llvm-project

08dd0406 - AMDGPU: Use minnum instead of maxnum for fmed3 src2-nan fold (#139531)

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.