diff options
| author | Richard Biener <rguenther@suse.de> | 2025-01-23 14:29:26 +0100 |
|---|---|---|
| committer | Richard Biener <rguenth@gcc.gnu.org> | 2025-08-21 13:53:02 +0200 |
| commit | 3182e95eda4a1d612b910b3248c997c8acc7add3 (patch) | |
| tree | 0c765a1cd2b00f4fcc6908bda478811c139c0c52 /libjava/gnu/java/net | |
| parent | 7dfb3a59fea4a9f423d62d8c604b6bd87cea6095 (diff) | |
| download | gcc-3182e95eda4a1d612b910b3248c997c8acc7add3.zip gcc-3182e95eda4a1d612b910b3248c997c8acc7add3.tar.gz gcc-3182e95eda4a1d612b910b3248c997c8acc7add3.tar.bz2 | |
tree-optimization/111494 - reduction vectorization with signed UB
The following makes sure to pun arithmetic that's used in vectorized
reduction to unsigned when overflow invokes undefined behavior.
PR tree-optimization/111494
* gimple-fold.h (arith_code_with_undefined_signed_overflow): Declare.
* gimple-fold.cc (arith_code_with_undefined_signed_overflow): Export.
* tree-vect-stmts.cc (vectorizable_operation): Use unsigned
arithmetic for operations participating in a reduction.
Diffstat (limited to 'libjava/gnu/java/net')
0 files changed, 0 insertions, 0 deletions