1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py 2; RUN: opt < %s -basic-aa -slp-vectorizer -dce -S -mtriple=i386-apple-macosx10.8.0 -mcpu=corei7-avx | FileCheck %s 3 4target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128-n8:16:32-S128" 5target triple = "i386-apple-macosx10.9.0" 6 7;int foo(int *A, int n, int k) { 8; for (int i=0; i < 10000; i+=4) { 9; A[i] += n; 10; A[i+1] += k; 11; A[i+2] += n; 12; A[i+3] += k; 13; } 14;} 15 16define i32 @foo(i32* nocapture %A, i32 %n, i32 %k) { 17; CHECK-LABEL: @foo( 18; CHECK-NEXT: entry: 19; CHECK-NEXT: [[TMP0:%.*]] = insertelement <2 x i32> undef, i32 [[N:%.*]], i32 0 20; CHECK-NEXT: [[TMP1:%.*]] = insertelement <2 x i32> [[TMP0]], i32 [[K:%.*]], i32 1 21; CHECK-NEXT: [[SHUFFLE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> undef, <4 x i32> <i32 0, i32 1, i32 0, i32 1> 22; CHECK-NEXT: br label [[FOR_BODY:%.*]] 23; CHECK: for.body: 24; CHECK-NEXT: [[I_024:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[ADD10:%.*]], [[FOR_BODY]] ] 25; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i32, i32* [[A:%.*]], i32 [[I_024]] 26; CHECK-NEXT: [[TMP2:%.*]] = bitcast i32* [[ARRAYIDX]] to <4 x i32>* 27; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i32>, <4 x i32>* [[TMP2]], align 4 28; CHECK-NEXT: [[TMP4:%.*]] = add nsw <4 x i32> [[TMP3]], [[SHUFFLE]] 29; CHECK-NEXT: [[TMP5:%.*]] = bitcast i32* [[ARRAYIDX]] to <4 x i32>* 30; CHECK-NEXT: store <4 x i32> [[TMP4]], <4 x i32>* [[TMP5]], align 4 31; CHECK-NEXT: [[ADD10]] = add nsw i32 [[I_024]], 4 32; CHECK-NEXT: [[CMP:%.*]] = icmp slt i32 [[ADD10]], 10000 33; CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_END:%.*]] 34; CHECK: for.end: 35; CHECK-NEXT: ret i32 undef 36; 37entry: 38 br label %for.body 39 40for.body: ; preds = %entry, %for.body 41 %i.024 = phi i32 [ 0, %entry ], [ %add10, %for.body ] 42 %arrayidx = getelementptr inbounds i32, i32* %A, i32 %i.024 43 %0 = load i32, i32* %arrayidx, align 4 44 %add = add nsw i32 %0, %n 45 store i32 %add, i32* %arrayidx, align 4 46 %add121 = or i32 %i.024, 1 47 %arrayidx2 = getelementptr inbounds i32, i32* %A, i32 %add121 48 %1 = load i32, i32* %arrayidx2, align 4 49 %add3 = add nsw i32 %1, %k 50 store i32 %add3, i32* %arrayidx2, align 4 51 %add422 = or i32 %i.024, 2 52 %arrayidx5 = getelementptr inbounds i32, i32* %A, i32 %add422 53 %2 = load i32, i32* %arrayidx5, align 4 54 %add6 = add nsw i32 %2, %n 55 store i32 %add6, i32* %arrayidx5, align 4 56 %add723 = or i32 %i.024, 3 57 %arrayidx8 = getelementptr inbounds i32, i32* %A, i32 %add723 58 %3 = load i32, i32* %arrayidx8, align 4 59 %add9 = add nsw i32 %3, %k 60 store i32 %add9, i32* %arrayidx8, align 4 61 %add10 = add nsw i32 %i.024, 4 62 %cmp = icmp slt i32 %add10, 10000 63 br i1 %cmp, label %for.body, label %for.end 64 65for.end: ; preds = %for.body 66 ret i32 undef 67} 68 69