From a87d04a759e60daba0dadf344aa54383860b31db Mon Sep 17 00:00:00 2001
From: Chandler Carruth <chandlerc@gmail.com>
Date: Thu, 25 Sep 2014 02:20:02 +0000
Subject: [PATCH] [x86] Re-work a bunch of the v32i8 test cases to actually
 involve byte shuffles rather than word shuffles.

As you might guess, these were built starting from the word shuffle test
cases and I failed to properly port a bunch of them and left them as
widened word shuffle test cases. We still have a couple of tests that
check our ability to widen shuffles, but now we will test the actual
byte shuffle quite a bit better.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@218438 91177308-0d34-0410-b5e6-96231b3b80d8
---
 test/CodeGen/X86/vector-shuffle-256-v32.ll | 346 +++++++++------------
 1 file changed, 150 insertions(+), 196 deletions(-)

diff --git a/test/CodeGen/X86/vector-shuffle-256-v32.ll b/test/CodeGen/X86/vector-shuffle-256-v32.ll
index b417eacb6c6..afd4848fea7 100644
--- a/test/CodeGen/X86/vector-shuffle-256-v32.ll
+++ b/test/CodeGen/X86/vector-shuffle-256-v32.ll
@@ -1055,108 +1055,90 @@ define <32 x i8> @shuffle_v32i8_32_32_02_02_36_36_06_06_40_40_10_10_44_44_14_14_
   ret <32 x i8> %shuffle
 }
 
-define <32 x i8> @shuffle_v32i8_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32(<32 x i8> %a, <32 x i8> %b) {
-; AVX1-LABEL: @shuffle_v32i8_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32
+define <32 x i8> @shuffle_v32i8_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32(<32 x i8> %a, <32 x i8> %b) {
+; AVX1-LABEL: @shuffle_v32i8_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vpshufb {{.*}} # xmm1 = xmm1[0,0,0,0,2,2,0,0,0,0,0,0,6,6,0,0]
+; AVX1-NEXT:    vpxor %xmm1, %xmm1, %xmm1
+; AVX1-NEXT:    vpshufb %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT:    vpshufd {{.*}} # xmm0 = xmm0[0,0,0,0]
-; AVX1-NEXT:    vpblendw {{.*}} # xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3],xmm0[4],xmm1[5],xmm0[6],xmm1[7]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
-; AVX2-LABEL: @shuffle_v32i8_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32
+; AVX2-LABEL: @shuffle_v32i8_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vpshufb {{.*}} # xmm1 = xmm1[0,0,0,0,2,2,0,0,0,0,0,0,6,6,0,0]
+; AVX2-NEXT:    vpxor %xmm1, %xmm1, %xmm1
+; AVX2-NEXT:    vpshufb %xmm1, %xmm0, %xmm0
 ; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX2-NEXT:    vpshufd {{.*}} # xmm0 = xmm0[0,0,0,0]
-; AVX2-NEXT:    vpblendw {{.*}} # xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3],xmm0[4],xmm1[5],xmm0[6],xmm1[7]
 ; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
-  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 32, i32 32, i32 0, i32 0, i32 32, i32 32, i32 0, i32 0, i32 32, i32 32, i32 0, i32 0, i32 32, i32 32, i32 0, i32 0, i32 32, i32 32, i32 0, i32 0, i32 32, i32 32, i32 0, i32 0, i32 32, i32 32, i32 0, i32 0, i32 32, i32 32>
+  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32>
   ret <32 x i8> %shuffle
 }
 
-define <32 x i8> @shuffle_v32i8_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_16_16_48_48_16_16_48_48_16_16_48_48_16_16_48_48(<32 x i8> %a, <32 x i8> %b) {
-; AVX1-LABEL: @shuffle_v32i8_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_16_16_48_48_16_16_48_48_16_16_48_48_16_16_48_48
+define <32 x i8> @shuffle_v32i8_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_16_48_16_48_16_48_16_48_16_48_16_48_16_48_16_48(<32 x i8> %a, <32 x i8> %b) {
+; AVX1-LABEL: @shuffle_v32i8_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_16_48_16_48_16_48_16_48_16_48_16_48_16_48_16_48
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm2
-; AVX1-NEXT:    vmovdqa {{.*}} # xmm3 = [0,0,0,0,2,2,0,0,0,0,0,0,6,6,0,0]
-; AVX1-NEXT:    vpshufb %xmm3, %xmm2, %xmm2
-; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm4
-; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm4 = xmm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT:    vpshufd {{.*}} # xmm4 = xmm4[0,0,0,0]
-; AVX1-NEXT:    vpblendw {{.*}} # xmm2 = xmm4[0],xmm2[1],xmm4[2],xmm2[3],xmm4[4],xmm2[5],xmm4[6],xmm2[7]
-; AVX1-NEXT:    vpshufb %xmm3, %xmm1, %xmm1
+; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm1
+; AVX1-NEXT:    vpxor %xmm2, %xmm2, %xmm2
+; AVX1-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; AVX1-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
 ; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT:    vpshufd {{.*}} # xmm0 = xmm0[0,0,0,0]
-; AVX1-NEXT:    vpblendw {{.*}} # xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3],xmm0[4],xmm1[5],xmm0[6],xmm1[7]
-; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
+; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
-; AVX2-LABEL: @shuffle_v32i8_00_00_32_32_00_00_32_32_00_00_32_32_00_00_32_32_16_16_48_48_16_16_48_48_16_16_48_48_16_16_48_48
+; AVX2-LABEL: @shuffle_v32i8_00_32_00_32_00_32_00_32_00_32_00_32_00_32_00_32_16_48_16_48_16_48_16_48_16_48_16_48_16_48_16_48
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm2
-; AVX2-NEXT:    vmovdqa {{.*}} # xmm3 = [0,0,0,0,2,2,0,0,0,0,0,0,6,6,0,0]
-; AVX2-NEXT:    vpshufb %xmm3, %xmm2, %xmm2
-; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm4
-; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm4 = xmm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX2-NEXT:    vpshufd {{.*}} # xmm4 = xmm4[0,0,0,0]
-; AVX2-NEXT:    vpblendw {{.*}} # xmm2 = xmm4[0],xmm2[1],xmm4[2],xmm2[3],xmm4[4],xmm2[5],xmm4[6],xmm2[7]
-; AVX2-NEXT:    vpshufb %xmm3, %xmm1, %xmm1
+; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
+; AVX2-NEXT:    vpxor %xmm2, %xmm2, %xmm2
+; AVX2-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; AVX2-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
 ; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX2-NEXT:    vpshufd {{.*}} # xmm0 = xmm0[0,0,0,0]
-; AVX2-NEXT:    vpblendw {{.*}} # xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3],xmm0[4],xmm1[5],xmm0[6],xmm1[7]
-; AVX2-NEXT:    vinserti128 $1, %xmm2, %ymm0, %ymm0
+; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
-  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 32, i32 32, i32 0, i32 0, i32 32, i32 32, i32 0, i32 0, i32 32, i32 32, i32 0, i32 0, i32 32, i32 32, i32 16, i32 16, i32 48, i32 48, i32 16, i32 16, i32 48, i32 48, i32 16, i32 16, i32 48, i32 48, i32 16, i32 16, i32 48, i32 48>
+  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 0, i32 32, i32 16, i32 48, i32 16, i32 48, i32 16, i32 48, i32 16, i32 48, i32 16, i32 48, i32 16, i32 48, i32 16, i32 48, i32 16, i32 48>
   ret <32 x i8> %shuffle
 }
 
-define <32 x i8> @shuffle_v32i8_32_32_32_32_32_32_32_32_08_08_10_10_12_12_14_14_48_48_48_48_48_48_48_48_24_24_26_26_28_28_30_30(<32 x i8> %a, <32 x i8> %b) {
-; AVX1-LABEL: @shuffle_v32i8_32_32_32_32_32_32_32_32_08_08_10_10_12_12_14_14_48_48_48_48_48_48_48_48_24_24_26_26_28_28_30_30
+define <32 x i8> @shuffle_v32i8_32_32_32_32_32_32_32_32_08_09_10_11_12_13_14_15_48_48_48_48_48_48_48_48_24_25_26_27_28_29_30_31(<32 x i8> %a, <32 x i8> %b) {
+; AVX1-LABEL: @shuffle_v32i8_32_32_32_32_32_32_32_32_08_09_10_11_12_13_14_15_48_48_48_48_48_48_48_48_24_25_26_27_28_29_30_31
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT:    vmovdqa {{.*}} # xmm3 = [8,8,10,10,10,10,11,11,8,8,10,10,12,12,14,14]
-; AVX1-NEXT:    vpshufb %xmm3, %xmm2, %xmm2
-; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm4
-; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm4 = xmm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT:    vpshuflw {{.*}} # xmm4 = xmm4[0,0,0,0,4,5,6,7]
-; AVX1-NEXT:    vpblendw {{.*}} # xmm2 = xmm4[0,1,2,3],xmm2[4,5,6,7]
-; AVX1-NEXT:    vpshufb %xmm3, %xmm0, %xmm0
+; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm3
+; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; AVX1-NEXT:    vpshuflw {{.*}} # xmm3 = xmm3[0,0,0,0,4,5,6,7]
+; AVX1-NEXT:    vpblendw {{.*}} # xmm2 = xmm3[0,1,2,3],xmm2[4,5,6,7]
 ; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; AVX1-NEXT:    vpshuflw {{.*}} # xmm1 = xmm1[0,0,0,0,4,5,6,7]
 ; AVX1-NEXT:    vpblendw {{.*}} # xmm0 = xmm1[0,1,2,3],xmm0[4,5,6,7]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
-; AVX2-LABEL: @shuffle_v32i8_32_32_32_32_32_32_32_32_08_08_10_10_12_12_14_14_48_48_48_48_48_48_48_48_24_24_26_26_28_28_30_30
+; AVX2-LABEL: @shuffle_v32i8_32_32_32_32_32_32_32_32_08_09_10_11_12_13_14_15_48_48_48_48_48_48_48_48_24_25_26_27_28_29_30_31
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm2
-; AVX2-NEXT:    vmovdqa {{.*}} # xmm3 = [8,8,10,10,10,10,11,11,8,8,10,10,12,12,14,14]
-; AVX2-NEXT:    vpshufb %xmm3, %xmm2, %xmm2
-; AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm4
-; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm4 = xmm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX2-NEXT:    vpshuflw {{.*}} # xmm4 = xmm4[0,0,0,0,4,5,6,7]
-; AVX2-NEXT:    vpblendd {{.*}} # xmm2 = xmm4[0,1],xmm2[2,3]
-; AVX2-NEXT:    vpshufb %xmm3, %xmm0, %xmm0
+; AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm3
+; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; AVX2-NEXT:    vpshuflw {{.*}} # xmm3 = xmm3[0,0,0,0,4,5,6,7]
+; AVX2-NEXT:    vpblendd {{.*}} # xmm2 = xmm3[0,1],xmm2[2,3]
 ; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; AVX2-NEXT:    vpshuflw {{.*}} # xmm1 = xmm1[0,0,0,0,4,5,6,7]
 ; AVX2-NEXT:    vpblendd {{.*}} # xmm0 = xmm1[0,1],xmm0[2,3]
 ; AVX2-NEXT:    vinserti128 $1, %xmm2, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
-  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 8, i32 8, i32 10, i32 10, i32 12, i32 12, i32 14, i32 14, i32 48, i32 48, i32 48, i32 48, i32 48, i32 48, i32 48, i32 48, i32 24, i32 24, i32 26, i32 26, i32 28, i32 28, i32 30, i32 30>
+  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 48, i32 48, i32 48, i32 48, i32 48, i32 48, i32 48, i32 48, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
   ret <32 x i8> %shuffle
 }
 
-define <32 x i8> @shuffle_v32i8_38_38_36_36_34_34_32_32_14_14_12_12_10_10_08_08_54_54_52_52_50_50_48_48_30_30_28_28_26_26_24_24(<32 x i8> %a, <32 x i8> %b) {
-; AVX1-LABEL: @shuffle_v32i8_38_38_36_36_34_34_32_32_14_14_12_12_10_10_08_08_54_54_52_52_50_50_48_48_30_30_28_28_26_26_24_24
+define <32 x i8> @shuffle_v32i8_39_38_37_36_35_34_33_32_15_14_13_12_11_10_09_08_55_54_53_52_51_50_49_48_31_30_29_28_27_26_25_24(<32 x i8> %a, <32 x i8> %b) {
+; AVX1-LABEL: @shuffle_v32i8_39_38_37_36_35_34_33_32_15_14_13_12_11_10_09_08_55_54_53_52_51_50_49_48_31_30_29_28_27_26_25_24
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT:    vmovdqa {{.*}} # xmm3 = [8,8,10,10,10,10,11,11,14,14,12,12,10,10,8,8]
+; AVX1-NEXT:    vmovdqa {{.*}} # xmm3 = <u,u,u,u,u,u,u,u,15,14,13,12,11,10,9,8>
 ; AVX1-NEXT:    vpshufb %xmm3, %xmm2, %xmm2
 ; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm4
-; AVX1-NEXT:    vmovdqa {{.*}} # xmm5 = [6,6,4,4,2,2,0,0,4,4,6,6,6,6,7,7]
+; AVX1-NEXT:    vmovdqa {{.*}} # xmm5 = <7,6,5,4,3,2,1,0,u,u,u,u,u,u,u,u>
 ; AVX1-NEXT:    vpshufb %xmm5, %xmm4, %xmm4
 ; AVX1-NEXT:    vpblendw {{.*}} # xmm2 = xmm4[0,1,2,3],xmm2[4,5,6,7]
 ; AVX1-NEXT:    vpshufb %xmm3, %xmm0, %xmm0
@@ -1165,13 +1147,13 @@ define <32 x i8> @shuffle_v32i8_38_38_36_36_34_34_32_32_14_14_12_12_10_10_08_08_
 ; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
-; AVX2-LABEL: @shuffle_v32i8_38_38_36_36_34_34_32_32_14_14_12_12_10_10_08_08_54_54_52_52_50_50_48_48_30_30_28_28_26_26_24_24
+; AVX2-LABEL: @shuffle_v32i8_39_38_37_36_35_34_33_32_15_14_13_12_11_10_09_08_55_54_53_52_51_50_49_48_31_30_29_28_27_26_25_24
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm2
-; AVX2-NEXT:    vmovdqa {{.*}} # xmm3 = [8,8,10,10,10,10,11,11,14,14,12,12,10,10,8,8]
+; AVX2-NEXT:    vmovdqa {{.*}} # xmm3 = <u,u,u,u,u,u,u,u,15,14,13,12,11,10,9,8>
 ; AVX2-NEXT:    vpshufb %xmm3, %xmm2, %xmm2
 ; AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm4
-; AVX2-NEXT:    vmovdqa {{.*}} # xmm5 = [6,6,4,4,2,2,0,0,4,4,6,6,6,6,7,7]
+; AVX2-NEXT:    vmovdqa {{.*}} # xmm5 = <7,6,5,4,3,2,1,0,u,u,u,u,u,u,u,u>
 ; AVX2-NEXT:    vpshufb %xmm5, %xmm4, %xmm4
 ; AVX2-NEXT:    vpblendd {{.*}} # xmm2 = xmm4[0,1],xmm2[2,3]
 ; AVX2-NEXT:    vpshufb %xmm3, %xmm0, %xmm0
@@ -1179,18 +1161,18 @@ define <32 x i8> @shuffle_v32i8_38_38_36_36_34_34_32_32_14_14_12_12_10_10_08_08_
 ; AVX2-NEXT:    vpblendd {{.*}} # xmm0 = xmm1[0,1],xmm0[2,3]
 ; AVX2-NEXT:    vinserti128 $1, %xmm2, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
-  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 38, i32 38, i32 36, i32 36, i32 34, i32 34, i32 32, i32 32, i32 14, i32 14, i32 12, i32 12, i32 10, i32 10, i32 8, i32 8, i32 54, i32 54, i32 52, i32 52, i32 50, i32 50, i32 48, i32 48, i32 30, i32 30, i32 28, i32 28, i32 26, i32 26, i32 24, i32 24>
+  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24>
   ret <32 x i8> %shuffle
 }
 
-define <32 x i8> @shuffle_v32i8_38_38_36_36_34_34_32_32_06_06_04_04_02_02_00_00_54_54_52_52_50_50_48_48_22_22_20_20_18_18_16_16(<32 x i8> %a, <32 x i8> %b) {
-; AVX1-LABEL: @shuffle_v32i8_38_38_36_36_34_34_32_32_06_06_04_04_02_02_00_00_54_54_52_52_50_50_48_48_22_22_20_20_18_18_16_16
+define <32 x i8> @shuffle_v32i8_39_38_37_36_35_34_33_32_07_06_05_04_03_02_01_00_55_54_53_52_51_50_49_48_23_22_21_20_19_18_17_16(<32 x i8> %a, <32 x i8> %b) {
+; AVX1-LABEL: @shuffle_v32i8_39_38_37_36_35_34_33_32_07_06_05_04_03_02_01_00_55_54_53_52_51_50_49_48_23_22_21_20_19_18_17_16
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT:    vmovdqa {{.*}} # xmm3 = [0,0,2,2,2,2,3,3,6,6,4,4,2,2,0,0]
+; AVX1-NEXT:    vmovdqa {{.*}} # xmm3 = <u,u,u,u,u,u,u,u,7,6,5,4,3,2,1,0>
 ; AVX1-NEXT:    vpshufb %xmm3, %xmm2, %xmm2
 ; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm4
-; AVX1-NEXT:    vmovdqa {{.*}} # xmm5 = [6,6,4,4,2,2,0,0,4,4,6,6,6,6,7,7]
+; AVX1-NEXT:    vmovdqa {{.*}} # xmm5 = <7,6,5,4,3,2,1,0,u,u,u,u,u,u,u,u>
 ; AVX1-NEXT:    vpshufb %xmm5, %xmm4, %xmm4
 ; AVX1-NEXT:    vpblendw {{.*}} # xmm2 = xmm4[0,1,2,3],xmm2[4,5,6,7]
 ; AVX1-NEXT:    vpshufb %xmm3, %xmm0, %xmm0
@@ -1199,13 +1181,13 @@ define <32 x i8> @shuffle_v32i8_38_38_36_36_34_34_32_32_06_06_04_04_02_02_00_00_
 ; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
-; AVX2-LABEL: @shuffle_v32i8_38_38_36_36_34_34_32_32_06_06_04_04_02_02_00_00_54_54_52_52_50_50_48_48_22_22_20_20_18_18_16_16
+; AVX2-LABEL: @shuffle_v32i8_39_38_37_36_35_34_33_32_07_06_05_04_03_02_01_00_55_54_53_52_51_50_49_48_23_22_21_20_19_18_17_16
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm2
-; AVX2-NEXT:    vmovdqa {{.*}} # xmm3 = [0,0,2,2,2,2,3,3,6,6,4,4,2,2,0,0]
+; AVX2-NEXT:    vmovdqa {{.*}} # xmm3 = <u,u,u,u,u,u,u,u,7,6,5,4,3,2,1,0>
 ; AVX2-NEXT:    vpshufb %xmm3, %xmm2, %xmm2
 ; AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm4
-; AVX2-NEXT:    vmovdqa {{.*}} # xmm5 = [6,6,4,4,2,2,0,0,4,4,6,6,6,6,7,7]
+; AVX2-NEXT:    vmovdqa {{.*}} # xmm5 = <7,6,5,4,3,2,1,0,u,u,u,u,u,u,u,u>
 ; AVX2-NEXT:    vpshufb %xmm5, %xmm4, %xmm4
 ; AVX2-NEXT:    vpblendd {{.*}} # xmm2 = xmm4[0,1],xmm2[2,3]
 ; AVX2-NEXT:    vpshufb %xmm3, %xmm0, %xmm0
@@ -1213,7 +1195,7 @@ define <32 x i8> @shuffle_v32i8_38_38_36_36_34_34_32_32_06_06_04_04_02_02_00_00_
 ; AVX2-NEXT:    vpblendd {{.*}} # xmm0 = xmm1[0,1],xmm0[2,3]
 ; AVX2-NEXT:    vinserti128 $1, %xmm2, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
-  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 38, i32 38, i32 36, i32 36, i32 34, i32 34, i32 32, i32 32, i32 6, i32 6, i32 4, i32 4, i32 2, i32 2, i32 0, i32 0, i32 54, i32 54, i32 52, i32 52, i32 50, i32 50, i32 48, i32 48, i32 22, i32 22, i32 20, i32 20, i32 18, i32 18, i32 16, i32 16>
+  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16>
   ret <32 x i8> %shuffle
 }
 
@@ -1351,139 +1333,111 @@ define <32 x i8> @shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_
   ret <32 x i8> %shuffle
 }
 
-define <32 x i8> @shuffle_v32i8_00_00_32_32_02_02_34_34_04_04_36_36_06_06_38_38_16_16_48_48_18_18_50_50_20_20_52_52_22_22_54_54(<32 x i8> %a, <32 x i8> %b) {
-; AVX1-LABEL: @shuffle_v32i8_00_00_32_32_02_02_34_34_04_04_36_36_06_06_38_38_16_16_48_48_18_18_50_50_20_20_52_52_22_22_54_54
+define <32 x i8> @shuffle_v32i8_00_32_01_33_02_34_03_35_04_36_05_37_06_38_07_39_16_48_17_49_18_50_19_51_20_52_21_53_22_54_23_55(<32 x i8> %a, <32 x i8> %b) {
+; AVX1-LABEL: @shuffle_v32i8_00_32_01_33_02_34_03_35_04_36_05_37_06_38_07_39_16_48_17_49_18_50_19_51_20_52_21_53_22_54_23_55
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm3
-; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT:    vpalignr {{.*}} # xmm3 = xmm3[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX1-NEXT:    vpblendw {{.*}} # xmm2 = xmm2[0],xmm3[1],xmm2[2],xmm3[3],xmm2[4],xmm3[5],xmm2[6],xmm3[7]
+; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm1
+; AVX1-NEXT:    vpmovzxbw %xmm1, %xmm1
+; AVX1-NEXT:    vmovdqa {{.*}} # xmm2 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
+; AVX1-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; AVX1-NEXT:    vpmovzxbw %xmm0, %xmm0
+; AVX1-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
 ; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT:    vpalignr {{.*}} # xmm1 = xmm1[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX1-NEXT:    vpblendw {{.*}} # xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3],xmm0[4],xmm1[5],xmm0[6],xmm1[7]
-; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
+; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
-; AVX2-LABEL: @shuffle_v32i8_00_00_32_32_02_02_34_34_04_04_36_36_06_06_38_38_16_16_48_48_18_18_50_50_20_20_52_52_22_22_54_54
+; AVX2-LABEL: @shuffle_v32i8_00_32_01_33_02_34_03_35_04_36_05_37_06_38_07_39_16_48_17_49_18_50_19_51_20_52_21_53_22_54_23_55
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm2
-; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm3
-; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX2-NEXT:    vpalignr {{.*}} # xmm3 = xmm3[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX2-NEXT:    vpblendw {{.*}} # xmm2 = xmm2[0],xmm3[1],xmm2[2],xmm3[3],xmm2[4],xmm3[5],xmm2[6],xmm3[7]
+; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
+; AVX2-NEXT:    vpmovzxbw %xmm1, %xmm1
+; AVX2-NEXT:    vmovdqa {{.*}} # xmm2 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
+; AVX2-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; AVX2-NEXT:    vpmovzxbw %xmm0, %xmm0
+; AVX2-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
 ; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX2-NEXT:    vpalignr {{.*}} # xmm1 = xmm1[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX2-NEXT:    vpblendw {{.*}} # xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3],xmm0[4],xmm1[5],xmm0[6],xmm1[7]
-; AVX2-NEXT:    vinserti128 $1, %xmm2, %ymm0, %ymm0
+; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
-  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 32, i32 32, i32 2, i32 2, i32 34, i32 34, i32 4, i32 4, i32 36, i32 36, i32 6, i32 6, i32 38, i32 38, i32 16, i32 16, i32 48, i32 48, i32 18, i32 18, i32 50, i32 50, i32 20, i32 20, i32 52, i32 52, i32 22, i32 22, i32 54, i32 54>
+  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 4, i32 36, i32 5, i32 37, i32 6, i32 38, i32 7, i32 39, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 20, i32 52, i32 21, i32 53, i32 22, i32 54, i32 23, i32 55>
   ret <32 x i8> %shuffle
 }
 
-define <32 x i8> @shuffle_v32i8_08_08_40_40_10_10_42_42_12_12_44_44_14_14_46_46_24_24_56_56_26_26_58_58_28_28_60_60_30_30_62_62(<32 x i8> %a, <32 x i8> %b) {
-; AVX1-LABEL: @shuffle_v32i8_08_08_40_40_10_10_42_42_12_12_44_44_14_14_46_46_24_24_56_56_26_26_58_58_28_28_60_60_30_30_62_62
+define <32 x i8> @shuffle_v32i8_08_40_09_41_10_42_11_43_12_44_13_45_14_46_15_47_24_56_25_57_26_58_27_59_28_60_29_61_30_62_31_63(<32 x i8> %a, <32 x i8> %b) {
+; AVX1-LABEL: @shuffle_v32i8_08_40_09_41_10_42_11_43_12_44_13_45_14_46_15_47_24_56_25_57_26_58_27_59_28_60_29_61_30_62_31_63
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT:    vpunpckhbw {{.*}} # xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm3
-; AVX1-NEXT:    vpunpckhbw {{.*}} # xmm3 = xmm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT:    vpalignr {{.*}} # xmm3 = xmm3[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX1-NEXT:    vpblendw {{.*}} # xmm2 = xmm2[0],xmm3[1],xmm2[2],xmm3[3],xmm2[4],xmm3[5],xmm2[6],xmm3[7]
-; AVX1-NEXT:    vpunpckhbw {{.*}} # xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT:    vpunpckhbw {{.*}} # xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT:    vpalignr {{.*}} # xmm1 = xmm1[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX1-NEXT:    vpblendw {{.*}} # xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3],xmm0[4],xmm1[5],xmm0[6],xmm1[7]
-; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: @shuffle_v32i8_08_08_40_40_10_10_42_42_12_12_44_44_14_14_46_46_24_24_56_56_26_26_58_58_28_28_60_60_30_30_62_62
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm2
-; AVX2-NEXT:    vpunpckhbw {{.*}} # xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm3
-; AVX2-NEXT:    vpunpckhbw {{.*}} # xmm3 = xmm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX2-NEXT:    vpalignr {{.*}} # xmm3 = xmm3[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX2-NEXT:    vpblendw {{.*}} # xmm2 = xmm2[0],xmm3[1],xmm2[2],xmm3[3],xmm2[4],xmm3[5],xmm2[6],xmm3[7]
-; AVX2-NEXT:    vpunpckhbw {{.*}} # xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX2-NEXT:    vpunpckhbw {{.*}} # xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX2-NEXT:    vpalignr {{.*}} # xmm1 = xmm1[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX2-NEXT:    vpblendw {{.*}} # xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3],xmm0[4],xmm1[5],xmm0[6],xmm1[7]
-; AVX2-NEXT:    vinserti128 $1, %xmm2, %ymm0, %ymm0
-; AVX2-NEXT:    retq
-  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 8, i32 8, i32 40, i32 40, i32 10, i32 10, i32 42, i32 42, i32 12, i32 12, i32 44, i32 44, i32 14, i32 14, i32 46, i32 46, i32 24, i32 24, i32 56, i32 56, i32 26, i32 26, i32 58, i32 58, i32 28, i32 28, i32 60, i32 60, i32 30, i32 30, i32 62, i32 62>
-  ret <32 x i8> %shuffle
-}
-
-define <32 x i8> @shuffle_v32i8_00_00_32_32_02_02_34_34_04_04_36_36_06_06_38_38_24_24_56_56_26_26_58_58_28_28_60_60_30_30_62_62(<32 x i8> %a, <32 x i8> %b) {
-; AVX1-LABEL: @shuffle_v32i8_00_00_32_32_02_02_34_34_04_04_36_36_06_06_38_38_24_24_56_56_26_26_58_58_28_28_60_60_30_30_62_62
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT:    vpunpckhbw {{.*}} # xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm3
-; AVX1-NEXT:    vpunpckhbw {{.*}} # xmm3 = xmm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT:    vpalignr {{.*}} # xmm3 = xmm3[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX1-NEXT:    vpblendw {{.*}} # xmm2 = xmm2[0],xmm3[1],xmm2[2],xmm3[3],xmm2[4],xmm3[5],xmm2[6],xmm3[7]
+; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm1
+; AVX1-NEXT:    vmovdqa {{.*}} # xmm2 = <8,9,10,11,12,13,14,15,u,u,u,u,u,u,u,u>
+; AVX1-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; AVX1-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
 ; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT:    vpalignr {{.*}} # xmm1 = xmm1[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX1-NEXT:    vpblendw {{.*}} # xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3],xmm0[4],xmm1[5],xmm0[6],xmm1[7]
-; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
+; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
-; AVX2-LABEL: @shuffle_v32i8_00_00_32_32_02_02_34_34_04_04_36_36_06_06_38_38_24_24_56_56_26_26_58_58_28_28_60_60_30_30_62_62
+; AVX2-LABEL: @shuffle_v32i8_08_40_09_41_10_42_11_43_12_44_13_45_14_46_15_47_24_56_25_57_26_58_27_59_28_60_29_61_30_62_31_63
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm2
-; AVX2-NEXT:    vpunpckhbw {{.*}} # xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm3
-; AVX2-NEXT:    vpunpckhbw {{.*}} # xmm3 = xmm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX2-NEXT:    vpalignr {{.*}} # xmm3 = xmm3[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX2-NEXT:    vpblendw {{.*}} # xmm2 = xmm2[0],xmm3[1],xmm2[2],xmm3[3],xmm2[4],xmm3[5],xmm2[6],xmm3[7]
+; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
+; AVX2-NEXT:    vmovdqa {{.*}} # xmm2 = <8,9,10,11,12,13,14,15,u,u,u,u,u,u,u,u>
+; AVX2-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; AVX2-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
 ; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX2-NEXT:    vpalignr {{.*}} # xmm1 = xmm1[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX2-NEXT:    vpblendw {{.*}} # xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3],xmm0[4],xmm1[5],xmm0[6],xmm1[7]
-; AVX2-NEXT:    vinserti128 $1, %xmm2, %ymm0, %ymm0
+; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
-  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 32, i32 32, i32 2, i32 2, i32 34, i32 34, i32 4, i32 4, i32 36, i32 36, i32 6, i32 6, i32 38, i32 38, i32 24, i32 24, i32 56, i32 56, i32 26, i32 26, i32 58, i32 58, i32 28, i32 28, i32 60, i32 60, i32 30, i32 30, i32 62, i32 62>
+  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 8, i32 40, i32 9, i32 41, i32 10, i32 42, i32 11, i32 43, i32 12, i32 44, i32 13, i32 45, i32 14, i32 46, i32 15, i32 47, i32 24, i32 56, i32 25, i32 57, i32 26, i32 58, i32 27, i32 59, i32 28, i32 60, i32 29, i32 61, i32 30, i32 62, i32 31, i32 63>
   ret <32 x i8> %shuffle
 }
 
-define <32 x i8> @shuffle_v32i8_08_08_40_40_10_10_42_42_12_12_44_44_14_14_46_46_16_16_48_48_18_18_50_50_20_20_52_52_22_22_54_54(<32 x i8> %a, <32 x i8> %b) {
-; AVX1-LABEL: @shuffle_v32i8_08_08_40_40_10_10_42_42_12_12_44_44_14_14_46_46_16_16_48_48_18_18_50_50_20_20_52_52_22_22_54_54
+define <32 x i8> @shuffle_v32i8_00_32_01_33_02_34_03_35_04_36_05_37_06_38_07_39_24_56_25_57_26_58_27_59_28_60_29_61_30_62_31_63(<32 x i8> %a, <32 x i8> %b) {
+; AVX1-LABEL: @shuffle_v32i8_00_32_01_33_02_34_03_35_04_36_05_37_06_38_07_39_24_56_25_57_26_58_27_59_28_60_29_61_30_62_31_63
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm3
-; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT:    vpalignr {{.*}} # xmm3 = xmm3[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX1-NEXT:    vpblendw {{.*}} # xmm2 = xmm2[0],xmm3[1],xmm2[2],xmm3[3],xmm2[4],xmm3[5],xmm2[6],xmm3[7]
-; AVX1-NEXT:    vpunpckhbw {{.*}} # xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT:    vpunpckhbw {{.*}} # xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT:    vpalignr {{.*}} # xmm1 = xmm1[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX1-NEXT:    vpblendw {{.*}} # xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3],xmm0[4],xmm1[5],xmm0[6],xmm1[7]
-; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
+; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm1
+; AVX1-NEXT:    vpshufb {{.*}} # xmm1 = xmm1[8,9,10,11,12,13,14,15,u,u,u,u,u,u,u,u]
+; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; AVX1-NEXT:    vpmovzxbw %xmm0, %xmm0
+; AVX1-NEXT:    vpshufb {{.*}} # xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
+; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
-; AVX2-LABEL: @shuffle_v32i8_08_08_40_40_10_10_42_42_12_12_44_44_14_14_46_46_16_16_48_48_18_18_50_50_20_20_52_52_22_22_54_54
+; AVX2-LABEL: @shuffle_v32i8_00_32_01_33_02_34_03_35_04_36_05_37_06_38_07_39_24_56_25_57_26_58_27_59_28_60_29_61_30_62_31_63
 ; AVX2:       # BB#0:
-; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm2
-; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm3
-; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX2-NEXT:    vpalignr {{.*}} # xmm3 = xmm3[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX2-NEXT:    vpblendw {{.*}} # xmm2 = xmm2[0],xmm3[1],xmm2[2],xmm3[3],xmm2[4],xmm3[5],xmm2[6],xmm3[7]
-; AVX2-NEXT:    vpunpckhbw {{.*}} # xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX2-NEXT:    vpunpckhbw {{.*}} # xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX2-NEXT:    vpalignr {{.*}} # xmm1 = xmm1[14,15,0,1,2,3,4,5,6,7,8,9,10,11,12,13]
-; AVX2-NEXT:    vpblendw {{.*}} # xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3],xmm0[4],xmm1[5],xmm0[6],xmm1[7]
-; AVX2-NEXT:    vinserti128 $1, %xmm2, %ymm0, %ymm0
+; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
+; AVX2-NEXT:    vpshufb {{.*}} # xmm1 = xmm1[8,9,10,11,12,13,14,15,u,u,u,u,u,u,u,u]
+; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; AVX2-NEXT:    vpmovzxbw %xmm0, %xmm0
+; AVX2-NEXT:    vpshufb {{.*}} # xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
+; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
-  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 8, i32 8, i32 40, i32 40, i32 10, i32 10, i32 42, i32 42, i32 12, i32 12, i32 44, i32 44, i32 14, i32 14, i32 46, i32 46, i32 16, i32 16, i32 48, i32 48, i32 18, i32 18, i32 50, i32 50, i32 20, i32 20, i32 52, i32 52, i32 22, i32 22, i32 54, i32 54>
+  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 4, i32 36, i32 5, i32 37, i32 6, i32 38, i32 7, i32 39, i32 24, i32 56, i32 25, i32 57, i32 26, i32 58, i32 27, i32 59, i32 28, i32 60, i32 29, i32 61, i32 30, i32 62, i32 31, i32 63>
+  ret <32 x i8> %shuffle
+}
+
+define <32 x i8> @shuffle_v32i8_08_40_09_41_10_42_11_43_12_44_13_45_14_46_15_47_16_48_17_49_18_50_19_51_20_52_21_53_22_54_23_55(<32 x i8> %a, <32 x i8> %b) {
+; AVX1-LABEL: @shuffle_v32i8_08_40_09_41_10_42_11_43_12_44_13_45_14_46_15_47_16_48_17_49_18_50_19_51_20_52_21_53_22_54_23_55
+; AVX1:       # BB#0:
+; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm1
+; AVX1-NEXT:    vpmovzxbw %xmm1, %xmm1
+; AVX1-NEXT:    vpshufb {{.*}} # xmm1 = xmm1[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
+; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; AVX1-NEXT:    vpshufb {{.*}} # xmm0 = xmm0[8,9,10,11,12,13,14,15,u,u,u,u,u,u,u,u]
+; AVX1-NEXT:    vpunpcklbw {{.*}} # xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: @shuffle_v32i8_08_40_09_41_10_42_11_43_12_44_13_45_14_46_15_47_16_48_17_49_18_50_19_51_20_52_21_53_22_54_23_55
+; AVX2:       # BB#0:
+; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
+; AVX2-NEXT:    vpmovzxbw %xmm1, %xmm1
+; AVX2-NEXT:    vpshufb {{.*}} # xmm1 = xmm1[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
+; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
+; AVX2-NEXT:    vpshufb {{.*}} # xmm0 = xmm0[8,9,10,11,12,13,14,15,u,u,u,u,u,u,u,u]
+; AVX2-NEXT:    vpunpcklbw {{.*}} # xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX2-NEXT:    retq
+  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 8, i32 40, i32 9, i32 41, i32 10, i32 42, i32 11, i32 43, i32 12, i32 44, i32 13, i32 45, i32 14, i32 46, i32 15, i32 47, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 20, i32 52, i32 21, i32 53, i32 22, i32 54, i32 23, i32 55>
   ret <32 x i8> %shuffle
 }
 
@@ -1651,61 +1605,61 @@ define <32 x i8> @shuffle_v32i8_08_08_08_08_08_08_08_08_00_00_00_00_00_00_00_00_
   ret <32 x i8> %shuffle
 }
 
-define <32 x i8> @shuffle_v32i8_00_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_16_16_16_16_uu_uu_uu_uu_uu_16_16_16_28_28_16_16(<32 x i8> %a, <32 x i8> %b) {
-; AVX1-LABEL: @shuffle_v32i8_00_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_16_16_16_16_uu_uu_uu_uu_uu_16_16_16_28_28_16_16
+define <32 x i8> @shuffle_v32i8_00_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_16_16_16_16_uu_uu_uu_uu_uu_16_16_16_16_16_30_16(<32 x i8> %a, <32 x i8> %b) {
+; AVX1-LABEL: @shuffle_v32i8_00_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_16_16_16_16_uu_uu_uu_uu_uu_16_16_16_16_16_30_16
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm1
-; AVX1-NEXT:    vpshufb {{.*}} # xmm1 = xmm1[0,0,0,0,12,12,13,13,0,0,0,0,12,12,0,0]
+; AVX1-NEXT:    vpshufb {{.*}} # xmm1 = xmm1[0,0,0,0,u,u,u,u,u,0,0,0,0,0,14,0]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
-; AVX2-LABEL: @shuffle_v32i8_00_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_16_16_16_16_uu_uu_uu_uu_uu_16_16_16_28_28_16_16
+; AVX2-LABEL: @shuffle_v32i8_00_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_16_16_16_16_uu_uu_uu_uu_uu_16_16_16_16_16_30_16
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
-; AVX2-NEXT:    vpshufb {{.*}} # xmm1 = xmm1[0,0,0,0,12,12,13,13,0,0,0,0,12,12,0,0]
+; AVX2-NEXT:    vpshufb {{.*}} # xmm1 = xmm1[0,0,0,0,u,u,u,u,u,0,0,0,0,0,14,0]
 ; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
-  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 16, i32 16, i32 16, i32 16, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 16, i32 16, i32 16, i32 28, i32 28, i32 16, i32 16>
+  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 16, i32 16, i32 16, i32 16, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 16, i32 16, i32 16, i32 16, i32 16, i32 30, i32 16>
   ret <32 x i8> %shuffle
 }
 
-define <32 x i8> @shuffle_v32i8_uu_14_uu_uu_00_00_00_00_00_00_00_00_00_00_00_00_16_16_uu_16_uu_uu_uu_uu_16_16_16_16_16_16_30_30(<32 x i8> %a, <32 x i8> %b) {
-; AVX1-LABEL: @shuffle_v32i8_uu_14_uu_uu_00_00_00_00_00_00_00_00_00_00_00_00_16_16_uu_16_uu_uu_uu_uu_16_16_16_16_16_16_30_30
+define <32 x i8> @shuffle_v32i8_uu_14_uu_uu_00_00_00_00_00_00_00_00_00_00_00_00_16_16_uu_16_uu_uu_uu_uu_16_16_16_16_16_16_16_31(<32 x i8> %a, <32 x i8> %b) {
+; AVX1-LABEL: @shuffle_v32i8_uu_14_uu_uu_00_00_00_00_00_00_00_00_00_00_00_00_16_16_uu_16_uu_uu_uu_uu_16_16_16_16_16_16_16_31
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vpshufb {{.*}} # xmm1 = xmm0[0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0]
 ; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm0
-; AVX1-NEXT:    vpshufb {{.*}} # xmm0 = xmm0[0,0,1,1,14,14,15,15,0,0,0,0,0,0,14,14]
+; AVX1-NEXT:    vpshufb {{.*}} # xmm0 = xmm0[0,0,u,0,u,u,u,u,0,0,0,0,0,0,0,15]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
-; AVX2-LABEL: @shuffle_v32i8_uu_14_uu_uu_00_00_00_00_00_00_00_00_00_00_00_00_16_16_uu_16_uu_uu_uu_uu_16_16_16_16_16_16_30_30
+; AVX2-LABEL: @shuffle_v32i8_uu_14_uu_uu_00_00_00_00_00_00_00_00_00_00_00_00_16_16_uu_16_uu_uu_uu_uu_16_16_16_16_16_16_16_31
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*}} # xmm1 = xmm0[0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0]
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm0
-; AVX2-NEXT:    vpshufb {{.*}} # xmm0 = xmm0[0,0,1,1,14,14,15,15,0,0,0,0,0,0,14,14]
+; AVX2-NEXT:    vpshufb {{.*}} # xmm0 = xmm0[0,0,u,0,u,u,u,u,0,0,0,0,0,0,0,15]
 ; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
-  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 undef, i32 14, i32 undef, i32 undef, 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 16, i32 16, i32 undef, i32 16, i32 undef, i32 undef, i32 undef, i32 undef, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 30, i32 30>
+  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 undef, i32 14, i32 undef, i32 undef, 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 16, i32 16, i32 undef, i32 16, i32 undef, i32 undef, i32 undef, i32 undef, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 31>
   ret <32 x i8> %shuffle
 }
 
-define <32 x i8> @shuffle_v32i8_00_00_00_uu_uu_uu_04_04_08_08_08_08_uu_uu_12_uu_28_28_28_28_uu_uu_24_24_20_20_20_20_16_16_16_16(<32 x i8> %a, <32 x i8> %b) {
-; AVX1-LABEL: @shuffle_v32i8_00_00_00_uu_uu_uu_04_04_08_08_08_08_uu_uu_12_uu_28_28_28_28_uu_uu_24_24_20_20_20_20_16_16_16_16
+define <32 x i8> @shuffle_v32i8_00_00_00_uu_uu_uu_04_uu_08_08_08_08_uu_uu_12_uu_28_28_28_28_uu_uu_uu_24_20_20_20_20_16_16_16_16(<32 x i8> %a, <32 x i8> %b) {
+; AVX1-LABEL: @shuffle_v32i8_00_00_00_uu_uu_uu_04_uu_08_08_08_08_uu_uu_12_uu_28_28_28_28_uu_uu_uu_24_20_20_20_20_16_16_16_16
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vpshufb {{.*}} # xmm1 = xmm0[0,0,0,0,4,4,4,4,8,8,8,8,12,12,12,12]
 ; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm0
-; AVX1-NEXT:    vpshufb {{.*}} # xmm0 = xmm0[12,12,12,12,8,8,8,8,4,4,4,4,0,0,0,0]
+; AVX1-NEXT:    vpshufb {{.*}} # xmm0 = xmm0[12,12,12,12,12,12,13,13,4,4,4,4,0,0,0,0]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
-; AVX2-LABEL: @shuffle_v32i8_00_00_00_uu_uu_uu_04_04_08_08_08_08_uu_uu_12_uu_28_28_28_28_uu_uu_24_24_20_20_20_20_16_16_16_16
+; AVX2-LABEL: @shuffle_v32i8_00_00_00_uu_uu_uu_04_uu_08_08_08_08_uu_uu_12_uu_28_28_28_28_uu_uu_uu_24_20_20_20_20_16_16_16_16
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufb {{.*}} # xmm1 = xmm0[0,0,0,0,4,4,4,4,8,8,8,8,12,12,12,12]
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm0
-; AVX2-NEXT:    vpshufb {{.*}} # xmm0 = xmm0[12,12,12,12,8,8,8,8,4,4,4,4,0,0,0,0]
+; AVX2-NEXT:    vpshufb {{.*}} # xmm0 = xmm0[12,12,12,12,12,12,13,13,4,4,4,4,0,0,0,0]
 ; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
-  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 undef, i32 undef, i32 undef, i32 4, i32 4, i32 8, i32 8, i32 8, i32 8, i32 undef, i32 undef, i32 12, i32 undef, i32 28, i32 28, i32 28, i32 28, i32 undef, i32 undef, i32 24, i32 24, i32 20, i32 20, i32 20, i32 20, i32 16, i32 16, i32 16, i32 16>
+  %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 undef, i32 undef, i32 undef, i32 4, i32 undef, i32 8, i32 8, i32 8, i32 8, i32 undef, i32 undef, i32 12, i32 undef, i32 28, i32 28, i32 28, i32 28, i32 undef, i32 undef, i32 undef, i32 24, i32 20, i32 20, i32 20, i32 20, i32 16, i32 16, i32 16, i32 16>
   ret <32 x i8> %shuffle
 }