[v2] LoongArch: Opitmize the cost of vec_construct.
Checks
Context |
Check |
Description |
linaro-tcwg-bot/tcwg_gcc_build--master-aarch64 |
success
|
Build passed
|
linaro-tcwg-bot/tcwg_gcc_check--master-aarch64 |
success
|
Test passed
|
linaro-tcwg-bot/tcwg_gcc_build--master-arm |
success
|
Build passed
|
Commit Message
When analyzing 525 on LoongArch architecture, it was found that the
for loop of hotspot function x264_pixel_satd_8x4 could not be quantized
256-bit due to the cost of vec_construct setting. After re-adjusting
vec_construct, the performance of 525 program was improved by 16.57%.
It was found that this function can be vectorized on the aarch64 and
x86 architectures, see [PR98138].
Co-Authored-By: Deng Jianbo <dengjianbo@loongson.cn>.
gcc/ChangeLog:
* config/loongarch/loongarch.cc
(loongarch_builtin_vectorization_cost): Modify the
construction cost of the vec_construct vector.
gcc/testsuite/ChangeLog:
* gcc.target/loongarch/vect-slp-two-operator.c: New test.
---
gcc/config/loongarch/loongarch.cc | 6 +--
.../loongarch/vect-slp-two-operator.c | 38 +++++++++++++++++++
2 files changed, 41 insertions(+), 3 deletions(-)
create mode 100644 gcc/testsuite/gcc.target/loongarch/vect-slp-two-operator.c
Comments
Pushed to r15-6755.
在 2025/1/7 下午9:04, chenxiaolong 写道:
> When analyzing 525 on LoongArch architecture, it was found that the
> for loop of hotspot function x264_pixel_satd_8x4 could not be quantized
> 256-bit due to the cost of vec_construct setting. After re-adjusting
> vec_construct, the performance of 525 program was improved by 16.57%.
> It was found that this function can be vectorized on the aarch64 and
> x86 architectures, see [PR98138].
>
> Co-Authored-By: Deng Jianbo <dengjianbo@loongson.cn>.
>
> gcc/ChangeLog:
>
> * config/loongarch/loongarch.cc
> (loongarch_builtin_vectorization_cost): Modify the
> construction cost of the vec_construct vector.
>
> gcc/testsuite/ChangeLog:
>
> * gcc.target/loongarch/vect-slp-two-operator.c: New test.
> ---
> gcc/config/loongarch/loongarch.cc | 6 +--
> .../loongarch/vect-slp-two-operator.c | 38 +++++++++++++++++++
> 2 files changed, 41 insertions(+), 3 deletions(-)
> create mode 100644 gcc/testsuite/gcc.target/loongarch/vect-slp-two-operator.c
>
> diff --git a/gcc/config/loongarch/loongarch.cc b/gcc/config/loongarch/loongarch.cc
> index 89237c377e7..ff27b96c31e 100644
> --- a/gcc/config/loongarch/loongarch.cc
> +++ b/gcc/config/loongarch/loongarch.cc
> @@ -4127,10 +4127,10 @@ loongarch_builtin_vectorization_cost (enum vect_cost_for_stmt type_of_cost,
>
> case vec_construct:
> elements = TYPE_VECTOR_SUBPARTS (vectype);
> - if (ISA_HAS_LASX)
> - return elements + 1;
> + if (LASX_SUPPORTED_MODE_P (mode) && !LSX_SUPPORTED_MODE_P (mode))
> + return elements / 2 + 3;
> else
> - return elements;
> + return elements / 2 + 1;
>
> default:
> gcc_unreachable ();
> diff --git a/gcc/testsuite/gcc.target/loongarch/vect-slp-two-operator.c b/gcc/testsuite/gcc.target/loongarch/vect-slp-two-operator.c
> new file mode 100644
> index 00000000000..43b46759902
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/loongarch/vect-slp-two-operator.c
> @@ -0,0 +1,38 @@
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -mlasx -ftree-vectorize -fdump-tree-vect -fdump-tree-vect-details" } */
> +
> +typedef unsigned char uint8_t;
> +typedef unsigned int uint32_t;
> +
> +#define HADAMARD4(d0, d1, d2, d3, s0, s1, s2, s3) \
> + { \
> + int t0 = s0 + s1; \
> + int t1 = s0 - s1; \
> + int t2 = s2 + s3; \
> + int t3 = s2 - s3; \
> + d0 = t0 + t2; \
> + d1 = t1 + t3; \
> + d2 = t0 - t2; \
> + d3 = t1 - t3; \
> + }
> +
> +void sink (uint32_t tmp[4][4]);
> +
> +void
> +x264_pixel_satd_8x4 (uint8_t *pix1, int i_pix1, uint8_t *pix2, int i_pix2)
> +{
> + uint32_t tmp[4][4];
> + int sum = 0;
> + for (int i = 0; i < 4; i++, pix1 += i_pix1, pix2 += i_pix2)
> + {
> + uint32_t a0 = (pix1[0] - pix2[0]) + ((pix1[4] - pix2[4]) << 16);
> + uint32_t a1 = (pix1[1] - pix2[1]) + ((pix1[5] - pix2[5]) << 16);
> + uint32_t a2 = (pix1[2] - pix2[2]) + ((pix1[6] - pix2[6]) << 16);
> + uint32_t a3 = (pix1[3] - pix2[3]) + ((pix1[7] - pix2[7]) << 16);
> + HADAMARD4 (tmp[i][0], tmp[i][1], tmp[i][2], tmp[i][3], a0, a1, a2, a3);
> + }
> + sink (tmp);
> +}
> +
> +/* { dg-final { scan-tree-dump "vectorizing stmts using SLP" "vect" } } */
> +/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" } } */
@@ -4127,10 +4127,10 @@ loongarch_builtin_vectorization_cost (enum vect_cost_for_stmt type_of_cost,
case vec_construct:
elements = TYPE_VECTOR_SUBPARTS (vectype);
- if (ISA_HAS_LASX)
- return elements + 1;
+ if (LASX_SUPPORTED_MODE_P (mode) && !LSX_SUPPORTED_MODE_P (mode))
+ return elements / 2 + 3;
else
- return elements;
+ return elements / 2 + 1;
default:
gcc_unreachable ();
new file mode 100644
@@ -0,0 +1,38 @@
+/* { dg-do compile } */
+/* { dg-options "-O2 -mlasx -ftree-vectorize -fdump-tree-vect -fdump-tree-vect-details" } */
+
+typedef unsigned char uint8_t;
+typedef unsigned int uint32_t;
+
+#define HADAMARD4(d0, d1, d2, d3, s0, s1, s2, s3) \
+ { \
+ int t0 = s0 + s1; \
+ int t1 = s0 - s1; \
+ int t2 = s2 + s3; \
+ int t3 = s2 - s3; \
+ d0 = t0 + t2; \
+ d1 = t1 + t3; \
+ d2 = t0 - t2; \
+ d3 = t1 - t3; \
+ }
+
+void sink (uint32_t tmp[4][4]);
+
+void
+x264_pixel_satd_8x4 (uint8_t *pix1, int i_pix1, uint8_t *pix2, int i_pix2)
+{
+ uint32_t tmp[4][4];
+ int sum = 0;
+ for (int i = 0; i < 4; i++, pix1 += i_pix1, pix2 += i_pix2)
+ {
+ uint32_t a0 = (pix1[0] - pix2[0]) + ((pix1[4] - pix2[4]) << 16);
+ uint32_t a1 = (pix1[1] - pix2[1]) + ((pix1[5] - pix2[5]) << 16);
+ uint32_t a2 = (pix1[2] - pix2[2]) + ((pix1[6] - pix2[6]) << 16);
+ uint32_t a3 = (pix1[3] - pix2[3]) + ((pix1[7] - pix2[7]) << 16);
+ HADAMARD4 (tmp[i][0], tmp[i][1], tmp[i][2], tmp[i][3], a0, a1, a2, a3);
+ }
+ sink (tmp);
+}
+
+/* { dg-final { scan-tree-dump "vectorizing stmts using SLP" "vect" } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" } } */