1 ; RUN: llc < %s -mcpu=x86-64 -mattr=+avx -x86-experimental-vector-shuffle-lowering | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX1
2 ; RUN: llc < %s -mcpu=x86-64 -mattr=+avx2 -x86-experimental-vector-shuffle-lowering | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX2
4 target triple = "x86_64-unknown-unknown"
6 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
7 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
9 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
10 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm0
11 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
14 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
16 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
17 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm0
18 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0
20 %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 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 0, i32 0, i32 0>
21 ret <32 x i8> %shuffle
24 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00(<32 x i8> %a, <32 x i8> %b) {
25 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00
27 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
28 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
29 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
30 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
33 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00
35 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
36 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
37 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
38 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
40 %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 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 0, i32 1, i32 0>
41 ret <32 x i8> %shuffle
44 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00(<32 x i8> %a, <32 x i8> %b) {
45 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00
47 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
48 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
49 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
50 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
53 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00
55 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
56 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
57 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
58 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
60 %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 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>
61 ret <32 x i8> %shuffle
64 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_03_00_00_00(<32 x i8> %a, <32 x i8> %b) {
65 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_03_00_00_00
67 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
68 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
69 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,3,0,0,0]
70 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
73 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_03_00_00_00
75 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
76 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
77 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,3,0,0,0]
78 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
80 %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 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 3, i32 0, i32 0, i32 0>
81 ret <32 x i8> %shuffle
84 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_04_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
85 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_04_00_00_00_00
87 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
88 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
89 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,4,0,0,0,0]
90 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
93 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_04_00_00_00_00
95 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
96 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
97 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,4,0,0,0,0]
98 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
100 %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 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 4, i32 0, i32 0, i32 0, i32 0>
101 ret <32 x i8> %shuffle
104 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_05_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
105 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_05_00_00_00_00_00
107 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
108 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
109 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,5,0,0,0,0,0]
110 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
113 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_05_00_00_00_00_00
115 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
116 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
117 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,5,0,0,0,0,0]
118 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
120 %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 0, i32 0, i32 0, i32 0, i32 0, 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>
121 ret <32 x i8> %shuffle
124 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_06_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
125 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_06_00_00_00_00_00_00
127 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
128 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
129 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,6,0,0,0,0,0,0]
130 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
133 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_06_00_00_00_00_00_00
135 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
136 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
137 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,6,0,0,0,0,0,0]
138 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
140 %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 0, i32 0, i32 0, i32 0, 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>
141 ret <32 x i8> %shuffle
144 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
145 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00
147 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
148 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
149 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
150 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
153 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00
155 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
156 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
157 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
158 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
160 %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 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>
161 ret <32 x i8> %shuffle
164 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
165 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00
167 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
168 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
169 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
170 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
173 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00
175 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
176 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
177 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
178 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
180 %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 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>
181 ret <32 x i8> %shuffle
184 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_09_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
185 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_09_00_00_00_00_00_00_00_00_00
187 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
188 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
189 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,9,0,0,0,0,0,0,0,0,0]
190 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
193 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_09_00_00_00_00_00_00_00_00_00
195 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
196 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
197 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,9,0,0,0,0,0,0,0,0,0]
198 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
200 %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 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 9, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
201 ret <32 x i8> %shuffle
204 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_10_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
205 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_10_00_00_00_00_00_00_00_00_00_00
207 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
208 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
209 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,10,0,0,0,0,0,0,0,0,0,0]
210 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
213 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_10_00_00_00_00_00_00_00_00_00_00
215 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
216 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
217 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,10,0,0,0,0,0,0,0,0,0,0]
218 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
220 %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 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 10, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
221 ret <32 x i8> %shuffle
224 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_11_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
225 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_11_00_00_00_00_00_00_00_00_00_00_00
227 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
228 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
229 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,11,0,0,0,0,0,0,0,0,0,0,0]
230 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
233 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_11_00_00_00_00_00_00_00_00_00_00_00
235 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
236 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
237 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,11,0,0,0,0,0,0,0,0,0,0,0]
238 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
240 %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 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 11, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
241 ret <32 x i8> %shuffle
244 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_12_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
245 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_12_00_00_00_00_00_00_00_00_00_00_00_00
247 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
248 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
249 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,12,0,0,0,0,0,0,0,0,0,0,0,0]
250 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
253 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_12_00_00_00_00_00_00_00_00_00_00_00_00
255 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
256 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
257 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,12,0,0,0,0,0,0,0,0,0,0,0,0]
258 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
260 %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 0, i32 0, i32 0, i32 0, i32 0, i32 12, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
261 ret <32 x i8> %shuffle
264 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_13_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
265 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_13_00_00_00_00_00_00_00_00_00_00_00_00_00
267 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
268 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
269 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,13,0,0,0,0,0,0,0,0,0,0,0,0,0]
270 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
273 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_13_00_00_00_00_00_00_00_00_00_00_00_00_00
275 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
276 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
277 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,13,0,0,0,0,0,0,0,0,0,0,0,0,0]
278 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
280 %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 0, i32 0, i32 0, i32 0, i32 13, 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>
281 ret <32 x i8> %shuffle
284 define <32 x i8> @shuffle_v32i8_00_00_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(<32 x i8> %a, <32 x i8> %b) {
285 ; AVX1-LABEL: @shuffle_v32i8_00_00_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
287 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
288 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
289 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
290 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
293 ; AVX2-LABEL: @shuffle_v32i8_00_00_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
295 ; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
296 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
297 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
298 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
300 %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 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>
301 ret <32 x i8> %shuffle
304 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
305 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
307 ; AVX1-NEXT: movl $15, %eax
308 ; AVX1-NEXT: vmovd %eax, %xmm1
309 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
310 ; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
311 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
312 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
315 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
317 ; AVX2-NEXT: movl $15, %eax
318 ; AVX2-NEXT: vmovd %eax, %xmm1
319 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm1
320 ; AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
321 ; AVX2-NEXT: vpshufb %xmm2, %xmm0, %xmm0
322 ; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
324 %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 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>
325 ret <32 x i8> %shuffle
328 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
329 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
331 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
332 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
333 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
334 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0]
335 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],zero
336 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
337 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
340 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
342 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
343 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm2
344 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
345 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm0
346 ; AVX2-NEXT: vmovdqa {{.*}} # ymm1 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
347 ; AVX2-NEXT: vpblendvb %ymm1, %ymm0, %ymm2, %ymm0
349 %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 0, i32 16, 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 0>
350 ret <32 x i8> %shuffle
353 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_17_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
354 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_17_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
356 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
357 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
358 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
359 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[1],zero
360 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0],zero,xmm0[0]
361 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
362 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
365 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_17_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
367 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
368 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
369 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
370 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,u,u,u,u,u,u,1,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
371 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,0,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
372 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
374 %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 17, 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 0, i32 0>
375 ret <32 x i8> %shuffle
378 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_18_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
379 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_18_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
381 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
382 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
383 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
384 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[2],zero,zero
385 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0],zero,xmm0[0,0]
386 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
387 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
390 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_18_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
392 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
393 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
394 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
395 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,u,u,u,u,u,2,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
396 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,128,128,128,0,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
397 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
399 %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 18, 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 0, i32 0, i32 0>
400 ret <32 x i8> %shuffle
403 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_19_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
404 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_19_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
406 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
407 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
408 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
409 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[3],zero,zero,zero
410 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0],zero,xmm0[0,0,0]
411 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
412 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
415 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_19_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
417 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
418 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
419 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
420 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,u,u,u,u,3,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
421 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,128,128,0,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
422 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
424 %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 19, 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 0, i32 0, i32 0, i32 0>
425 ret <32 x i8> %shuffle
428 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_20_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
429 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_20_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
431 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
432 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
433 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
434 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[4],zero,zero,zero,zero
435 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0],zero,xmm0[0,0,0,0]
436 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
437 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
440 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_20_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
442 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
443 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
444 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
445 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,u,u,u,4,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
446 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,128,0,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
447 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
449 %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 20, 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 0, i32 0, i32 0, i32 0, i32 0>
450 ret <32 x i8> %shuffle
453 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_21_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
454 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_21_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
456 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
457 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
458 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
459 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[5],zero,zero,zero,zero,zero
460 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0],zero,xmm0[0,0,0,0,0]
461 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
462 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
465 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_21_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
467 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
468 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
469 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
470 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,u,u,5,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
471 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,128,0,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
472 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
474 %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 21, 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 0, i32 0, i32 0, i32 0, i32 0, i32 0>
475 ret <32 x i8> %shuffle
478 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_22_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
479 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_22_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
481 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
482 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
483 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
484 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[6],zero,zero,zero,zero,zero,zero
485 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0],zero,xmm0[0,0,0,0,0,0]
486 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
487 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
490 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_22_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
492 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
493 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
494 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
495 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,u,6,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
496 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,128,0,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
497 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
499 %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 22, 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 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
500 ret <32 x i8> %shuffle
503 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_23_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
504 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_23_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
506 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
507 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
508 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
509 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,xmm2[7],zero,zero,zero,zero,zero,zero,zero
510 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0],zero,xmm0[0,0,0,0,0,0,0]
511 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
512 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
515 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_23_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
517 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
518 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
519 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
520 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,7,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
521 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,128,0,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
522 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
524 %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 23, 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 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
525 ret <32 x i8> %shuffle
528 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_24_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
529 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_24_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
531 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
532 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
533 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
534 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,zero,zero,zero,zero,xmm2[8],zero,zero,zero,zero,zero,zero,zero,zero
535 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0],zero,xmm0[0,0,0,0,0,0,0,0]
536 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
537 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
540 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_24_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
542 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
543 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
544 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
545 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,8,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
546 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,128,0,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
547 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
549 %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 24, 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 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
550 ret <32 x i8> %shuffle
553 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_25_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
554 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_25_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
556 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
557 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
558 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
559 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,zero,zero,zero,xmm2[9],zero,zero,zero,zero,zero,zero,zero,zero,zero
560 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0],zero,xmm0[0,0,0,0,0,0,0,0,0]
561 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
562 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
565 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_25_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
567 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
568 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
569 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
570 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,9,u,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
571 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,128,0,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
572 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
574 %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 25, 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 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
575 ret <32 x i8> %shuffle
578 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_26_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
579 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_26_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
581 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
582 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
583 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
584 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,zero,zero,xmm2[10],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
585 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0],zero,xmm0[0,0,0,0,0,0,0,0,0,0]
586 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
587 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
590 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_26_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
592 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
593 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
594 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
595 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,10,u,u,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
596 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,128,128,128,0,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
597 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
599 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 26, 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 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
600 ret <32 x i8> %shuffle
603 define <32 x i8> @shuffle_v32i8_00_00_00_00_27_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
604 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_27_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
606 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
607 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
608 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
609 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,zero,xmm2[11],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
610 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0],zero,xmm0[0,0,0,0,0,0,0,0,0,0,0]
611 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
612 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
615 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_27_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
617 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
618 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
619 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
620 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,11,u,u,u,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
621 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,128,128,0,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
622 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
624 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 27, 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 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
625 ret <32 x i8> %shuffle
628 define <32 x i8> @shuffle_v32i8_00_00_00_28_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
629 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_28_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
631 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
632 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
633 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
634 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,zero,xmm2[12],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
635 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0],zero,xmm0[0,0,0,0,0,0,0,0,0,0,0,0]
636 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
637 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
640 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_28_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
642 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
643 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
644 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
645 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,12,u,u,u,u,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
646 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,128,0,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
647 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
649 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 28, 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 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>
650 ret <32 x i8> %shuffle
653 define <32 x i8> @shuffle_v32i8_00_00_29_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
654 ; AVX1-LABEL: @shuffle_v32i8_00_00_29_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
656 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
657 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
658 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
659 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,zero,xmm2[13],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
660 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0],zero,xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0]
661 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
662 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
665 ; AVX2-LABEL: @shuffle_v32i8_00_00_29_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
667 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
668 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
669 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
670 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,13,u,u,u,u,u,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
671 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,128,0,128,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
672 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
674 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 29, 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 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>
675 ret <32 x i8> %shuffle
678 define <32 x i8> @shuffle_v32i8_00_30_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
679 ; AVX1-LABEL: @shuffle_v32i8_00_30_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
681 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
682 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
683 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
684 ; AVX1-NEXT: vpshufb {{.*}} # xmm2 = zero,xmm2[14],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
685 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0],zero,xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0]
686 ; AVX1-NEXT: vpor %xmm2, %xmm0, %xmm0
687 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
690 ; AVX2-LABEL: @shuffle_v32i8_00_30_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
692 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
693 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm1
694 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm0 = ymm0[2,3,0,1]
695 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,14,u,u,u,u,u,u,u,u,u,u,u,u,u,u,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
696 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [128,0,128,128,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
697 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
699 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 30, 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 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>
700 ret <32 x i8> %shuffle
703 define <32 x i8> @shuffle_v32i8_31_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
704 ; AVX1-LABEL: @shuffle_v32i8_31_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
706 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
707 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
708 ; AVX1-NEXT: movl $128, %eax
709 ; AVX1-NEXT: vmovd %eax, %xmm2
710 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm2
711 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
712 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
713 ; AVX1-NEXT: vpor %xmm0, %xmm2, %xmm0
714 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
717 ; AVX2-LABEL: @shuffle_v32i8_31_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
719 ; AVX2-NEXT: vperm2i128 {{.*}} # ymm1 = ymm0[2,3,0,1]
720 ; AVX2-NEXT: movl $15, %eax
721 ; AVX2-NEXT: vmovd %eax, %xmm2
722 ; AVX2-NEXT: vpxor %ymm3, %ymm3, %ymm3
723 ; AVX2-NEXT: vinserti128 $0, %xmm2, %ymm3, %ymm2
724 ; AVX2-NEXT: vpshufb %ymm2, %ymm1, %ymm1
725 ; AVX2-NEXT: vpshufb %ymm3, %ymm0, %ymm0
726 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [0,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
727 ; AVX2-NEXT: vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
729 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 31, 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 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 0>
730 ret <32 x i8> %shuffle
733 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_16_16_16_16_16_16_16_16(<32 x i8> %a, <32 x i8> %b) {
734 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_16_16_16_16_16_16_16_16
736 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
737 ; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
738 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
739 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
740 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
743 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_16_16_16_16_16_16_16_16
745 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
746 ; AVX2-NEXT: vpshufb %ymm1, %ymm0, %ymm0
748 %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 0, i32 0, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
749 ret <32 x i8> %shuffle
752 define <32 x i8> @shuffle_v32i8_15_15_15_15_15_15_15_15_15_15_15_15_15_15_15_15_31_31_31_31_31_31_31_31_31_31_31_31_31_31_31_31(<32 x i8> %a, <32 x i8> %b) {
753 ; AVX1-LABEL: @shuffle_v32i8_15_15_15_15_15_15_15_15_15_15_15_15_15_15_15_15_31_31_31_31_31_31_31_31_31_31_31_31_31_31_31_31
755 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
756 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15]
757 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
758 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
759 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
762 ; AVX2-LABEL: @shuffle_v32i8_15_15_15_15_15_15_15_15_15_15_15_15_15_15_15_15_31_31_31_31_31_31_31_31_31_31_31_31_31_31_31_31
764 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,31,31,31,31,31,31,31,31,31,31,31,31,31,31,31,31]
766 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31>
767 ret <32 x i8> %shuffle
770 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_08_08_08_08_08_08_08_08_16_16_16_16_16_16_16_16_24_24_24_24_24_24_24_24(<32 x i8> %a, <32 x i8> %b) {
771 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_08_08_08_08_08_08_08_08_16_16_16_16_16_16_16_16_24_24_24_24_24_24_24_24
773 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
774 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = [0,0,0,0,0,0,0,0,8,8,8,8,8,8,8,8]
775 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
776 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
777 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
780 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_08_08_08_08_08_08_08_08_16_16_16_16_16_16_16_16_24_24_24_24_24_24_24_24
782 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,0,0,0,0,0,0,8,8,8,8,8,8,8,8,16,16,16,16,16,16,16,16,24,24,24,24,24,24,24,24]
784 %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 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24>
785 ret <32 x i8> %shuffle
788 define <32 x i8> @shuffle_v32i8_07_07_07_07_07_07_07_07_15_15_15_15_15_15_15_15_23_23_23_23_23_23_23_23_31_31_31_31_31_31_31_31(<32 x i8> %a, <32 x i8> %b) {
789 ; AVX1-LABEL: @shuffle_v32i8_07_07_07_07_07_07_07_07_15_15_15_15_15_15_15_15_23_23_23_23_23_23_23_23_31_31_31_31_31_31_31_31
791 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
792 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = [7,7,7,7,7,7,7,7,15,15,15,15,15,15,15,15]
793 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
794 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
795 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
798 ; AVX2-LABEL: @shuffle_v32i8_07_07_07_07_07_07_07_07_15_15_15_15_15_15_15_15_23_23_23_23_23_23_23_23_31_31_31_31_31_31_31_31
800 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[7,7,7,7,7,7,7,7,15,15,15,15,15,15,15,15,23,23,23,23,23,23,23,23,31,31,31,31,31,31,31,31]
802 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31>
803 ret <32 x i8> %shuffle
806 define <32 x i8> @shuffle_v32i8_00_00_00_00_04_04_04_04_08_08_08_08_12_12_12_12_16_16_16_16_20_20_20_20_24_24_24_24_28_28_28_28(<32 x i8> %a, <32 x i8> %b) {
807 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_04_04_04_04_08_08_08_08_12_12_12_12_16_16_16_16_20_20_20_20_24_24_24_24_28_28_28_28
809 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
810 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = [0,0,0,0,4,4,4,4,8,8,8,8,12,12,12,12]
811 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
812 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
813 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
816 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_04_04_04_04_08_08_08_08_12_12_12_12_16_16_16_16_20_20_20_20_24_24_24_24_28_28_28_28
818 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,0,0,4,4,4,4,8,8,8,8,12,12,12,12,16,16,16,16,20,20,20,20,24,24,24,24,28,28,28,28]
820 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 4, i32 4, i32 4, i32 4, i32 8, i32 8, i32 8, i32 8, i32 12, i32 12, i32 12, i32 12, i32 16, i32 16, i32 16, i32 16, i32 20, i32 20, i32 20, i32 20, i32 24, i32 24, i32 24, i32 24, i32 28, i32 28, i32 28, i32 28>
821 ret <32 x i8> %shuffle
824 define <32 x i8> @shuffle_v32i8_03_03_03_03_07_07_07_07_11_11_11_11_15_15_15_15_19_19_19_19_23_23_23_23_27_27_27_27_31_31_31_31(<32 x i8> %a, <32 x i8> %b) {
825 ; AVX1-LABEL: @shuffle_v32i8_03_03_03_03_07_07_07_07_11_11_11_11_15_15_15_15_19_19_19_19_23_23_23_23_27_27_27_27_31_31_31_31
827 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
828 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = [3,3,3,3,7,7,7,7,11,11,11,11,15,15,15,15]
829 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
830 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
831 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
834 ; AVX2-LABEL: @shuffle_v32i8_03_03_03_03_07_07_07_07_11_11_11_11_15_15_15_15_19_19_19_19_23_23_23_23_27_27_27_27_31_31_31_31
836 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[3,3,3,3,7,7,7,7,11,11,11,11,15,15,15,15,19,19,19,19,23,23,23,23,27,27,27,27,31,31,31,31]
838 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 3, i32 3, i32 3, i32 3, i32 7, i32 7, i32 7, i32 7, i32 11, i32 11, i32 11, i32 11, i32 15, i32 15, i32 15, i32 15, i32 19, i32 19, i32 19, i32 19, i32 23, i32 23, i32 23, i32 23, i32 27, i32 27, i32 27, i32 27, i32 31, i32 31, i32 31, i32 31>
839 ret <32 x i8> %shuffle
842 define <32 x i8> @shuffle_v32i8_00_00_02_02_04_04_06_06_08_08_10_10_12_12_14_14_16_16_18_18_20_20_22_22_24_24_26_26_28_28_30_30(<32 x i8> %a, <32 x i8> %b) {
843 ; AVX1-LABEL: @shuffle_v32i8_00_00_02_02_04_04_06_06_08_08_10_10_12_12_14_14_16_16_18_18_20_20_22_22_24_24_26_26_28_28_30_30
845 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
846 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = [0,0,2,2,4,4,6,6,8,8,10,10,12,12,14,14]
847 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
848 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
849 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
852 ; AVX2-LABEL: @shuffle_v32i8_00_00_02_02_04_04_06_06_08_08_10_10_12_12_14_14_16_16_18_18_20_20_22_22_24_24_26_26_28_28_30_30
854 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,2,2,4,4,6,6,8,8,10,10,12,12,14,14,16,16,18,18,20,20,22,22,24,24,26,26,28,28,30,30]
856 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6, i32 8, i32 8, i32 10, i32 10, i32 12, i32 12, i32 14, i32 14, i32 16, i32 16, i32 18, i32 18, i32 20, i32 20, i32 22, i32 22, i32 24, i32 24, i32 26, i32 26, i32 28, i32 28, i32 30, i32 30>
857 ret <32 x i8> %shuffle
860 define <32 x i8> @shuffle_v32i8_01_01_03_03_05_05_07_07_09_09_11_11_13_13_15_15_17_17_19_19_21_21_23_23_25_25_27_27_29_29_31_31(<32 x i8> %a, <32 x i8> %b) {
861 ; AVX1-LABEL: @shuffle_v32i8_01_01_03_03_05_05_07_07_09_09_11_11_13_13_15_15_17_17_19_19_21_21_23_23_25_25_27_27_29_29_31_31
863 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
864 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = [1,1,3,3,5,5,7,7,9,9,11,11,13,13,15,15]
865 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
866 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
867 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
870 ; AVX2-LABEL: @shuffle_v32i8_01_01_03_03_05_05_07_07_09_09_11_11_13_13_15_15_17_17_19_19_21_21_23_23_25_25_27_27_29_29_31_31
872 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[1,1,3,3,5,5,7,7,9,9,11,11,13,13,15,15,17,17,19,19,21,21,23,23,25,25,27,27,29,29,31,31]
874 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7, i32 9, i32 9, i32 11, i32 11, i32 13, i32 13, i32 15, i32 15, i32 17, i32 17, i32 19, i32 19, i32 21, i32 21, i32 23, i32 23, i32 25, i32 25, i32 27, i32 27, i32 29, i32 29, i32 31, i32 31>
875 ret <32 x i8> %shuffle
878 define <32 x i8> @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(<32 x i8> %a, <32 x i8> %b) {
879 ; AVX1-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
881 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
882 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
885 ; 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
887 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
888 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0
890 %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>
891 ret <32 x i8> %shuffle
894 define <32 x i8> @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(<32 x i8> %a, <32 x i8> %b) {
895 ; AVX1-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
897 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
898 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
901 ; 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
903 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
904 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0
906 %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>
907 ret <32 x i8> %shuffle
910 define <32 x i8> @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(<32 x i8> %a, <32 x i8> %b) {
911 ; AVX1-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
913 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
914 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
917 ; 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
919 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
920 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0
922 %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>
923 ret <32 x i8> %shuffle
926 define <32 x i8> @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(<32 x i8> %a, <32 x i8> %b) {
927 ; AVX1-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
929 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
930 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
933 ; 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
935 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
936 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0
938 %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>
939 ret <32 x i8> %shuffle
942 define <32 x i8> @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(<32 x i8> %a, <32 x i8> %b) {
943 ; AVX1-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
945 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
946 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
949 ; 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
951 ; AVX2-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
952 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0
954 %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>
955 ret <32 x i8> %shuffle
958 define <32 x i8> @shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<32 x i8> %a, <32 x i8> %b) {
959 ; AVX1-LABEL: @shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
961 ; AVX1-NEXT: movl $15, %eax
962 ; AVX1-NEXT: vmovd %eax, %xmm1
963 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm0
964 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
967 ; AVX2-LABEL: @shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00
969 ; AVX2-NEXT: movl $15, %eax
970 ; AVX2-NEXT: vmovd %eax, %xmm1
971 ; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm0
972 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0
974 %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>
975 ret <32 x i8> %shuffle
978 define <32 x i8> @shuffle_v32i8_00_33_02_35_04_37_06_39_08_41_10_43_12_45_14_47_16_49_18_51_20_53_22_55_24_57_26_59_28_61_30_63(<32 x i8> %a, <32 x i8> %b) {
979 ; AVX1-LABEL: @shuffle_v32i8_00_33_02_35_04_37_06_39_08_41_10_43_12_45_14_47_16_49_18_51_20_53_22_55_24_57_26_59_28_61_30_63
981 ; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
982 ; AVX1-NEXT: vmovdqa {{.*}} # xmm3 = <1,3,5,7,9,11,13,15,u,u,u,u,u,u,u,u>
983 ; AVX1-NEXT: vpshufb %xmm3, %xmm2, %xmm2
984 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm4
985 ; AVX1-NEXT: vmovdqa {{.*}} # xmm5 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
986 ; AVX1-NEXT: vpshufb %xmm5, %xmm4, %xmm4
987 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm2 = xmm4[0],xmm2[0],xmm4[1],xmm2[1],xmm4[2],xmm2[2],xmm4[3],xmm2[3],xmm4[4],xmm2[4],xmm4[5],xmm2[5],xmm4[6],xmm2[6],xmm4[7],xmm2[7]
988 ; AVX1-NEXT: vpshufb %xmm3, %xmm1, %xmm1
989 ; AVX1-NEXT: vpshufb %xmm5, %xmm0, %xmm0
990 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
991 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
994 ; AVX2-LABEL: @shuffle_v32i8_00_33_02_35_04_37_06_39_08_41_10_43_12_45_14_47_16_49_18_51_20_53_22_55_24_57_26_59_28_61_30_63
996 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128]
997 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
999 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 33, i32 2, i32 35, i32 4, i32 37, i32 6, i32 39, i32 8, i32 41, i32 10, i32 43, i32 12, i32 45, i32 14, i32 47, i32 16, i32 49, i32 18, i32 51, i32 20, i32 53, i32 22, i32 55, i32 24, i32 57, i32 26, i32 59, i32 28, i32 61, i32 30, i32 63>
1000 ret <32 x i8> %shuffle
1003 define <32 x i8> @shuffle_v32i8_32_01_34_03_36_05_38_07_40_09_42_11_44_13_46_15_48_17_50_19_52_21_54_23_56_25_58_27_60_29_62_31(<32 x i8> %a, <32 x i8> %b) {
1004 ; AVX1-LABEL: @shuffle_v32i8_32_01_34_03_36_05_38_07_40_09_42_11_44_13_46_15_48_17_50_19_52_21_54_23_56_25_58_27_60_29_62_31
1006 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
1007 ; AVX1-NEXT: vmovdqa {{.*}} # xmm3 = <1,3,5,7,9,11,13,15,u,u,u,u,u,u,u,u>
1008 ; AVX1-NEXT: vpshufb %xmm3, %xmm2, %xmm2
1009 ; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm4
1010 ; AVX1-NEXT: vmovdqa {{.*}} # xmm5 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
1011 ; AVX1-NEXT: vpshufb %xmm5, %xmm4, %xmm4
1012 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm2 = xmm4[0],xmm2[0],xmm4[1],xmm2[1],xmm4[2],xmm2[2],xmm4[3],xmm2[3],xmm4[4],xmm2[4],xmm4[5],xmm2[5],xmm4[6],xmm2[6],xmm4[7],xmm2[7]
1013 ; AVX1-NEXT: vpshufb %xmm3, %xmm0, %xmm0
1014 ; AVX1-NEXT: vpshufb %xmm5, %xmm1, %xmm1
1015 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm0 = 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]
1016 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
1019 ; AVX2-LABEL: @shuffle_v32i8_32_01_34_03_36_05_38_07_40_09_42_11_44_13_46_15_48_17_50_19_52_21_54_23_56_25_58_27_60_29_62_31
1021 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128]
1022 ; AVX2-NEXT: vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
1024 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 32, i32 1, i32 34, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 29, i32 62, i32 31>
1025 ret <32 x i8> %shuffle
1028 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) {
1029 ; 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
1031 ; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
1032 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
1033 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
1034 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
1035 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
1038 ; 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
1040 ; AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
1041 ; AVX2-NEXT: vpshufb %xmm2, %xmm1, %xmm1
1042 ; AVX2-NEXT: vpshufb %xmm2, %xmm0, %xmm0
1043 ; AVX2-NEXT: vpunpcklbw {{.*}} # xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
1044 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0
1046 %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>
1047 ret <32 x i8> %shuffle
1050 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) {
1051 ; 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
1053 ; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
1054 ; AVX1-NEXT: vpxor %xmm3, %xmm3, %xmm3
1055 ; AVX1-NEXT: vpshufb %xmm3, %xmm2, %xmm2
1056 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm4
1057 ; AVX1-NEXT: vpshufb %xmm3, %xmm4, %xmm4
1058 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm2 = xmm4[0],xmm2[0],xmm4[1],xmm2[1],xmm4[2],xmm2[2],xmm4[3],xmm2[3],xmm4[4],xmm2[4],xmm4[5],xmm2[5],xmm4[6],xmm2[6],xmm4[7],xmm2[7]
1059 ; AVX1-NEXT: vpshufb %xmm3, %xmm1, %xmm1
1060 ; AVX1-NEXT: vpshufb %xmm3, %xmm0, %xmm0
1061 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
1062 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
1065 ; 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
1067 ; AVX2-NEXT: vpxor %ymm2, %ymm2, %ymm2
1068 ; AVX2-NEXT: vpshufb %ymm2, %ymm1, %ymm1
1069 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,1,0,1,0,1,0,1,0,1,0,1,0,1,0,1,16,17,16,17,16,17,16,17,16,17,16,17,16,17,16,17]
1070 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128]
1071 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
1073 %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>
1074 ret <32 x i8> %shuffle
1077 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) {
1078 ; 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
1080 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
1081 ; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3
1082 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
1083 ; AVX1-NEXT: vpshuflw {{.*}} # xmm3 = xmm3[0,0,0,0,4,5,6,7]
1084 ; AVX1-NEXT: vpblendw {{.*}} # xmm2 = xmm3[0,1,2,3],xmm2[4,5,6,7]
1085 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
1086 ; AVX1-NEXT: vpshuflw {{.*}} # xmm1 = xmm1[0,0,0,0,4,5,6,7]
1087 ; AVX1-NEXT: vpblendw {{.*}} # xmm0 = xmm1[0,1,2,3],xmm0[4,5,6,7]
1088 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
1091 ; 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
1093 ; AVX2-NEXT: vpxor %ymm2, %ymm2, %ymm2
1094 ; AVX2-NEXT: vpshufb %ymm2, %ymm1, %ymm1
1095 ; AVX2-NEXT: vpblendd {{.*}} # ymm0 = ymm1[0,1],ymm0[2,3],ymm1[4,5],ymm0[6,7]
1097 %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>
1098 ret <32 x i8> %shuffle
1101 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) {
1102 ; 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
1104 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
1105 ; AVX1-NEXT: vmovdqa {{.*}} # xmm3 = <u,u,u,u,u,u,u,u,15,14,13,12,11,10,9,8>
1106 ; AVX1-NEXT: vpshufb %xmm3, %xmm2, %xmm2
1107 ; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm4
1108 ; AVX1-NEXT: vmovdqa {{.*}} # xmm5 = <7,6,5,4,3,2,1,0,u,u,u,u,u,u,u,u>
1109 ; AVX1-NEXT: vpshufb %xmm5, %xmm4, %xmm4
1110 ; AVX1-NEXT: vpblendw {{.*}} # xmm2 = xmm4[0,1,2,3],xmm2[4,5,6,7]
1111 ; AVX1-NEXT: vpshufb %xmm3, %xmm0, %xmm0
1112 ; AVX1-NEXT: vpshufb %xmm5, %xmm1, %xmm1
1113 ; AVX1-NEXT: vpblendw {{.*}} # xmm0 = xmm1[0,1,2,3],xmm0[4,5,6,7]
1114 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
1117 ; 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
1119 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,15,14,13,12,11,10,9,8,u,u,u,u,u,u,u,u,31,30,29,28,27,26,25,24]
1120 ; AVX2-NEXT: vpshufb {{.*}} # ymm1 = ymm1[7,6,5,4,3,2,1,0,u,u,u,u,u,u,u,u,23,22,21,20,19,18,17,16,u,u,u,u,u,u,u,u]
1121 ; AVX2-NEXT: vpblendd {{.*}} # ymm0 = ymm1[0,1],ymm0[2,3],ymm1[4,5],ymm0[6,7]
1123 %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>
1124 ret <32 x i8> %shuffle
1127 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) {
1128 ; 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
1130 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
1131 ; AVX1-NEXT: vmovdqa {{.*}} # xmm3 = <u,u,u,u,u,u,u,u,7,6,5,4,3,2,1,0>
1132 ; AVX1-NEXT: vpshufb %xmm3, %xmm2, %xmm2
1133 ; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm4
1134 ; AVX1-NEXT: vmovdqa {{.*}} # xmm5 = <7,6,5,4,3,2,1,0,u,u,u,u,u,u,u,u>
1135 ; AVX1-NEXT: vpshufb %xmm5, %xmm4, %xmm4
1136 ; AVX1-NEXT: vpblendw {{.*}} # xmm2 = xmm4[0,1,2,3],xmm2[4,5,6,7]
1137 ; AVX1-NEXT: vpshufb %xmm3, %xmm0, %xmm0
1138 ; AVX1-NEXT: vpshufb %xmm5, %xmm1, %xmm1
1139 ; AVX1-NEXT: vpblendw {{.*}} # xmm0 = xmm1[0,1,2,3],xmm0[4,5,6,7]
1140 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
1143 ; 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
1145 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,u,u,u,u,u,u,u,7,6,5,4,3,2,1,0,u,u,u,u,u,u,u,u,23,22,21,20,19,18,17,16]
1146 ; AVX2-NEXT: vpshufb {{.*}} # ymm1 = ymm1[7,6,5,4,3,2,1,0,u,u,u,u,u,u,u,u,23,22,21,20,19,18,17,16,u,u,u,u,u,u,u,u]
1147 ; AVX2-NEXT: vpblendd {{.*}} # ymm0 = ymm1[0,1],ymm0[2,3],ymm1[4,5],ymm0[6,7]
1149 %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>
1150 ret <32 x i8> %shuffle
1153 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00_16_16_16_16_16_16_16_16_16_16_16_16_16_16_17_16(<32 x i8> %a, <32 x i8> %b) {
1154 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00_16_16_16_16_16_16_16_16_16_16_16_16_16_16_17_16
1156 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
1157 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
1158 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
1159 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
1160 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
1163 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00_16_16_16_16_16_16_16_16_16_16_16_16_16_16_17_16
1165 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,16,16,16,16,16,16,16,16,16,16,16,16,16,16,17,16]
1167 %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 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 17, i32 16>
1168 ret <32 x i8> %shuffle
1171 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00_16_16_16_16_16_16_16_16_16_16_16_16_16_18_16_16(<32 x i8> %a, <32 x i8> %b) {
1172 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00_16_16_16_16_16_16_16_16_16_16_16_16_16_18_16_16
1174 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
1175 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = [0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
1176 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
1177 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
1178 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
1181 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00_16_16_16_16_16_16_16_16_16_16_16_16_16_18_16_16
1183 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0,16,16,16,16,16,16,16,16,16,16,16,16,16,18,16,16]
1185 %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 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 18, i32 16, i32 16>
1186 ret <32 x i8> %shuffle
1189 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_23_16_16_16_16_16_16_16(<32 x i8> %a, <32 x i8> %b) {
1190 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_23_16_16_16_16_16_16_16
1192 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
1193 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = [0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
1194 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
1195 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
1196 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
1199 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_23_16_16_16_16_16_16_16
1201 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0,16,16,16,16,16,16,16,16,23,16,16,16,16,16,16,16]
1203 %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 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 23, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
1204 ret <32 x i8> %shuffle
1207 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_24_16_16_16_16_16_16_16_16(<32 x i8> %a, <32 x i8> %b) {
1208 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_24_16_16_16_16_16_16_16_16
1210 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
1211 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = [0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
1212 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
1213 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
1214 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
1217 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_24_16_16_16_16_16_16_16_16
1219 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0,16,16,16,16,16,16,16,24,16,16,16,16,16,16,16,16]
1221 %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 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 24, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
1222 ret <32 x i8> %shuffle
1225 define <32 x i8> @shuffle_v32i8_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_30_16_16_16_16_16_16_16_16_16_16_16_16_16_16(<32 x i8> %a, <32 x i8> %b) {
1226 ; AVX1-LABEL: @shuffle_v32i8_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_30_16_16_16_16_16_16_16_16_16_16_16_16_16_16
1228 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
1229 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = [0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
1230 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
1231 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
1232 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
1235 ; AVX2-LABEL: @shuffle_v32i8_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_30_16_16_16_16_16_16_16_16_16_16_16_16_16_16
1237 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0,16,30,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
1239 %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 16, i32 30, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
1240 ret <32 x i8> %shuffle
1243 define <32 x i8> @shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_31_16_16_16_16_16_16_16_16_16_16_16_16_16_16_16(<32 x i8> %a, <32 x i8> %b) {
1244 ; AVX1-LABEL: @shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_31_16_16_16_16_16_16_16_16_16_16_16_16_16_16_16
1246 ; AVX1-NEXT: movl $15, %eax
1247 ; AVX1-NEXT: vmovd %eax, %xmm1
1248 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
1249 ; AVX1-NEXT: vpshufb %xmm1, %xmm2, %xmm2
1250 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm0
1251 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
1254 ; AVX2-LABEL: @shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_31_16_16_16_16_16_16_16_16_16_16_16_16_16_16_16
1256 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[15,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,31,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
1258 %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 31, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
1259 ret <32 x i8> %shuffle
1262 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) {
1263 ; 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
1265 ; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
1266 ; AVX1-NEXT: vmovdqa {{.*}} # xmm3 = <0,1,2,3,4,5,6,7,u,u,u,u,u,u,u,u>
1267 ; AVX1-NEXT: vpshufb %xmm3, %xmm2, %xmm2
1268 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm4
1269 ; AVX1-NEXT: vpmovzxbw %xmm4, %xmm4
1270 ; AVX1-NEXT: vmovdqa {{.*}} # xmm5 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
1271 ; AVX1-NEXT: vpshufb %xmm5, %xmm4, %xmm4
1272 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm2 = xmm4[0],xmm2[0],xmm4[1],xmm2[1],xmm4[2],xmm2[2],xmm4[3],xmm2[3],xmm4[4],xmm2[4],xmm4[5],xmm2[5],xmm4[6],xmm2[6],xmm4[7],xmm2[7]
1273 ; AVX1-NEXT: vpshufb %xmm3, %xmm1, %xmm1
1274 ; AVX1-NEXT: vpmovzxbw %xmm0, %xmm0
1275 ; AVX1-NEXT: vpshufb %xmm5, %xmm0, %xmm0
1276 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
1277 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
1280 ; 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
1282 ; AVX2-NEXT: vpunpcklbw {{.*}} # ymm0 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[4],ymm1[4],ymm0[5],ymm1[5],ymm0[6],ymm1[6],ymm0[7],ymm1[7],ymm0[16],ymm1[16],ymm0[17],ymm1[17],ymm0[18],ymm1[18],ymm0[19],ymm1[19],ymm0[20],ymm1[20],ymm0[21],ymm1[21],ymm0[22],ymm1[22],ymm0[23],ymm1[23]
1284 %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>
1285 ret <32 x i8> %shuffle
1288 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) {
1289 ; 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
1291 ; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
1292 ; AVX1-NEXT: vmovdqa {{.*}} # xmm3 = <8,9,10,11,12,13,14,15,u,u,u,u,u,u,u,u>
1293 ; AVX1-NEXT: vpshufb %xmm3, %xmm2, %xmm2
1294 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm4
1295 ; AVX1-NEXT: vpshufb %xmm3, %xmm4, %xmm4
1296 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm2 = xmm4[0],xmm2[0],xmm4[1],xmm2[1],xmm4[2],xmm2[2],xmm4[3],xmm2[3],xmm4[4],xmm2[4],xmm4[5],xmm2[5],xmm4[6],xmm2[6],xmm4[7],xmm2[7]
1297 ; AVX1-NEXT: vpshufb %xmm3, %xmm1, %xmm1
1298 ; AVX1-NEXT: vpshufb %xmm3, %xmm0, %xmm0
1299 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
1300 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
1303 ; 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
1305 ; AVX2-NEXT: vpunpckhbw {{.*}} # ymm0 = ymm0[8],ymm1[8],ymm0[9],ymm1[9],ymm0[10],ymm1[10],ymm0[11],ymm1[11],ymm0[12],ymm1[12],ymm0[13],ymm1[13],ymm0[14],ymm1[14],ymm0[15],ymm1[15],ymm0[24],ymm1[24],ymm0[25],ymm1[25],ymm0[26],ymm1[26],ymm0[27],ymm1[27],ymm0[28],ymm1[28],ymm0[29],ymm1[29],ymm0[30],ymm1[30],ymm0[31],ymm1[31]
1307 %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>
1308 ret <32 x i8> %shuffle
1311 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) {
1312 ; 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
1314 ; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
1315 ; AVX1-NEXT: vmovdqa {{.*}} # xmm3 = <8,9,10,11,12,13,14,15,u,u,u,u,u,u,u,u>
1316 ; AVX1-NEXT: vpshufb %xmm3, %xmm2, %xmm2
1317 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm4
1318 ; AVX1-NEXT: vpshufb %xmm3, %xmm4, %xmm3
1319 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm2 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
1320 ; AVX1-NEXT: vpshufb {{.*}} # xmm1 = xmm1[0,1,2,3,4,5,6,7,u,u,u,u,u,u,u,u]
1321 ; AVX1-NEXT: vpmovzxbw %xmm0, %xmm0
1322 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
1323 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
1324 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
1327 ; 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
1329 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,u,1,u,2,u,3,u,4,u,5,u,6,u,7,u,24,u,25,u,26,u,27,u,28,u,29,u,30,u,31,u]
1330 ; AVX2-NEXT: vpshufb {{.*}} # ymm1 = ymm1[u,0,u,1,u,2,u,3,u,4,u,5,u,6,u,7,u,24,u,25,u,26,u,27,u,28,u,29,u,30,u,31]
1331 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128]
1332 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
1334 %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>
1335 ret <32 x i8> %shuffle
1338 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) {
1339 ; 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
1341 ; AVX1-NEXT: vmovdqa {{.*}} # xmm2 = <8,9,10,11,12,13,14,15,u,u,u,u,u,u,u,u>
1342 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm3
1343 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm2
1344 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm2 = xmm2[0],xmm3[0],xmm2[1],xmm3[1],xmm2[2],xmm3[2],xmm2[3],xmm3[3],xmm2[4],xmm3[4],xmm2[5],xmm3[5],xmm2[6],xmm3[6],xmm2[7],xmm3[7]
1345 ; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
1346 ; AVX1-NEXT: vpshufb {{.*}} # xmm1 = xmm1[0,1,2,3,4,5,6,7,u,u,u,u,u,u,u,u]
1347 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
1348 ; AVX1-NEXT: vpmovzxbw %xmm0, %xmm0
1349 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
1350 ; AVX1-NEXT: vpunpcklbw {{.*}} # xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
1351 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm2, %ymm0
1354 ; 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
1356 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[8,u,9,u,10,u,11,u,12,u,13,u,14,u,15,u,16,u,17,u,18,u,19,u,20,u,21,u,22,u,23,u]
1357 ; AVX2-NEXT: vpshufb {{.*}} # ymm1 = ymm1[u,8,u,9,u,10,u,11,u,12,u,13,u,14,u,15,u,16,u,17,u,18,u,19,u,20,u,21,u,22,u,23]
1358 ; AVX2-NEXT: vmovdqa {{.*}} # ymm2 = [0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128,0,128]
1359 ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0
1361 %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>
1362 ret <32 x i8> %shuffle
1365 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00_16_17_16_16_16_16_16_16_16_16_16_16_16_16_16_16(<32 x i8> %a, <32 x i8> %b) {
1366 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00_16_17_16_16_16_16_16_16_16_16_16_16_16_16_16_16
1368 ; AVX1-NEXT: vpshufb {{.*}} # xmm1 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
1369 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
1370 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
1371 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
1374 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00_16_17_16_16_16_16_16_16_16_16_16_16_16_16_16_16
1376 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,16,17,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
1378 %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 16, i32 17, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
1379 ret <32 x i8> %shuffle
1382 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00_16_16_18_16_16_16_16_16_16_16_16_16_16_16_16_16(<32 x i8> %a, <32 x i8> %b) {
1383 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00_16_16_18_16_16_16_16_16_16_16_16_16_16_16_16_16
1385 ; AVX1-NEXT: vpshufb {{.*}} # xmm1 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
1386 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
1387 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,2,0,0,0,0,0,0,0,0,0,0,0,0,0]
1388 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
1391 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00_16_16_18_16_16_16_16_16_16_16_16_16_16_16_16_16
1393 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0,16,16,18,16,16,16,16,16,16,16,16,16,16,16,16,16]
1395 %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 16, i32 16, i32 18, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
1396 ret <32 x i8> %shuffle
1399 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00_16_16_16_16_16_16_16_23_16_16_16_16_16_16_16_16(<32 x i8> %a, <32 x i8> %b) {
1400 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00_16_16_16_16_16_16_16_23_16_16_16_16_16_16_16_16
1402 ; AVX1-NEXT: vpshufb {{.*}} # xmm1 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
1403 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
1404 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,7,0,0,0,0,0,0,0,0]
1405 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
1408 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00_16_16_16_16_16_16_16_23_16_16_16_16_16_16_16_16
1410 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0,16,16,16,16,16,16,16,23,16,16,16,16,16,16,16,16]
1412 %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 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 23, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
1413 ret <32 x i8> %shuffle
1416 define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_24_16_16_16_16_16_16_16(<32 x i8> %a, <32 x i8> %b) {
1417 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_24_16_16_16_16_16_16_16
1419 ; AVX1-NEXT: vpshufb {{.*}} # xmm1 = xmm0[0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
1420 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
1421 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,8,0,0,0,0,0,0,0]
1422 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
1425 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_24_16_16_16_16_16_16_16
1427 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0,16,16,16,16,16,16,16,16,24,16,16,16,16,16,16,16]
1429 %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 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 24, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
1430 ret <32 x i8> %shuffle
1433 define <32 x i8> @shuffle_v32i8_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_16_16_16_16_16_16_30_16(<32 x i8> %a, <32 x i8> %b) {
1434 ; AVX1-LABEL: @shuffle_v32i8_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_16_16_16_16_16_16_30_16
1436 ; AVX1-NEXT: vpshufb {{.*}} # xmm1 = xmm0[0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
1437 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
1438 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,14,0]
1439 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
1442 ; AVX2-LABEL: @shuffle_v32i8_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_16_16_16_16_16_16_30_16
1444 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0,16,16,16,16,16,16,16,16,16,16,16,16,16,16,30,16]
1446 %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 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 30, i32 16>
1447 ret <32 x i8> %shuffle
1450 define <32 x i8> @shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_16_16_16_16_16_16_16_31(<32 x i8> %a, <32 x i8> %b) {
1451 ; AVX1-LABEL: @shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_16_16_16_16_16_16_16_31
1453 ; AVX1-NEXT: movl $15, %eax
1454 ; AVX1-NEXT: vmovd %eax, %xmm1
1455 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm1
1456 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
1457 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,15]
1458 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
1461 ; AVX2-LABEL: @shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_16_16_16_16_16_16_16_31
1463 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[15,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,31]
1465 %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 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 31>
1466 ret <32 x i8> %shuffle
1469 define <32 x i8> @shuffle_v32i8_00_00_00_00_04_04_04_04_08_08_08_08_12_12_12_12_28_28_28_28_24_24_24_24_20_20_20_20_16_16_16_16(<32 x i8> %a, <32 x i8> %b) {
1470 ; AVX1-LABEL: @shuffle_v32i8_00_00_00_00_04_04_04_04_08_08_08_08_12_12_12_12_28_28_28_28_24_24_24_24_20_20_20_20_16_16_16_16
1472 ; AVX1-NEXT: vpshufb {{.*}} # xmm1 = xmm0[0,0,0,0,4,4,4,4,8,8,8,8,12,12,12,12]
1473 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
1474 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[12,12,12,12,8,8,8,8,4,4,4,4,0,0,0,0]
1475 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
1478 ; AVX2-LABEL: @shuffle_v32i8_00_00_00_00_04_04_04_04_08_08_08_08_12_12_12_12_28_28_28_28_24_24_24_24_20_20_20_20_16_16_16_16
1480 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,0,0,4,4,4,4,8,8,8,8,12,12,12,12,28,28,28,28,24,24,24,24,20,20,20,20,16,16,16,16]
1482 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 4, i32 4, i32 4, i32 4, i32 8, i32 8, i32 8, i32 8, i32 12, i32 12, i32 12, i32 12, i32 28, i32 28, i32 28, i32 28, i32 24, i32 24, i32 24, i32 24, i32 20, i32 20, i32 20, i32 20, i32 16, i32 16, i32 16, i32 16>
1483 ret <32 x i8> %shuffle
1486 define <32 x i8> @shuffle_v32i8_08_08_08_08_08_08_08_08_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_24_24_24_24_24_24_24_24(<32 x i8> %a, <32 x i8> %b) {
1487 ; AVX1-LABEL: @shuffle_v32i8_08_08_08_08_08_08_08_08_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_24_24_24_24_24_24_24_24
1489 ; AVX1-NEXT: vpshufb {{.*}} # xmm1 = xmm0[8,8,8,8,8,8,8,8,0,0,0,0,0,0,0,0]
1490 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
1491 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,0,0,0,0,8,8,8,8,8,8,8,8]
1492 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
1495 ; AVX2-LABEL: @shuffle_v32i8_08_08_08_08_08_08_08_08_00_00_00_00_00_00_00_00_16_16_16_16_16_16_16_16_24_24_24_24_24_24_24_24
1497 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[8,8,8,8,8,8,8,8,0,0,0,0,0,0,0,0,16,16,16,16,16,16,16,16,24,24,24,24,24,24,24,24]
1499 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24>
1500 ret <32 x i8> %shuffle
1503 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) {
1504 ; 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
1506 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
1507 ; AVX1-NEXT: vpshufb {{.*}} # xmm1 = xmm1[0,0,0,0,u,u,u,u,u,0,0,0,0,0,14,0]
1508 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
1511 ; 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
1513 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,16,16,16,16,u,u,u,u,u,16,16,16,16,16,30,16]
1515 %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>
1516 ret <32 x i8> %shuffle
1519 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_16(<32 x i8> %a, <32 x i8> %b) {
1520 ; 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_16
1522 ; AVX1-NEXT: vpshufb {{.*}} # xmm1 = xmm0[14,14,1,1,0,0,0,0,0,0,0,0,0,0,0,0]
1523 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
1524 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,u,0,u,u,u,u,0,0,0,0,0,0,14,0]
1525 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
1528 ; 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_16
1530 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[u,14,u,u,0,0,0,0,0,0,0,0,0,0,0,0,16,16,u,16,u,u,u,u,16,16,16,16,16,16,30,16]
1532 %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 16>
1533 ret <32 x i8> %shuffle
1536 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) {
1537 ; 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
1539 ; AVX1-NEXT: vpshufb {{.*}} # xmm1 = xmm0[0,0,0,0,4,4,4,4,8,8,8,8,12,12,12,12]
1540 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
1541 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[12,12,12,12,8,8,8,8,4,4,4,4,0,0,0,0]
1542 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
1545 ; 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
1547 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[0,0,0,u,u,u,4,u,8,8,8,8,u,u,12,u,28,28,28,28,u,u,u,24,20,20,20,20,16,16,16,16]
1549 %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>
1550 ret <32 x i8> %shuffle
1553 define <32 x i8> @shuffle_v32i8_08_08_08_08_08_08_08_08_uu_uu_uu_uu_uu_uu_uu_uu_16_16_16_uu_uu_uu_uu_uu_uu_uu_24_24_24_24_24_24(<32 x i8> %a, <32 x i8> %b) {
1554 ; AVX1-LABEL: @shuffle_v32i8_08_08_08_08_08_08_08_08_uu_uu_uu_uu_uu_uu_uu_uu_16_16_16_uu_uu_uu_uu_uu_uu_uu_24_24_24_24_24_24
1556 ; AVX1-NEXT: vpunpckhbw {{.*}} # xmm1 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
1557 ; AVX1-NEXT: vpshuflw {{.*}} # xmm1 = xmm1[0,0,0,0,4,5,6,7]
1558 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
1559 ; AVX1-NEXT: vpshufb {{.*}} # xmm0 = xmm0[0,0,0,0,8,8,9,9,8,8,8,8,8,8,8,8]
1560 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
1563 ; AVX2-LABEL: @shuffle_v32i8_08_08_08_08_08_08_08_08_uu_uu_uu_uu_uu_uu_uu_uu_16_16_16_uu_uu_uu_uu_uu_uu_uu_24_24_24_24_24_24
1565 ; AVX2-NEXT: vpshufb {{.*}} # ymm0 = ymm0[8,8,8,8,8,8,8,8,u,u,u,u,u,u,u,u,16,16,16,u,u,u,u,u,u,u,24,24,24,24,24,24]
1567 %shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 16, i32 16, i32 16, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24>
1568 ret <32 x i8> %shuffle