[PATCH] D146523: [AMDGPU]: Add new intrinsic llvm.amdgcn.convergent.copy

[Jay Foad via Phabricator via llvm-commits]

foad added a comment.

> I am making sure that I am accessing only active lanes inside a for loop.

No you're not. In your example, all of LoopHeader, Loop, ForIf, ForElse run inside the part that only enables a single lane. (Why?) But they access arbitrary lanes of %sub.i.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D146523/new/

https://reviews.llvm.org/D146523