CodeGen peephole: fold redundant phys reg copies

[oota-llvm.git] / test / CodeGen / X86 / vec_split.ll

diff --git a/test/CodeGen/X86/vec_split.ll b/test/CodeGen/X86/vec_split.ll
index f9e7c20ba4e2e0dec4565ebb8c3d2c03f79bf3e8..1df4cf2b2325fb777d92216e66f9ba32d330034b 100644 (file)
--- a/test/CodeGen/X86/vec_split.ll
+++ b/test/CodeGen/X86/vec_split.ll
@@ -1,6 +1,6 @@
-; RUN: llc -march=x86-64 -mcpu=corei7 < %s | FileCheck %s -check-prefix=SSE4
-; RUN: llc -march=x86-64 -mcpu=corei7-avx < %s | FileCheck %s -check-prefix=AVX1
-; RUN: llc -march=x86-64 -mcpu=core-avx2 < %s | FileCheck %s -check-prefix=AVX2
+; RUN: llc -march=x86-64 -mattr=sse4.1 < %s | FileCheck %s -check-prefix=SSE4
+; RUN: llc -march=x86-64 -mattr=avx < %s | FileCheck %s -check-prefix=AVX1
+; RUN: llc -march=x86-64 -mattr=avx2 < %s | FileCheck %s -check-prefix=AVX2
 
 define <16 x i16> @split16(<16 x i16> %a, <16 x i16> %b, <16 x i8> %__mask) {
 ; SSE4-LABEL: split16:
@@ -40,3 +40,36 @@ define <32 x i16> @split32(<32 x i16> %a, <32 x i16> %b, <32 x i8> %__mask) {
   %2 = select <32 x i1> %1, <32 x i16> %a, <32 x i16> %b
   ret <32 x i16> %2
 }
+
+; PR19492
+define i128 @split128(<2 x i128> %a, <2 x i128> %b) {
+; SSE4-LABEL: split128:
+; SSE4: addq
+; SSE4: adcq
+; SSE4: addq
+; SSE4: adcq
+; SSE4: addq
+; SSE4: adcq
+; SSE4: ret
+; AVX1-LABEL: split128:
+; AVX1: addq
+; AVX1: adcq
+; AVX1: addq
+; AVX1: adcq
+; AVX1: addq
+; AVX1: adcq
+; AVX1: ret
+; AVX2-LABEL: split128:
+; AVX2: addq
+; AVX2: adcq
+; AVX2: addq
+; AVX2: adcq
+; AVX2: addq
+; AVX2: adcq
+; AVX2: ret
+  %add = add nsw <2 x i128> %a, %b
+  %rdx.shuf = shufflevector <2 x i128> %add, <2 x i128> undef, <2 x i32> <i32 undef, i32 0>
+  %bin.rdx = add <2 x i128> %add, %rdx.shuf
+  %e = extractelement <2 x i128> %bin.rdx, i32 1
+  ret i128 %e
+}