| 1 |
/* { dg-do compile { target { ! ia32 } } } */ |
| 2 |
/* { dg-options "-march=x86-64-v3 -mavx10.2 -mmovrs -O2" } */ |
| 3 |
/* { dg-final { scan-assembler-times "vmovrsb\[ \\t\]\+\\(%(?:r|e).x\\), %ymm\[0-9\]+" 3 } } */ |
| 4 |
/* { dg-final { scan-assembler-times "vmovrsb\[ \\t\]\+\\(%(?:r|e).x\\), %ymm\[0-9\]+{%k\[1-7\]}" 2 } } */ |
| 5 |
/* { dg-final { scan-assembler-times "vmovrsb\[ \\t\]\+\\(%(?:r|e).x\\), %ymm\[0-9\]+{%k\[1-7\]}{z}" 1 } } */ |
| 6 |
/* { dg-final { scan-assembler-times "vmovrsd\[ \\t\]\+\\(%(?:r|e).x\\), %ymm\[0-9\]+" 3 } } */ |
| 7 |
/* { dg-final { scan-assembler-times "vmovrsd\[ \\t\]\+\\(%(?:r|e).x\\), %ymm\[0-9\]+{%k\[1-7\]}" 2 } } */ |
| 8 |
/* { dg-final { scan-assembler-times "vmovrsd\[ \\t\]\+\\(%(?:r|e).x\\), %ymm\[0-9\]+{%k\[1-7\]}{z}" 1 } } */ |
| 9 |
/* { dg-final { scan-assembler-times "vmovrsq\[ \\t\]\+\\(%(?:r|e).x\\), %ymm\[0-9\]+" 3 } } */ |
| 10 |
/* { dg-final { scan-assembler-times "vmovrsq\[ \\t\]\+\\(%(?:r|e).x\\), %ymm\[0-9\]+{%k\[1-7\]}" 2 } } */ |
| 11 |
/* { dg-final { scan-assembler-times "vmovrsq\[ \\t\]\+\\(%(?:r|e).x\\), %ymm\[0-9\]+{%k\[1-7\]}{z}" 1 } } */ |
| 12 |
/* { dg-final { scan-assembler-times "vmovrsw\[ \\t\]\+\\(%(?:r|e).x\\), %ymm\[0-9\]+" 3 } } */ |
| 13 |
/* { dg-final { scan-assembler-times "vmovrsw\[ \\t\]\+\\(%(?:r|e).x\\), %ymm\[0-9\]+{%k\[1-7\]}" 2 } } */ |
| 14 |
/* { dg-final { scan-assembler-times "vmovrsw\[ \\t\]\+\\(%(?:r|e).x\\), %ymm\[0-9\]+{%k\[1-7\]}{z}" 1 } } */ |
| 15 |
/* { dg-final { scan-assembler-times "vmovrsb\[ \\t\]\+\\(%(?:r|e).x\\), %xmm\[0-9\]+" 3 } } */ |
| 16 |
/* { dg-final { scan-assembler-times "vmovrsb\[ \\t\]\+\\(%(?:r|e).x\\), %xmm\[0-9\]+{%k\[1-7\]}" 2 } } */ |
| 17 |
/* { dg-final { scan-assembler-times "vmovrsb\[ \\t\]\+\\(%(?:r|e).x\\), %xmm\[0-9\]+{%k\[1-7\]}{z}" 1 } } */ |
| 18 |
/* { dg-final { scan-assembler-times "vmovrsd\[ \\t\]\+\\(%(?:r|e).x\\), %xmm\[0-9\]+" 3 } } */ |
| 19 |
/* { dg-final { scan-assembler-times "vmovrsd\[ \\t\]\+\\(%(?:r|e).x\\), %xmm\[0-9\]+{%k\[1-7\]}" 2 } } */ |
| 20 |
/* { dg-final { scan-assembler-times "vmovrsd\[ \\t\]\+\\(%(?:r|e).x\\), %xmm\[0-9\]+{%k\[1-7\]}{z}" 1 } } */ |
| 21 |
/* { dg-final { scan-assembler-times "vmovrsq\[ \\t\]\+\\(%(?:r|e).x\\), %xmm\[0-9\]+" 3 } } */ |
| 22 |
/* { dg-final { scan-assembler-times "vmovrsq\[ \\t\]\+\\(%(?:r|e).x\\), %xmm\[0-9\]+{%k\[1-7\]}" 2 } } */ |
| 23 |
/* { dg-final { scan-assembler-times "vmovrsq\[ \\t\]\+\\(%(?:r|e).x\\), %xmm\[0-9\]+{%k\[1-7\]}{z}" 1 } } */ |
| 24 |
/* { dg-final { scan-assembler-times "vmovrsw\[ \\t\]\+\\(%(?:r|e).x\\), %xmm\[0-9\]+" 3 } } */ |
| 25 |
/* { dg-final { scan-assembler-times "vmovrsw\[ \\t\]\+\\(%(?:r|e).x\\), %xmm\[0-9\]+{%k\[1-7\]}" 2 } } */ |
| 26 |
/* { dg-final { scan-assembler-times "vmovrsw\[ \\t\]\+\\(%(?:r|e).x\\), %xmm\[0-9\]+{%k\[1-7\]}{z}" 1 } } */ |
| 27 |
|
| 28 |
#include <immintrin.h> |
| 29 |
|
| 30 |
__m256i *px1; |
| 31 |
volatile __m256i x1; |
| 32 |
__m128i *px2; |
| 33 |
volatile __m128i x2; |
| 34 |
volatile __mmask32 m1; |
| 35 |
volatile __mmask8 m2; |
| 36 |
volatile __mmask16 m3; |
| 37 |
|
| 38 |
|
| 39 |
void extern |
| 40 |
avx10_movrs_test (void) |
| 41 |
{ |
| 42 |
x1 = _mm256_loadrs_epi8(px1); |
| 43 |
x1 = _mm256_mask_loadrs_epi8(x1, m1, px1); |
| 44 |
x1 = _mm256_maskz_loadrs_epi8(m1, px1); |
| 45 |
x1 = _mm256_loadrs_epi32(px1); |
| 46 |
x1 = _mm256_mask_loadrs_epi32(x1, m2, px1); |
| 47 |
x1 = _mm256_maskz_loadrs_epi32(m2, px1); |
| 48 |
x1 = _mm256_loadrs_epi64(px1); |
| 49 |
x1 = _mm256_mask_loadrs_epi64(x1, m2, px1); |
| 50 |
x1 = _mm256_maskz_loadrs_epi64(m2, px1); |
| 51 |
x1 = _mm256_loadrs_epi16(px1); |
| 52 |
x1 = _mm256_mask_loadrs_epi16(x1, m3, px1); |
| 53 |
x1 = _mm256_maskz_loadrs_epi16(m3, px1); |
| 54 |
|
| 55 |
x2 = _mm_loadrs_epi8(px2); |
| 56 |
x2 = _mm_mask_loadrs_epi8(x2, m3, px2); |
| 57 |
x2 = _mm_maskz_loadrs_epi8(m3, px2); |
| 58 |
x2 = _mm_loadrs_epi32(px2); |
| 59 |
x2 = _mm_mask_loadrs_epi32(x2, m2, px2); |
| 60 |
x2 = _mm_maskz_loadrs_epi32(m2, px2); |
| 61 |
x2 = _mm_loadrs_epi64(px2); |
| 62 |
x2 = _mm_mask_loadrs_epi64(x2, m2, px2); |
| 63 |
x2 = _mm_maskz_loadrs_epi64(m2, px2); |
| 64 |
x2 = _mm_loadrs_epi16(px2); |
| 65 |
x2 = _mm_mask_loadrs_epi16(x2, m2, px2); |
| 66 |
x2 = _mm_maskz_loadrs_epi16(m2, px2); |
| 67 |
} |