| 1 |
/* { dg-do compile } */ |
| 2 |
/* { dg-additional-options "-O3 -march=armv9-a+fp8dot4 -moverride=tune=none" } */ |
| 3 |
/* { dg-final { check-function-bodies "**" "" "" { target *-*-* } {\.L[0-9]+} } } */ |
| 4 |
|
| 5 |
#include "arm_neon.h" |
| 6 |
|
| 7 |
/* |
| 8 |
** f1: |
| 9 |
** mrs (x[0-9]+), fpmr |
| 10 |
** cmp \1, x0 |
| 11 |
** beq ([^\n]+) |
| 12 |
** msr fpmr, x0 |
| 13 |
** ?\2: |
| 14 |
** fdot v0.2s, v1.8b, v2.8b |
| 15 |
** ret |
| 16 |
*/ |
| 17 |
float32x2_t |
| 18 |
f1 (float32x2_t a, mfloat8x8_t b, mfloat8x8_t c, fpm_t d) |
| 19 |
{ |
| 20 |
return vdot_f32_mf8_fpm (a, b, c, d); |
| 21 |
} |
| 22 |
|
| 23 |
/* |
| 24 |
** f2: |
| 25 |
** mrs (x[0-9]+), fpmr |
| 26 |
** cbz \1, ([^\n]+) |
| 27 |
** msr fpmr, xzr |
| 28 |
** ?\2: |
| 29 |
** fdot v0.2s, v1.8b, v2.8b |
| 30 |
** ret |
| 31 |
*/ |
| 32 |
float32x2_t |
| 33 |
f2 (float32x2_t a, mfloat8x8_t b, mfloat8x8_t c) |
| 34 |
{ |
| 35 |
return vdot_f32_mf8_fpm (a, b, c, 0); |
| 36 |
} |