diff options
| author | Juzhe-Zhong <juzhe.zhong@rivai.ai> | 2023-12-20 14:55:26 +0800 |
|---|---|---|
| committer | Pan Li <pan2.li@intel.com> | 2023-12-21 07:06:12 +0800 |
| commit | 008b80e42eb7cb55c6a2ef55728241b8733dfd17 (patch) | |
| tree | bb3a4cf1b625b83cd8ff92b4a966a0f44ca47772 /gcc/c-family | |
| parent | d82bb518fa372cc30cc3352e0a124d0bd6deb36f (diff) | |
| download | gcc-008b80e42eb7cb55c6a2ef55728241b8733dfd17.zip gcc-008b80e42eb7cb55c6a2ef55728241b8733dfd17.tar.gz gcc-008b80e42eb7cb55c6a2ef55728241b8733dfd17.tar.bz2 | |
RISC-V: Optimize SELECT_VL codegen when length is known as smaller than VF
While trying to fix bugs of PR113097, notice this following situation we
generate redundant vsetvli
_255 = SELECT_VL (3, POLY_INT_CST [4, 4]);
COND_LEN (..., _255)
Before this patch:
vsetivli a5, 3...
...
vadd.vv (use a5)
After this patch:
...
vadd.vv (use AVL = 3)
The reason we can do this is because known_ge (3, [4,4]) is true.
It's safe to apply such optimization
Tested on both RV32 and RV64 full coverage testing, no regression.
PR target/113087
gcc/ChangeLog:
* config/riscv/riscv-v.cc (expand_select_vl): Optimize SELECT_VL.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/autovec/pr113087-2.c: New test.
Diffstat (limited to 'gcc/c-family')
0 files changed, 0 insertions, 0 deletions