| 1 |
/* { dg-do compile } */ |
| 2 |
/* { dg-options "-mavx512f -O2" } */ |
| 3 |
/* { dg-final { scan-assembler-times {(?n)vcvtdq2ps[ \t]+32\(%.*%ymm} 1 } } */ |
| 4 |
/* { dg-final { scan-assembler-times {(?n)vcvtdq2ps[ \t]+16\(%.*%xmm} 1 } } */ |
| 5 |
/* { dg-final { scan-assembler-times {(?n)vmovq[ \t]+16\(%.*%xmm} 1 { target { ! ia32 } } } } */ |
| 6 |
/* { dg-final { scan-assembler-not {(?n)vpermd[ \t]+.*%zmm} } } */ |
| 7 |
|
| 8 |
typedef int v16si __attribute__((vector_size(64))); |
| 9 |
typedef float v8sf __attribute__((vector_size(32))); |
| 10 |
typedef float v4sf __attribute__((vector_size(16))); |
| 11 |
typedef float v2sf __attribute__((vector_size(8))); |
| 12 |
|
| 13 |
v8sf part (v16si *srcp) |
| 14 |
{ |
| 15 |
v16si src = *srcp; |
| 16 |
return (v8sf) { (float)src[8], (float) src[9], (float)src[10], (float)src[11], |
| 17 |
(float)src[12], (float)src[13], (float)src[14], (float)src[15] }; |
| 18 |
} |
| 19 |
|
| 20 |
v4sf part1 (v16si *srcp) |
| 21 |
{ |
| 22 |
v16si src = *srcp; |
| 23 |
return (v4sf) { (float)src[4], (float)src[5], (float)src[6], (float)src[7] }; |
| 24 |
} |
| 25 |
|
| 26 |
v2sf part2 (v16si *srcp) |
| 27 |
{ |
| 28 |
v16si src = *srcp; |
| 29 |
return (v2sf) { (float)src[4], (float)src[5] }; |
| 30 |
} |