| 1 |
/* { dg-do compile } */ |
| 2 |
/* { dg-options "-O -msve-vector-bits=256" } */ |
| 3 |
|
| 4 |
#include <stdint.h> |
| 5 |
|
| 6 |
typedef int64_t vnx2di __attribute__((vector_size (32))); |
| 7 |
typedef int32_t vnx4si __attribute__((vector_size (32))); |
| 8 |
typedef int16_t vnx8hi __attribute__((vector_size (32))); |
| 9 |
typedef int8_t vnx16qi __attribute__((vector_size (32))); |
| 10 |
typedef double vnx2df __attribute__((vector_size (32))); |
| 11 |
typedef float vnx4sf __attribute__((vector_size (32))); |
| 12 |
typedef _Float16 vnx8hf __attribute__((vector_size (32))); |
| 13 |
|
| 14 |
#define MASK_2(X) X, X |
| 15 |
#define MASK_4(X) MASK_2 (X), MASK_2 (X) |
| 16 |
#define MASK_8(X) MASK_4 (X), MASK_4 (X) |
| 17 |
#define MASK_16(X) MASK_8 (X), MASK_8 (X) |
| 18 |
#define MASK_32(X) MASK_16 (X), MASK_16 (X) |
| 19 |
|
| 20 |
#define INDEX_4 vnx2di |
| 21 |
#define INDEX_8 vnx4si |
| 22 |
#define INDEX_16 vnx8hi |
| 23 |
#define INDEX_32 vnx16qi |
| 24 |
|
| 25 |
#define DUP_LANE(TYPE, NUNITS, INDEX) \ |
| 26 |
TYPE dup_##INDEX##_##TYPE (TYPE values1, TYPE values2) \ |
| 27 |
{ \ |
| 28 |
return __builtin_shuffle (values1, values2, \ |
| 29 |
((INDEX_##NUNITS) { MASK_##NUNITS (INDEX) })); \ |
| 30 |
} |
| 31 |
|
| 32 |
#define TEST_ALL(T) \ |
| 33 |
T (vnx2di, 4, 0) \ |
| 34 |
T (vnx2di, 4, 2) \ |
| 35 |
T (vnx2di, 4, 3) \ |
| 36 |
T (vnx4si, 8, 0) \ |
| 37 |
T (vnx4si, 8, 5) \ |
| 38 |
T (vnx4si, 8, 7) \ |
| 39 |
T (vnx8hi, 16, 0) \ |
| 40 |
T (vnx8hi, 16, 6) \ |
| 41 |
T (vnx8hi, 16, 15) \ |
| 42 |
T (vnx16qi, 32, 0) \ |
| 43 |
T (vnx16qi, 32, 19) \ |
| 44 |
T (vnx16qi, 32, 31) \ |
| 45 |
T (vnx2df, 4, 0) \ |
| 46 |
T (vnx2df, 4, 2) \ |
| 47 |
T (vnx2df, 4, 3) \ |
| 48 |
T (vnx4sf, 8, 0) \ |
| 49 |
T (vnx4sf, 8, 5) \ |
| 50 |
T (vnx4sf, 8, 7) \ |
| 51 |
T (vnx8hf, 16, 0) \ |
| 52 |
T (vnx8hf, 16, 6) \ |
| 53 |
T (vnx8hf, 16, 15) \ |
| 54 |
|
| 55 |
TEST_ALL (DUP_LANE) |
| 56 |
|
| 57 |
/* { dg-final { scan-assembler-not {\ttbl\t} } } */ |
| 58 |
|
| 59 |
/* { dg-final { scan-assembler-times {\tmov\tz[0-9]+\.d, d[0-9]} 2 { target aarch64_little_endian } } } */ |
| 60 |
/* { dg-final { scan-assembler-times {\tdup\tz[0-9]+\.d, z[0-9]+\.d\[0\]} 2 { target aarch64_big_endian } } } */ |
| 61 |
/* { dg-final { scan-assembler-times {\tdup\tz[0-9]+\.d, z[0-9]+\.d\[2\]} 2 } } */ |
| 62 |
/* { dg-final { scan-assembler-times {\tdup\tz[0-9]+\.d, z[0-9]+\.d\[3\]} 2 } } */ |
| 63 |
/* { dg-final { scan-assembler-times {\tmov\tz[0-9]+\.s, s[0-9]} 2 { target aarch64_little_endian } } } */ |
| 64 |
/* { dg-final { scan-assembler-times {\tdup\tz[0-9]+\.s, z[0-9]+\.s\[0\]} 2 { target aarch64_big_endian } } } */ |
| 65 |
/* { dg-final { scan-assembler-times {\tdup\tz[0-9]+\.s, z[0-9]+\.s\[5\]} 2 } } */ |
| 66 |
/* { dg-final { scan-assembler-times {\tdup\tz[0-9]+\.s, z[0-9]+\.s\[7\]} 2 } } */ |
| 67 |
/* { dg-final { scan-assembler-times {\tmov\tz[0-9]+\.h, h[0-9]} 2 { target aarch64_little_endian } } } */ |
| 68 |
/* { dg-final { scan-assembler-times {\tdup\tz[0-9]+\.h, z[0-9]+\.h\[0\]} 2 { target aarch64_big_endian } } } */ |
| 69 |
/* { dg-final { scan-assembler-times {\tdup\tz[0-9]+\.h, z[0-9]+\.h\[6\]} 2 } } */ |
| 70 |
/* { dg-final { scan-assembler-times {\tdup\tz[0-9]+\.h, z[0-9]+\.h\[15\]} 2 } } */ |
| 71 |
/* { dg-final { scan-assembler-times {\tmov\tz[0-9]+\.b, b[0-9]} 1 { target aarch64_little_endian } } } */ |
| 72 |
/* { dg-final { scan-assembler-times {\tdup\tz[0-9]+\.b, z[0-9]+\.b\[0\]} 1 { target aarch64_big_endian } } } */ |
| 73 |
/* { dg-final { scan-assembler-times {\tdup\tz[0-9]+\.b, z[0-9]+\.b\[19\]} 1 } } */ |
| 74 |
/* { dg-final { scan-assembler-times {\tdup\tz[0-9]+\.b, z[0-9]+\.b\[31\]} 1 } } */ |