[PATCH] D135609: [AMDGPU] Fix True16 patterns for cmp on GFX11
Joe Nash via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Tue Oct 11 07:21:07 PDT 2022
Joe_Nash added inline comments.
================
Comment at: llvm/test/CodeGen/AMDGPU/v_cmp_gfx11.ll:7
+; CHECK: ; %bb.0: ; %entry
+; CHECK-NEXT: v_cmp_eq_u16_e64 s0, 0, 0
+; CHECK-NEXT: v_mov_b32_e32 v1, 0
----------------
Leonc wrote:
> Shouldn't this be `v_cmp_eq_u16_t16_e64` on gfx11?
The assembly mnemonics do not include references to true16, only the MachineInstr/MCInst names do. I checked that the True16 version of the instruction was used via print-after-all before asm emission in this case.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D135609/new/
https://reviews.llvm.org/D135609
More information about the llvm-commits
mailing list