr_pd(2., 3., 6., 7., 10., 11., 14., 15.); |
9651 | let r = _mm512_shuffle_f64x2::<0b00_00_00_00>(a, b); |
9652 | let e = _mm512_setr_pd(1., 4., 1., 4., 2., 3., 2., 3.); |
9653 | assert_eq_m512d(r, e); |
9654 | } |
9655 | |
9656 | #[simd_test(enable = "avx512f" )] |
9657 | unsafe fn test_mm512_mask_shuffle_f64x2() { |
9658 | let a = _mm512_setr_pd(1., 4., 5., 8., 9., 12., 13., 16.); |
9659 | let b = _mm512_setr_pd(2., 3., 6., 7., 10., 11., 14., 15.); |
9660 | let r = _mm512_mask_shuffle_f64x2::<0b00_00_00_00>(a, 0, a, b); |
9661 | assert_eq_m512d(r, a); |
9662 | let r = _mm512_mask_shuffle_f64x2::<0b00_00_00_00>(a, 0b11111111, a, b); |
9663 | let e = _mm512_setr_pd(1., 4., 1., 4., 2., 3., 2., 3.); |
9664 | assert_eq_m512d(r, e); |
9665 | } |
9666 | |
9667 | #[simd_test(enable = "avx512f" )] |
9668 | unsafe fn test_mm512_maskz_shuffle_f64x2() { |
9669 | let a = _mm512_setr_pd(1., 4., 5., 8., 9., 12., 13., 16.); |
9670 | let b = _mm512_setr_pd(2., 3., 6., 7., 10., 11., 14., 15.); |
9671 | let r = _mm512_maskz_shuffle_f64x2::<0b00_00_00_00>(0, a, b); |
9672 | assert_eq_m512d(r, _mm512_setzero_pd()); |
9673 | let r = _mm512_maskz_shuffle_f64x2::<0b00_00_00_00>(0b00001111, a, b); |
9674 | let e = _mm512_setr_pd(1., 4., 1., 4., 0., 0., 0., 0.); |
9675 | assert_eq_m512d(r, e); |
9676 | } |
9677 | |
9678 | #[simd_test(enable = "avx512f,avx512vl" )] |
9679 | unsafe fn test_mm256_shuffle_f64x2() { |
9680 | let a = _mm256_set_pd(1., 4., 5., 8.); |
9681 | let b = _mm256_set_pd(2., 3., 6., 7.); |
9682 | let r = _mm256_shuffle_f64x2::<0b00>(a, b); |
9683 | let e = _mm256_set_pd(6., 7., 5., 8.); |
9684 | assert_eq_m256d(r, e); |
9685 | } |
9686 | |
9687 | #[simd_test(enable = "avx512f,avx512vl" )] |
9688 | unsafe fn test_mm256_mask_shuffle_f64x2() { |
9689 | let a = _mm256_set_pd(1., 4., 5., 8.); |
9690 | let b = _mm256_set_pd(2., 3., 6., 7.); |
9691 | let r = _mm256_mask_shuffle_f64x2::<0b00>(a, 0, a, b); |
9692 | assert_eq_m256d(r, a); |
9693 | let r = _mm256_mask_shuffle_f64x2::<0b00>(a, 0b00001111, a, b); |
9694 | let e = _mm256_set_pd(6., 7., 5., 8.); |
9695 | assert_eq_m256d(r, e); |
9696 | } |
9697 | |
9698 | #[simd_test(enable = "avx512f,avx512vl" )] |
9699 | unsafe fn test_mm256_maskz_shuffle_f64x2() { |
9700 | let a = _mm256_set_pd(1., 4., 5., 8.); |
9701 | let b = _mm256_set_pd(2., 3., 6., 7.); |
9702 | let r = _mm256_maskz_shuffle_f64x2::<0b00>(0, a, b); |
9703 | assert_eq_m256d(r, _mm256_setzero_pd()); |
9704 | let r = _mm256_maskz_shuffle_f64x2::<0b00>(0b00001111, a, b); |
9705 | let e = _mm256_set_pd(6., 7., 5., 8.); |
9706 | assert_eq_m256d(r, e); |
9707 | } |
9708 | |
9709 | #[simd_test(enable = "avx512f" )] |
9710 | unsafe fn test_mm512_movedup_pd() { |
9711 | let a = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
9712 | let r = _mm512_movedup_pd(a); |
9713 | let e = _mm512_setr_pd(1., 1., 3., 3., 5., 5., 7., 7.); |
9714 | assert_eq_m512d(r, e); |
9715 | } |
9716 | |
9717 | #[simd_test(enable = "avx512f" )] |
9718 | unsafe fn test_mm512_mask_movedup_pd() { |
9719 | let a = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
9720 | let r = _mm512_mask_movedup_pd(a, 0, a); |
9721 | assert_eq_m512d(r, a); |
9722 | let r = _mm512_mask_movedup_pd(a, 0b11111111, a); |
9723 | let e = _mm512_setr_pd(1., 1., 3., 3., 5., 5., 7., 7.); |
9724 | assert_eq_m512d(r, e); |
9725 | } |
9726 | |
9727 | #[simd_test(enable = "avx512f" )] |
9728 | unsafe fn test_mm512_maskz_movedup_pd() { |
9729 | let a = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
9730 | let r = _mm512_maskz_movedup_pd(0, a); |
9731 | assert_eq_m512d(r, _mm512_setzero_pd()); |
9732 | let r = _mm512_maskz_movedup_pd(0b00001111, a); |
9733 | let e = _mm512_setr_pd(1., 1., 3., 3., 0., 0., 0., 0.); |
9734 | assert_eq_m512d(r, e); |
9735 | } |
9736 | |
9737 | #[simd_test(enable = "avx512f,avx512vl" )] |
9738 | unsafe fn test_mm256_mask_movedup_pd() { |
9739 | let a = _mm256_set_pd(1., 2., 3., 4.); |
9740 | let r = _mm256_mask_movedup_pd(a, 0, a); |
9741 | assert_eq_m256d(r, a); |
9742 | let r = _mm256_mask_movedup_pd(a, 0b00001111, a); |
9743 | let e = _mm256_set_pd(2., 2., 4., 4.); |
9744 | assert_eq_m256d(r, e); |
9745 | } |
9746 | |
9747 | #[simd_test(enable = "avx512f,avx512vl" )] |
9748 | unsafe fn test_mm256_maskz_movedup_pd() { |
9749 | let a = _mm256_set_pd(1., 2., 3., 4.); |
9750 | let r = _mm256_maskz_movedup_pd(0, a); |
9751 | assert_eq_m256d(r, _mm256_setzero_pd()); |
9752 | let r = _mm256_maskz_movedup_pd(0b00001111, a); |
9753 | let e = _mm256_set_pd(2., 2., 4., 4.); |
9754 | assert_eq_m256d(r, e); |
9755 | } |
9756 | |
9757 | #[simd_test(enable = "avx512f,avx512vl" )] |
9758 | unsafe fn test_mm_mask_movedup_pd() { |
9759 | let a = _mm_set_pd(1., 2.); |
9760 | let r = _mm_mask_movedup_pd(a, 0, a); |
9761 | assert_eq_m128d(r, a); |
9762 | let r = _mm_mask_movedup_pd(a, 0b00000011, a); |
9763 | let e = _mm_set_pd(2., 2.); |
9764 | assert_eq_m128d(r, e); |
9765 | } |
9766 | |
9767 | #[simd_test(enable = "avx512f,avx512vl" )] |
9768 | unsafe fn test_mm_maskz_movedup_pd() { |
9769 | let a = _mm_set_pd(1., 2.); |
9770 | let r = _mm_maskz_movedup_pd(0, a); |
9771 | assert_eq_m128d(r, _mm_setzero_pd()); |
9772 | let r = _mm_maskz_movedup_pd(0b00000011, a); |
9773 | let e = _mm_set_pd(2., 2.); |
9774 | assert_eq_m128d(r, e); |
9775 | } |
9776 | |
9777 | #[simd_test(enable = "avx512f" )] |
9778 | unsafe fn test_mm512_inserti64x4() { |
9779 | let a = _mm512_setr_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
9780 | let b = _mm256_setr_epi64x(17, 18, 19, 20); |
9781 | let r = _mm512_inserti64x4::<1>(a, b); |
9782 | let e = _mm512_setr_epi64(1, 2, 3, 4, 17, 18, 19, 20); |
9783 | assert_eq_m512i(r, e); |
9784 | } |
9785 | |
9786 | #[simd_test(enable = "avx512f" )] |
9787 | unsafe fn test_mm512_mask_inserti64x4() { |
9788 | let a = _mm512_setr_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
9789 | let b = _mm256_setr_epi64x(17, 18, 19, 20); |
9790 | let r = _mm512_mask_inserti64x4::<1>(a, 0, a, b); |
9791 | assert_eq_m512i(r, a); |
9792 | let r = _mm512_mask_inserti64x4::<1>(a, 0b11111111, a, b); |
9793 | let e = _mm512_setr_epi64(1, 2, 3, 4, 17, 18, 19, 20); |
9794 | assert_eq_m512i(r, e); |
9795 | } |
9796 | |
9797 | #[simd_test(enable = "avx512f" )] |
9798 | unsafe fn test_mm512_maskz_inserti64x4() { |
9799 | let a = _mm512_setr_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
9800 | let b = _mm256_setr_epi64x(17, 18, 19, 20); |
9801 | let r = _mm512_maskz_inserti64x4::<1>(0, a, b); |
9802 | assert_eq_m512i(r, _mm512_setzero_si512()); |
9803 | let r = _mm512_maskz_inserti64x4::<1>(0b00001111, a, b); |
9804 | let e = _mm512_setr_epi64(1, 2, 3, 4, 0, 0, 0, 0); |
9805 | assert_eq_m512i(r, e); |
9806 | } |
9807 | |
9808 | #[simd_test(enable = "avx512f" )] |
9809 | unsafe fn test_mm512_insertf64x4() { |
9810 | let a = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
9811 | let b = _mm256_setr_pd(17., 18., 19., 20.); |
9812 | let r = _mm512_insertf64x4::<1>(a, b); |
9813 | let e = _mm512_setr_pd(1., 2., 3., 4., 17., 18., 19., 20.); |
9814 | assert_eq_m512d(r, e); |
9815 | } |
9816 | |
9817 | #[simd_test(enable = "avx512f" )] |
9818 | unsafe fn test_mm512_mask_insertf64x4() { |
9819 | let a = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
9820 | let b = _mm256_setr_pd(17., 18., 19., 20.); |
9821 | let r = _mm512_mask_insertf64x4::<1>(a, 0, a, b); |
9822 | assert_eq_m512d(r, a); |
9823 | let r = _mm512_mask_insertf64x4::<1>(a, 0b11111111, a, b); |
9824 | let e = _mm512_setr_pd(1., 2., 3., 4., 17., 18., 19., 20.); |
9825 | assert_eq_m512d(r, e); |
9826 | } |
9827 | |
9828 | #[simd_test(enable = "avx512f" )] |
9829 | unsafe fn test_mm512_maskz_insertf64x4() { |
9830 | let a = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
9831 | let b = _mm256_setr_pd(17., 18., 19., 20.); |
9832 | let r = _mm512_maskz_insertf64x4::<1>(0, a, b); |
9833 | assert_eq_m512d(r, _mm512_setzero_pd()); |
9834 | let r = _mm512_maskz_insertf64x4::<1>(0b00001111, a, b); |
9835 | let e = _mm512_setr_pd(1., 2., 3., 4., 0., 0., 0., 0.); |
9836 | assert_eq_m512d(r, e); |
9837 | } |
9838 | |
9839 | #[simd_test(enable = "avx512f" )] |
9840 | unsafe fn test_mm512_castpd128_pd512() { |
9841 | let a = _mm_setr_pd(17., 18.); |
9842 | let r = _mm512_castpd128_pd512(a); |
9843 | let e = _mm512_setr_pd(17., 18., -1., -1., -1., -1., -1., -1.); |
9844 | assert_eq_m512d(r, e); |
9845 | } |
9846 | |
9847 | #[simd_test(enable = "avx512f" )] |
9848 | unsafe fn test_mm512_castpd256_pd512() { |
9849 | let a = _mm256_setr_pd(17., 18., 19., 20.); |
9850 | let r = _mm512_castpd256_pd512(a); |
9851 | let e = _mm512_setr_pd(17., 18., 19., 20., -1., -1., -1., -1.); |
9852 | assert_eq_m512d(r, e); |
9853 | } |
9854 | |
9855 | #[simd_test(enable = "avx512f" )] |
9856 | unsafe fn test_mm512_zextpd128_pd512() { |
9857 | let a = _mm_setr_pd(17., 18.); |
9858 | let r = _mm512_zextpd128_pd512(a); |
9859 | let e = _mm512_setr_pd(17., 18., 0., 0., 0., 0., 0., 0.); |
9860 | assert_eq_m512d(r, e); |
9861 | } |
9862 | |
9863 | #[simd_test(enable = "avx512f" )] |
9864 | unsafe fn test_mm512_zextpd256_pd512() { |
9865 | let a = _mm256_setr_pd(17., 18., 19., 20.); |
9866 | let r = _mm512_zextpd256_pd512(a); |
9867 | let e = _mm512_setr_pd(17., 18., 19., 20., 0., 0., 0., 0.); |
9868 | assert_eq_m512d(r, e); |
9869 | } |
9870 | |
9871 | #[simd_test(enable = "avx512f" )] |
9872 | unsafe fn test_mm512_castpd512_pd128() { |
9873 | let a = _mm512_setr_pd(17., 18., -1., -1., -1., -1., -1., -1.); |
9874 | let r = _mm512_castpd512_pd128(a); |
9875 | let e = _mm_setr_pd(17., 18.); |
9876 | assert_eq_m128d(r, e); |
9877 | } |
9878 | |
9879 | #[simd_test(enable = "avx512f" )] |
9880 | unsafe fn test_mm512_castpd512_pd256() { |
9881 | let a = _mm512_setr_pd(17., 18., 19., 20., -1., -1., -1., -1.); |
9882 | let r = _mm512_castpd512_pd256(a); |
9883 | let e = _mm256_setr_pd(17., 18., 19., 20.); |
9884 | assert_eq_m256d(r, e); |
9885 | } |
9886 | |
9887 | #[simd_test(enable = "avx512f" )] |
9888 | unsafe fn test_mm512_castpd_ps() { |
9889 | let a = _mm512_set1_pd(1.); |
9890 | let r = _mm512_castpd_ps(a); |
9891 | let e = _mm512_set_ps( |
9892 | 1.875, 0.0, 1.875, 0.0, 1.875, 0.0, 1.875, 0.0, 1.875, 0.0, 1.875, 0.0, 1.875, 0.0, |
9893 | 1.875, 0.0, |
9894 | ); |
9895 | assert_eq_m512(r, e); |
9896 | } |
9897 | |
9898 | #[simd_test(enable = "avx512f" )] |
9899 | unsafe fn test_mm512_castpd_si512() { |
9900 | let a = _mm512_set1_pd(1.); |
9901 | let r = _mm512_castpd_si512(a); |
9902 | let e = _mm512_set_epi32( |
9903 | 1072693248, 0, 1072693248, 0, 1072693248, 0, 1072693248, 0, 1072693248, 0, 1072693248, |
9904 | 0, 1072693248, 0, 1072693248, 0, |
9905 | ); |
9906 | assert_eq_m512i(r, e); |
9907 | } |
9908 | |
9909 | #[simd_test(enable = "avx512f" )] |
9910 | unsafe fn test_mm512_castsi128_si512() { |
9911 | let a = _mm_setr_epi64x(17, 18); |
9912 | let r = _mm512_castsi128_si512(a); |
9913 | let e = _mm512_setr_epi64(17, 18, -1, -1, -1, -1, -1, -1); |
9914 | assert_eq_m512i(r, e); |
9915 | } |
9916 | |
9917 | #[simd_test(enable = "avx512f" )] |
9918 | unsafe fn test_mm512_castsi256_si512() { |
9919 | let a = _mm256_setr_epi64x(17, 18, 19, 20); |
9920 | let r = _mm512_castsi256_si512(a); |
9921 | let e = _mm512_setr_epi64(17, 18, 19, 20, -1, -1, -1, -1); |
9922 | assert_eq_m512i(r, e); |
9923 | } |
9924 | |
9925 | #[simd_test(enable = "avx512f" )] |
9926 | unsafe fn test_mm512_zextsi128_si512() { |
9927 | let a = _mm_setr_epi64x(17, 18); |
9928 | let r = _mm512_zextsi128_si512(a); |
9929 | let e = _mm512_setr_epi64(17, 18, 0, 0, 0, 0, 0, 0); |
9930 | assert_eq_m512i(r, e); |
9931 | } |
9932 | |
9933 | #[simd_test(enable = "avx512f" )] |
9934 | unsafe fn test_mm512_zextsi256_si512() { |
9935 | let a = _mm256_setr_epi64x(17, 18, 19, 20); |
9936 | let r = _mm512_zextsi256_si512(a); |
9937 | let e = _mm512_setr_epi64(17, 18, 19, 20, 0, 0, 0, 0); |
9938 | assert_eq_m512i(r, e); |
9939 | } |
9940 | |
9941 | #[simd_test(enable = "avx512f" )] |
9942 | unsafe fn test_mm512_castsi512_si128() { |
9943 | let a = _mm512_setr_epi64(17, 18, -1, -1, -1, -1, -1, -1); |
9944 | let r = _mm512_castsi512_si128(a); |
9945 | let e = _mm_setr_epi64x(17, 18); |
9946 | assert_eq_m128i(r, e); |
9947 | } |
9948 | |
9949 | #[simd_test(enable = "avx512f" )] |
9950 | unsafe fn test_mm512_castsi512_si256() { |
9951 | let a = _mm512_setr_epi64(17, 18, 19, 20, -1, -1, -1, -1); |
9952 | let r = _mm512_castsi512_si256(a); |
9953 | let e = _mm256_setr_epi64x(17, 18, 19, 20); |
9954 | assert_eq_m256i(r, e); |
9955 | } |
9956 | |
9957 | #[simd_test(enable = "avx512f" )] |
9958 | unsafe fn test_mm512_castsi512_ps() { |
9959 | let a = _mm512_set1_epi64(1 << 62); |
9960 | let r = _mm512_castsi512_ps(a); |
9961 | let e = _mm512_set_ps( |
9962 | 2., 0., 2., 0., 2., 0., 2., 0., 2., 0., 2., 0., 2., 0., 2., 0., |
9963 | ); |
9964 | assert_eq_m512(r, e); |
9965 | } |
9966 | |
9967 | #[simd_test(enable = "avx512f" )] |
9968 | unsafe fn test_mm512_castsi512_pd() { |
9969 | let a = _mm512_set1_epi64(1 << 62); |
9970 | let r = _mm512_castsi512_pd(a); |
9971 | let e = _mm512_set_pd(2., 2., 2., 2., 2., 2., 2., 2.); |
9972 | assert_eq_m512d(r, e); |
9973 | } |
9974 | |
9975 | #[simd_test(enable = "avx512f" )] |
9976 | unsafe fn test_mm512_broadcastq_epi64() { |
9977 | let a = _mm_setr_epi64x(17, 18); |
9978 | let r = _mm512_broadcastq_epi64(a); |
9979 | let e = _mm512_set1_epi64(17); |
9980 | assert_eq_m512i(r, e); |
9981 | } |
9982 | |
9983 | #[simd_test(enable = "avx512f" )] |
9984 | unsafe fn test_mm512_mask_broadcastq_epi64() { |
9985 | let src = _mm512_set1_epi64(18); |
9986 | let a = _mm_setr_epi64x(17, 18); |
9987 | let r = _mm512_mask_broadcastq_epi64(src, 0, a); |
9988 | assert_eq_m512i(r, src); |
9989 | let r = _mm512_mask_broadcastq_epi64(src, 0b11111111, a); |
9990 | let e = _mm512_set1_epi64(17); |
9991 | assert_eq_m512i(r, e); |
9992 | } |
9993 | |
9994 | #[simd_test(enable = "avx512f" )] |
9995 | unsafe fn test_mm512_maskz_broadcastq_epi64() { |
9996 | let a = _mm_setr_epi64x(17, 18); |
9997 | let r = _mm512_maskz_broadcastq_epi64(0, a); |
9998 | assert_eq_m512i(r, _mm512_setzero_si512()); |
9999 | let r = _mm512_maskz_broadcastq_epi64(0b00001111, a); |
10000 | let e = _mm512_set_epi64(0, 0, 0, 0, 17, 17, 17, 17); |
10001 | assert_eq_m512i(r, e); |
10002 | } |
10003 | |
10004 | #[simd_test(enable = "avx512f,avx512vl" )] |
10005 | unsafe fn test_mm256_mask_broadcastq_epi64() { |
10006 | let src = _mm256_set1_epi64x(18); |
10007 | let a = _mm_set_epi64x(17, 18); |
10008 | let r = _mm256_mask_broadcastq_epi64(src, 0, a); |
10009 | assert_eq_m256i(r, src); |
10010 | let r = _mm256_mask_broadcastq_epi64(src, 0b00001111, a); |
10011 | let e = _mm256_set1_epi64x(18); |
10012 | assert_eq_m256i(r, e); |
10013 | } |
10014 | |
10015 | #[simd_test(enable = "avx512f,avx512vl" )] |
10016 | unsafe fn test_mm256_maskz_broadcastq_epi64() { |
10017 | let a = _mm_set_epi64x(17, 18); |
10018 | let r = _mm256_maskz_broadcastq_epi64(0, a); |
10019 | assert_eq_m256i(r, _mm256_setzero_si256()); |
10020 | let r = _mm256_maskz_broadcastq_epi64(0b00001111, a); |
10021 | let e = _mm256_set1_epi64x(18); |
10022 | assert_eq_m256i(r, e); |
10023 | } |
10024 | |
10025 | #[simd_test(enable = "avx512f,avx512vl" )] |
10026 | unsafe fn test_mm_mask_broadcastq_epi64() { |
10027 | let src = _mm_set1_epi64x(18); |
10028 | let a = _mm_set_epi64x(17, 18); |
10029 | let r = _mm_mask_broadcastq_epi64(src, 0, a); |
10030 | assert_eq_m128i(r, src); |
10031 | let r = _mm_mask_broadcastq_epi64(src, 0b00000011, a); |
10032 | let e = _mm_set1_epi64x(18); |
10033 | assert_eq_m128i(r, e); |
10034 | } |
10035 | |
10036 | #[simd_test(enable = "avx512f,avx512vl" )] |
10037 | unsafe fn test_mm_maskz_broadcastq_epi64() { |
10038 | let a = _mm_set_epi64x(17, 18); |
10039 | let r = _mm_maskz_broadcastq_epi64(0, a); |
10040 | assert_eq_m128i(r, _mm_setzero_si128()); |
10041 | let r = _mm_maskz_broadcastq_epi64(0b00000011, a); |
10042 | let e = _mm_set1_epi64x(18); |
10043 | assert_eq_m128i(r, e); |
10044 | } |
10045 | |
10046 | #[simd_test(enable = "avx512f" )] |
10047 | unsafe fn test_mm512_broadcastsd_pd() { |
10048 | let a = _mm_set_pd(17., 18.); |
10049 | let r = _mm512_broadcastsd_pd(a); |
10050 | let e = _mm512_set1_pd(18.); |
10051 | assert_eq_m512d(r, e); |
10052 | } |
10053 | |
10054 | #[simd_test(enable = "avx512f" )] |
10055 | unsafe fn test_mm512_mask_broadcastsd_pd() { |
10056 | let src = _mm512_set1_pd(18.); |
10057 | let a = _mm_set_pd(17., 18.); |
10058 | let r = _mm512_mask_broadcastsd_pd(src, 0, a); |
10059 | assert_eq_m512d(r, src); |
10060 | let r = _mm512_mask_broadcastsd_pd(src, 0b11111111, a); |
10061 | let e = _mm512_set1_pd(18.); |
10062 | assert_eq_m512d(r, e); |
10063 | } |
10064 | |
10065 | #[simd_test(enable = "avx512f" )] |
10066 | unsafe fn test_mm512_maskz_broadcastsd_pd() { |
10067 | let a = _mm_set_pd(17., 18.); |
10068 | let r = _mm512_maskz_broadcastsd_pd(0, a); |
10069 | assert_eq_m512d(r, _mm512_setzero_pd()); |
10070 | let r = _mm512_maskz_broadcastsd_pd(0b00001111, a); |
10071 | let e = _mm512_set_pd(0., 0., 0., 0., 18., 18., 18., 18.); |
10072 | assert_eq_m512d(r, e); |
10073 | } |
10074 | |
10075 | #[simd_test(enable = "avx512f,avx512vl" )] |
10076 | unsafe fn test_mm256_mask_broadcastsd_pd() { |
10077 | let src = _mm256_set1_pd(18.); |
10078 | let a = _mm_set_pd(17., 18.); |
10079 | let r = _mm256_mask_broadcastsd_pd(src, 0, a); |
10080 | assert_eq_m256d(r, src); |
10081 | let r = _mm256_mask_broadcastsd_pd(src, 0b00001111, a); |
10082 | let e = _mm256_set1_pd(18.); |
10083 | assert_eq_m256d(r, e); |
10084 | } |
10085 | |
10086 | #[simd_test(enable = "avx512f,avx512vl" )] |
10087 | unsafe fn test_mm256_maskz_broadcastsd_pd() { |
10088 | let a = _mm_set_pd(17., 18.); |
10089 | let r = _mm256_maskz_broadcastsd_pd(0, a); |
10090 | assert_eq_m256d(r, _mm256_setzero_pd()); |
10091 | let r = _mm256_maskz_broadcastsd_pd(0b00001111, a); |
10092 | let e = _mm256_set1_pd(18.); |
10093 | assert_eq_m256d(r, e); |
10094 | } |
10095 | |
10096 | #[simd_test(enable = "avx512f" )] |
10097 | unsafe fn test_mm512_broadcast_i64x4() { |
10098 | let a = _mm256_set_epi64x(17, 18, 19, 20); |
10099 | let r = _mm512_broadcast_i64x4(a); |
10100 | let e = _mm512_set_epi64(17, 18, 19, 20, 17, 18, 19, 20); |
10101 | assert_eq_m512i(r, e); |
10102 | } |
10103 | |
10104 | #[simd_test(enable = "avx512f" )] |
10105 | unsafe fn test_mm512_mask_broadcast_i64x4() { |
10106 | let src = _mm512_set1_epi64(18); |
10107 | let a = _mm256_set_epi64x(17, 18, 19, 20); |
10108 | let r = _mm512_mask_broadcast_i64x4(src, 0, a); |
10109 | assert_eq_m512i(r, src); |
10110 | let r = _mm512_mask_broadcast_i64x4(src, 0b11111111, a); |
10111 | let e = _mm512_set_epi64(17, 18, 19, 20, 17, 18, 19, 20); |
10112 | assert_eq_m512i(r, e); |
10113 | } |
10114 | |
10115 | #[simd_test(enable = "avx512f" )] |
10116 | unsafe fn test_mm512_maskz_broadcast_i64x4() { |
10117 | let a = _mm256_set_epi64x(17, 18, 19, 20); |
10118 | let r = _mm512_maskz_broadcast_i64x4(0, a); |
10119 | assert_eq_m512i(r, _mm512_setzero_si512()); |
10120 | let r = _mm512_maskz_broadcast_i64x4(0b00001111, a); |
10121 | let e = _mm512_set_epi64(0, 0, 0, 0, 17, 18, 19, 20); |
10122 | assert_eq_m512i(r, e); |
10123 | } |
10124 | |
10125 | #[simd_test(enable = "avx512f" )] |
10126 | unsafe fn test_mm512_broadcast_f64x4() { |
10127 | let a = _mm256_set_pd(17., 18., 19., 20.); |
10128 | let r = _mm512_broadcast_f64x4(a); |
10129 | let e = _mm512_set_pd(17., 18., 19., 20., 17., 18., 19., 20.); |
10130 | assert_eq_m512d(r, e); |
10131 | } |
10132 | |
10133 | #[simd_test(enable = "avx512f" )] |
10134 | unsafe fn test_mm512_mask_broadcast_f64x4() { |
10135 | let src = _mm512_set1_pd(18.); |
10136 | let a = _mm256_set_pd(17., 18., 19., 20.); |
10137 | let r = _mm512_mask_broadcast_f64x4(src, 0, a); |
10138 | assert_eq_m512d(r, src); |
10139 | let r = _mm512_mask_broadcast_f64x4(src, 0b11111111, a); |
10140 | let e = _mm512_set_pd(17., 18., 19., 20., 17., 18., 19., 20.); |
10141 | assert_eq_m512d(r, e); |
10142 | } |
10143 | |
10144 | #[simd_test(enable = "avx512f" )] |
10145 | unsafe fn test_mm512_maskz_broadcast_f64x4() { |
10146 | let a = _mm256_set_pd(17., 18., 19., 20.); |
10147 | let r = _mm512_maskz_broadcast_f64x4(0, a); |
10148 | assert_eq_m512d(r, _mm512_setzero_pd()); |
10149 | let r = _mm512_maskz_broadcast_f64x4(0b00001111, a); |
10150 | let e = _mm512_set_pd(0., 0., 0., 0., 17., 18., 19., 20.); |
10151 | assert_eq_m512d(r, e); |
10152 | } |
10153 | |
10154 | #[simd_test(enable = "avx512f" )] |
10155 | unsafe fn test_mm512_mask_blend_epi64() { |
10156 | let a = _mm512_set1_epi64(1); |
10157 | let b = _mm512_set1_epi64(2); |
10158 | let r = _mm512_mask_blend_epi64(0b11110000, a, b); |
10159 | let e = _mm512_set_epi64(2, 2, 2, 2, 1, 1, 1, 1); |
10160 | assert_eq_m512i(r, e); |
10161 | } |
10162 | |
10163 | #[simd_test(enable = "avx512f,avx512vl" )] |
10164 | unsafe fn test_mm256_mask_blend_epi64() { |
10165 | let a = _mm256_set1_epi64x(1); |
10166 | let b = _mm256_set1_epi64x(2); |
10167 | let r = _mm256_mask_blend_epi64(0b00001111, a, b); |
10168 | let e = _mm256_set1_epi64x(2); |
10169 | assert_eq_m256i(r, e); |
10170 | } |
10171 | |
10172 | #[simd_test(enable = "avx512f,avx512vl" )] |
10173 | unsafe fn test_mm_mask_blend_epi64() { |
10174 | let a = _mm_set1_epi64x(1); |
10175 | let b = _mm_set1_epi64x(2); |
10176 | let r = _mm_mask_blend_epi64(0b00000011, a, b); |
10177 | let e = _mm_set1_epi64x(2); |
10178 | assert_eq_m128i(r, e); |
10179 | } |
10180 | |
10181 | #[simd_test(enable = "avx512f" )] |
10182 | unsafe fn test_mm512_mask_blend_pd() { |
10183 | let a = _mm512_set1_pd(1.); |
10184 | let b = _mm512_set1_pd(2.); |
10185 | let r = _mm512_mask_blend_pd(0b11110000, a, b); |
10186 | let e = _mm512_set_pd(2., 2., 2., 2., 1., 1., 1., 1.); |
10187 | assert_eq_m512d(r, e); |
10188 | } |
10189 | |
10190 | #[simd_test(enable = "avx512f,avx512vl" )] |
10191 | unsafe fn test_mm256_mask_blend_pd() { |
10192 | let a = _mm256_set1_pd(1.); |
10193 | let b = _mm256_set1_pd(2.); |
10194 | let r = _mm256_mask_blend_pd(0b00001111, a, b); |
10195 | let e = _mm256_set1_pd(2.); |
10196 | assert_eq_m256d(r, e); |
10197 | } |
10198 | |
10199 | #[simd_test(enable = "avx512f,avx512vl" )] |
10200 | unsafe fn test_mm_mask_blend_pd() { |
10201 | let a = _mm_set1_pd(1.); |
10202 | let b = _mm_set1_pd(2.); |
10203 | let r = _mm_mask_blend_pd(0b00000011, a, b); |
10204 | let e = _mm_set1_pd(2.); |
10205 | assert_eq_m128d(r, e); |
10206 | } |
10207 | |
10208 | #[simd_test(enable = "avx512f" )] |
10209 | unsafe fn test_mm512_unpackhi_epi64() { |
10210 | let a = _mm512_set_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
10211 | let b = _mm512_set_epi64(17, 18, 19, 20, 21, 22, 23, 24); |
10212 | let r = _mm512_unpackhi_epi64(a, b); |
10213 | let e = _mm512_set_epi64(17, 1, 19, 3, 21, 5, 23, 7); |
10214 | assert_eq_m512i(r, e); |
10215 | } |
10216 | |
10217 | #[simd_test(enable = "avx512f" )] |
10218 | unsafe fn test_mm512_mask_unpackhi_epi64() { |
10219 | let a = _mm512_set_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
10220 | let b = _mm512_set_epi64(17, 18, 19, 20, 21, 22, 23, 24); |
10221 | let r = _mm512_mask_unpackhi_epi64(a, 0, a, b); |
10222 | assert_eq_m512i(r, a); |
10223 | let r = _mm512_mask_unpackhi_epi64(a, 0b11111111, a, b); |
10224 | let e = _mm512_set_epi64(17, 1, 19, 3, 21, 5, 23, 7); |
10225 | assert_eq_m512i(r, e); |
10226 | } |
10227 | |
10228 | #[simd_test(enable = "avx512f" )] |
10229 | unsafe fn test_mm512_maskz_unpackhi_epi64() { |
10230 | let a = _mm512_set_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
10231 | let b = _mm512_set_epi64(17, 18, 19, 20, 21, 22, 23, 24); |
10232 | let r = _mm512_maskz_unpackhi_epi64(0, a, b); |
10233 | assert_eq_m512i(r, _mm512_setzero_si512()); |
10234 | let r = _mm512_maskz_unpackhi_epi64(0b00001111, a, b); |
10235 | let e = _mm512_set_epi64(0, 0, 0, 0, 21, 5, 23, 7); |
10236 | assert_eq_m512i(r, e); |
10237 | } |
10238 | |
10239 | #[simd_test(enable = "avx512f,avx512vl" )] |
10240 | unsafe fn test_mm256_mask_unpackhi_epi64() { |
10241 | let a = _mm256_set_epi64x(1, 2, 3, 4); |
10242 | let b = _mm256_set_epi64x(17, 18, 19, 20); |
10243 | let r = _mm256_mask_unpackhi_epi64(a, 0, a, b); |
10244 | assert_eq_m256i(r, a); |
10245 | let r = _mm256_mask_unpackhi_epi64(a, 0b00001111, a, b); |
10246 | let e = _mm256_set_epi64x(17, 1, 19, 3); |
10247 | assert_eq_m256i(r, e); |
10248 | } |
10249 | |
10250 | #[simd_test(enable = "avx512f,avx512vl" )] |
10251 | unsafe fn test_mm256_maskz_unpackhi_epi64() { |
10252 | let a = _mm256_set_epi64x(1, 2, 3, 4); |
10253 | let b = _mm256_set_epi64x(17, 18, 19, 20); |
10254 | let r = _mm256_maskz_unpackhi_epi64(0, a, b); |
10255 | assert_eq_m256i(r, _mm256_setzero_si256()); |
10256 | let r = _mm256_maskz_unpackhi_epi64(0b00001111, a, b); |
10257 | let e = _mm256_set_epi64x(17, 1, 19, 3); |
10258 | assert_eq_m256i(r, e); |
10259 | } |
10260 | |
10261 | #[simd_test(enable = "avx512f,avx512vl" )] |
10262 | unsafe fn test_mm_mask_unpackhi_epi64() { |
10263 | let a = _mm_set_epi64x(1, 2); |
10264 | let b = _mm_set_epi64x(17, 18); |
10265 | let r = _mm_mask_unpackhi_epi64(a, 0, a, b); |
10266 | assert_eq_m128i(r, a); |
10267 | let r = _mm_mask_unpackhi_epi64(a, 0b00000011, a, b); |
10268 | let e = _mm_set_epi64x(17, 1); |
10269 | assert_eq_m128i(r, e); |
10270 | } |
10271 | |
10272 | #[simd_test(enable = "avx512f,avx512vl" )] |
10273 | unsafe fn test_mm_maskz_unpackhi_epi64() { |
10274 | let a = _mm_set_epi64x(1, 2); |
10275 | let b = _mm_set_epi64x(17, 18); |
10276 | let r = _mm_maskz_unpackhi_epi64(0, a, b); |
10277 | assert_eq_m128i(r, _mm_setzero_si128()); |
10278 | let r = _mm_maskz_unpackhi_epi64(0b00000011, a, b); |
10279 | let e = _mm_set_epi64x(17, 1); |
10280 | assert_eq_m128i(r, e); |
10281 | } |
10282 | |
10283 | #[simd_test(enable = "avx512f" )] |
10284 | unsafe fn test_mm512_unpackhi_pd() { |
10285 | let a = _mm512_set_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
10286 | let b = _mm512_set_pd(17., 18., 19., 20., 21., 22., 23., 24.); |
10287 | let r = _mm512_unpackhi_pd(a, b); |
10288 | let e = _mm512_set_pd(17., 1., 19., 3., 21., 5., 23., 7.); |
10289 | assert_eq_m512d(r, e); |
10290 | } |
10291 | |
10292 | #[simd_test(enable = "avx512f" )] |
10293 | unsafe fn test_mm512_mask_unpackhi_pd() { |
10294 | let a = _mm512_set_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
10295 | let b = _mm512_set_pd(17., 18., 19., 20., 21., 22., 23., 24.); |
10296 | let r = _mm512_mask_unpackhi_pd(a, 0, a, b); |
10297 | assert_eq_m512d(r, a); |
10298 | let r = _mm512_mask_unpackhi_pd(a, 0b11111111, a, b); |
10299 | let e = _mm512_set_pd(17., 1., 19., 3., 21., 5., 23., 7.); |
10300 | assert_eq_m512d(r, e); |
10301 | } |
10302 | |
10303 | #[simd_test(enable = "avx512f" )] |
10304 | unsafe fn test_mm512_maskz_unpackhi_pd() { |
10305 | let a = _mm512_set_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
10306 | let b = _mm512_set_pd(17., 18., 19., 20., 21., 22., 23., 24.); |
10307 | let r = _mm512_maskz_unpackhi_pd(0, a, b); |
10308 | assert_eq_m512d(r, _mm512_setzero_pd()); |
10309 | let r = _mm512_maskz_unpackhi_pd(0b00001111, a, b); |
10310 | let e = _mm512_set_pd(0., 0., 0., 0., 21., 5., 23., 7.); |
10311 | assert_eq_m512d(r, e); |
10312 | } |
10313 | |
10314 | #[simd_test(enable = "avx512f,avx512vl" )] |
10315 | unsafe fn test_mm256_mask_unpackhi_pd() { |
10316 | let a = _mm256_set_pd(1., 2., 3., 4.); |
10317 | let b = _mm256_set_pd(17., 18., 19., 20.); |
10318 | let r = _mm256_mask_unpackhi_pd(a, 0, a, b); |
10319 | assert_eq_m256d(r, a); |
10320 | let r = _mm256_mask_unpackhi_pd(a, 0b00001111, a, b); |
10321 | let e = _mm256_set_pd(17., 1., 19., 3.); |
10322 | assert_eq_m256d(r, e); |
10323 | } |
10324 | |
10325 | #[simd_test(enable = "avx512f,avx512vl" )] |
10326 | unsafe fn test_mm256_maskz_unpackhi_pd() { |
10327 | let a = _mm256_set_pd(1., 2., 3., 4.); |
10328 | let b = _mm256_set_pd(17., 18., 19., 20.); |
10329 | let r = _mm256_maskz_unpackhi_pd(0, a, b); |
10330 | assert_eq_m256d(r, _mm256_setzero_pd()); |
10331 | let r = _mm256_maskz_unpackhi_pd(0b00001111, a, b); |
10332 | let e = _mm256_set_pd(17., 1., 19., 3.); |
10333 | assert_eq_m256d(r, e); |
10334 | } |
10335 | |
10336 | #[simd_test(enable = "avx512f,avx512vl" )] |
10337 | unsafe fn test_mm_mask_unpackhi_pd() { |
10338 | let a = _mm_set_pd(1., 2.); |
10339 | let b = _mm_set_pd(17., 18.); |
10340 | let r = _mm_mask_unpackhi_pd(a, 0, a, b); |
10341 | assert_eq_m128d(r, a); |
10342 | let r = _mm_mask_unpackhi_pd(a, 0b00000011, a, b); |
10343 | let e = _mm_set_pd(17., 1.); |
10344 | assert_eq_m128d(r, e); |
10345 | } |
10346 | |
10347 | #[simd_test(enable = "avx512f,avx512vl" )] |
10348 | unsafe fn test_mm_maskz_unpackhi_pd() { |
10349 | let a = _mm_set_pd(1., 2.); |
10350 | let b = _mm_set_pd(17., 18.); |
10351 | let r = _mm_maskz_unpackhi_pd(0, a, b); |
10352 | assert_eq_m128d(r, _mm_setzero_pd()); |
10353 | let r = _mm_maskz_unpackhi_pd(0b00000011, a, b); |
10354 | let e = _mm_set_pd(17., 1.); |
10355 | assert_eq_m128d(r, e); |
10356 | } |
10357 | |
10358 | #[simd_test(enable = "avx512f" )] |
10359 | unsafe fn test_mm512_unpacklo_epi64() { |
10360 | let a = _mm512_set_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
10361 | let b = _mm512_set_epi64(17, 18, 19, 20, 21, 22, 23, 24); |
10362 | let r = _mm512_unpacklo_epi64(a, b); |
10363 | let e = _mm512_set_epi64(18, 2, 20, 4, 22, 6, 24, 8); |
10364 | assert_eq_m512i(r, e); |
10365 | } |
10366 | |
10367 | #[simd_test(enable = "avx512f" )] |
10368 | unsafe fn test_mm512_mask_unpacklo_epi64() { |
10369 | let a = _mm512_set_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
10370 | let b = _mm512_set_epi64(17, 18, 19, 20, 21, 22, 23, 24); |
10371 | let r = _mm512_mask_unpacklo_epi64(a, 0, a, b); |
10372 | assert_eq_m512i(r, a); |
10373 | let r = _mm512_mask_unpacklo_epi64(a, 0b11111111, a, b); |
10374 | let e = _mm512_set_epi64(18, 2, 20, 4, 22, 6, 24, 8); |
10375 | assert_eq_m512i(r, e); |
10376 | } |
10377 | |
10378 | #[simd_test(enable = "avx512f" )] |
10379 | unsafe fn test_mm512_maskz_unpacklo_epi64() { |
10380 | let a = _mm512_set_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
10381 | let b = _mm512_set_epi64(17, 18, 19, 20, 21, 22, 23, 24); |
10382 | let r = _mm512_maskz_unpacklo_epi64(0, a, b); |
10383 | assert_eq_m512i(r, _mm512_setzero_si512()); |
10384 | let r = _mm512_maskz_unpacklo_epi64(0b00001111, a, b); |
10385 | let e = _mm512_set_epi64(0, 0, 0, 0, 22, 6, 24, 8); |
10386 | assert_eq_m512i(r, e); |
10387 | } |
10388 | |
10389 | #[simd_test(enable = "avx512f,avx512vl" )] |
10390 | unsafe fn test_mm256_mask_unpacklo_epi64() { |
10391 | let a = _mm256_set_epi64x(1, 2, 3, 4); |
10392 | let b = _mm256_set_epi64x(17, 18, 19, 20); |
10393 | let r = _mm256_mask_unpacklo_epi64(a, 0, a, b); |
10394 | assert_eq_m256i(r, a); |
10395 | let r = _mm256_mask_unpacklo_epi64(a, 0b00001111, a, b); |
10396 | let e = _mm256_set_epi64x(18, 2, 20, 4); |
10397 | assert_eq_m256i(r, e); |
10398 | } |
10399 | |
10400 | #[simd_test(enable = "avx512f,avx512vl" )] |
10401 | unsafe fn test_mm256_maskz_unpacklo_epi64() { |
10402 | let a = _mm256_set_epi64x(1, 2, 3, 4); |
10403 | let b = _mm256_set_epi64x(17, 18, 19, 20); |
10404 | let r = _mm256_maskz_unpacklo_epi64(0, a, b); |
10405 | assert_eq_m256i(r, _mm256_setzero_si256()); |
10406 | let r = _mm256_maskz_unpacklo_epi64(0b00001111, a, b); |
10407 | let e = _mm256_set_epi64x(18, 2, 20, 4); |
10408 | assert_eq_m256i(r, e); |
10409 | } |
10410 | |
10411 | #[simd_test(enable = "avx512f,avx512vl" )] |
10412 | unsafe fn test_mm_mask_unpacklo_epi64() { |
10413 | let a = _mm_set_epi64x(1, 2); |
10414 | let b = _mm_set_epi64x(17, 18); |
10415 | let r = _mm_mask_unpacklo_epi64(a, 0, a, b); |
10416 | assert_eq_m128i(r, a); |
10417 | let r = _mm_mask_unpacklo_epi64(a, 0b00000011, a, b); |
10418 | let e = _mm_set_epi64x(18, 2); |
10419 | assert_eq_m128i(r, e); |
10420 | } |
10421 | |
10422 | #[simd_test(enable = "avx512f,avx512vl" )] |
10423 | unsafe fn test_mm_maskz_unpacklo_epi64() { |
10424 | let a = _mm_set_epi64x(1, 2); |
10425 | let b = _mm_set_epi64x(17, 18); |
10426 | let r = _mm_maskz_unpacklo_epi64(0, a, b); |
10427 | assert_eq_m128i(r, _mm_setzero_si128()); |
10428 | let r = _mm_maskz_unpacklo_epi64(0b00000011, a, b); |
10429 | let e = _mm_set_epi64x(18, 2); |
10430 | assert_eq_m128i(r, e); |
10431 | } |
10432 | |
10433 | #[simd_test(enable = "avx512f" )] |
10434 | unsafe fn test_mm512_unpacklo_pd() { |
10435 | let a = _mm512_set_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
10436 | let b = _mm512_set_pd(17., 18., 19., 20., 21., 22., 23., 24.); |
10437 | let r = _mm512_unpacklo_pd(a, b); |
10438 | let e = _mm512_set_pd(18., 2., 20., 4., 22., 6., 24., 8.); |
10439 | assert_eq_m512d(r, e); |
10440 | } |
10441 | |
10442 | #[simd_test(enable = "avx512f" )] |
10443 | unsafe fn test_mm512_mask_unpacklo_pd() { |
10444 | let a = _mm512_set_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
10445 | let b = _mm512_set_pd(17., 18., 19., 20., 21., 22., 23., 24.); |
10446 | let r = _mm512_mask_unpacklo_pd(a, 0, a, b); |
10447 | assert_eq_m512d(r, a); |
10448 | let r = _mm512_mask_unpacklo_pd(a, 0b11111111, a, b); |
10449 | let e = _mm512_set_pd(18., 2., 20., 4., 22., 6., 24., 8.); |
10450 | assert_eq_m512d(r, e); |
10451 | } |
10452 | |
10453 | #[simd_test(enable = "avx512f" )] |
10454 | unsafe fn test_mm512_maskz_unpacklo_pd() { |
10455 | let a = _mm512_set_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
10456 | let b = _mm512_set_pd(17., 18., 19., 20., 21., 22., 23., 24.); |
10457 | let r = _mm512_maskz_unpacklo_pd(0, a, b); |
10458 | assert_eq_m512d(r, _mm512_setzero_pd()); |
10459 | let r = _mm512_maskz_unpacklo_pd(0b00001111, a, b); |
10460 | let e = _mm512_set_pd(0., 0., 0., 0., 22., 6., 24., 8.); |
10461 | assert_eq_m512d(r, e); |
10462 | } |
10463 | |
10464 | #[simd_test(enable = "avx512f,avx512vl" )] |
10465 | unsafe fn test_mm256_mask_unpacklo_pd() { |
10466 | let a = _mm256_set_pd(1., 2., 3., 4.); |
10467 | let b = _mm256_set_pd(17., 18., 19., 20.); |
10468 | let r = _mm256_mask_unpacklo_pd(a, 0, a, b); |
10469 | assert_eq_m256d(r, a); |
10470 | let r = _mm256_mask_unpacklo_pd(a, 0b00001111, a, b); |
10471 | let e = _mm256_set_pd(18., 2., 20., 4.); |
10472 | assert_eq_m256d(r, e); |
10473 | } |
10474 | |
10475 | #[simd_test(enable = "avx512f,avx512vl" )] |
10476 | unsafe fn test_mm256_maskz_unpacklo_pd() { |
10477 | let a = _mm256_set_pd(1., 2., 3., 4.); |
10478 | let b = _mm256_set_pd(17., 18., 19., 20.); |
10479 | let r = _mm256_maskz_unpacklo_pd(0, a, b); |
10480 | assert_eq_m256d(r, _mm256_setzero_pd()); |
10481 | let r = _mm256_maskz_unpacklo_pd(0b00001111, a, b); |
10482 | let e = _mm256_set_pd(18., 2., 20., 4.); |
10483 | assert_eq_m256d(r, e); |
10484 | } |
10485 | |
10486 | #[simd_test(enable = "avx512f,avx512vl" )] |
10487 | unsafe fn test_mm_mask_unpacklo_pd() { |
10488 | let a = _mm_set_pd(1., 2.); |
10489 | let b = _mm_set_pd(17., 18.); |
10490 | let r = _mm_mask_unpacklo_pd(a, 0, a, b); |
10491 | assert_eq_m128d(r, a); |
10492 | let r = _mm_mask_unpacklo_pd(a, 0b00000011, a, b); |
10493 | let e = _mm_set_pd(18., 2.); |
10494 | assert_eq_m128d(r, e); |
10495 | } |
10496 | |
10497 | #[simd_test(enable = "avx512f,avx512vl" )] |
10498 | unsafe fn test_mm_maskz_unpacklo_pd() { |
10499 | let a = _mm_set_pd(1., 2.); |
10500 | let b = _mm_set_pd(17., 18.); |
10501 | let r = _mm_maskz_unpacklo_pd(0, a, b); |
10502 | assert_eq_m128d(r, _mm_setzero_pd()); |
10503 | let r = _mm_maskz_unpacklo_pd(0b00000011, a, b); |
10504 | let e = _mm_set_pd(18., 2.); |
10505 | assert_eq_m128d(r, e); |
10506 | } |
10507 | |
10508 | #[simd_test(enable = "avx512f" )] |
10509 | unsafe fn test_mm512_alignr_epi64() { |
10510 | let a = _mm512_set_epi64(8, 7, 6, 5, 4, 3, 2, 1); |
10511 | let b = _mm512_set_epi64(16, 15, 14, 13, 12, 11, 10, 9); |
10512 | let r = _mm512_alignr_epi64::<0>(a, b); |
10513 | assert_eq_m512i(r, b); |
10514 | let r = _mm512_alignr_epi64::<8>(a, b); |
10515 | assert_eq_m512i(r, b); |
10516 | let r = _mm512_alignr_epi64::<1>(a, b); |
10517 | let e = _mm512_set_epi64(1, 16, 15, 14, 13, 12, 11, 10); |
10518 | assert_eq_m512i(r, e); |
10519 | } |
10520 | |
10521 | #[simd_test(enable = "avx512f" )] |
10522 | unsafe fn test_mm512_mask_alignr_epi64() { |
10523 | let a = _mm512_set_epi64(8, 7, 6, 5, 4, 3, 2, 1); |
10524 | let b = _mm512_set_epi64(16, 15, 14, 13, 12, 11, 10, 9); |
10525 | let r = _mm512_mask_alignr_epi64::<1>(a, 0, a, b); |
10526 | assert_eq_m512i(r, a); |
10527 | let r = _mm512_mask_alignr_epi64::<1>(a, 0b11111111, a, b); |
10528 | let e = _mm512_set_epi64(1, 16, 15, 14, 13, 12, 11, 10); |
10529 | assert_eq_m512i(r, e); |
10530 | } |
10531 | |
10532 | #[simd_test(enable = "avx512f" )] |
10533 | unsafe fn test_mm512_maskz_alignr_epi64() { |
10534 | let a = _mm512_set_epi64(8, 7, 6, 5, 4, 3, 2, 1); |
10535 | let b = _mm512_set_epi64(16, 15, 14, 13, 12, 11, 10, 9); |
10536 | let r = _mm512_maskz_alignr_epi64::<1>(0, a, b); |
10537 | assert_eq_m512i(r, _mm512_setzero_si512()); |
10538 | let r = _mm512_maskz_alignr_epi64::<1>(0b00001111, a, b); |
10539 | let e = _mm512_set_epi64(0, 0, 0, 0, 13, 12, 11, 10); |
10540 | assert_eq_m512i(r, e); |
10541 | } |
10542 | |
10543 | #[simd_test(enable = "avx512f,avx512vl" )] |
10544 | unsafe fn test_mm256_alignr_epi64() { |
10545 | let a = _mm256_set_epi64x(4, 3, 2, 1); |
10546 | let b = _mm256_set_epi64x(8, 7, 6, 5); |
10547 | let r = _mm256_alignr_epi64::<0>(a, b); |
10548 | let e = _mm256_set_epi64x(8, 7, 6, 5); |
10549 | assert_eq_m256i(r, e); |
10550 | let r = _mm256_alignr_epi64::<6>(a, b); |
10551 | let e = _mm256_set_epi64x(6, 5, 4, 3); |
10552 | assert_eq_m256i(r, e); |
10553 | } |
10554 | |
10555 | #[simd_test(enable = "avx512f,avx512vl" )] |
10556 | unsafe fn test_mm256_mask_alignr_epi64() { |
10557 | let a = _mm256_set_epi64x(4, 3, 2, 1); |
10558 | let b = _mm256_set_epi64x(8, 7, 6, 5); |
10559 | let r = _mm256_mask_alignr_epi64::<1>(a, 0, a, b); |
10560 | assert_eq_m256i(r, a); |
10561 | let r = _mm256_mask_alignr_epi64::<0>(a, 0b00001111, a, b); |
10562 | let e = _mm256_set_epi64x(8, 7, 6, 5); |
10563 | assert_eq_m256i(r, e); |
10564 | } |
10565 | |
10566 | #[simd_test(enable = "avx512f,avx512vl" )] |
10567 | unsafe fn test_mm256_maskz_alignr_epi64() { |
10568 | let a = _mm256_set_epi64x(4, 3, 2, 1); |
10569 | let b = _mm256_set_epi64x(8, 7, 6, 5); |
10570 | let r = _mm256_maskz_alignr_epi64::<1>(0, a, b); |
10571 | assert_eq_m256i(r, _mm256_setzero_si256()); |
10572 | let r = _mm256_maskz_alignr_epi64::<0>(0b00001111, a, b); |
10573 | let e = _mm256_set_epi64x(8, 7, 6, 5); |
10574 | assert_eq_m256i(r, e); |
10575 | } |
10576 | |
10577 | #[simd_test(enable = "avx512f,avx512vl" )] |
10578 | unsafe fn test_mm_alignr_epi64() { |
10579 | let a = _mm_set_epi64x(2, 1); |
10580 | let b = _mm_set_epi64x(4, 3); |
10581 | let r = _mm_alignr_epi64::<0>(a, b); |
10582 | let e = _mm_set_epi64x(4, 3); |
10583 | assert_eq_m128i(r, e); |
10584 | } |
10585 | |
10586 | #[simd_test(enable = "avx512f,avx512vl" )] |
10587 | unsafe fn test_mm_mask_alignr_epi64() { |
10588 | let a = _mm_set_epi64x(2, 1); |
10589 | let b = _mm_set_epi64x(4, 3); |
10590 | let r = _mm_mask_alignr_epi64::<1>(a, 0, a, b); |
10591 | assert_eq_m128i(r, a); |
10592 | let r = _mm_mask_alignr_epi64::<0>(a, 0b00000011, a, b); |
10593 | let e = _mm_set_epi64x(4, 3); |
10594 | assert_eq_m128i(r, e); |
10595 | } |
10596 | |
10597 | #[simd_test(enable = "avx512f,avx512vl" )] |
10598 | unsafe fn test_mm_maskz_alignr_epi64() { |
10599 | let a = _mm_set_epi64x(2, 1); |
10600 | let b = _mm_set_epi64x(4, 3); |
10601 | let r = _mm_maskz_alignr_epi64::<1>(0, a, b); |
10602 | assert_eq_m128i(r, _mm_setzero_si128()); |
10603 | let r = _mm_maskz_alignr_epi64::<0>(0b00000011, a, b); |
10604 | let e = _mm_set_epi64x(4, 3); |
10605 | assert_eq_m128i(r, e); |
10606 | } |
10607 | |
10608 | #[simd_test(enable = "avx512f" )] |
10609 | unsafe fn test_mm512_and_epi64() { |
10610 | let a = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10611 | let b = _mm512_set_epi64(1 << 13, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10612 | let r = _mm512_and_epi64(a, b); |
10613 | let e = _mm512_set_epi64(0, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10614 | assert_eq_m512i(r, e); |
10615 | } |
10616 | |
10617 | #[simd_test(enable = "avx512f" )] |
10618 | unsafe fn test_mm512_mask_and_epi64() { |
10619 | let a = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10620 | let b = _mm512_set_epi64(1 << 13, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10621 | let r = _mm512_mask_and_epi64(a, 0, a, b); |
10622 | assert_eq_m512i(r, a); |
10623 | let r = _mm512_mask_and_epi64(a, 0b01111111, a, b); |
10624 | let e = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10625 | assert_eq_m512i(r, e); |
10626 | } |
10627 | |
10628 | #[simd_test(enable = "avx512f" )] |
10629 | unsafe fn test_mm512_maskz_and_epi64() { |
10630 | let a = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10631 | let b = _mm512_set_epi64(1 << 13, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10632 | let r = _mm512_maskz_and_epi64(0, a, b); |
10633 | assert_eq_m512i(r, _mm512_setzero_si512()); |
10634 | let r = _mm512_maskz_and_epi64(0b00001111, a, b); |
10635 | let e = _mm512_set_epi64(0, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10636 | assert_eq_m512i(r, e); |
10637 | } |
10638 | |
10639 | #[simd_test(enable = "avx512f,avx512vl" )] |
10640 | unsafe fn test_mm256_mask_and_epi64() { |
10641 | let a = _mm256_set1_epi64x(1 << 0 | 1 << 15); |
10642 | let b = _mm256_set1_epi64x(1 << 0); |
10643 | let r = _mm256_mask_and_epi64(a, 0, a, b); |
10644 | assert_eq_m256i(r, a); |
10645 | let r = _mm256_mask_and_epi64(a, 0b00001111, a, b); |
10646 | let e = _mm256_set1_epi64x(1 << 0); |
10647 | assert_eq_m256i(r, e); |
10648 | } |
10649 | |
10650 | #[simd_test(enable = "avx512f,avx512vl" )] |
10651 | unsafe fn test_mm256_maskz_and_epi64() { |
10652 | let a = _mm256_set1_epi64x(1 << 0 | 1 << 15); |
10653 | let b = _mm256_set1_epi64x(1 << 0); |
10654 | let r = _mm256_maskz_and_epi64(0, a, b); |
10655 | assert_eq_m256i(r, _mm256_setzero_si256()); |
10656 | let r = _mm256_maskz_and_epi64(0b00001111, a, b); |
10657 | let e = _mm256_set1_epi64x(1 << 0); |
10658 | assert_eq_m256i(r, e); |
10659 | } |
10660 | |
10661 | #[simd_test(enable = "avx512f,avx512vl" )] |
10662 | unsafe fn test_mm_mask_and_epi64() { |
10663 | let a = _mm_set1_epi64x(1 << 0 | 1 << 15); |
10664 | let b = _mm_set1_epi64x(1 << 0); |
10665 | let r = _mm_mask_and_epi64(a, 0, a, b); |
10666 | assert_eq_m128i(r, a); |
10667 | let r = _mm_mask_and_epi64(a, 0b00000011, a, b); |
10668 | let e = _mm_set1_epi64x(1 << 0); |
10669 | assert_eq_m128i(r, e); |
10670 | } |
10671 | |
10672 | #[simd_test(enable = "avx512f,avx512vl" )] |
10673 | unsafe fn test_mm_maskz_and_epi64() { |
10674 | let a = _mm_set1_epi64x(1 << 0 | 1 << 15); |
10675 | let b = _mm_set1_epi64x(1 << 0); |
10676 | let r = _mm_maskz_and_epi64(0, a, b); |
10677 | assert_eq_m128i(r, _mm_setzero_si128()); |
10678 | let r = _mm_maskz_and_epi64(0b00000011, a, b); |
10679 | let e = _mm_set1_epi64x(1 << 0); |
10680 | assert_eq_m128i(r, e); |
10681 | } |
10682 | |
10683 | #[simd_test(enable = "avx512f" )] |
10684 | unsafe fn test_mm512_and_si512() { |
10685 | let a = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10686 | let b = _mm512_set_epi64(1 << 13, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10687 | let r = _mm512_and_epi64(a, b); |
10688 | let e = _mm512_set_epi64(0, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10689 | assert_eq_m512i(r, e); |
10690 | } |
10691 | |
10692 | #[simd_test(enable = "avx512f" )] |
10693 | unsafe fn test_mm512_or_epi64() { |
10694 | let a = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10695 | let b = _mm512_set_epi64(1 << 13, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10696 | let r = _mm512_or_epi64(a, b); |
10697 | #[rustfmt::skip] |
10698 | let e = _mm512_set_epi64( |
10699 | 1 << 0 | 1 << 13 | 1 << 15, 0, 0, 0, |
10700 | 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3, |
10701 | ); |
10702 | assert_eq_m512i(r, e); |
10703 | } |
10704 | |
10705 | #[simd_test(enable = "avx512f" )] |
10706 | unsafe fn test_mm512_mask_or_epi64() { |
10707 | let a = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10708 | let b = _mm512_set_epi64(1 << 13, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10709 | let r = _mm512_mask_or_epi64(a, 0, a, b); |
10710 | assert_eq_m512i(r, a); |
10711 | let r = _mm512_mask_or_epi64(a, 0b11111111, a, b); |
10712 | #[rustfmt::skip] |
10713 | let e = _mm512_set_epi64( |
10714 | 1 << 0 | 1 << 13 | 1 << 15, 0, 0, 0, |
10715 | 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3, |
10716 | ); |
10717 | assert_eq_m512i(r, e); |
10718 | } |
10719 | |
10720 | #[simd_test(enable = "avx512f" )] |
10721 | unsafe fn test_mm512_maskz_or_epi64() { |
10722 | let a = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10723 | let b = _mm512_set_epi64(1 << 13, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10724 | let r = _mm512_maskz_or_epi64(0, a, b); |
10725 | assert_eq_m512i(r, _mm512_setzero_si512()); |
10726 | let r = _mm512_maskz_or_epi64(0b00001111, a, b); |
10727 | let e = _mm512_set_epi64(0, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10728 | assert_eq_m512i(r, e); |
10729 | } |
10730 | |
10731 | #[simd_test(enable = "avx512f,avx512vl" )] |
10732 | unsafe fn test_mm256_or_epi64() { |
10733 | let a = _mm256_set1_epi64x(1 << 0 | 1 << 15); |
10734 | let b = _mm256_set1_epi64x(1 << 13); |
10735 | let r = _mm256_or_epi64(a, b); |
10736 | let e = _mm256_set1_epi64x(1 << 0 | 1 << 13 | 1 << 15); |
10737 | assert_eq_m256i(r, e); |
10738 | } |
10739 | |
10740 | #[simd_test(enable = "avx512f,avx512vl" )] |
10741 | unsafe fn test_mm256_mask_or_epi64() { |
10742 | let a = _mm256_set1_epi64x(1 << 0 | 1 << 15); |
10743 | let b = _mm256_set1_epi64x(1 << 13); |
10744 | let r = _mm256_mask_or_epi64(a, 0, a, b); |
10745 | assert_eq_m256i(r, a); |
10746 | let r = _mm256_mask_or_epi64(a, 0b00001111, a, b); |
10747 | let e = _mm256_set1_epi64x(1 << 0 | 1 << 13 | 1 << 15); |
10748 | assert_eq_m256i(r, e); |
10749 | } |
10750 | |
10751 | #[simd_test(enable = "avx512f,avx512vl" )] |
10752 | unsafe fn test_mm256_maskz_or_epi64() { |
10753 | let a = _mm256_set1_epi64x(1 << 0 | 1 << 15); |
10754 | let b = _mm256_set1_epi64x(1 << 13); |
10755 | let r = _mm256_maskz_or_epi64(0, a, b); |
10756 | assert_eq_m256i(r, _mm256_setzero_si256()); |
10757 | let r = _mm256_maskz_or_epi64(0b00001111, a, b); |
10758 | let e = _mm256_set1_epi64x(1 << 0 | 1 << 13 | 1 << 15); |
10759 | assert_eq_m256i(r, e); |
10760 | } |
10761 | |
10762 | #[simd_test(enable = "avx512f,avx512vl" )] |
10763 | unsafe fn test_mm_or_epi64() { |
10764 | let a = _mm_set1_epi64x(1 << 0 | 1 << 15); |
10765 | let b = _mm_set1_epi64x(1 << 13); |
10766 | let r = _mm_or_epi64(a, b); |
10767 | let e = _mm_set1_epi64x(1 << 0 | 1 << 13 | 1 << 15); |
10768 | assert_eq_m128i(r, e); |
10769 | } |
10770 | |
10771 | #[simd_test(enable = "avx512f,avx512vl" )] |
10772 | unsafe fn test_mm_mask_or_epi64() { |
10773 | let a = _mm_set1_epi64x(1 << 0 | 1 << 15); |
10774 | let b = _mm_set1_epi64x(1 << 13); |
10775 | let r = _mm_mask_or_epi64(a, 0, a, b); |
10776 | assert_eq_m128i(r, a); |
10777 | let r = _mm_mask_or_epi64(a, 0b00000011, a, b); |
10778 | let e = _mm_set1_epi64x(1 << 0 | 1 << 13 | 1 << 15); |
10779 | assert_eq_m128i(r, e); |
10780 | } |
10781 | |
10782 | #[simd_test(enable = "avx512f,avx512vl" )] |
10783 | unsafe fn test_mm_maskz_or_epi64() { |
10784 | let a = _mm_set1_epi64x(1 << 0 | 1 << 15); |
10785 | let b = _mm_set1_epi64x(1 << 13); |
10786 | let r = _mm_maskz_or_epi64(0, a, b); |
10787 | assert_eq_m128i(r, _mm_setzero_si128()); |
10788 | let r = _mm_maskz_or_epi64(0b00000011, a, b); |
10789 | let e = _mm_set1_epi64x(1 << 0 | 1 << 13 | 1 << 15); |
10790 | assert_eq_m128i(r, e); |
10791 | } |
10792 | |
10793 | #[simd_test(enable = "avx512f" )] |
10794 | unsafe fn test_mm512_or_si512() { |
10795 | let a = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10796 | let b = _mm512_set_epi64(1 << 13, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10797 | let r = _mm512_or_epi64(a, b); |
10798 | #[rustfmt::skip] |
10799 | let e = _mm512_set_epi64( |
10800 | 1 << 0 | 1 << 13 | 1 << 15, 0, 0, 0, |
10801 | 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3, |
10802 | ); |
10803 | assert_eq_m512i(r, e); |
10804 | } |
10805 | |
10806 | #[simd_test(enable = "avx512f" )] |
10807 | unsafe fn test_mm512_xor_epi64() { |
10808 | let a = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10809 | let b = _mm512_set_epi64(1 << 13, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10810 | let r = _mm512_xor_epi64(a, b); |
10811 | let e = _mm512_set_epi64(1 << 0 | 1 << 13 | 1 << 15, 0, 0, 0, 0, 0, 0, 0); |
10812 | assert_eq_m512i(r, e); |
10813 | } |
10814 | |
10815 | #[simd_test(enable = "avx512f" )] |
10816 | unsafe fn test_mm512_mask_xor_epi64() { |
10817 | let a = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10818 | let b = _mm512_set_epi64(1 << 13, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10819 | let r = _mm512_mask_xor_epi64(a, 0, a, b); |
10820 | assert_eq_m512i(r, a); |
10821 | let r = _mm512_mask_xor_epi64(a, 0b11111111, a, b); |
10822 | let e = _mm512_set_epi64(1 << 0 | 1 << 13 | 1 << 15, 0, 0, 0, 0, 0, 0, 0); |
10823 | assert_eq_m512i(r, e); |
10824 | } |
10825 | |
10826 | #[simd_test(enable = "avx512f" )] |
10827 | unsafe fn test_mm512_maskz_xor_epi64() { |
10828 | let a = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10829 | let b = _mm512_set_epi64(1 << 13, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10830 | let r = _mm512_maskz_xor_epi64(0, a, b); |
10831 | assert_eq_m512i(r, _mm512_setzero_si512()); |
10832 | let r = _mm512_maskz_xor_epi64(0b00001111, a, b); |
10833 | let e = _mm512_set_epi64(0, 0, 0, 0, 0, 0, 0, 0); |
10834 | assert_eq_m512i(r, e); |
10835 | } |
10836 | |
10837 | #[simd_test(enable = "avx512f,avx512vl" )] |
10838 | unsafe fn test_mm256_xor_epi64() { |
10839 | let a = _mm256_set1_epi64x(1 << 0 | 1 << 15); |
10840 | let b = _mm256_set1_epi64x(1 << 13); |
10841 | let r = _mm256_xor_epi64(a, b); |
10842 | let e = _mm256_set1_epi64x(1 << 0 | 1 << 13 | 1 << 15); |
10843 | assert_eq_m256i(r, e); |
10844 | } |
10845 | |
10846 | #[simd_test(enable = "avx512f,avx512vl" )] |
10847 | unsafe fn test_mm256_mask_xor_epi64() { |
10848 | let a = _mm256_set1_epi64x(1 << 0 | 1 << 15); |
10849 | let b = _mm256_set1_epi64x(1 << 13); |
10850 | let r = _mm256_mask_xor_epi64(a, 0, a, b); |
10851 | assert_eq_m256i(r, a); |
10852 | let r = _mm256_mask_xor_epi64(a, 0b00001111, a, b); |
10853 | let e = _mm256_set1_epi64x(1 << 0 | 1 << 13 | 1 << 15); |
10854 | assert_eq_m256i(r, e); |
10855 | } |
10856 | |
10857 | #[simd_test(enable = "avx512f,avx512vl" )] |
10858 | unsafe fn test_mm256_maskz_xor_epi64() { |
10859 | let a = _mm256_set1_epi64x(1 << 0 | 1 << 15); |
10860 | let b = _mm256_set1_epi64x(1 << 13); |
10861 | let r = _mm256_maskz_xor_epi64(0, a, b); |
10862 | assert_eq_m256i(r, _mm256_setzero_si256()); |
10863 | let r = _mm256_maskz_xor_epi64(0b00001111, a, b); |
10864 | let e = _mm256_set1_epi64x(1 << 0 | 1 << 13 | 1 << 15); |
10865 | assert_eq_m256i(r, e); |
10866 | } |
10867 | |
10868 | #[simd_test(enable = "avx512f,avx512vl" )] |
10869 | unsafe fn test_mm_xor_epi64() { |
10870 | let a = _mm_set1_epi64x(1 << 0 | 1 << 15); |
10871 | let b = _mm_set1_epi64x(1 << 13); |
10872 | let r = _mm_xor_epi64(a, b); |
10873 | let e = _mm_set1_epi64x(1 << 0 | 1 << 13 | 1 << 15); |
10874 | assert_eq_m128i(r, e); |
10875 | } |
10876 | |
10877 | #[simd_test(enable = "avx512f,avx512vl" )] |
10878 | unsafe fn test_mm_mask_xor_epi64() { |
10879 | let a = _mm_set1_epi64x(1 << 0 | 1 << 15); |
10880 | let b = _mm_set1_epi64x(1 << 13); |
10881 | let r = _mm_mask_xor_epi64(a, 0, a, b); |
10882 | assert_eq_m128i(r, a); |
10883 | let r = _mm_mask_xor_epi64(a, 0b00000011, a, b); |
10884 | let e = _mm_set1_epi64x(1 << 0 | 1 << 13 | 1 << 15); |
10885 | assert_eq_m128i(r, e); |
10886 | } |
10887 | |
10888 | #[simd_test(enable = "avx512f,avx512vl" )] |
10889 | unsafe fn test_mm_maskz_xor_epi64() { |
10890 | let a = _mm_set1_epi64x(1 << 0 | 1 << 15); |
10891 | let b = _mm_set1_epi64x(1 << 13); |
10892 | let r = _mm_maskz_xor_epi64(0, a, b); |
10893 | assert_eq_m128i(r, _mm_setzero_si128()); |
10894 | let r = _mm_maskz_xor_epi64(0b00000011, a, b); |
10895 | let e = _mm_set1_epi64x(1 << 0 | 1 << 13 | 1 << 15); |
10896 | assert_eq_m128i(r, e); |
10897 | } |
10898 | |
10899 | #[simd_test(enable = "avx512f" )] |
10900 | unsafe fn test_mm512_xor_si512() { |
10901 | let a = _mm512_set_epi64(1 << 0 | 1 << 15, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10902 | let b = _mm512_set_epi64(1 << 13, 0, 0, 0, 0, 0, 0, 1 << 1 | 1 << 2 | 1 << 3); |
10903 | let r = _mm512_xor_epi64(a, b); |
10904 | let e = _mm512_set_epi64(1 << 0 | 1 << 13 | 1 << 15, 0, 0, 0, 0, 0, 0, 0); |
10905 | assert_eq_m512i(r, e); |
10906 | } |
10907 | |
10908 | #[simd_test(enable = "avx512f" )] |
10909 | unsafe fn test_mm512_andnot_epi64() { |
10910 | let a = _mm512_set1_epi64(0); |
10911 | let b = _mm512_set1_epi64(1 << 3 | 1 << 4); |
10912 | let r = _mm512_andnot_epi64(a, b); |
10913 | let e = _mm512_set1_epi64(1 << 3 | 1 << 4); |
10914 | assert_eq_m512i(r, e); |
10915 | } |
10916 | |
10917 | #[simd_test(enable = "avx512f" )] |
10918 | unsafe fn test_mm512_mask_andnot_epi64() { |
10919 | let a = _mm512_set1_epi64(1 << 1 | 1 << 2); |
10920 | let b = _mm512_set1_epi64(1 << 3 | 1 << 4); |
10921 | let r = _mm512_mask_andnot_epi64(a, 0, a, b); |
10922 | assert_eq_m512i(r, a); |
10923 | let r = _mm512_mask_andnot_epi64(a, 0b11111111, a, b); |
10924 | let e = _mm512_set1_epi64(1 << 3 | 1 << 4); |
10925 | assert_eq_m512i(r, e); |
10926 | } |
10927 | |
10928 | #[simd_test(enable = "avx512f" )] |
10929 | unsafe fn test_mm512_maskz_andnot_epi64() { |
10930 | let a = _mm512_set1_epi64(1 << 1 | 1 << 2); |
10931 | let b = _mm512_set1_epi64(1 << 3 | 1 << 4); |
10932 | let r = _mm512_maskz_andnot_epi64(0, a, b); |
10933 | assert_eq_m512i(r, _mm512_setzero_si512()); |
10934 | let r = _mm512_maskz_andnot_epi64(0b00001111, a, b); |
10935 | #[rustfmt::skip] |
10936 | let e = _mm512_set_epi64( |
10937 | 0, 0, 0, 0, |
10938 | 1 << 3 | 1 << 4, 1 << 3 | 1 << 4, 1 << 3 | 1 << 4, 1 << 3 | 1 << 4, |
10939 | ); |
10940 | assert_eq_m512i(r, e); |
10941 | } |
10942 | |
10943 | #[simd_test(enable = "avx512f,avx512vl" )] |
10944 | unsafe fn test_mm256_mask_andnot_epi64() { |
10945 | let a = _mm256_set1_epi64x(1 << 1 | 1 << 2); |
10946 | let b = _mm256_set1_epi64x(1 << 3 | 1 << 4); |
10947 | let r = _mm256_mask_andnot_epi64(a, 0, a, b); |
10948 | assert_eq_m256i(r, a); |
10949 | let r = _mm256_mask_andnot_epi64(a, 0b00001111, a, b); |
10950 | let e = _mm256_set1_epi64x(1 << 3 | 1 << 4); |
10951 | assert_eq_m256i(r, e); |
10952 | } |
10953 | |
10954 | #[simd_test(enable = "avx512f,avx512vl" )] |
10955 | unsafe fn test_mm256_maskz_andnot_epi64() { |
10956 | let a = _mm256_set1_epi64x(1 << 1 | 1 << 2); |
10957 | let b = _mm256_set1_epi64x(1 << 3 | 1 << 4); |
10958 | let r = _mm256_maskz_andnot_epi64(0, a, b); |
10959 | assert_eq_m256i(r, _mm256_setzero_si256()); |
10960 | let r = _mm256_maskz_andnot_epi64(0b00001111, a, b); |
10961 | let e = _mm256_set1_epi64x(1 << 3 | 1 << 4); |
10962 | assert_eq_m256i(r, e); |
10963 | } |
10964 | |
10965 | #[simd_test(enable = "avx512f,avx512vl" )] |
10966 | unsafe fn test_mm_mask_andnot_epi64() { |
10967 | let a = _mm_set1_epi64x(1 << 1 | 1 << 2); |
10968 | let b = _mm_set1_epi64x(1 << 3 | 1 << 4); |
10969 | let r = _mm_mask_andnot_epi64(a, 0, a, b); |
10970 | assert_eq_m128i(r, a); |
10971 | let r = _mm_mask_andnot_epi64(a, 0b00000011, a, b); |
10972 | let e = _mm_set1_epi64x(1 << 3 | 1 << 4); |
10973 | assert_eq_m128i(r, e); |
10974 | } |
10975 | |
10976 | #[simd_test(enable = "avx512f,avx512vl" )] |
10977 | unsafe fn test_mm_maskz_andnot_epi64() { |
10978 | let a = _mm_set1_epi64x(1 << 1 | 1 << 2); |
10979 | let b = _mm_set1_epi64x(1 << 3 | 1 << 4); |
10980 | let r = _mm_maskz_andnot_epi64(0, a, b); |
10981 | assert_eq_m128i(r, _mm_setzero_si128()); |
10982 | let r = _mm_maskz_andnot_epi64(0b00000011, a, b); |
10983 | let e = _mm_set1_epi64x(1 << 3 | 1 << 4); |
10984 | assert_eq_m128i(r, e); |
10985 | } |
10986 | |
10987 | #[simd_test(enable = "avx512f" )] |
10988 | unsafe fn test_mm512_andnot_si512() { |
10989 | let a = _mm512_set1_epi64(0); |
10990 | let b = _mm512_set1_epi64(1 << 3 | 1 << 4); |
10991 | let r = _mm512_andnot_si512(a, b); |
10992 | let e = _mm512_set1_epi64(1 << 3 | 1 << 4); |
10993 | assert_eq_m512i(r, e); |
10994 | } |
10995 | |
10996 | #[simd_test(enable = "avx512f" )] |
10997 | unsafe fn test_mm512_reduce_add_epi64() { |
10998 | let a = _mm512_set1_epi64(1); |
10999 | let e: i64 = _mm512_reduce_add_epi64(a); |
11000 | assert_eq!(8, e); |
11001 | } |
11002 | |
11003 | #[simd_test(enable = "avx512f" )] |
11004 | unsafe fn test_mm512_mask_reduce_add_epi64() { |
11005 | let a = _mm512_set1_epi64(1); |
11006 | let e: i64 = _mm512_mask_reduce_add_epi64(0b11110000, a); |
11007 | assert_eq!(4, e); |
11008 | } |
11009 | |
11010 | #[simd_test(enable = "avx512f" )] |
11011 | unsafe fn test_mm512_reduce_add_pd() { |
11012 | let a = _mm512_set1_pd(1.); |
11013 | let e: f64 = _mm512_reduce_add_pd(a); |
11014 | assert_eq!(8., e); |
11015 | } |
11016 | |
11017 | #[simd_test(enable = "avx512f" )] |
11018 | unsafe fn test_mm512_mask_reduce_add_pd() { |
11019 | let a = _mm512_set1_pd(1.); |
11020 | let e: f64 = _mm512_mask_reduce_add_pd(0b11110000, a); |
11021 | assert_eq!(4., e); |
11022 | } |
11023 | |
11024 | #[simd_test(enable = "avx512f" )] |
11025 | unsafe fn test_mm512_reduce_mul_epi64() { |
11026 | let a = _mm512_set1_epi64(2); |
11027 | let e: i64 = _mm512_reduce_mul_epi64(a); |
11028 | assert_eq!(256, e); |
11029 | } |
11030 | |
11031 | #[simd_test(enable = "avx512f" )] |
11032 | unsafe fn test_mm512_mask_reduce_mul_epi64() { |
11033 | let a = _mm512_set1_epi64(2); |
11034 | let e: i64 = _mm512_mask_reduce_mul_epi64(0b11110000, a); |
11035 | assert_eq!(16, e); |
11036 | } |
11037 | |
11038 | #[simd_test(enable = "avx512f" )] |
11039 | unsafe fn test_mm512_reduce_mul_pd() { |
11040 | let a = _mm512_set1_pd(2.); |
11041 | let e: f64 = _mm512_reduce_mul_pd(a); |
11042 | assert_eq!(256., e); |
11043 | } |
11044 | |
11045 | #[simd_test(enable = "avx512f" )] |
11046 | unsafe fn test_mm512_mask_reduce_mul_pd() { |
11047 | let a = _mm512_set1_pd(2.); |
11048 | let e: f64 = _mm512_mask_reduce_mul_pd(0b11110000, a); |
11049 | assert_eq!(16., e); |
11050 | } |
11051 | |
11052 | #[simd_test(enable = "avx512f" )] |
11053 | unsafe fn test_mm512_reduce_max_epi64() { |
11054 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
11055 | let e: i64 = _mm512_reduce_max_epi64(a); |
11056 | assert_eq!(7, e); |
11057 | } |
11058 | |
11059 | #[simd_test(enable = "avx512f" )] |
11060 | unsafe fn test_mm512_mask_reduce_max_epi64() { |
11061 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
11062 | let e: i64 = _mm512_mask_reduce_max_epi64(0b11110000, a); |
11063 | assert_eq!(3, e); |
11064 | } |
11065 | |
11066 | #[simd_test(enable = "avx512f" )] |
11067 | unsafe fn test_mm512_reduce_max_epu64() { |
11068 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
11069 | let e: u64 = _mm512_reduce_max_epu64(a); |
11070 | assert_eq!(7, e); |
11071 | } |
11072 | |
11073 | #[simd_test(enable = "avx512f" )] |
11074 | unsafe fn test_mm512_mask_reduce_max_epu64() { |
11075 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
11076 | let e: u64 = _mm512_mask_reduce_max_epu64(0b11110000, a); |
11077 | assert_eq!(3, e); |
11078 | } |
11079 | |
11080 | #[simd_test(enable = "avx512f" )] |
11081 | unsafe fn test_mm512_reduce_max_pd() { |
11082 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
11083 | let e: f64 = _mm512_reduce_max_pd(a); |
11084 | assert_eq!(7., e); |
11085 | } |
11086 | |
11087 | #[simd_test(enable = "avx512f" )] |
11088 | unsafe fn test_mm512_mask_reduce_max_pd() { |
11089 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
11090 | let e: f64 = _mm512_mask_reduce_max_pd(0b11110000, a); |
11091 | assert_eq!(3., e); |
11092 | } |
11093 | |
11094 | #[simd_test(enable = "avx512f" )] |
11095 | unsafe fn test_mm512_reduce_min_epi64() { |
11096 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
11097 | let e: i64 = _mm512_reduce_min_epi64(a); |
11098 | assert_eq!(0, e); |
11099 | } |
11100 | |
11101 | #[simd_test(enable = "avx512f" )] |
11102 | unsafe fn test_mm512_mask_reduce_min_epi64() { |
11103 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
11104 | let e: i64 = _mm512_mask_reduce_min_epi64(0b11110000, a); |
11105 | assert_eq!(0, e); |
11106 | } |
11107 | |
11108 | #[simd_test(enable = "avx512f" )] |
11109 | unsafe fn test_mm512_reduce_min_epu64() { |
11110 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
11111 | let e: u64 = _mm512_reduce_min_epu64(a); |
11112 | assert_eq!(0, e); |
11113 | } |
11114 | |
11115 | #[simd_test(enable = "avx512f" )] |
11116 | unsafe fn test_mm512_mask_reduce_min_epu64() { |
11117 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
11118 | let e: u64 = _mm512_mask_reduce_min_epu64(0b11110000, a); |
11119 | assert_eq!(0, e); |
11120 | } |
11121 | |
11122 | #[simd_test(enable = "avx512f" )] |
11123 | unsafe fn test_mm512_reduce_min_pd() { |
11124 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
11125 | let e: f64 = _mm512_reduce_min_pd(a); |
11126 | assert_eq!(0., e); |
11127 | } |
11128 | |
11129 | #[simd_test(enable = "avx512f" )] |
11130 | unsafe fn test_mm512_mask_reduce_min_pd() { |
11131 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
11132 | let e: f64 = _mm512_mask_reduce_min_pd(0b11110000, a); |
11133 | assert_eq!(0., e); |
11134 | } |
11135 | |
11136 | #[simd_test(enable = "avx512f" )] |
11137 | unsafe fn test_mm512_reduce_and_epi64() { |
11138 | let a = _mm512_set_epi64(1, 1, 1, 1, 2, 2, 2, 2); |
11139 | let e: i64 = _mm512_reduce_and_epi64(a); |
11140 | assert_eq!(0, e); |
11141 | } |
11142 | |
11143 | #[simd_test(enable = "avx512f" )] |
11144 | unsafe fn test_mm512_mask_reduce_and_epi64() { |
11145 | let a = _mm512_set_epi64(1, 1, 1, 1, 2, 2, 2, 2); |
11146 | let e: i64 = _mm512_mask_reduce_and_epi64(0b11110000, a); |
11147 | assert_eq!(1, e); |
11148 | } |
11149 | |
11150 | #[simd_test(enable = "avx512f" )] |
11151 | unsafe fn test_mm512_reduce_or_epi64() { |
11152 | let a = _mm512_set_epi64(1, 1, 1, 1, 2, 2, 2, 2); |
11153 | let e: i64 = _mm512_reduce_or_epi64(a); |
11154 | assert_eq!(3, e); |
11155 | } |
11156 | |
11157 | #[simd_test(enable = "avx512f" )] |
11158 | unsafe fn test_mm512_mask_reduce_or_epi64() { |
11159 | let a = _mm512_set_epi64(1, 1, 1, 1, 2, 2, 2, 2); |
11160 | let e: i64 = _mm512_mask_reduce_or_epi64(0b11110000, a); |
11161 | assert_eq!(1, e); |
11162 | } |
11163 | |
11164 | #[simd_test(enable = "avx512f" )] |
11165 | unsafe fn test_mm512_extractf64x4_pd() { |
11166 | let a = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
11167 | let r = _mm512_extractf64x4_pd::<1>(a); |
11168 | let e = _mm256_setr_pd(5., 6., 7., 8.); |
11169 | assert_eq_m256d(r, e); |
11170 | } |
11171 | |
11172 | #[simd_test(enable = "avx512f" )] |
11173 | unsafe fn test_mm512_mask_extractf64x4_pd() { |
11174 | let a = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
11175 | let src = _mm256_set1_pd(100.); |
11176 | let r = _mm512_mask_extractf64x4_pd::<1>(src, 0, a); |
11177 | assert_eq_m256d(r, src); |
11178 | let r = _mm512_mask_extractf64x4_pd::<1>(src, 0b11111111, a); |
11179 | let e = _mm256_setr_pd(5., 6., 7., 8.); |
11180 | assert_eq_m256d(r, e); |
11181 | } |
11182 | |
11183 | #[simd_test(enable = "avx512f" )] |
11184 | unsafe fn test_mm512_maskz_extractf64x4_pd() { |
11185 | let a = _mm512_setr_pd(1., 2., 3., 4., 5., 6., 7., 8.); |
11186 | let r = _mm512_maskz_extractf64x4_pd::<1>(0, a); |
11187 | assert_eq_m256d(r, _mm256_setzero_pd()); |
11188 | let r = _mm512_maskz_extractf64x4_pd::<1>(0b00000001, a); |
11189 | let e = _mm256_setr_pd(5., 0., 0., 0.); |
11190 | assert_eq_m256d(r, e); |
11191 | } |
11192 | |
11193 | #[simd_test(enable = "avx512f" )] |
11194 | unsafe fn test_mm512_extracti64x4_epi64() { |
11195 | let a = _mm512_setr_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
11196 | let r = _mm512_extracti64x4_epi64::<0x1>(a); |
11197 | let e = _mm256_setr_epi64x(5, 6, 7, 8); |
11198 | assert_eq_m256i(r, e); |
11199 | } |
11200 | |
11201 | #[simd_test(enable = "avx512f" )] |
11202 | unsafe fn test_mm512_mask_extracti64x4_epi64() { |
11203 | let a = _mm512_setr_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
11204 | let src = _mm256_set1_epi64x(100); |
11205 | let r = _mm512_mask_extracti64x4_epi64::<0x1>(src, 0, a); |
11206 | assert_eq_m256i(r, src); |
11207 | let r = _mm512_mask_extracti64x4_epi64::<0x1>(src, 0b11111111, a); |
11208 | let e = _mm256_setr_epi64x(5, 6, 7, 8); |
11209 | assert_eq_m256i(r, e); |
11210 | } |
11211 | |
11212 | #[simd_test(enable = "avx512f" )] |
11213 | unsafe fn test_mm512_maskz_extracti64x4_epi64() { |
11214 | let a = _mm512_setr_epi64(1, 2, 3, 4, 5, 6, 7, 8); |
11215 | let r = _mm512_maskz_extracti64x4_epi64::<0x1>(0, a); |
11216 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11217 | let r = _mm512_maskz_extracti64x4_epi64::<0x1>(0b00000001, a); |
11218 | let e = _mm256_setr_epi64x(5, 0, 0, 0); |
11219 | assert_eq_m256i(r, e); |
11220 | } |
11221 | |
11222 | #[simd_test(enable = "avx512f" )] |
11223 | unsafe fn test_mm512_mask_compress_epi64() { |
11224 | let src = _mm512_set1_epi64(200); |
11225 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
11226 | let r = _mm512_mask_compress_epi64(src, 0, a); |
11227 | assert_eq_m512i(r, src); |
11228 | let r = _mm512_mask_compress_epi64(src, 0b01010101, a); |
11229 | let e = _mm512_set_epi64(200, 200, 200, 200, 1, 3, 5, 7); |
11230 | assert_eq_m512i(r, e); |
11231 | } |
11232 | |
11233 | #[simd_test(enable = "avx512f" )] |
11234 | unsafe fn test_mm512_maskz_compress_epi64() { |
11235 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
11236 | let r = _mm512_maskz_compress_epi64(0, a); |
11237 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11238 | let r = _mm512_maskz_compress_epi64(0b01010101, a); |
11239 | let e = _mm512_set_epi64(0, 0, 0, 0, 1, 3, 5, 7); |
11240 | assert_eq_m512i(r, e); |
11241 | } |
11242 | |
11243 | #[simd_test(enable = "avx512f,avx512vl" )] |
11244 | unsafe fn test_mm256_mask_compress_epi64() { |
11245 | let src = _mm256_set1_epi64x(200); |
11246 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
11247 | let r = _mm256_mask_compress_epi64(src, 0, a); |
11248 | assert_eq_m256i(r, src); |
11249 | let r = _mm256_mask_compress_epi64(src, 0b00000101, a); |
11250 | let e = _mm256_set_epi64x(200, 200, 1, 3); |
11251 | assert_eq_m256i(r, e); |
11252 | } |
11253 | |
11254 | #[simd_test(enable = "avx512f,avx512vl" )] |
11255 | unsafe fn test_mm256_maskz_compress_epi64() { |
11256 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
11257 | let r = _mm256_maskz_compress_epi64(0, a); |
11258 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11259 | let r = _mm256_maskz_compress_epi64(0b00000101, a); |
11260 | let e = _mm256_set_epi64x(0, 0, 1, 3); |
11261 | assert_eq_m256i(r, e); |
11262 | } |
11263 | |
11264 | #[simd_test(enable = "avx512f,avx512vl" )] |
11265 | unsafe fn test_mm_mask_compress_epi64() { |
11266 | let src = _mm_set1_epi64x(200); |
11267 | let a = _mm_set_epi64x(0, 1); |
11268 | let r = _mm_mask_compress_epi64(src, 0, a); |
11269 | assert_eq_m128i(r, src); |
11270 | let r = _mm_mask_compress_epi64(src, 0b00000001, a); |
11271 | let e = _mm_set_epi64x(200, 1); |
11272 | assert_eq_m128i(r, e); |
11273 | } |
11274 | |
11275 | #[simd_test(enable = "avx512f,avx512vl" )] |
11276 | unsafe fn test_mm_maskz_compress_epi64() { |
11277 | let a = _mm_set_epi64x(0, 1); |
11278 | let r = _mm_maskz_compress_epi64(0, a); |
11279 | assert_eq_m128i(r, _mm_setzero_si128()); |
11280 | let r = _mm_maskz_compress_epi64(0b00000001, a); |
11281 | let e = _mm_set_epi64x(0, 1); |
11282 | assert_eq_m128i(r, e); |
11283 | } |
11284 | |
11285 | #[simd_test(enable = "avx512f" )] |
11286 | unsafe fn test_mm512_mask_compress_pd() { |
11287 | let src = _mm512_set1_pd(200.); |
11288 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
11289 | let r = _mm512_mask_compress_pd(src, 0, a); |
11290 | assert_eq_m512d(r, src); |
11291 | let r = _mm512_mask_compress_pd(src, 0b01010101, a); |
11292 | let e = _mm512_set_pd(200., 200., 200., 200., 1., 3., 5., 7.); |
11293 | assert_eq_m512d(r, e); |
11294 | } |
11295 | |
11296 | #[simd_test(enable = "avx512f" )] |
11297 | unsafe fn test_mm512_maskz_compress_pd() { |
11298 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
11299 | let r = _mm512_maskz_compress_pd(0, a); |
11300 | assert_eq_m512d(r, _mm512_setzero_pd()); |
11301 | let r = _mm512_maskz_compress_pd(0b01010101, a); |
11302 | let e = _mm512_set_pd(0., 0., 0., 0., 1., 3., 5., 7.); |
11303 | assert_eq_m512d(r, e); |
11304 | } |
11305 | |
11306 | #[simd_test(enable = "avx512f,avx512vl" )] |
11307 | unsafe fn test_mm256_mask_compress_pd() { |
11308 | let src = _mm256_set1_pd(200.); |
11309 | let a = _mm256_set_pd(0., 1., 2., 3.); |
11310 | let r = _mm256_mask_compress_pd(src, 0, a); |
11311 | assert_eq_m256d(r, src); |
11312 | let r = _mm256_mask_compress_pd(src, 0b00000101, a); |
11313 | let e = _mm256_set_pd(200., 200., 1., 3.); |
11314 | assert_eq_m256d(r, e); |
11315 | } |
11316 | |
11317 | #[simd_test(enable = "avx512f,avx512vl" )] |
11318 | unsafe fn test_mm256_maskz_compress_pd() { |
11319 | let a = _mm256_set_pd(0., 1., 2., 3.); |
11320 | let r = _mm256_maskz_compress_pd(0, a); |
11321 | assert_eq_m256d(r, _mm256_setzero_pd()); |
11322 | let r = _mm256_maskz_compress_pd(0b00000101, a); |
11323 | let e = _mm256_set_pd(0., 0., 1., 3.); |
11324 | assert_eq_m256d(r, e); |
11325 | } |
11326 | |
11327 | #[simd_test(enable = "avx512f,avx512vl" )] |
11328 | unsafe fn test_mm_mask_compress_pd() { |
11329 | let src = _mm_set1_pd(200.); |
11330 | let a = _mm_set_pd(0., 1.); |
11331 | let r = _mm_mask_compress_pd(src, 0, a); |
11332 | assert_eq_m128d(r, src); |
11333 | let r = _mm_mask_compress_pd(src, 0b00000001, a); |
11334 | let e = _mm_set_pd(200., 1.); |
11335 | assert_eq_m128d(r, e); |
11336 | } |
11337 | |
11338 | #[simd_test(enable = "avx512f,avx512vl" )] |
11339 | unsafe fn test_mm_maskz_compress_pd() { |
11340 | let a = _mm_set_pd(0., 1.); |
11341 | let r = _mm_maskz_compress_pd(0, a); |
11342 | assert_eq_m128d(r, _mm_setzero_pd()); |
11343 | let r = _mm_maskz_compress_pd(0b00000001, a); |
11344 | let e = _mm_set_pd(0., 1.); |
11345 | assert_eq_m128d(r, e); |
11346 | } |
11347 | |
11348 | #[simd_test(enable = "avx512f" )] |
11349 | unsafe fn test_mm512_mask_expand_epi64() { |
11350 | let src = _mm512_set1_epi64(200); |
11351 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
11352 | let r = _mm512_mask_expand_epi64(src, 0, a); |
11353 | assert_eq_m512i(r, src); |
11354 | let r = _mm512_mask_expand_epi64(src, 0b01010101, a); |
11355 | let e = _mm512_set_epi64(200, 4, 200, 5, 200, 6, 200, 7); |
11356 | assert_eq_m512i(r, e); |
11357 | } |
11358 | |
11359 | #[simd_test(enable = "avx512f" )] |
11360 | unsafe fn test_mm512_maskz_expand_epi64() { |
11361 | let a = _mm512_set_epi64(0, 1, 2, 3, 4, 5, 6, 7); |
11362 | let r = _mm512_maskz_expand_epi64(0, a); |
11363 | assert_eq_m512i(r, _mm512_setzero_si512()); |
11364 | let r = _mm512_maskz_expand_epi64(0b01010101, a); |
11365 | let e = _mm512_set_epi64(0, 4, 0, 5, 0, 6, 0, 7); |
11366 | assert_eq_m512i(r, e); |
11367 | } |
11368 | |
11369 | #[simd_test(enable = "avx512f,avx512vl" )] |
11370 | unsafe fn test_mm256_mask_expand_epi64() { |
11371 | let src = _mm256_set1_epi64x(200); |
11372 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
11373 | let r = _mm256_mask_expand_epi64(src, 0, a); |
11374 | assert_eq_m256i(r, src); |
11375 | let r = _mm256_mask_expand_epi64(src, 0b00000101, a); |
11376 | let e = _mm256_set_epi64x(200, 2, 200, 3); |
11377 | assert_eq_m256i(r, e); |
11378 | } |
11379 | |
11380 | #[simd_test(enable = "avx512f,avx512vl" )] |
11381 | unsafe fn test_mm256_maskz_expand_epi64() { |
11382 | let a = _mm256_set_epi64x(0, 1, 2, 3); |
11383 | let r = _mm256_maskz_expand_epi64(0, a); |
11384 | assert_eq_m256i(r, _mm256_setzero_si256()); |
11385 | let r = _mm256_maskz_expand_epi64(0b00000101, a); |
11386 | let e = _mm256_set_epi64x(0, 2, 0, 3); |
11387 | assert_eq_m256i(r, e); |
11388 | } |
11389 | |
11390 | #[simd_test(enable = "avx512f,avx512vl" )] |
11391 | unsafe fn test_mm_mask_expand_epi64() { |
11392 | let src = _mm_set1_epi64x(200); |
11393 | let a = _mm_set_epi64x(0, 1); |
11394 | let r = _mm_mask_expand_epi64(src, 0, a); |
11395 | assert_eq_m128i(r, src); |
11396 | let r = _mm_mask_expand_epi64(src, 0b00000001, a); |
11397 | let e = _mm_set_epi64x(200, 1); |
11398 | assert_eq_m128i(r, e); |
11399 | } |
11400 | |
11401 | #[simd_test(enable = "avx512f,avx512vl" )] |
11402 | unsafe fn test_mm_maskz_expand_epi64() { |
11403 | let a = _mm_set_epi64x(0, 1); |
11404 | let r = _mm_maskz_expand_epi64(0, a); |
11405 | assert_eq_m128i(r, _mm_setzero_si128()); |
11406 | let r = _mm_maskz_expand_epi64(0b00000001, a); |
11407 | let e = _mm_set_epi64x(0, 1); |
11408 | assert_eq_m128i(r, e); |
11409 | } |
11410 | |
11411 | #[simd_test(enable = "avx512f" )] |
11412 | unsafe fn test_mm512_mask_expand_pd() { |
11413 | let src = _mm512_set1_pd(200.); |
11414 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
11415 | let r = _mm512_mask_expand_pd(src, 0, a); |
11416 | assert_eq_m512d(r, src); |
11417 | let r = _mm512_mask_expand_pd(src, 0b01010101, a); |
11418 | let e = _mm512_set_pd(200., 4., 200., 5., 200., 6., 200., 7.); |
11419 | assert_eq_m512d(r, e); |
11420 | } |
11421 | |
11422 | #[simd_test(enable = "avx512f" )] |
11423 | unsafe fn test_mm512_maskz_expand_pd() { |
11424 | let a = _mm512_set_pd(0., 1., 2., 3., 4., 5., 6., 7.); |
11425 | let r = _mm512_maskz_expand_pd(0, a); |
11426 | assert_eq_m512d(r, _mm512_setzero_pd()); |
11427 | let r = _mm512_maskz_expand_pd(0b01010101, a); |
11428 | let e = _mm512_set_pd(0., 4., 0., 5., 0., 6., 0., 7.); |
11429 | assert_eq_m512d(r, e); |
11430 | } |
11431 | |
11432 | #[simd_test(enable = "avx512f,avx512vl" )] |
11433 | unsafe fn test_mm256_mask_expand_pd() { |
11434 | let src = _mm256_set1_pd(200.); |
11435 | let a = _mm256_set_pd(0., 1., 2., 3.); |
11436 | let r = _mm256_mask_expand_pd(src, 0, a); |
11437 | assert_eq_m256d(r, src); |
11438 | let r = _mm256_mask_expand_pd(src, 0b00000101, a); |
11439 | let e = _mm256_set_pd(200., 2., 200., 3.); |
11440 | assert_eq_m256d(r, e); |
11441 | } |
11442 | |
11443 | #[simd_test(enable = "avx512f,avx512vl" )] |
11444 | unsafe fn test_mm256_maskz_expand_pd() { |
11445 | let a = _mm256_set_pd(0., 1., 2., 3.); |
11446 | let r = _mm256_maskz_expand_pd(0, a); |
11447 | assert_eq_m256d(r, _mm256_setzero_pd()); |
11448 | let r = _mm256_maskz_expand_pd(0b00000101, a); |
11449 | let e = _mm256_set_pd(0., 2., 0., 3.); |
11450 | assert_eq_m256d(r, e); |
11451 | } |
11452 | |
11453 | #[simd_test(enable = "avx512f,avx512vl" )] |
11454 | unsafe fn test_mm_mask_expand_pd() { |
11455 | let src = _mm_set1_pd(200.); |
11456 | let a = _mm_set_pd(0., 1.); |
11457 | let r = _mm_mask_expand_pd(src, 0, a); |
11458 | assert_eq_m128d(r, src); |
11459 | let r = _mm_mask_expand_pd(src, 0b00000001, a); |
11460 | let e = _mm_set_pd(200., 1.); |
11461 | assert_eq_m128d(r, e); |
11462 | } |
11463 | |
11464 | #[simd_test(enable = "avx512f,avx512vl" )] |
11465 | unsafe fn test_mm_maskz_expand_pd() { |
11466 | let a = _mm_set_pd(0., 1.); |
11467 | let r = _mm_maskz_expand_pd(0, a); |
11468 | assert_eq_m128d(r, _mm_setzero_pd()); |
11469 | let r = _mm_maskz_expand_pd(0b00000001, a); |
11470 | let e = _mm_set_pd(0., 1.); |
11471 | assert_eq_m128d(r, e); |
11472 | } |
11473 | |
11474 | #[simd_test(enable = "avx512f" )] |
11475 | unsafe fn test_mm512_loadu_epi64() { |
11476 | let a = &[4, 3, 2, 5, -8, -9, -64, -50]; |
11477 | let p = a.as_ptr(); |
11478 | let r = _mm512_loadu_epi64(black_box(p)); |
11479 | let e = _mm512_setr_epi64(4, 3, 2, 5, -8, -9, -64, -50); |
11480 | assert_eq_m512i(r, e); |
11481 | } |
11482 | |
11483 | #[simd_test(enable = "avx512f,avx512vl" )] |
11484 | unsafe fn test_mm256_loadu_epi64() { |
11485 | let a = &[4, 3, 2, 5]; |
11486 | let p = a.as_ptr(); |
11487 | let r = _mm256_loadu_epi64(black_box(p)); |
11488 | let e = _mm256_setr_epi64x(4, 3, 2, 5); |
11489 | assert_eq_m256i(r, e); |
11490 | } |
11491 | |
11492 | #[simd_test(enable = "avx512f,avx512vl" )] |
11493 | unsafe fn test_mm_loadu_epi64() { |
11494 | let a = &[4, 3]; |
11495 | let p = a.as_ptr(); |
11496 | let r = _mm_loadu_epi64(black_box(p)); |
11497 | let e = _mm_setr_epi64x(4, 3); |
11498 | assert_eq_m128i(r, e); |
11499 | } |
11500 | |
11501 | #[simd_test(enable = "avx512f" )] |
11502 | unsafe fn test_mm512_mask_cvtepi64_storeu_epi16() { |
11503 | let a = _mm512_set1_epi64(9); |
11504 | let mut r = _mm_undefined_si128(); |
11505 | _mm512_mask_cvtepi64_storeu_epi16(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11506 | let e = _mm_set1_epi16(9); |
11507 | assert_eq_m128i(r, e); |
11508 | } |
11509 | |
11510 | #[simd_test(enable = "avx512f,avx512vl" )] |
11511 | unsafe fn test_mm256_mask_cvtepi64_storeu_epi16() { |
11512 | let a = _mm256_set1_epi64x(9); |
11513 | let mut r = _mm_set1_epi16(0); |
11514 | _mm256_mask_cvtepi64_storeu_epi16(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11515 | let e = _mm_set_epi16(0, 0, 0, 0, 9, 9, 9, 9); |
11516 | assert_eq_m128i(r, e); |
11517 | } |
11518 | |
11519 | #[simd_test(enable = "avx512f,avx512vl" )] |
11520 | unsafe fn test_mm_mask_cvtepi64_storeu_epi16() { |
11521 | let a = _mm_set1_epi64x(9); |
11522 | let mut r = _mm_set1_epi16(0); |
11523 | _mm_mask_cvtepi64_storeu_epi16(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11524 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, 9, 9); |
11525 | assert_eq_m128i(r, e); |
11526 | } |
11527 | |
11528 | #[simd_test(enable = "avx512f" )] |
11529 | unsafe fn test_mm512_mask_cvtsepi64_storeu_epi16() { |
11530 | let a = _mm512_set1_epi64(i64::MAX); |
11531 | let mut r = _mm_undefined_si128(); |
11532 | _mm512_mask_cvtsepi64_storeu_epi16(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11533 | let e = _mm_set1_epi16(i16::MAX); |
11534 | assert_eq_m128i(r, e); |
11535 | } |
11536 | |
11537 | #[simd_test(enable = "avx512f,avx512vl" )] |
11538 | unsafe fn test_mm256_mask_cvtsepi64_storeu_epi16() { |
11539 | let a = _mm256_set1_epi64x(i64::MAX); |
11540 | let mut r = _mm_set1_epi16(0); |
11541 | _mm256_mask_cvtsepi64_storeu_epi16(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11542 | let e = _mm_set_epi16(0, 0, 0, 0, i16::MAX, i16::MAX, i16::MAX, i16::MAX); |
11543 | assert_eq_m128i(r, e); |
11544 | } |
11545 | |
11546 | #[simd_test(enable = "avx512f,avx512vl" )] |
11547 | unsafe fn test_mm_mask_cvtsepi64_storeu_epi16() { |
11548 | let a = _mm_set1_epi64x(i64::MAX); |
11549 | let mut r = _mm_set1_epi16(0); |
11550 | _mm_mask_cvtsepi64_storeu_epi16(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11551 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, i16::MAX, i16::MAX); |
11552 | assert_eq_m128i(r, e); |
11553 | } |
11554 | |
11555 | #[simd_test(enable = "avx512f" )] |
11556 | unsafe fn test_mm512_mask_cvtusepi64_storeu_epi16() { |
11557 | let a = _mm512_set1_epi64(i64::MAX); |
11558 | let mut r = _mm_undefined_si128(); |
11559 | _mm512_mask_cvtusepi64_storeu_epi16(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11560 | let e = _mm_set1_epi16(u16::MAX as i16); |
11561 | assert_eq_m128i(r, e); |
11562 | } |
11563 | |
11564 | #[simd_test(enable = "avx512f,avx512vl" )] |
11565 | unsafe fn test_mm256_mask_cvtusepi64_storeu_epi16() { |
11566 | let a = _mm256_set1_epi64x(i64::MAX); |
11567 | let mut r = _mm_set1_epi16(0); |
11568 | _mm256_mask_cvtusepi64_storeu_epi16(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11569 | let e = _mm_set_epi16( |
11570 | 0, |
11571 | 0, |
11572 | 0, |
11573 | 0, |
11574 | u16::MAX as i16, |
11575 | u16::MAX as i16, |
11576 | u16::MAX as i16, |
11577 | u16::MAX as i16, |
11578 | ); |
11579 | assert_eq_m128i(r, e); |
11580 | } |
11581 | |
11582 | #[simd_test(enable = "avx512f,avx512vl" )] |
11583 | unsafe fn test_mm_mask_cvtusepi64_storeu_epi16() { |
11584 | let a = _mm_set1_epi64x(i64::MAX); |
11585 | let mut r = _mm_set1_epi16(0); |
11586 | _mm_mask_cvtusepi64_storeu_epi16(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11587 | let e = _mm_set_epi16(0, 0, 0, 0, 0, 0, u16::MAX as i16, u16::MAX as i16); |
11588 | assert_eq_m128i(r, e); |
11589 | } |
11590 | |
11591 | #[simd_test(enable = "avx512f" )] |
11592 | unsafe fn test_mm512_mask_cvtepi64_storeu_epi8() { |
11593 | let a = _mm512_set1_epi64(9); |
11594 | let mut r = _mm_set1_epi8(0); |
11595 | _mm512_mask_cvtepi64_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11596 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 9, 9, 9, 9, 9, 9, 9, 9); |
11597 | assert_eq_m128i(r, e); |
11598 | } |
11599 | |
11600 | #[simd_test(enable = "avx512f,avx512vl" )] |
11601 | unsafe fn test_mm256_mask_cvtepi64_storeu_epi8() { |
11602 | let a = _mm256_set1_epi64x(9); |
11603 | let mut r = _mm_set1_epi8(0); |
11604 | _mm256_mask_cvtepi64_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11605 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 9, 9, 9, 9); |
11606 | assert_eq_m128i(r, e); |
11607 | } |
11608 | |
11609 | #[simd_test(enable = "avx512f,avx512vl" )] |
11610 | unsafe fn test_mm_mask_cvtepi64_storeu_epi8() { |
11611 | let a = _mm_set1_epi64x(9); |
11612 | let mut r = _mm_set1_epi8(0); |
11613 | _mm_mask_cvtepi64_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11614 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 9, 9); |
11615 | assert_eq_m128i(r, e); |
11616 | } |
11617 | |
11618 | #[simd_test(enable = "avx512f" )] |
11619 | unsafe fn test_mm512_mask_cvtsepi64_storeu_epi8() { |
11620 | let a = _mm512_set1_epi64(i64::MAX); |
11621 | let mut r = _mm_set1_epi8(0); |
11622 | _mm512_mask_cvtsepi64_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11623 | #[rustfmt::skip] |
11624 | let e = _mm_set_epi8( |
11625 | 0, 0, 0, 0, |
11626 | 0, 0, 0, 0, |
11627 | i8::MAX, i8::MAX, i8::MAX, i8::MAX, |
11628 | i8::MAX, i8::MAX, i8::MAX, i8::MAX, |
11629 | ); |
11630 | assert_eq_m128i(r, e); |
11631 | } |
11632 | |
11633 | #[simd_test(enable = "avx512f,avx512vl" )] |
11634 | unsafe fn test_mm256_mask_cvtsepi64_storeu_epi8() { |
11635 | let a = _mm256_set1_epi64x(i64::MAX); |
11636 | let mut r = _mm_set1_epi8(0); |
11637 | _mm256_mask_cvtsepi64_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11638 | #[rustfmt::skip] |
11639 | let e = _mm_set_epi8( |
11640 | 0, 0, 0, 0, |
11641 | 0, 0, 0, 0, |
11642 | 0, 0, 0, 0, |
11643 | i8::MAX, i8::MAX, i8::MAX, i8::MAX, |
11644 | ); |
11645 | assert_eq_m128i(r, e); |
11646 | } |
11647 | |
11648 | #[simd_test(enable = "avx512f,avx512vl" )] |
11649 | unsafe fn test_mm_mask_cvtsepi64_storeu_epi8() { |
11650 | let a = _mm_set1_epi64x(i64::MAX); |
11651 | let mut r = _mm_set1_epi8(0); |
11652 | _mm_mask_cvtsepi64_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11653 | let e = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, i8::MAX, i8::MAX); |
11654 | assert_eq_m128i(r, e); |
11655 | } |
11656 | |
11657 | #[simd_test(enable = "avx512f" )] |
11658 | unsafe fn test_mm512_mask_cvtusepi64_storeu_epi8() { |
11659 | let a = _mm512_set1_epi64(i64::MAX); |
11660 | let mut r = _mm_set1_epi8(0); |
11661 | _mm512_mask_cvtusepi64_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11662 | #[rustfmt::skip] |
11663 | let e = _mm_set_epi8( |
11664 | 0, 0, 0, 0, |
11665 | 0, 0, 0, 0, |
11666 | u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, |
11667 | u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, |
11668 | ); |
11669 | assert_eq_m128i(r, e); |
11670 | } |
11671 | |
11672 | #[simd_test(enable = "avx512f,avx512vl" )] |
11673 | unsafe fn test_mm256_mask_cvtusepi64_storeu_epi8() { |
11674 | let a = _mm256_set1_epi64x(i64::MAX); |
11675 | let mut r = _mm_set1_epi8(0); |
11676 | _mm256_mask_cvtusepi64_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11677 | #[rustfmt::skip] |
11678 | let e = _mm_set_epi8( |
11679 | 0, 0, 0, 0, |
11680 | 0, 0, 0, 0, |
11681 | 0, 0, 0, 0, |
11682 | u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, u8::MAX as i8, |
11683 | ); |
11684 | assert_eq_m128i(r, e); |
11685 | } |
11686 | |
11687 | #[simd_test(enable = "avx512f,avx512vl" )] |
11688 | unsafe fn test_mm_mask_cvtusepi64_storeu_epi8() { |
11689 | let a = _mm_set1_epi64x(i64::MAX); |
11690 | let mut r = _mm_set1_epi8(0); |
11691 | _mm_mask_cvtusepi64_storeu_epi8(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11692 | #[rustfmt::skip] |
11693 | let e = _mm_set_epi8( |
11694 | 0, 0, 0, 0, |
11695 | 0, 0, 0, 0, |
11696 | 0, 0, 0, 0, |
11697 | 0, 0, u8::MAX as i8, u8::MAX as i8, |
11698 | ); |
11699 | assert_eq_m128i(r, e); |
11700 | } |
11701 | |
11702 | #[simd_test(enable = "avx512f" )] |
11703 | unsafe fn test_mm512_mask_cvtepi64_storeu_epi32() { |
11704 | let a = _mm512_set1_epi64(9); |
11705 | let mut r = _mm256_undefined_si256(); |
11706 | _mm512_mask_cvtepi64_storeu_epi32(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11707 | let e = _mm256_set1_epi32(9); |
11708 | assert_eq_m256i(r, e); |
11709 | } |
11710 | |
11711 | #[simd_test(enable = "avx512f,avx512vl" )] |
11712 | unsafe fn test_mm256_mask_cvtepi64_storeu_epi32() { |
11713 | let a = _mm256_set1_epi64x(9); |
11714 | let mut r = _mm_set1_epi32(0); |
11715 | _mm256_mask_cvtepi64_storeu_epi32(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11716 | let e = _mm_set_epi32(9, 9, 9, 9); |
11717 | assert_eq_m128i(r, e); |
11718 | } |
11719 | |
11720 | #[simd_test(enable = "avx512f,avx512vl" )] |
11721 | unsafe fn test_mm_mask_cvtepi64_storeu_epi32() { |
11722 | let a = _mm_set1_epi64x(9); |
11723 | let mut r = _mm_set1_epi16(0); |
11724 | _mm_mask_cvtepi64_storeu_epi32(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11725 | let e = _mm_set_epi32(0, 0, 9, 9); |
11726 | assert_eq_m128i(r, e); |
11727 | } |
11728 | |
11729 | #[simd_test(enable = "avx512f" )] |
11730 | unsafe fn test_mm512_mask_cvtsepi64_storeu_epi32() { |
11731 | let a = _mm512_set1_epi64(i64::MAX); |
11732 | let mut r = _mm256_undefined_si256(); |
11733 | _mm512_mask_cvtsepi64_storeu_epi32(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11734 | let e = _mm256_set1_epi32(i32::MAX); |
11735 | assert_eq_m256i(r, e); |
11736 | } |
11737 | |
11738 | #[simd_test(enable = "avx512f,avx512vl" )] |
11739 | unsafe fn test_mm256_mask_cvtsepi64_storeu_epi32() { |
11740 | let a = _mm256_set1_epi64x(i64::MAX); |
11741 | let mut r = _mm_set1_epi32(0); |
11742 | _mm256_mask_cvtsepi64_storeu_epi32(&mut r as *mut _ as *mut i8, 0b00001111, a); |
11743 | let e = _mm_set1_epi32(i32::MAX); |
11744 | assert_eq_m128i(r, e); |
11745 | } |
11746 | |
11747 | #[simd_test(enable = "avx512f,avx512vl" )] |
11748 | unsafe fn test_mm_mask_cvtsepi64_storeu_epi32() { |
11749 | let a = _mm_set1_epi64x(i64::MAX); |
11750 | let mut r = _mm_set1_epi16(0); |
11751 | _mm_mask_cvtsepi64_storeu_epi32(&mut r as *mut _ as *mut i8, 0b00000011, a); |
11752 | let e = _mm_set_epi32(0, 0, i32::MAX, i32::MAX); |
11753 | assert_eq_m128i(r, e); |
11754 | } |
11755 | |
11756 | #[simd_test(enable = "avx512f" )] |
11757 | unsafe fn test_mm512_mask_cvtusepi64_storeu_epi32() { |
11758 | let a = _mm512_set1_epi64(i64::MAX); |
11759 | let mut r = _mm256_undefined_si256(); |
11760 | _mm512_mask_cvtusepi64_storeu_epi32(&mut r as *mut _ as *mut i8, 0b11111111, a); |
11761 | let e = _mm256_set1_epi32(u32::MAX as i32); |
11762 | assert_eq_m256i(r, e); |
11763 | } |
11764 | |
11765 | #[simd_test(enable = "avx512f,avx512vl" )] |
11766 | unsafe fn test_mm256_mask_cvtusepi64_storeu_epi32() { |
11767 | let a = _mm256_set1_epi64x(i64::MAX); |
11768 | let mut r = _mm_set1_epi32(0); |
11769 | _mm256_mask_cvtusepi64_storeu_epi32(&mut r as *mut _ as *mut i8, 0b00001111, a); |
11770 | let e = _mm_set1_epi32(u32::MAX as i32); |
11771 | assert_eq_m128i(r, e); |
11772 | } |
11773 | |
11774 | #[simd_test(enable = "avx512f,avx512vl" )] |
11775 | unsafe fn test_mm_mask_cvtusepi64_storeu_epi32() { |
11776 | let a = _mm_set1_epi64x(i64::MAX); |
11777 | let mut r = _mm_set1_epi16(0); |
11778 | _mm_mask_cvtusepi64_storeu_epi32(&mut r as *mut _ as *mut i8, 0b00000011, a); |
11779 | let e = _mm_set_epi32(0, 0, u32::MAX as i32, u32::MAX as i32); |
11780 | assert_eq_m128i(r, e); |
11781 | } |
11782 | |
11783 | #[simd_test(enable = "avx512f" )] |
11784 | unsafe fn test_mm512_storeu_epi64() { |
11785 | let a = _mm512_set1_epi64(9); |
11786 | let mut r = _mm512_set1_epi64(0); |
11787 | _mm512_storeu_epi64(&mut r as *mut _ as *mut i64, a); |
11788 | assert_eq_m512i(r, a); |
11789 | } |
11790 | |
11791 | #[simd_test(enable = "avx512f,avx512vl" )] |
11792 | unsafe fn test_mm256_storeu_epi64() { |
11793 | let a = _mm256_set1_epi64x(9); |
11794 | let mut r = _mm256_set1_epi64x(0); |
11795 | _mm256_storeu_epi64(&mut r as *mut _ as *mut i64, a); |
11796 | assert_eq_m256i(r, a); |
11797 | } |
11798 | |
11799 | #[simd_test(enable = "avx512f,avx512vl" )] |
11800 | unsafe fn test_mm_storeu_epi64() { |
11801 | let a = _mm_set1_epi64x(9); |
11802 | let mut r = _mm_set1_epi64x(0); |
11803 | _mm_storeu_epi64(&mut r as *mut _ as *mut i64, a); |
11804 | assert_eq_m128i(r, a); |
11805 | } |
11806 | |
11807 | #[simd_test(enable = "avx512f" )] |
11808 | unsafe fn test_mm512_load_epi64() { |
11809 | #[repr (align(64))] |
11810 | struct Align { |
11811 | data: [i64; 8], // 64 bytes |
11812 | } |
11813 | let a = Align { |
11814 | data: [4, 3, 2, 5, -8, -9, -64, -50], |
11815 | }; |
11816 | let p = (a.data).as_ptr(); |
11817 | let r = _mm512_load_epi64(black_box(p)); |
11818 | let e = _mm512_setr_epi64(4, 3, 2, 5, -8, -9, -64, -50); |
11819 | assert_eq_m512i(r, e); |
11820 | } |
11821 | |
11822 | #[simd_test(enable = "avx512f,avx512vl" )] |
11823 | unsafe fn test_mm256_load_epi64() { |
11824 | #[repr (align(64))] |
11825 | struct Align { |
11826 | data: [i64; 4], |
11827 | } |
11828 | let a = Align { data: [4, 3, 2, 5] }; |
11829 | let p = (a.data).as_ptr(); |
11830 | let r = _mm256_load_epi64(black_box(p)); |
11831 | let e = _mm256_set_epi64x(5, 2, 3, 4); |
11832 | assert_eq_m256i(r, e); |
11833 | } |
11834 | |
11835 | #[simd_test(enable = "avx512f,avx512vl" )] |
11836 | unsafe fn test_mm_load_epi64() { |
11837 | #[repr (align(64))] |
11838 | struct Align { |
11839 | data: [i64; 2], |
11840 | } |
11841 | let a = Align { data: [4, 3] }; |
11842 | let p = (a.data).as_ptr(); |
11843 | let r = _mm_load_epi64(black_box(p)); |
11844 | let e = _mm_set_epi64x(3, 4); |
11845 | assert_eq_m128i(r, e); |
11846 | } |
11847 | |
11848 | #[simd_test(enable = "avx512f" )] |
11849 | unsafe fn test_mm512_store_epi64() { |
11850 | let a = _mm512_set1_epi64(9); |
11851 | let mut r = _mm512_set1_epi64(0); |
11852 | _mm512_store_epi64(&mut r as *mut _ as *mut i64, a); |
11853 | assert_eq_m512i(r, a); |
11854 | } |
11855 | |
11856 | #[simd_test(enable = "avx512f,avx512vl" )] |
11857 | unsafe fn test_mm256_store_epi64() { |
11858 | let a = _mm256_set1_epi64x(9); |
11859 | let mut r = _mm256_set1_epi64x(0); |
11860 | _mm256_store_epi64(&mut r as *mut _ as *mut i64, a); |
11861 | assert_eq_m256i(r, a); |
11862 | } |
11863 | |
11864 | #[simd_test(enable = "avx512f,avx512vl" )] |
11865 | unsafe fn test_mm_store_epi64() { |
11866 | let a = _mm_set1_epi64x(9); |
11867 | let mut r = _mm_set1_epi64x(0); |
11868 | _mm_store_epi64(&mut r as *mut _ as *mut i64, a); |
11869 | assert_eq_m128i(r, a); |
11870 | } |
11871 | |
11872 | #[simd_test(enable = "avx512f" )] |
11873 | unsafe fn test_mm512_load_pd() { |
11874 | #[repr (align(64))] |
11875 | struct Align { |
11876 | data: [f64; 8], // 64 bytes |
11877 | } |
11878 | let a = Align { |
11879 | data: [4., 3., 2., 5., -8., -9., -64., -50.], |
11880 | }; |
11881 | let p = (a.data).as_ptr(); |
11882 | let r = _mm512_load_pd(black_box(p)); |
11883 | let e = _mm512_setr_pd(4., 3., 2., 5., -8., -9., -64., -50.); |
11884 | assert_eq_m512d(r, e); |
11885 | } |
11886 | |
11887 | #[simd_test(enable = "avx512f" )] |
11888 | unsafe fn test_mm512_store_pd() { |
11889 | let a = _mm512_set1_pd(9.); |
11890 | let mut r = _mm512_undefined_pd(); |
11891 | _mm512_store_pd(&mut r as *mut _ as *mut f64, a); |
11892 | assert_eq_m512d(r, a); |
11893 | } |
11894 | |
11895 | #[simd_test(enable = "avx512f" )] |
11896 | unsafe fn test_mm512_test_epi64_mask() { |
11897 | let a = _mm512_set1_epi64(1 << 0); |
11898 | let b = _mm512_set1_epi64(1 << 0 | 1 << 1); |
11899 | let r = _mm512_test_epi64_mask(a, b); |
11900 | let e: __mmask8 = 0b11111111; |
11901 | assert_eq!(r, e); |
11902 | } |
11903 | |
11904 | #[simd_test(enable = "avx512f" )] |
11905 | unsafe fn test_mm512_mask_test_epi64_mask() { |
11906 | let a = _mm512_set1_epi64(1 << 0); |
11907 | let b = _mm512_set1_epi64(1 << 0 | 1 << 1); |
11908 | let r = _mm512_mask_test_epi64_mask(0, a, b); |
11909 | assert_eq!(r, 0); |
11910 | let r = _mm512_mask_test_epi64_mask(0b11111111, a, b); |
11911 | let e: __mmask8 = 0b11111111; |
11912 | assert_eq!(r, e); |
11913 | } |
11914 | |
11915 | #[simd_test(enable = "avx512f,avx512vl" )] |
11916 | unsafe fn test_mm256_test_epi64_mask() { |
11917 | let a = _mm256_set1_epi64x(1 << 0); |
11918 | let b = _mm256_set1_epi64x(1 << 0 | 1 << 1); |
11919 | let r = _mm256_test_epi64_mask(a, b); |
11920 | let e: __mmask8 = 0b00001111; |
11921 | assert_eq!(r, e); |
11922 | } |
11923 | |
11924 | #[simd_test(enable = "avx512f,avx512vl" )] |
11925 | unsafe fn test_mm256_mask_test_epi64_mask() { |
11926 | let a = _mm256_set1_epi64x(1 << 0); |
11927 | let b = _mm256_set1_epi64x(1 << 0 | 1 << 1); |
11928 | let r = _mm256_mask_test_epi64_mask(0, a, b); |
11929 | assert_eq!(r, 0); |
11930 | let r = _mm256_mask_test_epi64_mask(0b00001111, a, b); |
11931 | let e: __mmask8 = 0b00001111; |
11932 | assert_eq!(r, e); |
11933 | } |
11934 | |
11935 | #[simd_test(enable = "avx512f,avx512vl" )] |
11936 | unsafe fn test_mm_test_epi64_mask() { |
11937 | let a = _mm_set1_epi64x(1 << 0); |
11938 | let b = _mm_set1_epi64x(1 << 0 | 1 << 1); |
11939 | let r = _mm_test_epi64_mask(a, b); |
11940 | let e: __mmask8 = 0b00000011; |
11941 | assert_eq!(r, e); |
11942 | } |
11943 | |
11944 | #[simd_test(enable = "avx512f,avx512vl" )] |
11945 | unsafe fn test_mm_mask_test_epi64_mask() { |
11946 | let a = _mm_set1_epi64x(1 << 0); |
11947 | let b = _mm_set1_epi64x(1 << 0 | 1 << 1); |
11948 | let r = _mm_mask_test_epi64_mask(0, a, b); |
11949 | assert_eq!(r, 0); |
11950 | let r = _mm_mask_test_epi64_mask(0b00000011, a, b); |
11951 | let e: __mmask8 = 0b00000011; |
11952 | assert_eq!(r, e); |
11953 | } |
11954 | |
11955 | #[simd_test(enable = "avx512f" )] |
11956 | unsafe fn test_mm512_testn_epi64_mask() { |
11957 | let a = _mm512_set1_epi64(1 << 0); |
11958 | let b = _mm512_set1_epi64(1 << 0 | 1 << 1); |
11959 | let r = _mm512_testn_epi64_mask(a, b); |
11960 | let e: __mmask8 = 0b00000000; |
11961 | assert_eq!(r, e); |
11962 | } |
11963 | |
11964 | #[simd_test(enable = "avx512f" )] |
11965 | unsafe fn test_mm512_mask_testn_epi64_mask() { |
11966 | let a = _mm512_set1_epi64(1 << 0); |
11967 | let b = _mm512_set1_epi64(1 << 1); |
11968 | let r = _mm512_mask_testn_epi64_mask(0, a, b); |
11969 | assert_eq!(r, 0); |
11970 | let r = _mm512_mask_testn_epi64_mask(0b11111111, a, b); |
11971 | let e: __mmask8 = 0b11111111; |
11972 | assert_eq!(r, e); |
11973 | } |
11974 | |
11975 | #[simd_test(enable = "avx512f,avx512vl" )] |
11976 | unsafe fn test_mm256_testn_epi64_mask() { |
11977 | let a = _mm256_set1_epi64x(1 << 0); |
11978 | let b = _mm256_set1_epi64x(1 << 1); |
11979 | let r = _mm256_testn_epi64_mask(a, b); |
11980 | let e: __mmask8 = 0b00001111; |
11981 | assert_eq!(r, e); |
11982 | } |
11983 | |
11984 | #[simd_test(enable = "avx512f,avx512vl" )] |
11985 | unsafe fn test_mm256_mask_testn_epi64_mask() { |
11986 | let a = _mm256_set1_epi64x(1 << 0); |
11987 | let b = _mm256_set1_epi64x(1 << 1); |
11988 | let r = _mm256_mask_testn_epi64_mask(0, a, b); |
11989 | assert_eq!(r, 0); |
11990 | let r = _mm256_mask_testn_epi64_mask(0b11111111, a, b); |
11991 | let e: __mmask8 = 0b00001111; |
11992 | assert_eq!(r, e); |
11993 | } |
11994 | |
11995 | #[simd_test(enable = "avx512f,avx512vl" )] |
11996 | unsafe fn test_mm_testn_epi64_mask() { |
11997 | let a = _mm_set1_epi64x(1 << 0); |
11998 | let b = _mm_set1_epi64x(1 << 1); |
11999 | let r = _mm_testn_epi64_mask(a, b); |
12000 | let e: __mmask8 = 0b00000011; |
12001 | assert_eq!(r, e); |
12002 | } |
12003 | |
12004 | #[simd_test(enable = "avx512f,avx512vl" )] |
12005 | unsafe fn test_mm_mask_testn_epi64_mask() { |
12006 | let a = _mm_set1_epi64x(1 << 0); |
12007 | let b = _mm_set1_epi64x(1 << 1); |
12008 | let r = _mm_mask_testn_epi64_mask(0, a, b); |
12009 | assert_eq!(r, 0); |
12010 | let r = _mm_mask_testn_epi64_mask(0b11111111, a, b); |
12011 | let e: __mmask8 = 0b00000011; |
12012 | assert_eq!(r, e); |
12013 | } |
12014 | |
12015 | #[simd_test(enable = "avx512f" )] |
12016 | unsafe fn test_mm512_stream_pd() { |
12017 | #[repr (align(64))] |
12018 | struct Memory { |
12019 | pub data: [f64; 8], |
12020 | } |
12021 | let a = _mm512_set1_pd(7.0); |
12022 | let mut mem = Memory { data: [-1.0; 8] }; |
12023 | |
12024 | _mm512_stream_pd(&mut mem.data[0] as *mut f64, a); |
12025 | for i in 0..8 { |
12026 | assert_eq!(mem.data[i], get_m512d(a, i)); |
12027 | } |
12028 | } |
12029 | |
12030 | #[simd_test(enable = "avx512f" )] |
12031 | unsafe fn test_mm512_stream_si512() { |
12032 | #[repr (align(64))] |
12033 | struct Memory { |
12034 | pub data: [i64; 8], |
12035 | } |
12036 | let a = _mm512_set1_epi64(7); |
12037 | let mut mem = Memory { data: [-1; 8] }; |
12038 | |
12039 | _mm512_stream_si512(&mut mem.data[0] as *mut i64, a); |
12040 | for i in 0..8 { |
12041 | assert_eq!(mem.data[i], get_m512i(a, i)); |
12042 | } |
12043 | } |
12044 | |
12045 | #[simd_test(enable = "avx512f" )] |
12046 | unsafe fn test_mm512_mask_set1_epi64() { |
12047 | let src = _mm512_set1_epi64(2); |
12048 | let a: i64 = 11; |
12049 | let r = _mm512_mask_set1_epi64(src, 0, a); |
12050 | assert_eq_m512i(r, src); |
12051 | let r = _mm512_mask_set1_epi64(src, 0b11111111, a); |
12052 | let e = _mm512_set1_epi64(11); |
12053 | assert_eq_m512i(r, e); |
12054 | } |
12055 | |
12056 | #[simd_test(enable = "avx512f" )] |
12057 | unsafe fn test_mm512_maskz_set1_epi64() { |
12058 | let a: i64 = 11; |
12059 | let r = _mm512_maskz_set1_epi64(0, a); |
12060 | assert_eq_m512i(r, _mm512_setzero_si512()); |
12061 | let r = _mm512_maskz_set1_epi64(0b11111111, a); |
12062 | let e = _mm512_set1_epi64(11); |
12063 | assert_eq_m512i(r, e); |
12064 | } |
12065 | |
12066 | #[simd_test(enable = "avx512f,avx512vl" )] |
12067 | unsafe fn test_mm256_mask_set1_epi64() { |
12068 | let src = _mm256_set1_epi64x(2); |
12069 | let a: i64 = 11; |
12070 | let r = _mm256_mask_set1_epi64(src, 0, a); |
12071 | assert_eq_m256i(r, src); |
12072 | let r = _mm256_mask_set1_epi64(src, 0b00001111, a); |
12073 | let e = _mm256_set1_epi64x(11); |
12074 | assert_eq_m256i(r, e); |
12075 | } |
12076 | |
12077 | #[simd_test(enable = "avx512f,avx512vl" )] |
12078 | unsafe fn test_mm256_maskz_set1_epi64() { |
12079 | let a: i64 = 11; |
12080 | let r = _mm256_maskz_set1_epi64(0, a); |
12081 | assert_eq_m256i(r, _mm256_setzero_si256()); |
12082 | let r = _mm256_maskz_set1_epi64(0b00001111, a); |
12083 | let e = _mm256_set1_epi64x(11); |
12084 | assert_eq_m256i(r, e); |
12085 | } |
12086 | |
12087 | #[simd_test(enable = "avx512f,avx512vl" )] |
12088 | unsafe fn test_mm_mask_set1_epi64() { |
12089 | let src = _mm_set1_epi64x(2); |
12090 | let a: i64 = 11; |
12091 | let r = _mm_mask_set1_epi64(src, 0, a); |
12092 | assert_eq_m128i(r, src); |
12093 | let r = _mm_mask_set1_epi64(src, 0b00000011, a); |
12094 | let e = _mm_set1_epi64x(11); |
12095 | assert_eq_m128i(r, e); |
12096 | } |
12097 | |
12098 | #[simd_test(enable = "avx512f,avx512vl" )] |
12099 | unsafe fn test_mm_maskz_set1_epi64() { |
12100 | let a: i64 = 11; |
12101 | let r = _mm_maskz_set1_epi64(0, a); |
12102 | assert_eq_m128i(r, _mm_setzero_si128()); |
12103 | let r = _mm_maskz_set1_epi64(0b00000011, a); |
12104 | let e = _mm_set1_epi64x(11); |
12105 | assert_eq_m128i(r, e); |
12106 | } |
12107 | |
12108 | #[simd_test(enable = "avx512f" )] |
12109 | unsafe fn test_mm_cvtsd_i64() { |
12110 | let a = _mm_set_pd(1., -1.5); |
12111 | let r = _mm_cvtsd_i64(a); |
12112 | let e: i64 = -2; |
12113 | assert_eq!(r, e); |
12114 | } |
12115 | |
12116 | #[simd_test(enable = "avx512f" )] |
12117 | unsafe fn test_mm_cvtss_i64() { |
12118 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12119 | let r = _mm_cvtss_i64(a); |
12120 | let e: i64 = -2; |
12121 | assert_eq!(r, e); |
12122 | } |
12123 | |
12124 | #[simd_test(enable = "avx512f" )] |
12125 | unsafe fn test_mm_cvt_roundi64_ss() { |
12126 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12127 | let b: i64 = 9; |
12128 | let r = _mm_cvt_roundi64_ss::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b); |
12129 | let e = _mm_set_ps(0., -0.5, 1., 9.); |
12130 | assert_eq_m128(r, e); |
12131 | } |
12132 | |
12133 | #[simd_test(enable = "avx512f" )] |
12134 | unsafe fn test_mm_cvt_roundsi64_ss() { |
12135 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12136 | let b: i64 = 9; |
12137 | let r = _mm_cvt_roundsi64_ss::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b); |
12138 | let e = _mm_set_ps(0., -0.5, 1., 9.); |
12139 | assert_eq_m128(r, e); |
12140 | } |
12141 | |
12142 | #[simd_test(enable = "avx512f" )] |
12143 | unsafe fn test_mm_cvti64_ss() { |
12144 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12145 | let b: i64 = 9; |
12146 | let r = _mm_cvti64_ss(a, b); |
12147 | let e = _mm_set_ps(0., -0.5, 1., 9.); |
12148 | assert_eq_m128(r, e); |
12149 | } |
12150 | |
12151 | #[simd_test(enable = "avx512f" )] |
12152 | unsafe fn test_mm_cvti64_sd() { |
12153 | let a = _mm_set_pd(1., -1.5); |
12154 | let b: i64 = 9; |
12155 | let r = _mm_cvti64_sd(a, b); |
12156 | let e = _mm_set_pd(1., 9.); |
12157 | assert_eq_m128d(r, e); |
12158 | } |
12159 | |
12160 | #[simd_test(enable = "avx512f" )] |
12161 | unsafe fn test_mm_cvt_roundsd_si64() { |
12162 | let a = _mm_set_pd(1., -1.5); |
12163 | let r = _mm_cvt_roundsd_si64::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a); |
12164 | let e: i64 = -1; |
12165 | assert_eq!(r, e); |
12166 | } |
12167 | |
12168 | #[simd_test(enable = "avx512f" )] |
12169 | unsafe fn test_mm_cvt_roundsd_i64() { |
12170 | let a = _mm_set_pd(1., -1.5); |
12171 | let r = _mm_cvt_roundsd_i64::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a); |
12172 | let e: i64 = -1; |
12173 | assert_eq!(r, e); |
12174 | } |
12175 | |
12176 | #[simd_test(enable = "avx512f" )] |
12177 | unsafe fn test_mm_cvt_roundsd_u64() { |
12178 | let a = _mm_set_pd(1., f64::MAX); |
12179 | let r = _mm_cvt_roundsd_u64::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a); |
12180 | let e: u64 = u64::MAX; |
12181 | assert_eq!(r, e); |
12182 | } |
12183 | |
12184 | #[simd_test(enable = "avx512f" )] |
12185 | unsafe fn test_mm_cvtsd_u64() { |
12186 | let a = _mm_set_pd(1., -1.5); |
12187 | let r = _mm_cvtsd_u64(a); |
12188 | let e: u64 = u64::MAX; |
12189 | assert_eq!(r, e); |
12190 | } |
12191 | |
12192 | #[simd_test(enable = "avx512f" )] |
12193 | unsafe fn test_mm_cvt_roundss_i64() { |
12194 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12195 | let r = _mm_cvt_roundss_i64::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a); |
12196 | let e: i64 = -1; |
12197 | assert_eq!(r, e); |
12198 | } |
12199 | |
12200 | #[simd_test(enable = "avx512f" )] |
12201 | unsafe fn test_mm_cvt_roundss_si64() { |
12202 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12203 | let r = _mm_cvt_roundss_si64::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a); |
12204 | let e: i64 = -1; |
12205 | assert_eq!(r, e); |
12206 | } |
12207 | |
12208 | #[simd_test(enable = "avx512f" )] |
12209 | unsafe fn test_mm_cvt_roundss_u64() { |
12210 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12211 | let r = _mm_cvt_roundss_u64::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a); |
12212 | let e: u64 = u64::MAX; |
12213 | assert_eq!(r, e); |
12214 | } |
12215 | |
12216 | #[simd_test(enable = "avx512f" )] |
12217 | unsafe fn test_mm_cvtss_u64() { |
12218 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12219 | let r = _mm_cvtss_u64(a); |
12220 | let e: u64 = u64::MAX; |
12221 | assert_eq!(r, e); |
12222 | } |
12223 | |
12224 | #[simd_test(enable = "avx512f" )] |
12225 | unsafe fn test_mm_cvttsd_i64() { |
12226 | let a = _mm_set_pd(1., -1.5); |
12227 | let r = _mm_cvttsd_i64(a); |
12228 | let e: i64 = -2; |
12229 | assert_eq!(r, e); |
12230 | } |
12231 | |
12232 | #[simd_test(enable = "avx512f" )] |
12233 | unsafe fn test_mm_cvtt_roundsd_i64() { |
12234 | let a = _mm_set_pd(1., -1.5); |
12235 | let r = _mm_cvtt_roundsd_i64::<_MM_FROUND_CUR_DIRECTION>(a); |
12236 | let e: i64 = -2; |
12237 | assert_eq!(r, e); |
12238 | } |
12239 | |
12240 | #[simd_test(enable = "avx512f" )] |
12241 | unsafe fn test_mm_cvtt_roundsd_si64() { |
12242 | let a = _mm_set_pd(1., -1.5); |
12243 | let r = _mm_cvtt_roundsd_si64::<_MM_FROUND_CUR_DIRECTION>(a); |
12244 | let e: i64 = -2; |
12245 | assert_eq!(r, e); |
12246 | } |
12247 | |
12248 | #[simd_test(enable = "avx512f" )] |
12249 | unsafe fn test_mm_cvtt_roundsd_u64() { |
12250 | let a = _mm_set_pd(1., -1.5); |
12251 | let r = _mm_cvtt_roundsd_u64::<_MM_FROUND_CUR_DIRECTION>(a); |
12252 | let e: u64 = u64::MAX; |
12253 | assert_eq!(r, e); |
12254 | } |
12255 | |
12256 | #[simd_test(enable = "avx512f" )] |
12257 | unsafe fn test_mm_cvttsd_u64() { |
12258 | let a = _mm_set_pd(1., -1.5); |
12259 | let r = _mm_cvttsd_u64(a); |
12260 | let e: u64 = u64::MAX; |
12261 | assert_eq!(r, e); |
12262 | } |
12263 | |
12264 | #[simd_test(enable = "avx512f" )] |
12265 | unsafe fn test_mm_cvttss_i64() { |
12266 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12267 | let r = _mm_cvttss_i64(a); |
12268 | let e: i64 = -2; |
12269 | assert_eq!(r, e); |
12270 | } |
12271 | |
12272 | #[simd_test(enable = "avx512f" )] |
12273 | unsafe fn test_mm_cvtt_roundss_i64() { |
12274 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12275 | let r = _mm_cvtt_roundss_i64::<_MM_FROUND_CUR_DIRECTION>(a); |
12276 | let e: i64 = -2; |
12277 | assert_eq!(r, e); |
12278 | } |
12279 | |
12280 | #[simd_test(enable = "avx512f" )] |
12281 | unsafe fn test_mm_cvtt_roundss_si64() { |
12282 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12283 | let r = _mm_cvtt_roundss_si64::<_MM_FROUND_CUR_DIRECTION>(a); |
12284 | let e: i64 = -2; |
12285 | assert_eq!(r, e); |
12286 | } |
12287 | |
12288 | #[simd_test(enable = "avx512f" )] |
12289 | unsafe fn test_mm_cvtt_roundss_u64() { |
12290 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12291 | let r = _mm_cvtt_roundss_u64::<_MM_FROUND_CUR_DIRECTION>(a); |
12292 | let e: u64 = u64::MAX; |
12293 | assert_eq!(r, e); |
12294 | } |
12295 | |
12296 | #[simd_test(enable = "avx512f" )] |
12297 | unsafe fn test_mm_cvttss_u64() { |
12298 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12299 | let r = _mm_cvttss_u64(a); |
12300 | let e: u64 = u64::MAX; |
12301 | assert_eq!(r, e); |
12302 | } |
12303 | |
12304 | #[simd_test(enable = "avx512f" )] |
12305 | unsafe fn test_mm_cvtu64_ss() { |
12306 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12307 | let b: u64 = 9; |
12308 | let r = _mm_cvtu64_ss(a, b); |
12309 | let e = _mm_set_ps(0., -0.5, 1., 9.); |
12310 | assert_eq_m128(r, e); |
12311 | } |
12312 | |
12313 | #[simd_test(enable = "avx512f" )] |
12314 | unsafe fn test_mm_cvtu64_sd() { |
12315 | let a = _mm_set_pd(1., -1.5); |
12316 | let b: u64 = 9; |
12317 | let r = _mm_cvtu64_sd(a, b); |
12318 | let e = _mm_set_pd(1., 9.); |
12319 | assert_eq_m128d(r, e); |
12320 | } |
12321 | |
12322 | #[simd_test(enable = "avx512f" )] |
12323 | unsafe fn test_mm_cvt_roundu64_ss() { |
12324 | let a = _mm_set_ps(0., -0.5, 1., -1.5); |
12325 | let b: u64 = 9; |
12326 | let r = _mm_cvt_roundu64_ss::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b); |
12327 | let e = _mm_set_ps(0., -0.5, 1., 9.); |
12328 | assert_eq_m128(r, e); |
12329 | } |
12330 | |
12331 | #[simd_test(enable = "avx512f" )] |
12332 | unsafe fn test_mm_cvt_roundu64_sd() { |
12333 | let a = _mm_set_pd(1., -1.5); |
12334 | let b: u64 = 9; |
12335 | let r = _mm_cvt_roundu64_sd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b); |
12336 | let e = _mm_set_pd(1., 9.); |
12337 | assert_eq_m128d(r, e); |
12338 | } |
12339 | |
12340 | #[simd_test(enable = "avx512f" )] |
12341 | unsafe fn test_mm_cvt_roundi64_sd() { |
12342 | let a = _mm_set_pd(1., -1.5); |
12343 | let b: i64 = 9; |
12344 | let r = _mm_cvt_roundi64_sd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b); |
12345 | let e = _mm_set_pd(1., 9.); |
12346 | assert_eq_m128d(r, e); |
12347 | } |
12348 | |
12349 | #[simd_test(enable = "avx512f" )] |
12350 | unsafe fn test_mm_cvt_roundsi64_sd() { |
12351 | let a = _mm_set_pd(1., -1.5); |
12352 | let b: i64 = 9; |
12353 | let r = _mm_cvt_roundsi64_sd::<{ _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC }>(a, b); |
12354 | let e = _mm_set_pd(1., 9.); |
12355 | assert_eq_m128d(r, e); |
12356 | } |
12357 | } |
12358 | |