| 1 |
/* { dg-do compile } */ |
| 2 |
/* { dg-options "-mrvv-vector-bits=scalable -march=rv32gcv -mabi=ilp32 -fno-schedule-insns -fno-schedule-insns2 -fno-tree-vectorize" } */ |
| 3 |
|
| 4 |
#include "riscv_vector.h" |
| 5 |
|
| 6 |
void f (int8_t * restrict in, int8_t * restrict out, int n, int m, int cond) |
| 7 |
{ |
| 8 |
size_t vl = 101; |
| 9 |
vbool64_t mask = *(vbool64_t*) (in + 1000000); |
| 10 |
for (size_t j = 0; j < m; j++){ |
| 11 |
for (size_t i = 0; i < cond; i++) |
| 12 |
{ |
| 13 |
vint8mf8_t v = __riscv_vle8_v_i8mf8 (in + i + j, vl); |
| 14 |
__riscv_vse8_v_i8mf8 (out + i, v, vl); |
| 15 |
} |
| 16 |
|
| 17 |
for (size_t i = 0; i < n * n; i++) |
| 18 |
out[i] = out[i] * out[i]; |
| 19 |
for (size_t i = 0; i < n * n * n; i++) |
| 20 |
out[i] = out[i] + out[i]; |
| 21 |
for (size_t i = 0; i < n * n * n * n; i++) |
| 22 |
out[i] = out[i] + 2; |
| 23 |
for (size_t i = 0; i < n * n * n * n * n; i++) |
| 24 |
out[i] = out[i] * 100; |
| 25 |
for (size_t i = 0; i < n * n * n * n * n * n; i++) |
| 26 |
out[i] = out[i] - 77; |
| 27 |
|
| 28 |
for (size_t i = 0; i < n; i++) |
| 29 |
{ |
| 30 |
vfloat32mf2_t v = __riscv_vle32_v_f32mf2 ((float *)(in + i + j + 200), vl); |
| 31 |
__riscv_vse32_v_f32mf2 ((float *)(out + i + j + 200), v, vl); |
| 32 |
|
| 33 |
vfloat32mf2_t v2 = __riscv_vle32_v_f32mf2_tumu (mask, v, (float *)(in + i + j + 300), vl); |
| 34 |
__riscv_vse32_v_f32mf2_m (mask, (float *)(out + i + j + 300), v2, vl); |
| 35 |
} |
| 36 |
} |
| 37 |
} |
| 38 |
|
| 39 |
/* { dg-final { scan-assembler-times {vsetvli\s+zero,\s*[a-x0-9]+,\s*e32,\s*mf2,\s*tu,\s*mu} 1 { target { no-opts "-O0" no-opts "-g" no-opts "-funroll-loops" } } } } */ |
| 40 |
/* { dg-final { scan-assembler-times {vsetvli} 2 { target { no-opts "-O0" no-opts "-g" no-opts "-funroll-loops" } } } } */ |