This way, they can be detected later, e.g. by the
SIOptimizeVGPRLiveRange pass.
Details
Details
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
llvm/lib/Target/AMDGPU/SILowerControlFlow.cpp | ||
---|---|---|
425 | Can you use MachineInstr::setDesc to mutate MI into an S_CBRANCH_EXECNZ, instead of creating a new instruction and erasing the old one? | |
llvm/lib/Target/AMDGPU/SILowerI1Copies.cpp | ||
174 ↗ | (On Diff #356637) | Is this a change in behaviour? Previously we used an S_CBRANCH_EXECNZ which would not set Divergent here. |
Can you use MachineInstr::setDesc to mutate MI into an S_CBRANCH_EXECNZ, instead of creating a new instruction and erasing the old one?