[PATCH] D57059: [SLP] Initial support for the vectorization of the non-power-of-2 vectors.
Simon Pilgrim via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Tue Sep 22 10:58:44 PDT 2020
RKSimon added a comment.
some very minor style comments - a general comment would be to try and pre-commit the style/NFC refactor/cleanup changes so the size of this patch is smaller
================
Comment at: llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp:2748
+ for (Value *V : InstructionsOnly)
+ for (unsigned I = 0, E = PH->getNumIncomingValues(); I < E; ++I) {
+ auto *Term =
----------------
Do these trivial style refactors separately now to reduce the size of the patch?
================
Comment at: llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp:2768
SmallVector<ValueList, 2> OperandsVec;
- for (unsigned i = 0, e = PH->getNumIncomingValues(); i < e; ++i) {
+ for (unsigned I = 0, E = PH->getNumIncomingValues(); I < E; ++I) {
ValueList Operands;
----------------
Do these trivial style refactors separately now to reduce the size of the patch?
================
Comment at: llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp:3579
+ Instruction::ExtractElement,
+ cast<ExtractElementInst>(V)->getVectorOperandType(),
+ IO->getZExtValue());
----------------
duplicate cast
================
Comment at: llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp:3610
+ Instruction::ExtractElement,
+ cast<ExtractElementInst>(VL[I])->getVectorOperandType(), Idx);
} else {
----------------
duplicate cast
================
Comment at: llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp:3625
+ Instruction::ExtractElement,
+ cast<ExtractElementInst>(V)->getVectorOperandType(), Idx);
} else {
----------------
duplicate cast
================
Comment at: llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp:3680
+ Instruction::ExtractElement,
+ cast<ExtractElementInst>(EI)->getVectorOperandType(), Idx);
+ } else {
----------------
duplicate cast
================
Comment at: llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp:4254
APInt DemandedElts = APInt::getNullValue(NumElts);
- for (unsigned i = 0; i < NumElts; ++i)
- if (!ShuffledIndices.count(i))
- DemandedElts.setBit(i);
+ for (unsigned I = 0; I < NumElts; ++I) {
+ if (IgnoredIndices.test(I))
----------------
trivial style refactor - pull out of patch?
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D57059/new/
https://reviews.llvm.org/D57059
More information about the llvm-commits
mailing list