1 /* { dg-do assemble { target aarch64_asm_sve_ok } } */
2 /* { dg-options "-O -msve-vector-bits=256 --save-temps" } */
6 typedef int8_t vnx16qi
__attribute__((vector_size (32)));
7 typedef int16_t vnx8hi
__attribute__((vector_size (32)));
8 typedef int32_t vnx4si
__attribute__((vector_size (32)));
9 typedef int64_t vnx2di
__attribute__((vector_size (32)));
10 typedef _Float16 vnx8hf
__attribute__((vector_size (32)));
11 typedef float vnx4sf
__attribute__((vector_size (32)));
12 typedef double vnx2df
__attribute__((vector_size (32)));
14 #define MASK_2(X, Y) (Y) - 1 - (X), (Y) - 2 - (X)
15 #define MASK_4(X, Y) MASK_2 (X, Y), MASK_2 (X + 2, Y)
16 #define MASK_8(X, Y) MASK_4 (X, Y), MASK_4 (X + 4, Y)
17 #define MASK_16(X, Y) MASK_8 (X, Y), MASK_8 (X + 8, Y)
18 #define MASK_32(X, Y) MASK_16 (X, Y), MASK_16 (X + 16, Y)
20 #define INDEX_32 vnx16qi
21 #define INDEX_16 vnx8hi
22 #define INDEX_8 vnx4si
23 #define INDEX_4 vnx2di
25 #define PERMUTE(TYPE, NUNITS) \
26 TYPE permute_##TYPE (TYPE values1, TYPE values2) \
28 return __builtin_shuffle \
30 ((INDEX_##NUNITS) { MASK_##NUNITS (0, NUNITS) })); \
44 /* { dg-final { scan-assembler-not {\ttbl\t} } } */
46 /* { dg-final { scan-assembler-times {\trev\tz[0-9]+\.b, z[0-9]+\.b\n} 1 } } */
47 /* { dg-final { scan-assembler-times {\trev\tz[0-9]+\.h, z[0-9]+\.h\n} 2 } } */
48 /* { dg-final { scan-assembler-times {\trev\tz[0-9]+\.s, z[0-9]+\.s\n} 2 } } */
49 /* { dg-final { scan-assembler-times {\trev\tz[0-9]+\.d, z[0-9]+\.d\n} 2 } } */