diff options
| author | Juzhe-Zhong <juzhe.zhong@rivai.ai> | 2023-09-26 17:58:47 +0800 |
|---|---|---|
| committer | Pan Li <pan2.li@intel.com> | 2023-09-26 20:20:58 +0800 |
| commit | 277456180ea4d002675a9498a7b1ec9d483ded73 (patch) | |
| tree | 152bfc72f441940adbb45bb18c840fd6e5526973 /gcc | |
| parent | dd0197fb4cdee8cd5f78fea9a965c96d7ca47229 (diff) | |
| download | gcc-277456180ea4d002675a9498a7b1ec9d483ded73.zip gcc-277456180ea4d002675a9498a7b1ec9d483ded73.tar.gz gcc-277456180ea4d002675a9498a7b1ec9d483ded73.tar.bz2 | |
MATCH: Optimize COND_ADD reduction pattern
Current COND_ADD reduction pattern can't optimize floating-point vector.
As Richard suggested: https://gcc.gnu.org/pipermail/gcc-patches/2023-September/631336.html
Allow COND_ADD reduction pattern to optimize floating-point vector.
Bootstrap and Regression is running.
Ok for trunk if tests pass ?
gcc/ChangeLog:
* match.pd: Optimize COND_ADD reduction pattern.
Diffstat (limited to 'gcc')
| -rw-r--r-- | gcc/match.pd | 7 |
1 files changed, 5 insertions, 2 deletions
diff --git a/gcc/match.pd b/gcc/match.pd index 3ce90c3..790d956 100644 --- a/gcc/match.pd +++ b/gcc/match.pd @@ -8863,8 +8863,11 @@ and, c = mask1 && mask2 ? d + b : d. */ (simplify - (IFN_COND_ADD @0 @1 (vec_cond @2 @3 integer_zerop) @1) - (IFN_COND_ADD (bit_and @0 @2) @1 @3 @1)) + (IFN_COND_ADD @0 @1 (vec_cond @2 @3 zerop@4) @1) + (if (ANY_INTEGRAL_TYPE_P (type) + || (FLOAT_TYPE_P (type) + && fold_real_zero_addition_p (type, NULL_TREE, @4, 0))) + (IFN_COND_ADD (bit_and @0 @2) @1 @3 @1))) /* Detect simplication for a conditional length reduction where |