diff options
author | Ezra Sitorus <ezra.sitorus@arm.com> | 2023-12-07 15:21:55 +0000 |
---|---|---|
committer | Richard Earnshaw <rearnsha@arm.com> | 2023-12-07 17:13:53 +0000 |
commit | 2514a331835e055a963fd059dc5770e5ae500af0 (patch) | |
tree | e179288021c073aea057aa1cf6278940aae5ed99 /gcc/testsuite | |
parent | a1a0cdf21bb6a076e98658d815645d8ad1193840 (diff) | |
download | gcc-2514a331835e055a963fd059dc5770e5ae500af0.zip gcc-2514a331835e055a963fd059dc5770e5ae500af0.tar.gz gcc-2514a331835e055a963fd059dc5770e5ae500af0.tar.bz2 |
arm: vld1q_types_x3 ACLE intrinsics
This patch is part of a series of patches implementing the _xN
variants of the vld1q intrinsic for the arm port. This patch adds the
_x3 variants of the vld1q 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
(vld1q_u8_x3, vld1q_u16_x3, vld1q_u32_x3, vld1q_u64_x3): New.
(vld1q_s8_x3, vld1q_s16_x3, vld1q_s32_x3, vld1q_s64_x3): New.
(vld1q_f16_x3, vld1q_f32_x3): New.
(vld1q_p8_x3, vld1q_p16_x3, vld1q_p64_x3): New.
(vld1q_bf16_x3): New.
* config/arm/arm_neon_builtins.def (vld1_x3): New entries.
* config/arm/neon.md (vld1_x3<mode>): New.
gcc/testsuite/ChangeLog:
* gcc.target/arm/simd/vld1q_base_xN_1.c: Add new tests.
* gcc.target/arm/simd/vld1q_bf16_xN_1.c: Add new tests.
* gcc.target/arm/simd/vld1q_fp16_xN_1.c: Add new tests.
* gcc.target/arm/simd/vld1q_p64_xN_1.c: Add new tests.
Diffstat (limited to 'gcc/testsuite')
4 files changed, 80 insertions, 3 deletions
diff --git a/gcc/testsuite/gcc.target/arm/simd/vld1q_base_xN_1.c b/gcc/testsuite/gcc.target/arm/simd/vld1q_base_xN_1.c index 1d31777..bfad282 100644 --- a/gcc/testsuite/gcc.target/arm/simd/vld1q_base_xN_1.c +++ b/gcc/testsuite/gcc.target/arm/simd/vld1q_base_xN_1.c @@ -60,8 +60,69 @@ poly16x8x2_t test_vld1q_p16_x2 (poly16_t * a) return vld1q_p16_x2 (a); } +uint8x16x3_t test_vld1q_u8_x3 (uint8_t * a) +{ + return vld1q_u8_x3 (a); +} + +uint16x8x3_t test_vld1q_u16_x3 (uint16_t * a) +{ + return vld1q_u16_x3 (a); +} + +uint32x4x3_t test_vld1q_u32_x3 (uint32_t * a) +{ + return vld1q_u32_x3 (a); +} + +uint64x2x3_t test_vld1q_u64_x3 (uint64_t * a) +{ + return vld1q_u64_x3 (a); +} + +int8x16x3_t test_vld1q_s8_x3 (int8_t * a) +{ + return vld1q_s8_x3 (a); +} + +int16x8x3_t test_vld1q_s16_x3 (int16_t * a) +{ + return vld1q_s16_x3 (a); +} + +int32x4x3_t test_vld1q_s32_x3 (int32_t * a) +{ + return vld1q_s32_x3 (a); +} + +int64x2x3_t test_vld1q_s64_x3 (int64_t * a) +{ + return vld1q_s64_x3 (a); +} + +float32x4x3_t test_vld1q_f32_x3 (float32_t * a) +{ + return vld1q_f32_x3 (a); +} + +poly8x16x3_t test_vld1q_p8_x3 (poly8_t * a) +{ + return vld1q_p8_x3 (a); +} + +poly16x8x3_t test_vld1q_p16_x3 (poly16_t * a) +{ + return vld1q_p16_x3 (a); +} + /* { dg-final { scan-assembler-times {vld1.8\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 3 } } */ +/* { dg-final { scan-assembler-times {vld1.8\t\{d[0-9]+, d[0-9]+, d[0-9]+\}, \[r[0-9]+\]\n} 6 } } */ + /* { dg-final { scan-assembler-times {vld1.16\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 3 } } */ +/* { dg-final { scan-assembler-times {vld1.16\t\{d[0-9]+, d[0-9]+, d[0-9]+\}, \[r[0-9]+\]\n} 6 } } */ + /* { dg-final { scan-assembler-times {vld1.32\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 3 } } */ -/* { dg-final { scan-assembler-times {vld1.64\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+:64\]\n} 2 } } */ +/* { dg-final { scan-assembler-times {vld1.32\t\{d[0-9]+, d[0-9]+, d[0-9]+\}, \[r[0-9]+\]\n} 6 } } */ +/* { dg-final { scan-assembler-times {vld1.64\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+:64\]\n} 2 } } */ +/* { dg-final { scan-assembler-times {vld1.64\t\{d[0-9]+, d[0-9]+, d[0-9]+\}, \[r[0-9]+:64\]\n} 4 } } */ diff --git a/gcc/testsuite/gcc.target/arm/simd/vld1q_bf16_xN_1.c b/gcc/testsuite/gcc.target/arm/simd/vld1q_bf16_xN_1.c index 5f6fc98..4138fe9 100644 --- a/gcc/testsuite/gcc.target/arm/simd/vld1q_bf16_xN_1.c +++ b/gcc/testsuite/gcc.target/arm/simd/vld1q_bf16_xN_1.c @@ -10,4 +10,10 @@ bfloat16x8x2_t test_vld1q_bf16_x2 (bfloat16_t * a) return vld1q_bf16_x2 (a); } +bfloat16x8x3_t test_vld1q_bf16_x3 (bfloat16_t * a) +{ + return vld1q_bf16_x3 (a); +} + /* { dg-final { scan-assembler-times {vld1.16\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 1 } } */ +/* { dg-final { scan-assembler-times {vld1.16\t\{d[0-9]+, d[0-9]+, d[0-9]+\}, \[r[0-9]+\]\n} 2 } } */ diff --git a/gcc/testsuite/gcc.target/arm/simd/vld1q_fp16_xN_1.c b/gcc/testsuite/gcc.target/arm/simd/vld1q_fp16_xN_1.c index aecf491..01640d7 100644 --- a/gcc/testsuite/gcc.target/arm/simd/vld1q_fp16_xN_1.c +++ b/gcc/testsuite/gcc.target/arm/simd/vld1q_fp16_xN_1.c @@ -10,5 +10,10 @@ float16x8x2_t test_vld1q_f16_x2 (float16_t * a) return vld1q_f16_x2 (a); } -/* { dg-final { scan-assembler-times {vld1.16\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 1 } } */ +float16x8x3_t test_vld1q_f16_x3 (float16_t * a) +{ + return vld1q_f16_x3 (a); +} +/* { dg-final { scan-assembler-times {vld1.16\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+\]\n} 1 } } */ +/* { dg-final { scan-assembler-times {vld1.16\t\{d[0-9]+, d[0-9]+, d[0-9]+\}, \[r[0-9]+\]\n} 2 } } */ diff --git a/gcc/testsuite/gcc.target/arm/simd/vld1q_p64_xN_1.c b/gcc/testsuite/gcc.target/arm/simd/vld1q_p64_xN_1.c index 04ceb5e..ae2ab36 100644 --- a/gcc/testsuite/gcc.target/arm/simd/vld1q_p64_xN_1.c +++ b/gcc/testsuite/gcc.target/arm/simd/vld1q_p64_xN_1.c @@ -10,5 +10,10 @@ poly64x2x2_t test_vld1q_p64_x2 (poly64_t * a) return vld1q_p64_x2 (a); } -/* { dg-final { scan-assembler-times {vld1.64\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+:64\]\n} 1 } } */ +poly64x2x3_t test_vld1q_p64_x3 (poly64_t * a) +{ + return vld1q_p64_x3 (a); +} +/* { dg-final { scan-assembler-times {vld1.64\t\{d[0-9]+-d[0-9]+\}, \[r[0-9]+:64\]\n} 1 } } */ +/* { dg-final { scan-assembler-times {vld1.64\t\{d[0-9]+, d[0-9]+, d[0-9]+\}, \[r[0-9]+:64\]\n} 2 } } */ |