| 1 |
/* { dg-do compile } */ |
| 2 |
/* { dg-options "-mdejagnu-cpu=power8 -mvsx -O3" } */ |
| 3 |
/* { dg-require-effective-target powerpc_vsx } */ |
| 4 |
/* { dg-final { scan-assembler "lxvd2x" } } */ |
| 5 |
/* { dg-final { scan-assembler "stxvd2x" } } */ |
| 6 |
/* { dg-final { scan-assembler "xxspltw" } } */ |
| 7 |
|
| 8 |
/* Currently the analyze_swaps phase cannot optimize this loop because |
| 9 |
of the presence of an UNSPEC_VSX_CVDPSPN. At such time as this is |
| 10 |
handled, we need to add a 'scan-assembler-not "xxpermdi"' directive to |
| 11 |
this test. */ |
| 12 |
#include <altivec.h> |
| 13 |
void abort(); |
| 14 |
|
| 15 |
#define N 4096 |
| 16 |
#define M 10000000 |
| 17 |
vector float ca[N][4] = {0}; |
| 18 |
vector float cb[N][4] = {0}; |
| 19 |
vector float cc[N][4] = {0}; |
| 20 |
|
| 21 |
__attribute__((noinline)) void foo () |
| 22 |
{ |
| 23 |
int i; |
| 24 |
for (i = 0; i < N; i++) { |
| 25 |
cc[i][0] = vec_mul(vec_splats(cb[i][0][0]), ca[i][0]); |
| 26 |
cc[i][0] = vec_madd(cc[i][0],vec_splats(cb[i][0][1]), ca[i][1]); |
| 27 |
cc[i][0] = vec_madd(cc[i][0],vec_splats(cb[i][0][2]), ca[i][2]); |
| 28 |
cc[i][0] = vec_madd(cc[i][0],vec_splats(cb[i][0][3]), ca[i][3]); |
| 29 |
|
| 30 |
cc[i][1] = vec_mul(vec_splats(cb[i][1][0]), ca[i][0]); |
| 31 |
cc[i][1] = vec_madd(cc[i][0],vec_splats(cb[i][1][1]), ca[i][1]); |
| 32 |
cc[i][1] = vec_madd(cc[i][0],vec_splats(cb[i][1][2]), ca[i][2]); |
| 33 |
cc[i][1] = vec_madd(cc[i][0],vec_splats(cb[i][1][3]), ca[i][3]); |
| 34 |
|
| 35 |
cc[i][2] = vec_mul(vec_splats(cb[i][2][0]), ca[i][0]); |
| 36 |
cc[i][2] = vec_madd(cc[i][0],vec_splats(cb[i][2][1]), ca[i][1]); |
| 37 |
cc[i][2] = vec_madd(cc[i][0],vec_splats(cb[i][2][2]), ca[i][2]); |
| 38 |
cc[i][2] = vec_madd(cc[i][0],vec_splats(cb[i][2][3]), ca[i][3]); |
| 39 |
|
| 40 |
cc[i][3] = vec_mul(vec_splats(cb[i][3][0]), ca[i][0]); |
| 41 |
cc[i][3] = vec_madd(cc[i][0],vec_splats(cb[i][3][1]), ca[i][1]); |
| 42 |
cc[i][3] = vec_madd(cc[i][0],vec_splats(cb[i][3][2]), ca[i][2]); |
| 43 |
cc[i][3] = vec_madd(cc[i][0],vec_splats(cb[i][3][3]), ca[i][3]); |
| 44 |
} |
| 45 |
} |
| 46 |
|
| 47 |
int main () |
| 48 |
{ |
| 49 |
foo (); |
| 50 |
return 0; |
| 51 |
} |