CodeGen peephole: fold redundant phys reg copies

[oota-llvm.git] / test / CodeGen / X86 / vector-shuffle-128-v2.ll

diff --git a/test/CodeGen/X86/vector-shuffle-128-v2.ll b/test/CodeGen/X86/vector-shuffle-128-v2.ll
index ad33ef7645bec484c236588fbae474478784bea9..624b785db195d37f4f1391b7d667cb9e4fb17ec9 100644 (file)
--- a/test/CodeGen/X86/vector-shuffle-128-v2.ll
+++ b/test/CodeGen/X86/vector-shuffle-128-v2.ll
@@ -1,3 +1,4 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+sse3 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE3
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+ssse3 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSSE3
@@ -142,20 +143,10 @@ define <2 x double> @shuffle_v2f64_10(<2 x double> %a, <2 x double> %b) {
 ; SSE-NEXT:    shufpd {{.*#+}} xmm0 = xmm0[1,0]
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_v2f64_10:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v2f64_10:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v2f64_10:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vpermilpd $1, %xmm0, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_v2f64_10:
+; AVX:       # BB#0:
+; AVX-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
+; AVX-NEXT:    retq
 
   %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 1, i32 0>
   ret <2 x double> %shuffle
@@ -166,20 +157,10 @@ define <2 x double> @shuffle_v2f64_11(<2 x double> %a, <2 x double> %b) {
 ; SSE-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_v2f64_11:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v2f64_11:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v2f64_11:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vmovhlps %xmm0, %xmm0, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_v2f64_11:
+; AVX:       # BB#0:
+; AVX-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX-NEXT:    retq
   %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 1, i32 1>
   ret <2 x double> %shuffle
 }
@@ -219,20 +200,10 @@ define <2 x double> @shuffle_v2f64_32(<2 x double> %a, <2 x double> %b) {
 ; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_v2f64_32:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm1[1,0]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v2f64_32:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm1[1,0]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v2f64_32:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vpermilpd $1, %xmm1, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_v2f64_32:
+; AVX:       # BB#0:
+; AVX-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm1[1,0]
+; AVX-NEXT:    retq
 
   %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 3, i32 2>
   ret <2 x double> %shuffle
@@ -244,20 +215,10 @@ define <2 x double> @shuffle_v2f64_33(<2 x double> %a, <2 x double> %b) {
 ; SSE-NEXT:    movaps %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_v2f64_33:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm1[1,1]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v2f64_33:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm1[1,1]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v2f64_33:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vmovhlps %xmm1, %xmm1, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_v2f64_33:
+; AVX:       # BB#0:
+; AVX-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm1[1,1]
+; AVX-NEXT:    retq
   %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 3, i32 3>
   ret <2 x double> %shuffle
 }
@@ -328,20 +289,10 @@ define <2 x i64> @shuffle_v2i64_02(<2 x i64> %a, <2 x i64> %b) {
 ; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_v2i64_02:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v2i64_02:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v2i64_02:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vpunpcklqdq %xmm1, %xmm0, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_v2i64_02:
+; AVX:       # BB#0:
+; AVX-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT:    retq
   %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> <i32 0, i32 2>
   ret <2 x i64> %shuffle
 }
@@ -352,20 +303,10 @@ define <2 x i64> @shuffle_v2i64_02_copy(<2 x i64> %nonce, <2 x i64> %a, <2 x i64
 ; SSE-NEXT:    movdqa %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_v2i64_02_copy:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm2[0]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v2i64_02_copy:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm2[0]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v2i64_02_copy:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vpunpcklqdq %xmm2, %xmm1, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_v2i64_02_copy:
+; AVX:       # BB#0:
+; AVX-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm2[0]
+; AVX-NEXT:    retq
   %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> <i32 0, i32 2>
   ret <2 x i64> %shuffle
 }
@@ -520,20 +461,10 @@ define <2 x i64> @shuffle_v2i64_13(<2 x i64> %a, <2 x i64> %b) {
 ; SSE-NEXT:    punpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1]
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_v2i64_13:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vpunpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v2i64_13:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vpunpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v2i64_13:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vpunpckhqdq %xmm1, %xmm0, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_v2i64_13:
+; AVX:       # BB#0:
+; AVX-NEXT:    vpunpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1]
+; AVX-NEXT:    retq
   %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> <i32 1, i32 3>
   ret <2 x i64> %shuffle
 }
@@ -544,20 +475,10 @@ define <2 x i64> @shuffle_v2i64_13_copy(<2 x i64> %nonce, <2 x i64> %a, <2 x i64
 ; SSE-NEXT:    movdqa %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_v2i64_13_copy:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm2[1]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v2i64_13_copy:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm2[1]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v2i64_13_copy:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vpunpckhqdq %xmm2, %xmm1, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_v2i64_13_copy:
+; AVX:       # BB#0:
+; AVX-NEXT:    vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm2[1]
+; AVX-NEXT:    retq
   %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> <i32 1, i32 3>
   ret <2 x i64> %shuffle
 }
@@ -568,20 +489,10 @@ define <2 x i64> @shuffle_v2i64_20(<2 x i64> %a, <2 x i64> %b) {
 ; SSE-NEXT:    movdqa %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_v2i64_20:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v2i64_20:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v2i64_20:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vpunpcklqdq %xmm0, %xmm1, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_v2i64_20:
+; AVX:       # BB#0:
+; AVX-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
+; AVX-NEXT:    retq
   %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> <i32 2, i32 0>
   ret <2 x i64> %shuffle
 }
@@ -592,20 +503,10 @@ define <2 x i64> @shuffle_v2i64_20_copy(<2 x i64> %nonce, <2 x i64> %a, <2 x i64
 ; SSE-NEXT:    movdqa %xmm2, %xmm0
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_v2i64_20_copy:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm2[0],xmm1[0]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v2i64_20_copy:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm2[0],xmm1[0]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v2i64_20_copy:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vpunpcklqdq %xmm1, %xmm2, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_v2i64_20_copy:
+; AVX:       # BB#0:
+; AVX-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm2[0],xmm1[0]
+; AVX-NEXT:    retq
   %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> <i32 2, i32 0>
   ret <2 x i64> %shuffle
 }
@@ -758,20 +659,10 @@ define <2 x i64> @shuffle_v2i64_31(<2 x i64> %a, <2 x i64> %b) {
 ; SSE-NEXT:    movdqa %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_v2i64_31:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm0[1]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v2i64_31:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm0[1]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v2i64_31:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vpunpckhqdq %xmm0, %xmm1, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_v2i64_31:
+; AVX:       # BB#0:
+; AVX-NEXT:    vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm0[1]
+; AVX-NEXT:    retq
   %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> <i32 3, i32 1>
   ret <2 x i64> %shuffle
 }
@@ -782,20 +673,10 @@ define <2 x i64> @shuffle_v2i64_31_copy(<2 x i64> %nonce, <2 x i64> %a, <2 x i64
 ; SSE-NEXT:    movdqa %xmm2, %xmm0
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_v2i64_31_copy:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vpunpckhqdq {{.*#+}} xmm0 = xmm2[1],xmm1[1]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v2i64_31_copy:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vpunpckhqdq {{.*#+}} xmm0 = xmm2[1],xmm1[1]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v2i64_31_copy:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vpunpckhqdq %xmm1, %xmm2, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_v2i64_31_copy:
+; AVX:       # BB#0:
+; AVX-NEXT:    vpunpckhqdq {{.*#+}} xmm0 = xmm2[1],xmm1[1]
+; AVX-NEXT:    retq
   %shuffle = shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> <i32 3, i32 1>
   ret <2 x i64> %shuffle
 }
@@ -944,7 +825,7 @@ define <2 x double> @shuffle_v2f64_1z(<2 x double> %a) {
 ; AVX512VL-LABEL: shuffle_v2f64_1z:
 ; AVX512VL:       # BB#0:
 ; AVX512VL-NEXT:    vxorps %xmm1, %xmm1, %xmm1
-; AVX512VL-NEXT:    vunpckhpd %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    vunpckhpd {{.*#+}} xmm0 = xmm0[1],xmm1[1]
 ; AVX512VL-NEXT:    retq
   %shuffle = shufflevector <2 x double> %a, <2 x double> zeroinitializer, <2 x i32> <i32 1, i32 3>
   ret <2 x double> %shuffle
@@ -973,7 +854,7 @@ define <2 x double> @shuffle_v2f64_z0(<2 x double> %a) {
 ; AVX512VL-LABEL: shuffle_v2f64_z0:
 ; AVX512VL:       # BB#0:
 ; AVX512VL-NEXT:    vxorps %xmm1, %xmm1, %xmm1
-; AVX512VL-NEXT:    vunpcklpd %xmm0, %xmm1, %xmm0
+; AVX512VL-NEXT:    vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0]
 ; AVX512VL-NEXT:    retq
   %shuffle = shufflevector <2 x double> %a, <2 x double> zeroinitializer, <2 x i32> <i32 2, i32 0>
   ret <2 x double> %shuffle
@@ -1035,7 +916,7 @@ define <2 x double> @shuffle_v2f64_bitcast_1z(<2 x double> %a) {
 ; AVX512VL-LABEL: shuffle_v2f64_bitcast_1z:
 ; AVX512VL:       # BB#0:
 ; AVX512VL-NEXT:    vpxor %xmm1, %xmm1, %xmm1
-; AVX512VL-NEXT:    vshufpd $1, %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    vshufpd {{.*#+}} xmm0 = xmm0[1],xmm1[0]
 ; AVX512VL-NEXT:    retq
   %shuffle64 = shufflevector <2 x double> %a, <2 x double> zeroinitializer, <2 x i32> <i32 2, i32 1>
   %bitcast32 = bitcast <2 x double> %shuffle64 to <4 x float>
@@ -1295,23 +1176,11 @@ define <2 x i64> @insert_reg_hi_v2i64(i64 %a, <2 x i64> %b) {
 ; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: insert_reg_hi_v2i64:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vmovq %rdi, %xmm1
-; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: insert_reg_hi_v2i64:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vmovq %rdi, %xmm1
-; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: insert_reg_hi_v2i64:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vmovq %rdi, %xmm1
-; AVX512VL-NEXT:    vpunpcklqdq %xmm1, %xmm0, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: insert_reg_hi_v2i64:
+; AVX:       # BB#0:
+; AVX-NEXT:    vmovq %rdi, %xmm1
+; AVX-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT:    retq
   %v = insertelement <2 x i64> undef, i64 %a, i32 0
   %shuffle = shufflevector <2 x i64> %v, <2 x i64> %b, <2 x i32> <i32 2, i32 0>
   ret <2 x i64> %shuffle
@@ -1339,7 +1208,7 @@ define <2 x i64> @insert_mem_hi_v2i64(i64* %ptr, <2 x i64> %b) {
 ; AVX512VL-LABEL: insert_mem_hi_v2i64:
 ; AVX512VL:       # BB#0:
 ; AVX512VL-NEXT:    vmovq (%rdi), %xmm1
-; AVX512VL-NEXT:    vpunpcklqdq %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; AVX512VL-NEXT:    retq
   %a = load i64, i64* %ptr
   %v = insertelement <2 x i64> undef, i64 %a, i32 0
@@ -1396,20 +1265,10 @@ define <2 x double> @insert_reg_hi_v2f64(double %a, <2 x double> %b) {
 ; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: insert_reg_hi_v2f64:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: insert_reg_hi_v2f64:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: insert_reg_hi_v2f64:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vunpcklpd %xmm0, %xmm1, %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: insert_reg_hi_v2f64:
+; AVX:       # BB#0:
+; AVX-NEXT:    vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0]
+; AVX-NEXT:    retq
   %v = insertelement <2 x double> undef, double %a, i32 0
   %shuffle = shufflevector <2 x double> %v, <2 x double> %b, <2 x i32> <i32 2, i32 0>
   ret <2 x double> %shuffle
@@ -1483,27 +1342,48 @@ define <2 x double> @insert_dup_mem_v2f64(double* %ptr) {
 ; SSE41-NEXT:    movddup {{.*#+}} xmm0 = mem[0,0]
 ; SSE41-NEXT:    retq
 ;
-; AVX1-LABEL: insert_dup_mem_v2f64:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vmovddup {{.*#+}} xmm0 = mem[0,0]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: insert_dup_mem_v2f64:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vmovddup {{.*#+}} xmm0 = mem[0,0]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: insert_dup_mem_v2f64:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vmovsd (%rdi), %xmm0
-; AVX512VL-NEXT:    vmovddup {{.*#+}} xmm0 = xmm0[0,0]
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: insert_dup_mem_v2f64:
+; AVX:       # BB#0:
+; AVX-NEXT:    vmovddup {{.*#+}} xmm0 = mem[0,0]
+; AVX-NEXT:    retq
   %a = load double, double* %ptr
   %v = insertelement <2 x double> undef, double %a, i32 0
   %shuffle = shufflevector <2 x double> %v, <2 x double> undef, <2 x i32> <i32 0, i32 0>
   ret <2 x double> %shuffle
 }
 
+define <2 x double> @insert_dup_mem128_v2f64(<2 x double>* %ptr) nounwind {
+; SSE2-LABEL: insert_dup_mem128_v2f64:
+; SSE2:       # BB#0:
+; SSE2-NEXT:    movaps (%rdi), %xmm0
+; SSE2-NEXT:    movlhps {{.*#+}} xmm0 = xmm0[0,0]
+; SSE2-NEXT:    retq
+;
+; SSE3-LABEL: insert_dup_mem128_v2f64:
+; SSE3:       # BB#0:
+; SSE3-NEXT:    movddup {{.*#+}} xmm0 = mem[0,0]
+; SSE3-NEXT:    retq
+;
+; SSSE3-LABEL: insert_dup_mem128_v2f64:
+; SSSE3:       # BB#0:
+; SSSE3-NEXT:    movddup {{.*#+}} xmm0 = mem[0,0]
+; SSSE3-NEXT:    retq
+;
+; SSE41-LABEL: insert_dup_mem128_v2f64:
+; SSE41:       # BB#0:
+; SSE41-NEXT:    movddup {{.*#+}} xmm0 = mem[0,0]
+; SSE41-NEXT:    retq
+;
+; AVX-LABEL: insert_dup_mem128_v2f64:
+; AVX:       # BB#0:
+; AVX-NEXT:    vmovddup {{.*#+}} xmm0 = mem[0,0]
+; AVX-NEXT:    retq
+  %v = load  <2 x double>,  <2 x double>* %ptr
+  %shuffle = shufflevector <2 x double> %v, <2 x double> undef, <2 x i32> <i32 0, i32 0>
+  ret <2 x double> %shuffle
+}
+
+
 define <2 x i64> @insert_dup_mem_v2i64(i64* %ptr) {
 ; SSE-LABEL: insert_dup_mem_v2i64:
 ; SSE:       # BB#0:
@@ -1539,20 +1419,10 @@ define <2 x double> @shuffle_mem_v2f64_10(<2 x double>* %ptr) {
 ; SSE-NEXT:    shufpd {{.*#+}} xmm0 = xmm0[1,0]
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: shuffle_mem_v2f64_10:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vpermilpd {{.*#+}} xmm0 = mem[1,0]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_mem_v2f64_10:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vpermilpd {{.*#+}} xmm0 = mem[1,0]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_mem_v2f64_10:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vpermilpd $1, (%rdi), %xmm0
-; AVX512VL-NEXT:    retq
+; AVX-LABEL: shuffle_mem_v2f64_10:
+; AVX:       # BB#0:
+; AVX-NEXT:    vpermilpd {{.*#+}} xmm0 = mem[1,0]
+; AVX-NEXT:    retq
 
   %a = load <2 x double>, <2 x double>* %ptr
   %shuffle = shufflevector <2 x double> %a, <2 x double> undef, <2 x i32> <i32 1, i32 0>