[PATCH] D141355: [AMDGPUUnifyDivergentExitNodes] Add NewPM support

[Jay Foad via Phabricator via llvm-commits]

foad added a comment.

The commit message should make it clear that you're also changing it to use UniformityAnalysis instead of (Legacy?)DivergenceAnalysis.



================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUUnifyDivergentExitNodes.h:57
+
+class AMDGPUUnifyDivergentExitNodesPass
+    : public PassInfoMixin<AMDGPUUnifyDivergentExitNodesPass> {
----------------
It looks like only this class needs to be declared in the `.h` file. The others could stay in the `.cpp` file.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D141355/new/

https://reviews.llvm.org/D141355