| 1 |
/* { dg-do compile } */ |
| 2 |
/* { dg-additional-options "-march=rv64gcv_zvfh -mabi=lp64d -Wno-pedantic" } */ |
| 3 |
|
| 4 |
#include <stdint-gcc.h> |
| 5 |
|
| 6 |
typedef int64_t vnx16di __attribute__((vector_size (128))); |
| 7 |
typedef int32_t vnx32si __attribute__((vector_size (128))); |
| 8 |
typedef int16_t vnx64hi __attribute__((vector_size (128))); |
| 9 |
typedef int8_t vnx128qi __attribute__((vector_size (128))); |
| 10 |
typedef _Float16 vnx64hf __attribute__((vector_size (128))); |
| 11 |
typedef float vnx32sf __attribute__((vector_size (128))); |
| 12 |
typedef double vnx16df __attribute__((vector_size (128))); |
| 13 |
|
| 14 |
#define VEC_SET(S,V,IDX) \ |
| 15 |
V \ |
| 16 |
__attribute__((noipa)) \ |
| 17 |
vec_set_##V##_##IDX (V v, S s) \ |
| 18 |
{ \ |
| 19 |
v[IDX] = s; \ |
| 20 |
return v; \ |
| 21 |
} |
| 22 |
|
| 23 |
#define VEC_SET_VAR4(S,V) \ |
| 24 |
V \ |
| 25 |
__attribute__((noipa)) \ |
| 26 |
vec_set_var_##V (V v, int64_t idx, S s) \ |
| 27 |
{ \ |
| 28 |
v[idx] = s; \ |
| 29 |
return v; \ |
| 30 |
} \ |
| 31 |
|
| 32 |
#define TEST_ALL4(T) \ |
| 33 |
T (_Float16, vnx64hf, 0) \ |
| 34 |
T (_Float16, vnx64hf, 3) \ |
| 35 |
T (_Float16, vnx64hf, 7) \ |
| 36 |
T (_Float16, vnx64hf, 8) \ |
| 37 |
T (_Float16, vnx64hf, 16) \ |
| 38 |
T (_Float16, vnx64hf, 31) \ |
| 39 |
T (_Float16, vnx64hf, 42) \ |
| 40 |
T (_Float16, vnx64hf, 63) \ |
| 41 |
T (float, vnx32sf, 0) \ |
| 42 |
T (float, vnx32sf, 3) \ |
| 43 |
T (float, vnx32sf, 12) \ |
| 44 |
T (float, vnx32sf, 17) \ |
| 45 |
T (float, vnx32sf, 14) \ |
| 46 |
T (double, vnx16df, 0) \ |
| 47 |
T (double, vnx16df, 4) \ |
| 48 |
T (double, vnx16df, 8) \ |
| 49 |
T (double, vnx16df, 12) \ |
| 50 |
T (int64_t, vnx16di, 0) \ |
| 51 |
T (int64_t, vnx16di, 4) \ |
| 52 |
T (int64_t, vnx16di, 8) \ |
| 53 |
T (int64_t, vnx16di, 12) \ |
| 54 |
T (int32_t, vnx32si, 0) \ |
| 55 |
T (int32_t, vnx32si, 4) \ |
| 56 |
T (int32_t, vnx32si, 12) \ |
| 57 |
T (int32_t, vnx32si, 16) \ |
| 58 |
T (int32_t, vnx32si, 28) \ |
| 59 |
T (int16_t, vnx64hi, 0) \ |
| 60 |
T (int16_t, vnx64hi, 4) \ |
| 61 |
T (int16_t, vnx64hi, 28) \ |
| 62 |
T (int16_t, vnx64hi, 32) \ |
| 63 |
T (int16_t, vnx64hi, 60) \ |
| 64 |
T (int8_t, vnx128qi, 0) \ |
| 65 |
T (int8_t, vnx128qi, 4) \ |
| 66 |
T (int8_t, vnx128qi, 30) \ |
| 67 |
T (int8_t, vnx128qi, 60) \ |
| 68 |
T (int8_t, vnx128qi, 64) \ |
| 69 |
T (int8_t, vnx128qi, 127) \ |
| 70 |
|
| 71 |
#define TEST_ALL_VAR4(T) \ |
| 72 |
T (_Float16, vnx64hf) \ |
| 73 |
T (float, vnx32sf) \ |
| 74 |
T (double, vnx16df) \ |
| 75 |
T (int64_t, vnx16di) \ |
| 76 |
T (int32_t, vnx32si) \ |
| 77 |
T (int16_t, vnx64hi) \ |
| 78 |
T (int8_t, vnx128qi) \ |
| 79 |
|
| 80 |
TEST_ALL4 (VEC_SET) |
| 81 |
TEST_ALL_VAR4 (VEC_SET_VAR4) |
| 82 |
|
| 83 |
/* { dg-final { scan-assembler-times {vset[i]*vli\s+[a-z0-9,]+,\s*e8,\s*m8,\s*tu,\s*ma} 7 } } */ |
| 84 |
/* { dg-final { scan-assembler-times {vset[i]*vli\s+[a-z0-9,]+,\s*e16,\s*m8,\s*tu,\s*ma} 15 } } */ |
| 85 |
/* { dg-final { scan-assembler-times {vset[i]*vli\s+[a-z0-9,]+,\s*e32,\s*m8,\s*tu,\s*ma} 12 } } */ |
| 86 |
/* { dg-final { scan-assembler-times {vset[i]*vli\s+[a-z0-9,]+,\s*e64,\s*m8,\s*tu,\s*ma} 10 } } */ |
| 87 |
|
| 88 |
/* { dg-final { scan-assembler-times {\tvmv.v.x} 20 } } */ |
| 89 |
/* { dg-final { scan-assembler-times {\tvfmv.v.f} 17 } } */ |
| 90 |
/* { dg-final { scan-assembler-times {\tvslideup.vi} 23 } } */ |
| 91 |
/* { dg-final { scan-assembler-times {\tvslideup.vx} 14 } } */ |
| 92 |
|
| 93 |
/* { dg-final { scan-assembler-times {\tvfmv.s.f} 3 } } */ |
| 94 |
/* { dg-final { scan-assembler-times {\tvmv.s.x} 4 } } */ |