| 1 |
/* { dg-do compile { target { ! ia32 } } } */ |
| 2 |
/* { dg-options "-march=x86-64-v3 -mavx10.2 -O2" } */ |
| 3 |
/* { dg-final { scan-assembler-times "vmulbf16\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 2 } } */ |
| 4 |
/* { dg-final { scan-assembler-times "vaddbf16\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 2 } } */ |
| 5 |
/* { dg-final { scan-assembler-times "vdivbf16\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 2 } } */ |
| 6 |
/* { dg-final { scan-assembler-times "vsubbf16\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 2 } } */ |
| 7 |
|
| 8 |
typedef __bf16 v4bf __attribute__ ((__vector_size__ (8))); |
| 9 |
typedef __bf16 v2bf __attribute__ ((__vector_size__ (4))); |
| 10 |
|
| 11 |
v4bf |
| 12 |
foo_mul_4 (v4bf a, v4bf b) |
| 13 |
{ |
| 14 |
return a * b; |
| 15 |
} |
| 16 |
|
| 17 |
v4bf |
| 18 |
foo_add_4 (v4bf a, v4bf b) |
| 19 |
{ |
| 20 |
return a + b; |
| 21 |
} |
| 22 |
|
| 23 |
v4bf |
| 24 |
foo_div_4 (v4bf a, v4bf b) |
| 25 |
{ |
| 26 |
return a / b; |
| 27 |
} |
| 28 |
|
| 29 |
v4bf |
| 30 |
foo_sub_4 (v4bf a, v4bf b) |
| 31 |
{ |
| 32 |
return a - b; |
| 33 |
} |
| 34 |
|
| 35 |
v2bf |
| 36 |
foo_mul_2 (v2bf a, v2bf b) |
| 37 |
{ |
| 38 |
return a * b; |
| 39 |
} |
| 40 |
|
| 41 |
v2bf |
| 42 |
foo_add_2 (v2bf a, v2bf b) |
| 43 |
{ |
| 44 |
return a + b; |
| 45 |
} |
| 46 |
|
| 47 |
v2bf |
| 48 |
foo_div_2 (v2bf a, v2bf b) |
| 49 |
{ |
| 50 |
return a / b; |
| 51 |
} |
| 52 |
|
| 53 |
v2bf |
| 54 |
foo_sub_2 (v2bf a, v2bf b) |
| 55 |
{ |
| 56 |
return a - b; |
| 57 |
} |