X-Git-Url: http://plrg.eecs.uci.edu/git/?a=blobdiff_plain;f=test%2FTransforms%2FSLPVectorizer%2FX86%2Fdiamond.ll;h=5135a92a7bdbd0fac5535980a6ba9cac1d7b72b7;hb=39f4e8d9cce22b60a3417a5f17c847fa5b1daebf;hp=8e85cb6c9b8f84da9db5741218befc344bd67c2a;hpb=1f098af3648f7714dd0501f8ba97601e99471806;p=oota-llvm.git

diff --git a/test/Transforms/SLPVectorizer/X86/diamond.ll b/test/Transforms/SLPVectorizer/X86/diamond.ll
index 8e85cb6c9b8..5135a92a7bd 100644
--- a/test/Transforms/SLPVectorizer/X86/diamond.ll
+++ b/test/Transforms/SLPVectorizer/X86/diamond.ll
@@ -1,4 +1,4 @@
-; RUN: opt < %s -basicaa -slp-vectorizer -dce -S -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx | FileCheck %s
+; 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"
@@ -11,7 +11,7 @@ target triple = "x86_64-apple-macosx10.8.0"
 ;   return 0;
 ; }
 
-; CHECK: @foo
+; CHECK-LABEL: @foo(
 ; CHECK: load <4 x i32>
 ; CHECK: mul <4 x i32>
 ; CHECK: store <4 x i32>
@@ -41,7 +41,7 @@ entry:
 }
 
 
-; int foo_fail(int * restrict B,  int * restrict A, int n, int m) {
+; int extr_user(int * restrict B,  int * restrict A, int n, int m) {
 ;   B[0] = n * A[0] + m * A[0];
 ;   B[1] = n * A[1] + m * A[1];
 ;   B[2] = n * A[2] + m * A[2];
@@ -49,10 +49,12 @@ entry:
 ;   return A[0];
 ; }
 
-; CHECK: @foo_fail
-; CHECK-NOT: load <4 x i32>
-; CHECK: ret
-define i32 @foo_fail(i32* noalias nocapture %B, i32* noalias nocapture %A, i32 %n, i32 %m) {
+; CHECK-LABEL: @extr_user(
+; CHECK: load <4 x i32>
+; CHECK: store <4 x i32>
+; CHECK: extractelement <4 x i32>
+; CHECK-NEXT: ret
+define i32 @extr_user(i32* noalias nocapture %B, i32* noalias nocapture %A, i32 %n, i32 %m) {
 entry:
   %0 = load i32* %A, align 4
   %mul238 = add i32 %m, %n
@@ -73,6 +75,35 @@ entry:
   %add20 = mul i32 %3, %mul238
   %arrayidx21 = getelementptr inbounds i32* %B, i64 3
   store i32 %add20, i32* %arrayidx21, align 4
-  ret i32 %0  ;<--------- This value has multiple users and can't be vectorized.
+  ret i32 %0  ;<--------- This value has multiple users
 }
 
+; In this example we have an external user that is not the first element in the vector.
+; CHECK-LABEL: @extr_user1(
+; CHECK: load <4 x i32>
+; CHECK: store <4 x i32>
+; CHECK: extractelement <4 x i32>
+; CHECK-NEXT: ret
+define i32 @extr_user1(i32* noalias nocapture %B, i32* noalias nocapture %A, i32 %n, i32 %m) {
+entry:
+  %0 = load i32* %A, align 4
+  %mul238 = add i32 %m, %n
+  %add = mul i32 %0, %mul238
+  store i32 %add, i32* %B, align 4
+  %arrayidx4 = getelementptr inbounds i32* %A, i64 1
+  %1 = load i32* %arrayidx4, align 4
+  %add8 = mul i32 %1, %mul238
+  %arrayidx9 = getelementptr inbounds i32* %B, i64 1
+  store i32 %add8, i32* %arrayidx9, align 4
+  %arrayidx10 = getelementptr inbounds i32* %A, i64 2
+  %2 = load i32* %arrayidx10, align 4
+  %add14 = mul i32 %2, %mul238
+  %arrayidx15 = getelementptr inbounds i32* %B, i64 2
+  store i32 %add14, i32* %arrayidx15, align 4
+  %arrayidx16 = getelementptr inbounds i32* %A, i64 3
+  %3 = load i32* %arrayidx16, align 4
+  %add20 = mul i32 %3, %mul238
+  %arrayidx21 = getelementptr inbounds i32* %B, i64 3
+  store i32 %add20, i32* %arrayidx21, align 4
+  ret i32 %1  ;<--------- This value has multiple users
+}