[X86] Add extra rules for combining vselect dag nodes into movsd.

This improves the fix committed at revision 199683 adding the
following new target specific combine rules:

1) fold (v4i32: vselect <0,0,-1,-1>, A, B) ->
        (v4i32 (bitcast (movsd (v2i64 (bitcast A)), (v2i64 (bitcast B))) ))

2) fold (v4f32: vselect <0,0,-1,-1>, A, B) ->
        (v4f32 (bitcast (movsd (v2f64 (bitcast A)), (v2f64 (bitcast B))) ))

3) fold (v4i32: vselect <-1,-1,0,0>, A, B) ->
        (v4i32 (bitcast (movsd (v2i64 (bitcast B)), (v2i64 (bitcast A))) ))

4) fold (v4f32: vselect <-1,-1,0,0>, A, B) ->
        (v4f32 (bitcast (movsd (v2i64 (bitcast B)), (v2i64 (bitcast A))) ))

llvm-svn: 200324

2 files changed, 35 insertions, 2 deletions

diff --git a/llvm/test/CodeGen/X86/sse41-blend.ll b/llvm/test/CodeGen/X86/sse41-blend.ll
index 597852c3690..4681fde7548 100644
--- a/llvm/test/CodeGen/X86/sse41-blend.ll
+++ b/llvm/test/CodeGen/X86/sse41-blend.ll
@@ -13,7 +13,7 @@ define <4 x float> @vsel_float(<4 x float> %v1, <4 x float> %v2) {
 ;CHECK: blendvps
 ;CHECK: ret
 define <4 x i8> @vsel_4xi8(<4 x i8> %v1, <4 x i8> %v2) {
-  %vsel = select <4 x i1> <i1 true, i1 true, i1 false, i1 false>, <4 x i8> %v1, <4 x i8> %v2
+  %vsel = select <4 x i1> <i1 true, i1 true, i1 false, i1 true>, <4 x i8> %v1, <4 x i8> %v2
   ret <4 x i8> %vsel
 }
 
@@ -30,7 +30,7 @@ define <4 x i16> @vsel_4xi16(<4 x i16> %v1, <4 x i16> %v2) {
 ;CHECK: blendvps
 ;CHECK: ret
 define <4 x i32> @vsel_i32(<4 x i32> %v1, <4 x i32> %v2) {
-  %vsel = select <4 x i1> <i1 true, i1 true, i1 false, i1 false>, <4 x i32> %v1, <4 x i32> %v2
+  %vsel = select <4 x i1> <i1 true, i1 true, i1 false, i1 true>, <4 x i32> %v1, <4 x i32> %v2
   ret <4 x i32> %vsel
 }
 
diff --git a/llvm/test/CodeGen/X86/vselect-2.ll b/llvm/test/CodeGen/X86/vselect-2.ll
new file mode 100644
index 00000000000..50da32c67a3
--- /dev/null
+++ b/llvm/test/CodeGen/X86/vselect-2.ll
@@ -0,0 +1,33 @@
+; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=corei7 -mattr=sse2 | FileCheck %s
+
+define <4 x i32> @test1(<4 x i32> %A, <4 x i32> %B) {
+  %select = select <4 x i1><i1 true, i1 true, i1 false, i1 false>, <4 x i32> %A, <4 x i32> %B
+  ret <4 x i32> %select
+}
+; CHECK-LABEL: test1
+; CHECK: movsd
+; CHECK: ret
+
+define <4 x i32> @test2(<4 x i32> %A, <4 x i32> %B) {
+  %select = select <4 x i1><i1 false, i1 false, i1 true, i1 true>, <4 x i32> %A, <4 x i32> %B
+  ret <4 x i32> %select
+}
+; CHECK-LABEL: test2
+; CHECK: movsd
+; CHECK-NEXT: ret
+
+define <4 x float> @test3(<4 x float> %A, <4 x float> %B) {
+  %select = select <4 x i1><i1 true, i1 true, i1 false, i1 false>, <4 x float> %A, <4 x float> %B
+  ret <4 x float> %select
+}
+; CHECK-LABEL: test3
+; CHECK: movsd
+; CHECK: ret
+
+define <4 x float> @test4(<4 x float> %A, <4 x float> %B) {
+  %select = select <4 x i1><i1 false, i1 false, i1 true, i1 true>, <4 x float> %A, <4 x float> %B
+  ret <4 x float> %select
+}
+; CHECK-LABEL: test4
+; CHECK: movsd
+; CHECK-NEXT: ret