From 606660f1f3aade24be4946c520af4c4b8230100b Mon Sep 17 00:00:00 2001 From: Yi Jiang Date: Mon, 5 May 2014 17:59:14 +0000 Subject: [PATCH] Always set alignment of vectorized LD/ST in SLP-Vectorizer. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207983 91177308-0d34-0410-b5e6-96231b3b80d8 --- lib/Transforms/Vectorize/SLPVectorizer.cpp | 4 ++++ test/Transforms/SLPVectorizer/X86/align.ll | 27 ++++++++++++++++++++++ 2 files changed, 31 insertions(+) create mode 100644 test/Transforms/SLPVectorizer/X86/align.ll diff --git a/lib/Transforms/Vectorize/SLPVectorizer.cpp b/lib/Transforms/Vectorize/SLPVectorizer.cpp index 895826f1d87..0eb045e628b 100644 --- a/lib/Transforms/Vectorize/SLPVectorizer.cpp +++ b/lib/Transforms/Vectorize/SLPVectorizer.cpp @@ -1622,6 +1622,8 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { VecTy->getPointerTo(AS)); unsigned Alignment = LI->getAlignment(); LI = Builder.CreateLoad(VecPtr); + if (!Alignment) + Alignment = DL->getABITypeAlignment(LI->getPointerOperand()->getType()); LI->setAlignment(Alignment); E->VectorizedValue = LI; return propagateMetadata(LI, E->Scalars); @@ -1641,6 +1643,8 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { Value *VecPtr = Builder.CreateBitCast(SI->getPointerOperand(), VecTy->getPointerTo(AS)); StoreInst *S = Builder.CreateStore(VecValue, VecPtr); + if (!Alignment) + Alignment = DL->getABITypeAlignment(SI->getPointerOperand()->getType()); S->setAlignment(Alignment); E->VectorizedValue = S; return propagateMetadata(S, E->Scalars); diff --git a/test/Transforms/SLPVectorizer/X86/align.ll b/test/Transforms/SLPVectorizer/X86/align.ll new file mode 100644 index 00000000000..f5865733ccb --- /dev/null +++ b/test/Transforms/SLPVectorizer/X86/align.ll @@ -0,0 +1,27 @@ +; RUN: opt < %s -basicaa -slp-vectorizer -S -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx | FileCheck %s + +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-apple-macosx10.8.0" + +; Simple 3-pair chain with loads and stores +; CHECK: test1 +define void @test1(double* %a, double* %b, double* %c) { +entry: + %agg.tmp.i.i.sroa.0 = alloca [3 x double], align 16 +; CHECK: %[[V0:[0-9]+]] = load <2 x double>* %[[V2:[0-9]+]], align 8 + %i0 = load double* %a + %i1 = load double* %b + %mul = fmul double %i0, %i1 + %store1 = getelementptr inbounds [3 x double]* %agg.tmp.i.i.sroa.0, i64 0, i64 1 + %store2 = getelementptr inbounds [3 x double]* %agg.tmp.i.i.sroa.0, i64 0, i64 2 + %arrayidx3 = getelementptr inbounds double* %a, i64 1 + %i3 = load double* %arrayidx3, align 8 + %arrayidx4 = getelementptr inbounds double* %b, i64 1 + %i4 = load double* %arrayidx4, align 8 + %mul5 = fmul double %i3, %i4 +; CHECK: store <2 x double> %[[V1:[0-9]+]], <2 x double>* %[[V2:[0-9]+]], align 8 + store double %mul, double* %store1 + store double %mul5, double* %store2, align 16 +; CHECK: ret + ret void +}