[PATCH] D97560: [AMDGPU] Add selection pattern for v_xnor_b32

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added a comment.

In D97560#2590473 <https://reviews.llvm.org/D97560#2590473>, @foad wrote:

> Is it worth disabling the pattern for SelectionDAG? Or maybe restricting it to divergent cases, so we can select v_xnor directly in that case without having to move it to VALU in SIFixSGPRCopies?

Not really. We don't have a mechanism for disabling patterns per selector. The current patterns are already a hodge-podge of cases that happen to work in one selector or the other, and I don't think it's worth putting effort into making selectiondag nicer


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D97560/new/

https://reviews.llvm.org/D97560