From: Jonathan Wright Date: Fri, 23 Jul 2021 12:41:39 +0000 (+0100) Subject: aarch64: Use memcpy to copy vector tables in vst1[q]_x2 intrinsics X-Git-Tag: upstream/12.2.0~6200 X-Git-Url: http://review.tizen.org/git/?a=commitdiff_plain;h=50752b751fff56e7e2c74024bae659d5e9dea50f;p=platform%2Fupstream%2Fgcc.git aarch64: Use memcpy to copy vector tables in vst1[q]_x2 intrinsics Use __builtin_memcpy to copy vector structures instead of building a new opaque structure one vector at a time in each of the vst1[q]_x2 Neon intrinsics in arm_neon.h. This simplifies the header file and also improves code generation - superfluous move instructions were emitted for every register extraction/set in this additional structure. Add new code generation tests to verify that superfluous move instructions are not generated for the vst1q_x2 intrinsics. gcc/ChangeLog: 2021-07-23 Jonathan Wright * config/aarch64/arm_neon.h (vst1_s64_x2): Use __builtin_memcpy instead of constructing __builtin_aarch64_simd_oi one vector at a time. (vst1_u64_x2): Likewise. (vst1_f64_x2): Likewise. (vst1_s8_x2): Likewise. (vst1_p8_x2): Likewise. (vst1_s16_x2): Likewise. (vst1_p16_x2): Likewise. (vst1_s32_x2): Likewise. (vst1_u8_x2): Likewise. (vst1_u16_x2): Likewise. (vst1_u32_x2): Likewise. (vst1_f16_x2): Likewise. (vst1_f32_x2): Likewise. (vst1_p64_x2): Likewise. (vst1q_s8_x2): Likewise. (vst1q_p8_x2): Likewise. (vst1q_s16_x2): Likewise. (vst1q_p16_x2): Likewise. (vst1q_s32_x2): Likewise. (vst1q_s64_x2): Likewise. (vst1q_u8_x2): Likewise. (vst1q_u16_x2): Likewise. (vst1q_u32_x2): Likewise. (vst1q_u64_x2): Likewise. (vst1q_f16_x2): Likewise. (vst1q_f32_x2): Likewise. (vst1q_f64_x2): Likewise. (vst1q_p64_x2): Likewise. gcc/testsuite/ChangeLog: * gcc.target/aarch64/vector_structure_intrinsics.c: Add new tests. --- diff --git a/gcc/config/aarch64/arm_neon.h b/gcc/config/aarch64/arm_neon.h index 47bb94c..7523974 100644 --- a/gcc/config/aarch64/arm_neon.h +++ b/gcc/config/aarch64/arm_neon.h @@ -26276,8 +26276,7 @@ vst1_s64_x2 (int64_t * __a, int64x1x2_t __val) = vcombine_s64 (__val.val[0], vcreate_s64 (__AARCH64_INT64_C (0))); __temp.val[1] = vcombine_s64 (__val.val[1], vcreate_s64 (__AARCH64_INT64_C (0))); - __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2di ((__builtin_aarch64_simd_di *) __a, __o); } @@ -26291,8 +26290,7 @@ vst1_u64_x2 (uint64_t * __a, uint64x1x2_t __val) = vcombine_u64 (__val.val[0], vcreate_u64 (__AARCH64_UINT64_C (0))); __temp.val[1] = vcombine_u64 (__val.val[1], vcreate_u64 (__AARCH64_UINT64_C (0))); - __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2di ((__builtin_aarch64_simd_di *) __a, __o); } @@ -26306,8 +26304,7 @@ vst1_f64_x2 (float64_t * __a, float64x1x2_t __val) = vcombine_f64 (__val.val[0], vcreate_f64 (__AARCH64_UINT64_C (0))); __temp.val[1] = vcombine_f64 (__val.val[1], vcreate_f64 (__AARCH64_UINT64_C (0))); - __o = __builtin_aarch64_set_qregoiv2df (__o, (float64x2_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv2df (__o, (float64x2_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2df ((__builtin_aarch64_simd_df *) __a, __o); } @@ -26321,8 +26318,7 @@ vst1_s8_x2 (int8_t * __a, int8x8x2_t __val) = vcombine_s8 (__val.val[0], vcreate_s8 (__AARCH64_INT64_C (0))); __temp.val[1] = vcombine_s8 (__val.val[1], vcreate_s8 (__AARCH64_INT64_C (0))); - __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2v8qi ((__builtin_aarch64_simd_qi *) __a, __o); } @@ -26336,8 +26332,7 @@ vst1_p8_x2 (poly8_t * __a, poly8x8x2_t __val) = vcombine_p8 (__val.val[0], vcreate_p8 (__AARCH64_UINT64_C (0))); __temp.val[1] = vcombine_p8 (__val.val[1], vcreate_p8 (__AARCH64_UINT64_C (0))); - __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2v8qi ((__builtin_aarch64_simd_qi *) __a, __o); } @@ -26351,8 +26346,7 @@ vst1_s16_x2 (int16_t * __a, int16x4x2_t __val) = vcombine_s16 (__val.val[0], vcreate_s16 (__AARCH64_INT64_C (0))); __temp.val[1] = vcombine_s16 (__val.val[1], vcreate_s16 (__AARCH64_INT64_C (0))); - __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2v4hi ((__builtin_aarch64_simd_hi *) __a, __o); } @@ -26366,8 +26360,7 @@ vst1_p16_x2 (poly16_t * __a, poly16x4x2_t __val) = vcombine_p16 (__val.val[0], vcreate_p16 (__AARCH64_UINT64_C (0))); __temp.val[1] = vcombine_p16 (__val.val[1], vcreate_p16 (__AARCH64_UINT64_C (0))); - __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2v4hi ((__builtin_aarch64_simd_hi *) __a, __o); } @@ -26381,8 +26374,7 @@ vst1_s32_x2 (int32_t * __a, int32x2x2_t __val) = vcombine_s32 (__val.val[0], vcreate_s32 (__AARCH64_INT64_C (0))); __temp.val[1] = vcombine_s32 (__val.val[1], vcreate_s32 (__AARCH64_INT64_C (0))); - __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2v2si ((__builtin_aarch64_simd_si *) __a, __o); } @@ -26394,8 +26386,7 @@ vst1_u8_x2 (uint8_t * __a, uint8x8x2_t __val) uint8x16x2_t __temp; __temp.val[0] = vcombine_u8 (__val.val[0], vcreate_u8 (__AARCH64_UINT64_C (0))); __temp.val[1] = vcombine_u8 (__val.val[1], vcreate_u8 (__AARCH64_UINT64_C (0))); - __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2v8qi ((__builtin_aarch64_simd_qi *) __a, __o); } @@ -26407,8 +26398,7 @@ vst1_u16_x2 (uint16_t * __a, uint16x4x2_t __val) uint16x8x2_t __temp; __temp.val[0] = vcombine_u16 (__val.val[0], vcreate_u16 (__AARCH64_UINT64_C (0))); __temp.val[1] = vcombine_u16 (__val.val[1], vcreate_u16 (__AARCH64_UINT64_C (0))); - __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2v4hi ((__builtin_aarch64_simd_hi *) __a, __o); } @@ -26420,8 +26410,7 @@ vst1_u32_x2 (uint32_t * __a, uint32x2x2_t __val) uint32x4x2_t __temp; __temp.val[0] = vcombine_u32 (__val.val[0], vcreate_u32 (__AARCH64_UINT64_C (0))); __temp.val[1] = vcombine_u32 (__val.val[1], vcreate_u32 (__AARCH64_UINT64_C (0))); - __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2v2si ((__builtin_aarch64_simd_si *) __a, __o); } @@ -26433,8 +26422,7 @@ vst1_f16_x2 (float16_t * __a, float16x4x2_t __val) float16x8x2_t __temp; __temp.val[0] = vcombine_f16 (__val.val[0], vcreate_f16 (__AARCH64_UINT64_C (0))); __temp.val[1] = vcombine_f16 (__val.val[1], vcreate_f16 (__AARCH64_UINT64_C (0))); - __o = __builtin_aarch64_set_qregoiv8hf (__o, __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv8hf (__o, __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2v4hf (__a, __o); } @@ -26446,8 +26434,7 @@ vst1_f32_x2 (float32_t * __a, float32x2x2_t __val) float32x4x2_t __temp; __temp.val[0] = vcombine_f32 (__val.val[0], vcreate_f32 (__AARCH64_UINT64_C (0))); __temp.val[1] = vcombine_f32 (__val.val[1], vcreate_f32 (__AARCH64_UINT64_C (0))); - __o = __builtin_aarch64_set_qregoiv4sf (__o, (float32x4_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv4sf (__o, (float32x4_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2v2sf ((__builtin_aarch64_simd_sf *) __a, __o); } @@ -26459,10 +26446,7 @@ vst1_p64_x2 (poly64_t * __a, poly64x1x2_t __val) poly64x2x2_t __temp; __temp.val[0] = vcombine_p64 (__val.val[0], vcreate_p64 (__AARCH64_UINT64_C (0))); __temp.val[1] = vcombine_p64 (__val.val[1], vcreate_p64 (__AARCH64_UINT64_C (0))); - __o = __builtin_aarch64_set_qregoiv2di_ssps (__o, - (poly64x2_t) __temp.val[0], 0); - __o = __builtin_aarch64_set_qregoiv2di_ssps (__o, - (poly64x2_t) __temp.val[1], 1); + __builtin_memcpy (&__o, &__temp, sizeof (__temp)); __builtin_aarch64_st1x2di ((__builtin_aarch64_simd_di *) __a, __o); } @@ -26471,8 +26455,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_s8_x2 (int8_t * __a, int8x16x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v16qi ((__builtin_aarch64_simd_qi *) __a, __o); } @@ -26481,8 +26464,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_p8_x2 (poly8_t * __a, poly8x16x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v16qi ((__builtin_aarch64_simd_qi *) __a, __o); } @@ -26491,8 +26473,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_s16_x2 (int16_t * __a, int16x8x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v8hi ((__builtin_aarch64_simd_hi *) __a, __o); } @@ -26501,8 +26482,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_p16_x2 (poly16_t * __a, poly16x8x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v8hi ((__builtin_aarch64_simd_hi *) __a, __o); } @@ -26511,8 +26491,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_s32_x2 (int32_t * __a, int32x4x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v4si ((__builtin_aarch64_simd_si *) __a, __o); } @@ -26521,8 +26500,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_s64_x2 (int64_t * __a, int64x2x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v2di ((__builtin_aarch64_simd_di *) __a, __o); } @@ -26531,8 +26509,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_u8_x2 (uint8_t * __a, uint8x16x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v16qi ((__builtin_aarch64_simd_qi *) __a, __o); } @@ -26541,8 +26518,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_u16_x2 (uint16_t * __a, uint16x8x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v8hi ((__builtin_aarch64_simd_hi *) __a, __o); } @@ -26551,8 +26527,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_u32_x2 (uint32_t * __a, uint32x4x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v4si ((__builtin_aarch64_simd_si *) __a, __o); } @@ -26561,8 +26536,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_u64_x2 (uint64_t * __a, uint64x2x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v2di ((__builtin_aarch64_simd_di *) __a, __o); } @@ -26571,8 +26545,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_f16_x2 (float16_t * __a, float16x8x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv8hf (__o, __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv8hf (__o, __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v8hf (__a, __o); } @@ -26581,8 +26554,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_f32_x2 (float32_t * __a, float32x4x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv4sf (__o, (float32x4_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv4sf (__o, (float32x4_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v4sf ((__builtin_aarch64_simd_sf *) __a, __o); } @@ -26591,8 +26563,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_f64_x2 (float64_t * __a, float64x2x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv2df (__o, (float64x2_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv2df (__o, (float64x2_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v2df ((__builtin_aarch64_simd_df *) __a, __o); } @@ -26601,10 +26572,7 @@ __attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) vst1q_p64_x2 (poly64_t * __a, poly64x2x2_t __val) { __builtin_aarch64_simd_oi __o; - __o = __builtin_aarch64_set_qregoiv2di_ssps (__o, - (poly64x2_t) __val.val[0], 0); - __o = __builtin_aarch64_set_qregoiv2di_ssps (__o, - (poly64x2_t) __val.val[1], 1); + __builtin_memcpy (&__o, &__val, sizeof (__val)); __builtin_aarch64_st1x2v2di ((__builtin_aarch64_simd_di *) __a, __o); } diff --git a/gcc/testsuite/gcc.target/aarch64/vector_structure_intrinsics.c b/gcc/testsuite/gcc.target/aarch64/vector_structure_intrinsics.c index 9dff6d8..60c53bc 100644 --- a/gcc/testsuite/gcc.target/aarch64/vector_structure_intrinsics.c +++ b/gcc/testsuite/gcc.target/aarch64/vector_structure_intrinsics.c @@ -160,6 +160,21 @@ TEST_ST1xN (vst1q, uint64x2x4_t, uint64_t*, u64, x4); TEST_ST1xN (vst1q, poly64x2x4_t, poly64_t*, p64, x4); TEST_ST1xN (vst1q, float64x2x4_t, float64_t*, f64, x4); +TEST_ST1xN (vst1q, int8x16x2_t, int8_t*, s8, x2); +TEST_ST1xN (vst1q, uint8x16x2_t, uint8_t*, u8, x2); +TEST_ST1xN (vst1q, poly8x16x2_t, poly8_t*, p8, x2); +TEST_ST1xN (vst1q, int16x8x2_t, int16_t*, s16, x2); +TEST_ST1xN (vst1q, uint16x8x2_t, uint16_t*, u16, x2); +TEST_ST1xN (vst1q, poly16x8x2_t, poly16_t*, p16, x2); +TEST_ST1xN (vst1q, float16x8x2_t, float16_t*, f16, x2); +TEST_ST1xN (vst1q, int32x4x2_t, int32_t*, s32, x2); +TEST_ST1xN (vst1q, uint32x4x2_t, uint32_t*, u32, x2); +TEST_ST1xN (vst1q, float32x4x2_t, float32_t*, f32, x2); +TEST_ST1xN (vst1q, int64x2x2_t, int64_t*, s64, x2); +TEST_ST1xN (vst1q, uint64x2x2_t, uint64_t*, u64, x2); +TEST_ST1xN (vst1q, poly64x2x2_t, poly64_t*, p64, x2); +TEST_ST1xN (vst1q, float64x2x2_t, float64_t*, f64, x2); + #define TEST_ST1x3(name, tbltype, ptrtype, ts, xn) \ void test_ ## name ## _ ## ts ## _ ## xn (ptrtype a, int8x8_t dummy, \ tbltype b) \ @@ -189,4 +204,4 @@ TEST_ST1x3 (vst1q, float64x2x3_t, float64_t*, f64, x3); /* { dg-final { scan-assembler-times "st4\\t" 14} } */ /* { dg-final { scan-assembler-times "st3\\t" 14} } */ /* { dg-final { scan-assembler-times "st2\\t" 14} } */ -/* { dg-final { scan-assembler-times "st1\\t" 28} } */ +/* { dg-final { scan-assembler-times "st1\\t" 42} } */