Matt Arsenault 37c341df28 Revert "AMDGPU: Don't canonicalize fminnum/fmaxnum if targets support IEEE fminimum(maximum)_num (#127711)" ...

This reverts commit 36eaf0daf5.

This is not a sound approach to dealing with this instruction change.
The new behavior is a different opcode pair, not a modifier on the
existing opcode.

2025-02-20 10:19:14 +07:00

486 KiB

Raw Blame History

View Raw