[PATCH] D117075: [AMDGPU][WQM] Set the lanes that should be deactivated with VCC.

[Carl Ritson via Phabricator via llvm-commits]

critson added a comment.

You can potentially rebase this change on top of D117455 <https://reviews.llvm.org/D117455>.
I will submit that when it has had enough time for all reviewers to respond.



================
Comment at: llvm/test/CodeGen/AMDGPU/wqm.ll:2562
+; Check if the correct register is selected. WQM pass incorrectly uses VCC for vector comparisons in Wave32 mode. Make register selection wave size dependent.
+define amdgpu_ps void @test_for_deactivating_lanes_in_wave32(float addrspace(6)* inreg noalias align 4 dereferenceable(18446744073709551615) %0) #8 {
+; CHECK-WAVE32-LABEL: test_for_deactivating_lanes_in_wave32:
----------------
This can be reduced to:
```define amdgpu_ps void @test_for_deactivating_lanes_in_wave32(float addrspace(6)* inreg %0)```

and the associated attribute (8) removed without effecting code generation.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D117075/new/

https://reviews.llvm.org/D117075