Index: llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
===================================================================
--- llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
+++ llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
@@ -10586,6 +10586,10 @@
     // subvector.
     bool OnlyUsesV1 = isShuffleEquivalent(V1, V2, Mask, {0, 1, 0, 1});
     if (OnlyUsesV1 || isShuffleEquivalent(V1, V2, Mask, {0, 1, 4, 5})) {
+      // With AVX2 we should use VPERMQ/VPERMPD to allow memory folding.
+      if (Subtarget.hasAVX2() && isSingleInputShuffleMask(Mask))
+        return SDValue();
+
       MVT SubVT = MVT::getVectorVT(VT.getVectorElementType(),
                                    VT.getVectorNumElements() / 2);
       SDValue LoV = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, SubVT, V1,
@@ -11038,8 +11042,9 @@
 
   SmallVector<int, 4> WidenedMask;
   if (canWidenShuffleElements(Mask, WidenedMask))
-    return lowerV2X128VectorShuffle(DL, MVT::v4f64, V1, V2, Mask, Subtarget,
-                                    DAG);
+    if (SDValue V = lowerV2X128VectorShuffle(DL, MVT::v4f64, V1, V2, Mask,
+                                             Subtarget, DAG))
+      return V;
 
   if (isSingleInputShuffleMask(Mask)) {
     // Check for being able to broadcast a single element.
@@ -11133,8 +11138,9 @@
 
   SmallVector<int, 4> WidenedMask;
   if (canWidenShuffleElements(Mask, WidenedMask))
-    return lowerV2X128VectorShuffle(DL, MVT::v4i64, V1, V2, Mask, Subtarget,
-                                    DAG);
+    if (SDValue V = lowerV2X128VectorShuffle(DL, MVT::v4i64, V1, V2, Mask,
+                                             Subtarget, DAG))
+      return V;
 
   if (SDValue Blend = lowerVectorShuffleAsBlend(DL, MVT::v4i64, V1, V2, Mask,
                                                 Subtarget, DAG))
Index: llvm/trunk/test/CodeGen/X86/avx-vperm2x128.ll
===================================================================
--- llvm/trunk/test/CodeGen/X86/avx-vperm2x128.ll
+++ llvm/trunk/test/CodeGen/X86/avx-vperm2x128.ll
@@ -35,21 +35,31 @@
 }
 
 define <8 x float> @shuffle_v8f32_01230123(<8 x float> %a, <8 x float> %b) nounwind uwtable readnone ssp {
-; ALL-LABEL: shuffle_v8f32_01230123:
-; ALL:       ## BB#0: ## %entry
-; ALL-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; ALL-NEXT:    retq
+; AVX1-LABEL: shuffle_v8f32_01230123:
+; AVX1:       ## BB#0: ## %entry
+; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: shuffle_v8f32_01230123:
+; AVX2:       ## BB#0: ## %entry
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT:    retq
 entry:
   %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
   ret <8 x float> %shuffle
 }
 
 define <8 x float> @shuffle_v8f32_01230123_mem(<8 x float>* %pa, <8 x float>* %pb) nounwind uwtable readnone ssp {
-; ALL-LABEL: shuffle_v8f32_01230123_mem:
-; ALL:       ## BB#0: ## %entry
-; ALL-NEXT:    vmovaps (%rdi), %ymm0
-; ALL-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; ALL-NEXT:    retq
+; AVX1-LABEL: shuffle_v8f32_01230123_mem:
+; AVX1:       ## BB#0: ## %entry
+; AVX1-NEXT:    vmovaps (%rdi), %ymm0
+; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: shuffle_v8f32_01230123_mem:
+; AVX2:       ## BB#0: ## %entry
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = mem[0,1,0,1]
+; AVX2-NEXT:    retq
 entry:
   %a = load <8 x float>, <8 x float>* %pa
   %b = load <8 x float>, <8 x float>* %pb
Index: llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v16.ll
===================================================================
--- llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v16.ll
+++ llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v16.ll
@@ -455,7 +455,7 @@
 ; AVX2-LABEL: shuffle_v16i16_00_00_00_00_00_00_01_00_00_00_00_00_00_00_01_00:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,1,0,1,0,1,0,1,0,1,0,1,2,3,0,1]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0>
   ret <16 x i16> %shuffle
@@ -471,7 +471,7 @@
 ; AVX2-LABEL: shuffle_v16i16_00_00_00_00_00_02_00_00_00_00_00_00_00_02_00_00:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,1,0,1,0,1,0,1,0,1,4,5,0,1,0,1]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 2, i32 0, i32 0>
   ret <16 x i16> %shuffle
@@ -487,7 +487,7 @@
 ; AVX2-LABEL: shuffle_v16i16_00_00_00_00_03_00_00_00_00_00_00_00_03_00_00_00:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,1,0,1,0,1,0,1,6,7,0,1,0,1,0,1]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 3, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 3, i32 0, i32 0, i32 0>
   ret <16 x i16> %shuffle
@@ -503,7 +503,7 @@
 ; AVX2-LABEL: shuffle_v16i16_00_00_00_04_00_00_00_00_00_00_00_04_00_00_00_00:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,1,0,1,0,1,8,9,0,1,0,1,0,1,0,1]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 0, i32 0, i32 4, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 4, i32 0, i32 0, i32 0, i32 0>
   ret <16 x i16> %shuffle
@@ -519,7 +519,7 @@
 ; AVX2-LABEL: shuffle_v16i16_00_00_05_00_00_00_00_00_00_00_05_00_00_00_00_00:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,1,0,1,10,11,0,1,0,1,0,1,0,1,0,1]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 0, i32 5, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 5, i32 0, i32 0, i32 0, i32 0, i32 0>
   ret <16 x i16> %shuffle
@@ -535,7 +535,7 @@
 ; AVX2-LABEL: shuffle_v16i16_00_06_00_00_00_00_00_00_00_06_00_00_00_00_00_00:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,1,12,13,0,1,0,1,0,1,0,1,0,1,0,1]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 6, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 6, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   ret <16 x i16> %shuffle
@@ -551,7 +551,7 @@
 ; AVX2-LABEL: shuffle_v16i16_07_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[14,15,0,1,0,1,0,1,0,1,0,1,0,1,0,1]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 7, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 7, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   ret <16 x i16> %shuffle
Index: llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v32.ll
===================================================================
--- llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v32.ll
+++ llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v32.ll
@@ -818,7 +818,7 @@
 ; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0>
   ret <32 x i8> %shuffle
@@ -834,7 +834,7 @@
 ; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 2, i32 0, i32 0>
   ret <32 x i8> %shuffle
@@ -850,7 +850,7 @@
 ; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 7, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 7, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   ret <32 x i8> %shuffle
@@ -866,7 +866,7 @@
 ; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 8, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 8, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   ret <32 x i8> %shuffle
@@ -882,7 +882,7 @@
 ; AVX2-LABEL: shuffle_v32i8_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 14, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 14, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   ret <32 x i8> %shuffle
@@ -902,7 +902,7 @@
 ; AVX2-NEXT:    movl $15, %eax
 ; AVX2-NEXT:    vmovd %eax, %xmm1
 ; AVX2-NEXT:    vpshufb %xmm1, %xmm0, %xmm0
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 15, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 15, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   ret <32 x i8> %shuffle
Index: llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v4.ll
===================================================================
--- llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v4.ll
+++ llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v4.ll
@@ -849,15 +849,15 @@
 ;
 ; AVX2-LABEL: shuffle_v4i64_0451:
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,0,1,3]
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,2,1]
 ; AVX2-NEXT:    vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5],ymm0[6,7]
 ; AVX2-NEXT:    retq
 ;
 ; AVX512VL-LABEL: shuffle_v4i64_0451:
 ; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vinserti32x4 $1, %xmm0, %ymm0, %ymm0
 ; AVX512VL-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,0,1,3]
+; AVX512VL-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,2,1]
 ; AVX512VL-NEXT:    vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5],ymm0[6,7]
 ; AVX512VL-NEXT:    retq
   %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 4, i32 5, i32 1>
@@ -893,14 +893,14 @@
 ;
 ; AVX2-LABEL: shuffle_v4i64_4015:
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm1, %ymm1
+; AVX2-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,1,2,1]
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,0,1,3]
 ; AVX2-NEXT:    vpblendd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3,4,5],ymm1[6,7]
 ; AVX2-NEXT:    retq
 ;
 ; AVX512VL-LABEL: shuffle_v4i64_4015:
 ; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vinserti32x4 $1, %xmm1, %ymm1, %ymm1
+; AVX512VL-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,1,2,1]
 ; AVX512VL-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,0,1,3]
 ; AVX512VL-NEXT:    vpblendd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3,4,5],ymm1[6,7]
 ; AVX512VL-NEXT:    retq
Index: llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v8.ll
===================================================================
--- llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v8.ll
+++ llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v8.ll
@@ -679,11 +679,17 @@
 }
 
 define <8 x float> @shuffle_v8f32_32103210(<8 x float> %a, <8 x float> %b) {
-; ALL-LABEL: shuffle_v8f32_32103210:
-; ALL:       # BB#0:
-; ALL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
-; ALL-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; ALL-NEXT:    retq
+; AVX1-LABEL: shuffle_v8f32_32103210:
+; AVX1:       # BB#0:
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: shuffle_v8f32_32103210:
+; AVX2:       # BB#0:
+; AVX2-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0>
   ret <8 x float> %shuffle
 }
@@ -1764,7 +1770,7 @@
 ; AVX2-LABEL: shuffle_v8i32_32103210:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[3,2,1,0]
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0>
   ret <8 x i32> %shuffle