[AMDGPU] Restrict packed math FP32 instructions to read only one SGPR per operand on gfx12+ #100783
premerge.yaml
on: pull_request
Build and Test Linux
0s
Build and Test Windows
MacOS Premerge Checks
0s