Fix inserting new elements in a specified location.

[oota-llvm.git] / test / Transforms / ScalarRepl / vector_promote.ll

diff --git a/test/Transforms/ScalarRepl/vector_promote.ll b/test/Transforms/ScalarRepl/vector_promote.ll
index 67bbf71b362d7ca28f03e9d93a693e4b5f6b944c..8ca1ed5005b33cb8716f47d7ed7e41b65b14292d 100644 (file)
--- a/test/Transforms/ScalarRepl/vector_promote.ll
+++ b/test/Transforms/ScalarRepl/vector_promote.ll
@@ -14,7 +14,7 @@ entry:
        %tmp6 = fadd <4 x float> %tmp4, %tmp4           ; <<4 x float>> [#uses=1]
        store <4 x float> %tmp6, <4 x float>* %F
        ret void
-; CHECK: @test1
+; CHECK-LABEL: @test1(
 ; CHECK-NOT: alloca
 ; CHECK: %tmp = load <4 x float>* %F
 ; CHECK: fadd <4 x float> %tmp, %tmp
@@ -33,7 +33,7 @@ entry:
        %tmp6 = fadd <4 x float> %tmp4, %tmp4           ; <<4 x float>> [#uses=1]
        store <4 x float> %tmp6, <4 x float>* %F
        ret void
-; CHECK: @test2
+; CHECK-LABEL: @test2(
 ; CHECK-NOT: alloca
 ; CHECK: %tmp = load <4 x float>* %F
 ; CHECK: fadd <4 x float> %tmp, %tmp
@@ -50,7 +50,7 @@ entry:
        %tmp.upgrd.4 = load float* %tmp.upgrd.3         ; <float> [#uses=1]
        store float %tmp.upgrd.4, float* %f
        ret void
-; CHECK: @test3
+; CHECK-LABEL: @test3(
 ; CHECK-NOT: alloca
 ; CHECK: %tmp = load <4 x float>* %F
 ; CHECK: fadd <4 x float> %tmp, %tmp
@@ -67,7 +67,7 @@ entry:
        %tmp.upgrd.6 = load float* %G.upgrd.5           ; <float> [#uses=1]
        store float %tmp.upgrd.6, float* %f
        ret void
-; CHECK: @test4
+; CHECK-LABEL: @test4(
 ; CHECK-NOT: alloca
 ; CHECK: %tmp = load <4 x float>* %F
 ; CHECK: fadd <4 x float> %tmp, %tmp
@@ -81,109 +81,57 @@ define i32 @test5(float %X) {  ;; should turn into bitcast.
        %a = bitcast float* %X1 to i32*
        %tmp = load i32* %a
        ret i32 %tmp
-; CHECK: @test5
+; CHECK-LABEL: @test5(
 ; CHECK-NEXT: bitcast float %X to i32
 ; CHECK-NEXT: ret i32
 }
 
-
 define i64 @test6(<2 x float> %X) {
        %X_addr = alloca <2 x float>
         store <2 x float> %X, <2 x float>* %X_addr
        %P = bitcast <2 x float>* %X_addr to i64*
        %tmp = load i64* %P
        ret i64 %tmp
-; CHECK: @test6
-; CHECK: bitcast <2 x float> %X to <1 x i64>
+; CHECK-LABEL: @test6(
+; CHECK: bitcast <2 x float> %X to i64
 ; CHECK: ret i64
 }
 
-define float @test7(<4 x float> %x) {
-       %a = alloca <4 x float>
-       store <4 x float> %x, <4 x float>* %a
-       %p = bitcast <4 x float>* %a to <2 x float>*
-       %b = load <2 x float>* %p
-       %q = getelementptr <4 x float>* %a, i32 0, i32 2
-       %c = load float* %q
-       ret float %c
-; CHECK: @test7
-; CHECK-NOT: alloca
-; CHECK: bitcast <4 x float> %x to <2 x double>
-; CHECK-NEXT: extractelement <2 x double>
-; CHECK-NEXT: bitcast double %tmp4 to <2 x float>
-; CHECK-NEXT: extractelement <4 x float>
-}
+%struct.test7 = type { [6 x i32] }
 
-define void @test8(<4 x float> %x, <2 x float> %y) {
-       %a = alloca <4 x float>
-       store <4 x float> %x, <4 x float>* %a
-       %p = bitcast <4 x float>* %a to <2 x float>*
-       store <2 x float> %y, <2 x float>* %p
-       ret void
-; CHECK: @test8
-; CHECK-NOT: alloca
-; CHECK: bitcast <4 x float> %x to <2 x double>
-; CHECK-NEXT: bitcast <2 x float> %y to double
-; CHECK-NEXT: insertelement <2 x double>
-; CHECK-NEXT: bitcast <2 x double> %tmp2 to <4 x float>
-}
-
-define i256 @test9(<4 x i256> %x) {
-       %a = alloca <4 x i256>
-       store <4 x i256> %x, <4 x i256>* %a
-       %p = bitcast <4 x i256>* %a to <2 x i256>*
-       %b = load <2 x i256>* %p
-       %q = getelementptr <4 x i256>* %a, i32 0, i32 2
-       %c = load i256* %q
-       ret i256 %c
-; CHECK: @test9
-; CHECK-NOT: alloca
-; CHECK: bitcast <4 x i256> %x to <2 x i512>
-; CHECK-NEXT: extractelement <2 x i512>
-; CHECK-NEXT: bitcast i512 %tmp4 to <2 x i256>
-; CHECK-NEXT: extractelement <4 x i256>
-}
-
-define void @test10(<4 x i256> %x, <2 x i256> %y) {
-       %a = alloca <4 x i256>
-       store <4 x i256> %x, <4 x i256>* %a
-       %p = bitcast <4 x i256>* %a to <2 x i256>*
-       store <2 x i256> %y, <2 x i256>* %p
-       ret void
-; CHECK: @test10
-; CHECK-NOT: alloca
-; CHECK: bitcast <4 x i256> %x to <2 x i512>
-; CHECK-NEXT: bitcast <2 x i256> %y to i512
-; CHECK-NEXT: insertelement <2 x i512>
-; CHECK-NEXT: bitcast <2 x i512> %tmp2 to <4 x i256>
-}
-
-%union.v = type { <2 x i64> }
-
-define void @test11(<2 x i64> %x) {
-  %a = alloca %union.v
-  %p = getelementptr inbounds %union.v* %a, i32 0, i32 0
-  store <2 x i64> %x, <2 x i64>* %p, align 16
-  %q = getelementptr inbounds %union.v* %a, i32 0, i32 0
-  %r = bitcast <2 x i64>* %q to <4 x float>*
-  %b = load <4 x float>* %r, align 16
+define void @test7() {
+entry:
+  %memtmp = alloca %struct.test7, align 16
+  %0 = bitcast %struct.test7* %memtmp to <4 x i32>*
+  store <4 x i32> zeroinitializer, <4 x i32>* %0, align 16
+  %1 = getelementptr inbounds %struct.test7* %memtmp, i64 0, i32 0, i64 5
+  store i32 0, i32* %1, align 4
   ret void
-; CHECK: @test11
+; CHECK-LABEL: @test7(
 ; CHECK-NOT: alloca
+; CHECK: and i192
 }
 
-define void @test12() {
+; When promoting an alloca to a 1-element vector type, instructions that
+; produce that same vector type should not be changed to insert one element
+; into a new vector. <rdar://problem/14249078>
+define <1 x i64> @test8(<1 x i64> %a) {
 entry:
-  %a = alloca <64 x i8>, align 64
-  store <64 x i8> undef, <64 x i8>* %a, align 64
-  %p = bitcast <64 x i8>* %a to <16 x i8>*
-  %0 = load <16 x i8>* %p, align 64
-  store <16 x i8> undef, <16 x i8>* %p, align 64
-  %q = bitcast <16 x i8>* %p to <64 x i8>*
-  %1 = load <64 x i8>* %q, align 64
-  ret void
-; CHECK: @test12
+  %a.addr = alloca <1 x i64>, align 8
+  %__a = alloca <1 x i64>, align 8
+  %tmp = alloca <1 x i64>, align 8
+  store <1 x i64> %a, <1 x i64>* %a.addr, align 8
+  %0 = load <1 x i64>* %a.addr, align 8
+  store <1 x i64> %0, <1 x i64>* %__a, align 8
+  %1 = load <1 x i64>* %__a, align 8
+  %2 = bitcast <1 x i64> %1 to <8 x i8>
+  %3 = bitcast <8 x i8> %2 to <1 x i64>
+  %vshl_n = shl <1 x i64> %3, <i64 4>
+  store <1 x i64> %vshl_n, <1 x i64>* %tmp
+  %4 = load <1 x i64>* %tmp
+  ret <1 x i64> %4
+; CHECK-LABEL: @test8(
 ; CHECK-NOT: alloca
-; CHECK: extractelement <4 x i128>
-; CHECK: insertelement <4 x i128>
+; CHECK-NOT: insertelement
+; CHECK: ret <1 x i64>
 }