diff options
| author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2025-05-12 20:26:29 +0200 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-05-12 20:26:29 +0200 |
| commit | 08dd0406c6c8b5418cd05528057bc64f6fa1d52f (patch) | |
| tree | 4ddc59f81e16c1a8a4268b3d7430cd2bc09b6141 /clang | |
| parent | bfd2ef765957307d8a02f72cf6d67de5d3977f10 (diff) | |
| download | llvm-08dd0406c6c8b5418cd05528057bc64f6fa1d52f.zip llvm-08dd0406c6c8b5418cd05528057bc64f6fa1d52f.tar.gz llvm-08dd0406c6c8b5418cd05528057bc64f6fa1d52f.tar.bz2 | |
AMDGPU: Use minnum instead of maxnum for fmed3 src2-nan fold (#139531)
By the pseudocode in the ISA manual, if any input is a nan it acts
like min3, which will fold to min2 of the other operands. The other
cases fold to min, I'm not sure how this one was wrong.
Diffstat (limited to 'clang')
0 files changed, 0 insertions, 0 deletions