[PATCH] D119696: [AMDGPU] Improve v_cmpx usage on GFX10.3.
Jay Foad via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Fri Feb 25 03:46:23 PST 2022
foad added inline comments.
================
Comment at: llvm/lib/Target/AMDGPU/SIOptimizeExecMasking.cpp:344-345
+
+ // Try to find the last v_cmp instruction that defs the saveexec input
+ // operand.
+ VCmp = findInstrBackwards(
----------------
The code just finds the last v_cmp instruction. Then you hope that it defs the saveexec input operand? Do you ever actually check that?
================
Comment at: llvm/lib/Target/AMDGPU/SIOptimizeExecMasking.cpp:374
+
+ if (!Src0 || !Src1)
+ return nullptr;
----------------
I don't think you need this check (all vcmp instructions have src0 and src1 operands, don't they?).
================
Comment at: llvm/lib/Target/AMDGPU/SIOptimizeExecMasking.cpp:377
+
+ SmallVector<MCRegister, 3> NonDefRegs{Exec};
+ if (Src0->isReg())
----------------
You could have checked for Exec in the first call to findInstrBackwards, just to bail out a bit earlier.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D119696/new/
https://reviews.llvm.org/D119696
More information about the llvm-commits
mailing list