1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
|
/* { dg-do run } */
/* { dg-require-effective-target avx } */
/* { dg-options "-O2 -mavx" } */
#include "avx-check.h"
static void
avx_test (void)
{
__m256i src;
#ifdef __x86_64__
char reg_save[16][32];
char d[16][32];
#else
char reg_save[8][32];
char d[8][32];
#endif
int s[8] = {1, 2, 3, 4, 5, 6, 7, 8};
__builtin_memset (d, 0, sizeof d);
__builtin_memset (reg_save, -1, sizeof reg_save);
src = _mm256_loadu_si256 ((__m256i*) s);
_mm256_zeroall ();
__asm__ __volatile__ ("vmovdqu %%ymm0,%0":"=m"(reg_save[0]));
__asm__ __volatile__ ("vmovdqu %%ymm1,%0":"=m"(reg_save[1]));
__asm__ __volatile__ ("vmovdqu %%ymm2,%0":"=m"(reg_save[2]));
__asm__ __volatile__ ("vmovdqu %%ymm3,%0":"=m"(reg_save[3]));
__asm__ __volatile__ ("vmovdqu %%ymm4,%0":"=m"(reg_save[4]));
__asm__ __volatile__ ("vmovdqu %%ymm5,%0":"=m"(reg_save[5]));
__asm__ __volatile__ ("vmovdqu %%ymm6,%0":"=m"(reg_save[6]));
__asm__ __volatile__ ("vmovdqu %%ymm7,%0":"=m"(reg_save[7]));
#ifdef __x86_64__
__asm__ __volatile__ ("vmovdqu %%ymm8,%0":"=m"(reg_save[8]));
__asm__ __volatile__ ("vmovdqu %%ymm9,%0":"=m"(reg_save[9]));
__asm__ __volatile__ ("vmovdqu %%ymm10,%0":"=m"(reg_save[10]));
__asm__ __volatile__ ("vmovdqu %%ymm11,%0":"=m"(reg_save[11]));
__asm__ __volatile__ ("vmovdqu %%ymm12,%0":"=m"(reg_save[12]));
__asm__ __volatile__ ("vmovdqu %%ymm13,%0":"=m"(reg_save[13]));
__asm__ __volatile__ ("vmovdqu %%ymm14,%0":"=m"(reg_save[14]));
__asm__ __volatile__ ("vmovdqu %%ymm15,%0":"=m"(reg_save[15]));
#endif
if (__builtin_memcmp (reg_save, d, sizeof d))
abort ();
_mm256_storeu_si256 ((__m256i*) d, src);
}
|
