[PATCH] D136592: [AMDGPU][GISel] Select llvm.amdgcn.fcmp intrinsics

[Jay Foad via Phabricator via llvm-commits]

foad added a comment.

> Add a RegBankCombine to transform the TRUE/FALSE FP CCs into simple constants for less complicated codegen.

Isn't this already done at the IR level in GCNTTIImpl::instCombineIntrinsic?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D136592/new/

https://reviews.llvm.org/D136592