| 1 |
/* i?86 does not have V2SF, x32 does though. */ |
| 2 |
/* { dg-do compile { target { ! ia32 } } } */ |
| 3 |
/* { dg-options "-O3 -mavx -mfma" } */ |
| 4 |
|
| 5 |
struct Matrix |
| 6 |
{ |
| 7 |
float m11; |
| 8 |
float m12; |
| 9 |
float m21; |
| 10 |
float m22; |
| 11 |
float dx; |
| 12 |
float dy; |
| 13 |
}; |
| 14 |
|
| 15 |
struct Matrix multiply(const struct Matrix *a, const struct Matrix *b) |
| 16 |
{ |
| 17 |
struct Matrix out; |
| 18 |
out.m11 = a->m11*b->m11 + a->m12*b->m21; |
| 19 |
out.m12 = a->m11*b->m12 + a->m12*b->m22; |
| 20 |
out.m21 = a->m21*b->m11 + a->m22*b->m21; |
| 21 |
out.m22 = a->m21*b->m12 + a->m22*b->m22; |
| 22 |
|
| 23 |
out.dx = a->dx*b->m11 + a->dy*b->m21 + b->dx; |
| 24 |
out.dy = a->dx*b->m12 + a->dy*b->m22 + b->dy; |
| 25 |
return out; |
| 26 |
} |
| 27 |
|
| 28 |
/* The whole kernel should be vectorized with V4SF and V2SF operations. */ |
| 29 |
/* { dg-final { scan-assembler-times "vadd" 1 } } */ |
| 30 |
/* { dg-final { scan-assembler-times "vmul" 2 } } */ |
| 31 |
/* { dg-final { scan-assembler-times "vfma" 2 } } */ |