6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2026-04-24 13:18:17 +00:00
Code Issues Projects Releases Wiki Activity

[Reassociate] Canonicalize operands of vector binary operators.

Browse Source 
Prior to this commit fmul and fadd binary operators were being canonicalized for
both scalar and vector versions.  We now canonicalize add, mul, and, or, and xor
vector instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@222006 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Chad Rosier
2014-11-14 17:08:15 +00:00
parent 7e61b1fb62
commit 1298a29c64
2 changed files with 51 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files
lib/Transforms/Scalar/Reassociate.cpp
+1 -6
View File
@@ -2083,14 +2083,9 @@ void Reassociate::OptimizeInst(Instruction *I) {
// writing other transformations simpler.
if (I->getType()->isFloatingPointTy() || I->getType()->isVectorTy()) {
// FAdd and FMul can be commuted.
unsigned Opcode = I->getOpcode();
if (Opcode == Instruction::FMul || Opcode == Instruction::FAdd)
if (I->isCommutative())
canonicalizeOperands(I);
// FIXME: We should commute vector instructions as well. However, this
// requires further analysis to determine the effect on later passes.
// Don't try to optimize vector instructions or anything that doesn't have
// unsafe algebra.
if (I->getType()->isVectorTy() || !I->hasUnsafeAlgebra())