[PATCH] D156414: [AMDGPU] Break Large PHIs: Take whole PHI chains into account

[Pierre van Houtryve via Phabricator via llvm-commits]

Pierre-vh added a comment.

I've also been wondering if this transform should just be a separate pass that runs very late, right before ISel, to avoid any interference/unintended side effects such as the Structurizer adding branches if we split a PHI.
Then it could be generalized and made smarter to think about more than just PHIs.

Is that close to what you're thinking, or are you thinking about something else entirely?
I've also been wondering if it's possible to just do this


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D156414/new/

https://reviews.llvm.org/D156414