6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2025-06-14 14:24:05 +00:00
Code Issues Projects Releases Wiki Activity

Add a convenience method to copy wrapping, exact, and fast-math flags (NFC).

Browse Source 
The loop vectorizer preserves wrapping, exact, and fast-math properties of scalar instructions.
This patch adds a convenience method to make that operation easier because we need to do this
in the loop vectorizer, SLP vectorizer, and possibly other places.

Although this is a 'no functional change' patch, I've added a testcase to verify that the exact
flag is preserved by the loop vectorizer. The wrapping and fast-math flags are already checked
in existing testcases.

Differential Revision: http://reviews.llvm.org/D5138



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@216886 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Sanjay Patel
2014-09-01 18:44:57 +00:00
parent 88350bf61d
commit 73f8eff19e
4 changed files with 47 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
lib/Transforms/Vectorize/LoopVectorize.cpp
View File

@ -3248,19 +3248,9 @@ void InnerLoopVectorizer::vectorizeBlockInLoop(BasicBlock *BB, PhiVector *PV) {
for (unsigned Part = 0; Part < UF; ++Part) {
Value *V = Builder.CreateBinOp(BinOp->getOpcode(), A[Part], B[Part]);
// Update the NSW, NUW and Exact flags. Notice: V can be an Undef.
BinaryOperator *VecOp = dyn_cast<BinaryOperator>(V);
if (VecOp && isa<OverflowingBinaryOperator>(BinOp)) {
VecOp->setHasNoSignedWrap(BinOp->hasNoSignedWrap());
VecOp->setHasNoUnsignedWrap(BinOp->hasNoUnsignedWrap());
}
if (VecOp && isa<PossiblyExactOperator>(VecOp))
VecOp->setIsExact(BinOp->isExact());
// Copy the fast-math flags.
if (VecOp && isa<FPMathOperator>(V))
VecOp->setFastMathFlags(it->getFastMathFlags());
if (BinaryOperator *VecOp = dyn_cast<BinaryOperator>(V))
VecOp->copyFlags(BinOp);
Entry[Part] = V;
}