diff options
author | Ezra Sitorus <ezra.sitorus@arm.com> | 2023-12-07 15:28:44 +0000 |
---|---|---|
committer | Richard Earnshaw <rearnsha@arm.com> | 2023-12-07 17:14:40 +0000 |
commit | 2f48d846c794ba091b266133f73717361096d454 (patch) | |
tree | f444f00a63186e377059f13ee24492826f6c271f /gcc/testsuite | |
parent | ef07ae652c25ec04c2e3ef8cec14b0771a809861 (diff) | |
download | gcc-2f48d846c794ba091b266133f73717361096d454.zip gcc-2f48d846c794ba091b266133f73717361096d454.tar.gz gcc-2f48d846c794ba091b266133f73717361096d454.tar.bz2 |
arm: vst1_types_x4 ACLE intrinsics
This patch is part of a series of patches implementing the _xN
variants of the vst1 intrinsic for the arm port. This patch adds the
_x4 variants of the vst1 intrinsic.
ACLE documents:
https://developer.arm.com/documentation/ihi0053/latest/
ISA documents:
https://developer.arm.com/documentation/ddi0487/latest/
gcc/ChangeLog:
* config/arm/arm_neon.h
(vst1_u8_x4, vst1_u16_x4, vst1_u32_x4, vst1_u64_x4): New.
(vst1_s8_x4, vst1_s16_x4, vst1_s32_x4, vst1_s64_x4): New.
(vst1_f16_x4, vst1_f32_x4): New.
(vst1_p8_x4, vst1_p16_x4, vst1_p64_x4): New.
(vst1_bf16_x4): New.
* config/arm/arm_neon_builtins.def (vst1_x4): New entries.
* config/arm/neon.md (vst1_x4<mode>): New.
gcc/testsuite/ChangeLog:
* gcc.target/arm/simd/vst1_base_xN_1.c: Add new test.
* gcc.target/arm/simd/vst1_bf16_xN_1.c: Add new test.
* gcc.target/arm/simd/vst1_fp16_xN_1.c: Add new test.
* gcc.target/arm/simd/vst1_p64_xN_1.c: Add new test.
Diffstat (limited to 'gcc/testsuite')
4 files changed, 75 insertions, 7 deletions
diff --git a/gcc/testsuite/gcc.target/arm/simd/vst1_base_xN_1.c b/gcc/testsuite/gcc.target/arm/simd/vst1_base_xN_1.c index 5f820a6..04ca6583 100644 --- a/gcc/testsuite/gcc.target/arm/simd/vst1_base_xN_1.c +++ b/gcc/testsuite/gcc.target/arm/simd/vst1_base_xN_1.c @@ -115,8 +115,62 @@ void test_vst1_p16_x3 (poly16_t * ptr, poly16x4x3_t val) vst1_p16_x3 (ptr, val); } +void test_vst1_u8_x4 (uint8_t * ptr, uint8x8x4_t val) +{ + vst1_u8_x4 (ptr, val); +} + +void test_vst1_u16_x4 (uint16_t * ptr, uint16x4x4_t val) +{ + vst1_u16_x4 (ptr, val); +} + +void test_vst1_u32_x4 (uint32_t * ptr, uint32x2x4_t val) +{ + vst1_u32_x4 (ptr, val); +} + +void test_vst1_u64_x4 (uint64_t * ptr, uint64x1x4_t val) +{ + vst1_u64_x4 (ptr, val); +} + +void test_vst1_s8_x4 (int8_t * ptr, int8x8x4_t val) +{ + vst1_s8_x4 (ptr, val); +} + +void test_vst1_s16_x4 (int16_t * ptr, int16x4x4_t val) +{ + vst1_s16_x4 (ptr, val); +} + +void test_vst1_s32_x4 (int32_t * ptr, int32x2x4_t val) +{ + vst1_s32_x4 (ptr, val); +} + +void test_vst1_s64_x4 (int64_t * ptr, int64x1x4_t val) +{ + vst1_s64_x4 (ptr, val); +} + +void test_vst1_f32_x4 (float32_t * ptr, float32x2x4_t val) +{ + vst1_f32_x4 (ptr, val); +} + +void test_vst1_p8_x4 (poly8_t * ptr, poly8x8x4_t val) +{ + vst1_p8_x4 (ptr, val); +} + +void test_vst1_p16_x4 (poly16_t * ptr, poly16x4x4_t val) +{ + vst1_p16_x4 (ptr, val); +} -/* { dg-final { scan-assembler-times {vst1.8\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 6 } } */ -/* { dg-final { scan-assembler-times {vst1.16\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 6 } } */ -/* { dg-final { scan-assembler-times {vst1.32\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 6 } } */ -/* { dg-final { scan-assembler-times {vst1.64\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+:64\]\n} 4 } } */ +/* { dg-final { scan-assembler-times {vst1.8\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 9 } } */ +/* { dg-final { scan-assembler-times {vst1.16\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 9 } } */ +/* { dg-final { scan-assembler-times {vst1.32\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 9 } } */ +/* { dg-final { scan-assembler-times {vst1.64\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+:64\]\n} 6 } } */ diff --git a/gcc/testsuite/gcc.target/arm/simd/vst1_bf16_xN_1.c b/gcc/testsuite/gcc.target/arm/simd/vst1_bf16_xN_1.c index a3a00ea..d919c7d 100644 --- a/gcc/testsuite/gcc.target/arm/simd/vst1_bf16_xN_1.c +++ b/gcc/testsuite/gcc.target/arm/simd/vst1_bf16_xN_1.c @@ -15,4 +15,8 @@ void test_vst1_bf16_x3 (bfloat16_t * ptr, bfloat16x4x3_t val) vst1_bf16_x3 (ptr, val); } -/* { dg-final { scan-assembler-times {vst1.16\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 2 } } */ +void test_vst1_bf16_x4 (bfloat16_t * ptr, bfloat16x4x4_t val) +{ + vst1_bf16_x4 (ptr, val); +} +/* { dg-final { scan-assembler-times {vst1.16\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 3 } } */ diff --git a/gcc/testsuite/gcc.target/arm/simd/vst1_fp16_xN_1.c b/gcc/testsuite/gcc.target/arm/simd/vst1_fp16_xN_1.c index 0a6863e..3d1d1eb7 100644 --- a/gcc/testsuite/gcc.target/arm/simd/vst1_fp16_xN_1.c +++ b/gcc/testsuite/gcc.target/arm/simd/vst1_fp16_xN_1.c @@ -15,4 +15,9 @@ void test_vst1_f16_x3 (float16_t * ptr, float16x4x3_t val) vst1_f16_x3 (ptr, val); } -/* { dg-final { scan-assembler-times {vst1.16\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 2 } } */ +void test_vst1_f16_x4 (float16_t * ptr, float16x4x4_t val) +{ + vst1_f16_x4 (ptr, val); +} + +/* { dg-final { scan-assembler-times {vst1.16\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 3 } } */ diff --git a/gcc/testsuite/gcc.target/arm/simd/vst1_p64_xN_1.c b/gcc/testsuite/gcc.target/arm/simd/vst1_p64_xN_1.c index 5dbd604..6291214 100644 --- a/gcc/testsuite/gcc.target/arm/simd/vst1_p64_xN_1.c +++ b/gcc/testsuite/gcc.target/arm/simd/vst1_p64_xN_1.c @@ -15,4 +15,9 @@ void test_vst1_p64_x3 (poly64_t * ptr, poly64x1x3_t val) vst1_p64_x3 (ptr, val); } -/* { dg-final { scan-assembler-times {vst1.64\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+:64\]\n} 2 } } */
\ No newline at end of file +void test_vst1_p64_x4 (poly64_t * ptr, poly64x1x4_t val) +{ + vst1_p64_x4 (ptr, val); +} + +/* { dg-final { scan-assembler-times {vst1.64\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+:64\]\n} 3 } } */
\ No newline at end of file |