[PATCH] D142990: [AMDGPU][NFC] More precise predicates on GFX9 f16 insts

[Jay Foad via Phabricator via llvm-commits]

foad accepted this revision.
foad added a comment.
This revision is now accepted and ready to land.

LGTM.



================
Comment at: llvm/lib/Target/AMDGPU/VOP3Instructions.td:306
 let FPDPRounding = 1 in {
   let Predicates = [Has16BitInsts, isGFX8Only] in {
     defm V_DIV_FIXUP_F16 : VOP3Inst <"v_div_fixup_f16", VOP3_Profile<VOP_F16_F16_F16_F16>, AMDGPUdiv_fixup>;
----------------
Joe_Nash wrote:
> foad wrote:
> > Can you make the same change here?
> No. We explicitly want to override predicates on the pseudo instruction but not the real, so assembly of V_DIV_FIXUP_LEGACY_F16_gfx9 works. SubtargetPredicate would be copied to the real.
OK. I still think this can be cleaned up a little, but it needn't hold up your patch.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142990/new/

https://reviews.llvm.org/D142990