| blob | 57a3c95f31717f5e42b54b5591ec9e84f61843b3 |
1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown | FileCheck %s --check-prefixes=ALL,SSE,SSE2
3 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+ssse3 | FileCheck %s --check-prefixes=ALL,SSE,SSSE3
4 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=ALL,SSE,SSE41
5 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=ALL,AVX,AVX1OR2,AVX1
6 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=ALL,AVX,AVX1OR2,AVX2OR512VL,AVX2,AVX2-SLOW
7 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2,+fast-variable-crosslane-shuffle,+fast-variable-perlane-shuffle | FileCheck %s --check-prefixes=ALL,AVX,AVX1OR2,AVX2OR512VL,AVX2,AVX2-FAST
8 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2,+fast-variable-perlane-shuffle | FileCheck %s --check-prefixes=ALL,AVX,AVX1OR2,AVX2OR512VL,AVX2,AVX2-FAST
9 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+avx512bw,+fast-variable-crosslane-shuffle,+fast-variable-perlane-shuffle | FileCheck %s --check-prefixes=ALL,AVX,AVX2OR512VL,AVX512VL,AVX512VLBW
10 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+avx512bw,+fast-variable-perlane-shuffle | FileCheck %s --check-prefixes=ALL,AVX,AVX2OR512VL,AVX512VL,AVX512VLBW
11 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+avx512bw,+avx512vbmi,+fast-variable-crosslane-shuffle,+fast-variable-perlane-shuffle | FileCheck %s --check-prefixes=ALL,AVX,AVX2OR512VL,AVX512VL,AVX512VLVBMI
12 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+avx512bw,+avx512vbmi,+fast-variable-perlane-shuffle | FileCheck %s --check-prefixes=ALL,AVX,AVX2OR512VL,AVX512VL,AVX512VLVBMI
13 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=ALL,AVX,AVX1OR2,XOP,XOPAVX1
14 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=ALL,AVX,AVX1OR2,XOP,XOPAVX2
16 define <16 x i8> @shuffle_v16i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<16 x i8> %a, <16 x i8> %b) {
17 ; SSE2-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
18 ; SSE2: # %bb.0:
19 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
20 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
21 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
22 ; SSE2-NEXT: retq
23 ;
24 ; SSSE3-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
25 ; SSSE3: # %bb.0:
26 ; SSSE3-NEXT: pxor %xmm1, %xmm1
27 ; SSSE3-NEXT: pshufb %xmm1, %xmm0
28 ; SSSE3-NEXT: retq
29 ;
30 ; SSE41-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
31 ; SSE41: # %bb.0:
32 ; SSE41-NEXT: pxor %xmm1, %xmm1
33 ; SSE41-NEXT: pshufb %xmm1, %xmm0
34 ; SSE41-NEXT: retq
35 ;
36 ; AVX1-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
37 ; AVX1: # %bb.0:
38 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
39 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm0
40 ; AVX1-NEXT: retq
41 ;
42 ; AVX2OR512VL-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
43 ; AVX2OR512VL: # %bb.0:
44 ; AVX2OR512VL-NEXT: vpbroadcastb %xmm0, %xmm0
45 ; AVX2OR512VL-NEXT: retq
46 ;
47 ; XOPAVX1-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
48 ; XOPAVX1: # %bb.0:
49 ; XOPAVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
50 ; XOPAVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm0
51 ; XOPAVX1-NEXT: retq
52 ;
53 ; XOPAVX2-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
54 ; XOPAVX2: # %bb.0:
55 ; XOPAVX2-NEXT: vpbroadcastb %xmm0, %xmm0
56 ; XOPAVX2-NEXT: retq
57 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 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>
58 ret <16 x i8> %shuffle
59 }
61 define <16 x i8> @shuffle_v16i8_00_00_00_00_00_00_00_00_01_01_01_01_01_01_01_01(<16 x i8> %a, <16 x i8> %b) {
62 ; SSE2-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_01_01_01_01_01_01_01_01:
63 ; SSE2: # %bb.0:
64 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
65 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,1,1,4,5,6,7]
66 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
67 ; SSE2-NEXT: retq
68 ;
69 ; SSSE3-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_01_01_01_01_01_01_01_01:
70 ; SSSE3: # %bb.0:
71 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1]
72 ; SSSE3-NEXT: retq
73 ;
74 ; SSE41-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_01_01_01_01_01_01_01_01:
75 ; SSE41: # %bb.0:
76 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1]
77 ; SSE41-NEXT: retq
78 ;
79 ; AVX-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_01_01_01_01_01_01_01_01:
80 ; AVX: # %bb.0:
81 ; AVX-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1]
82 ; AVX-NEXT: retq
83 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
84 ret <16 x i8> %shuffle
85 }
87 define <16 x i8> @shuffle_v16i8_00_00_00_00_00_00_00_00_08_08_08_08_08_08_08_08(<16 x i8> %a, <16 x i8> %b) {
88 ; SSE2-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_08_08_08_08_08_08_08_08:
89 ; SSE2: # %bb.0:
90 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
91 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
92 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
93 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,2,2,4,5,6,7]
94 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
95 ; SSE2-NEXT: retq
96 ;
97 ; SSSE3-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_08_08_08_08_08_08_08_08:
98 ; SSSE3: # %bb.0:
99 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,8,8,8,8,8,8,8,8]
100 ; SSSE3-NEXT: retq
101 ;
102 ; SSE41-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_08_08_08_08_08_08_08_08:
103 ; SSE41: # %bb.0:
104 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,8,8,8,8,8,8,8,8]
105 ; SSE41-NEXT: retq
106 ;
107 ; AVX-LABEL: shuffle_v16i8_00_00_00_00_00_00_00_00_08_08_08_08_08_08_08_08:
108 ; AVX: # %bb.0:
109 ; AVX-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,8,8,8,8,8,8,8,8]
110 ; AVX-NEXT: retq
111 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 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>
112 ret <16 x i8> %shuffle
113 }
115 define <16 x i8> @shuffle_v16i8_00_00_00_00_01_01_01_01_02_02_02_02_03_03_03_03(<16 x i8> %a, <16 x i8> %b) {
116 ; SSE-LABEL: shuffle_v16i8_00_00_00_00_01_01_01_01_02_02_02_02_03_03_03_03:
117 ; SSE: # %bb.0:
118 ; SSE-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
119 ; SSE-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
120 ; SSE-NEXT: retq
121 ;
122 ; AVX1-LABEL: shuffle_v16i8_00_00_00_00_01_01_01_01_02_02_02_02_03_03_03_03:
123 ; AVX1: # %bb.0:
124 ; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
125 ; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
126 ; AVX1-NEXT: retq
127 ;
128 ; AVX2-SLOW-LABEL: shuffle_v16i8_00_00_00_00_01_01_01_01_02_02_02_02_03_03_03_03:
129 ; AVX2-SLOW: # %bb.0:
130 ; AVX2-SLOW-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
131 ; AVX2-SLOW-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
132 ; AVX2-SLOW-NEXT: retq
133 ;
134 ; AVX2-FAST-LABEL: shuffle_v16i8_00_00_00_00_01_01_01_01_02_02_02_02_03_03_03_03:
135 ; AVX2-FAST: # %bb.0:
136 ; AVX2-FAST-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,1,1,1,1,2,2,2,2,3,3,3,3]
137 ; AVX2-FAST-NEXT: retq
138 ;
139 ; AVX512VL-LABEL: shuffle_v16i8_00_00_00_00_01_01_01_01_02_02_02_02_03_03_03_03:
140 ; AVX512VL: # %bb.0:
141 ; AVX512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,1,1,1,1,2,2,2,2,3,3,3,3]
142 ; AVX512VL-NEXT: retq
143 ;
144 ; XOP-LABEL: shuffle_v16i8_00_00_00_00_01_01_01_01_02_02_02_02_03_03_03_03:
145 ; XOP: # %bb.0:
146 ; XOP-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
147 ; XOP-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
148 ; XOP-NEXT: retq
149 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1, i32 2, i32 2, i32 2, i32 2, i32 3, i32 3, i32 3, i32 3>
150 ret <16 x i8> %shuffle
151 }
153 define <16 x i8> @shuffle_v16i8_04_04_04_04_05_05_05_05_06_06_06_06_07_07_07_07(<16 x i8> %a, <16 x i8> %b) {
154 ; SSE-LABEL: shuffle_v16i8_04_04_04_04_05_05_05_05_06_06_06_06_07_07_07_07:
155 ; SSE: # %bb.0:
156 ; SSE-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
157 ; SSE-NEXT: punpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
158 ; SSE-NEXT: retq
159 ;
160 ; AVX1-LABEL: shuffle_v16i8_04_04_04_04_05_05_05_05_06_06_06_06_07_07_07_07:
161 ; AVX1: # %bb.0:
162 ; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
163 ; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
164 ; AVX1-NEXT: retq
165 ;
166 ; AVX2-SLOW-LABEL: shuffle_v16i8_04_04_04_04_05_05_05_05_06_06_06_06_07_07_07_07:
167 ; AVX2-SLOW: # %bb.0:
168 ; AVX2-SLOW-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
169 ; AVX2-SLOW-NEXT: vpunpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
170 ; AVX2-SLOW-NEXT: retq
171 ;
172 ; AVX2-FAST-LABEL: shuffle_v16i8_04_04_04_04_05_05_05_05_06_06_06_06_07_07_07_07:
173 ; AVX2-FAST: # %bb.0:
174 ; AVX2-FAST-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[4,4,4,4,5,5,5,5,6,6,6,6,7,7,7,7]
175 ; AVX2-FAST-NEXT: retq
176 ;
177 ; AVX512VL-LABEL: shuffle_v16i8_04_04_04_04_05_05_05_05_06_06_06_06_07_07_07_07:
178 ; AVX512VL: # %bb.0:
179 ; AVX512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[4,4,4,4,5,5,5,5,6,6,6,6,7,7,7,7]
180 ; AVX512VL-NEXT: retq
181 ;
182 ; XOP-LABEL: shuffle_v16i8_04_04_04_04_05_05_05_05_06_06_06_06_07_07_07_07:
183 ; XOP: # %bb.0:
184 ; XOP-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
185 ; XOP-NEXT: vpunpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
186 ; XOP-NEXT: retq
187 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 4, i32 4, i32 4, i32 4, i32 5, i32 5, i32 5, i32 5, i32 6, i32 6, i32 6, i32 6, i32 7, i32 7, i32 7, i32 7>
188 ret <16 x i8> %shuffle
189 }
191 define <16 x i8> @shuffle_v16i8_00_00_00_00_04_04_04_04_08_08_08_08_12_12_12_12(<16 x i8> %a, <16 x i8> %b) {
192 ; SSE2-LABEL: shuffle_v16i8_00_00_00_00_04_04_04_04_08_08_08_08_12_12_12_12:
193 ; SSE2: # %bb.0:
194 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
195 ; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
196 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
197 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
198 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,2,2,4,5,6,7]
199 ; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,6,6]
200 ; SSE2-NEXT: retq
201 ;
202 ; SSSE3-LABEL: shuffle_v16i8_00_00_00_00_04_04_04_04_08_08_08_08_12_12_12_12:
203 ; SSSE3: # %bb.0:
204 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,4,4,4,4,8,8,8,8,12,12,12,12]
205 ; SSSE3-NEXT: retq
206 ;
207 ; SSE41-LABEL: shuffle_v16i8_00_00_00_00_04_04_04_04_08_08_08_08_12_12_12_12:
208 ; SSE41: # %bb.0:
209 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,4,4,4,4,8,8,8,8,12,12,12,12]
210 ; SSE41-NEXT: retq
211 ;
212 ; AVX-LABEL: shuffle_v16i8_00_00_00_00_04_04_04_04_08_08_08_08_12_12_12_12:
213 ; AVX: # %bb.0:
214 ; AVX-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,4,4,4,4,8,8,8,8,12,12,12,12]
215 ; AVX-NEXT: retq
216 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 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>
217 ret <16 x i8> %shuffle
218 }
220 define <16 x i8> @shuffle_v16i8_00_00_01_01_02_02_03_03_04_04_05_05_06_06_07_07(<16 x i8> %a, <16 x i8> %b) {
221 ; SSE-LABEL: shuffle_v16i8_00_00_01_01_02_02_03_03_04_04_05_05_06_06_07_07:
222 ; SSE: # %bb.0:
223 ; SSE-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
224 ; SSE-NEXT: retq
225 ;
226 ; AVX-LABEL: shuffle_v16i8_00_00_01_01_02_02_03_03_04_04_05_05_06_06_07_07:
227 ; AVX: # %bb.0:
228 ; AVX-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
229 ; AVX-NEXT: retq
230 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 0, i32 1, i32 1, i32 2, i32 2, i32 3, i32 3, i32 4, i32 4, i32 5, i32 5, i32 6, i32 6, i32 7, i32 7>
231 ret <16 x i8> %shuffle
232 }
234 define <16 x i8> @shuffle_v16i8_0101010101010101(<16 x i8> %a, <16 x i8> %b) {
235 ; SSE-LABEL: shuffle_v16i8_0101010101010101:
236 ; SSE: # %bb.0:
237 ; SSE-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
238 ; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
239 ; SSE-NEXT: retq
240 ;
241 ; AVX1-LABEL: shuffle_v16i8_0101010101010101:
242 ; AVX1: # %bb.0:
243 ; AVX1-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
244 ; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
245 ; AVX1-NEXT: retq
246 ;
247 ; AVX2OR512VL-LABEL: shuffle_v16i8_0101010101010101:
248 ; AVX2OR512VL: # %bb.0:
249 ; AVX2OR512VL-NEXT: vpbroadcastw %xmm0, %xmm0
250 ; AVX2OR512VL-NEXT: retq
251 ;
252 ; XOPAVX1-LABEL: shuffle_v16i8_0101010101010101:
253 ; XOPAVX1: # %bb.0:
254 ; XOPAVX1-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
255 ; XOPAVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
256 ; XOPAVX1-NEXT: retq
257 ;
258 ; XOPAVX2-LABEL: shuffle_v16i8_0101010101010101:
259 ; XOPAVX2: # %bb.0:
260 ; XOPAVX2-NEXT: vpbroadcastw %xmm0, %xmm0
261 ; XOPAVX2-NEXT: retq
262 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>
263 ret <16 x i8> %shuffle
264 }
266 define <16 x i8> @shuffle_v16i8_00_16_01_17_02_18_03_19_04_20_05_21_06_22_07_23(<16 x i8> %a, <16 x i8> %b) {
267 ; SSE-LABEL: shuffle_v16i8_00_16_01_17_02_18_03_19_04_20_05_21_06_22_07_23:
268 ; SSE: # %bb.0:
269 ; SSE-NEXT: punpcklbw {{.*#+}} 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]
270 ; SSE-NEXT: retq
271 ;
272 ; AVX-LABEL: shuffle_v16i8_00_16_01_17_02_18_03_19_04_20_05_21_06_22_07_23:
273 ; AVX: # %bb.0:
274 ; AVX-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]
275 ; AVX-NEXT: retq
276 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
277 ret <16 x i8> %shuffle
278 }
280 define <16 x i8> @shuffle_v16i8_08_24_09_25_10_26_11_27_12_28_13_29_14_30_15_31(<16 x i8> %a, <16 x i8> %b) {
281 ; SSE-LABEL: shuffle_v16i8_08_24_09_25_10_26_11_27_12_28_13_29_14_30_15_31:
282 ; SSE: # %bb.0:
283 ; SSE-NEXT: punpckhbw {{.*#+}} xmm0 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
284 ; SSE-NEXT: retq
285 ;
286 ; AVX-LABEL: shuffle_v16i8_08_24_09_25_10_26_11_27_12_28_13_29_14_30_15_31:
287 ; AVX: # %bb.0:
288 ; AVX-NEXT: vpunpckhbw {{.*#+}} xmm0 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
289 ; AVX-NEXT: retq
290 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
291 ret <16 x i8> %shuffle
292 }
294 define <16 x i8> @shuffle_v16i8_16_00_16_01_16_02_16_03_16_04_16_05_16_06_16_07(<16 x i8> %a, <16 x i8> %b) {
295 ; SSE-LABEL: shuffle_v16i8_16_00_16_01_16_02_16_03_16_04_16_05_16_06_16_07:
296 ; SSE: # %bb.0:
297 ; SSE-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
298 ; SSE-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,0,0,0,4,5,6,7]
299 ; SSE-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
300 ; SSE-NEXT: movdqa %xmm1, %xmm0
301 ; SSE-NEXT: retq
302 ;
303 ; AVX1-LABEL: shuffle_v16i8_16_00_16_01_16_02_16_03_16_04_16_05_16_06_16_07:
304 ; AVX1: # %bb.0:
305 ; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
306 ; AVX1-NEXT: vpshuflw {{.*#+}} xmm1 = xmm1[0,0,0,0,4,5,6,7]
307 ; 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]
308 ; AVX1-NEXT: retq
309 ;
310 ; AVX2OR512VL-LABEL: shuffle_v16i8_16_00_16_01_16_02_16_03_16_04_16_05_16_06_16_07:
311 ; AVX2OR512VL: # %bb.0:
312 ; AVX2OR512VL-NEXT: vpbroadcastb %xmm1, %xmm1
313 ; AVX2OR512VL-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]
314 ; AVX2OR512VL-NEXT: retq
315 ;
316 ; XOPAVX1-LABEL: shuffle_v16i8_16_00_16_01_16_02_16_03_16_04_16_05_16_06_16_07:
317 ; XOPAVX1: # %bb.0:
318 ; XOPAVX1-NEXT: vpperm {{.*#+}} xmm0 = xmm1[0],xmm0[0],xmm1[0],xmm0[1],xmm1[0],xmm0[2],xmm1[0],xmm0[3],xmm1[0],xmm0[4],xmm1[0],xmm0[5],xmm1[0],xmm0[6],xmm1[0],xmm0[7]
319 ; XOPAVX1-NEXT: retq
320 ;
321 ; XOPAVX2-LABEL: shuffle_v16i8_16_00_16_01_16_02_16_03_16_04_16_05_16_06_16_07:
322 ; XOPAVX2: # %bb.0:
323 ; XOPAVX2-NEXT: vpbroadcastb %xmm1, %xmm1
324 ; XOPAVX2-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]
325 ; XOPAVX2-NEXT: retq
326 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 16, i32 0, i32 16, i32 1, i32 16, i32 2, i32 16, i32 3, i32 16, i32 4, i32 16, i32 5, i32 16, i32 6, i32 16, i32 7>
327 ret <16 x i8> %shuffle
328 }
330 define <16 x i8> @shuffle_v16i8_15_14_13_12_11_10_09_08_07_06_05_04_03_02_01_00(<16 x i8> %a, <16 x i8> %b) {
331 ; SSE2-LABEL: shuffle_v16i8_15_14_13_12_11_10_09_08_07_06_05_04_03_02_01_00:
332 ; SSE2: # %bb.0:
333 ; SSE2-NEXT: pxor %xmm1, %xmm1
334 ; SSE2-NEXT: movdqa %xmm0, %xmm2
335 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3],xmm2[4],xmm1[4],xmm2[5],xmm1[5],xmm2[6],xmm1[6],xmm2[7],xmm1[7]
336 ; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1]
337 ; SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[3,2,1,0,4,5,6,7]
338 ; SSE2-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,7,6,5,4]
339 ; SSE2-NEXT: punpckhbw {{.*#+}} xmm0 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
340 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
341 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7]
342 ; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,6,5,4]
343 ; SSE2-NEXT: packuswb %xmm2, %xmm0
344 ; SSE2-NEXT: retq
345 ;
346 ; SSSE3-LABEL: shuffle_v16i8_15_14_13_12_11_10_09_08_07_06_05_04_03_02_01_00:
347 ; SSSE3: # %bb.0:
348 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0]
349 ; SSSE3-NEXT: retq
350 ;
351 ; SSE41-LABEL: shuffle_v16i8_15_14_13_12_11_10_09_08_07_06_05_04_03_02_01_00:
352 ; SSE41: # %bb.0:
353 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0]
354 ; SSE41-NEXT: retq
355 ;
356 ; AVX-LABEL: shuffle_v16i8_15_14_13_12_11_10_09_08_07_06_05_04_03_02_01_00:
357 ; AVX: # %bb.0:
358 ; AVX-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0]
359 ; AVX-NEXT: retq
360 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
361 ret <16 x i8> %shuffle
362 }
364 define <16 x i8> @shuffle_v16i8_07_06_05_04_03_02_01_00_15_14_13_12_11_10_09_08(<16 x i8> %a, <16 x i8> %b) {
365 ; SSE2-LABEL: shuffle_v16i8_07_06_05_04_03_02_01_00_15_14_13_12_11_10_09_08:
366 ; SSE2: # %bb.0:
367 ; SSE2-NEXT: pxor %xmm1, %xmm1
368 ; SSE2-NEXT: movdqa %xmm0, %xmm2
369 ; SSE2-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8],xmm1[8],xmm2[9],xmm1[9],xmm2[10],xmm1[10],xmm2[11],xmm1[11],xmm2[12],xmm1[12],xmm2[13],xmm1[13],xmm2[14],xmm1[14],xmm2[15],xmm1[15]
370 ; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1]
371 ; SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[3,2,1,0,4,5,6,7]
372 ; SSE2-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,7,6,5,4]
373 ; SSE2-NEXT: punpcklbw {{.*#+}} 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]
374 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
375 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7]
376 ; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,6,5,4]
377 ; SSE2-NEXT: packuswb %xmm2, %xmm0
378 ; SSE2-NEXT: retq
379 ;
380 ; SSSE3-LABEL: shuffle_v16i8_07_06_05_04_03_02_01_00_15_14_13_12_11_10_09_08:
381 ; SSSE3: # %bb.0:
382 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[7,6,5,4,3,2,1,0,15,14,13,12,11,10,9,8]
383 ; SSSE3-NEXT: retq
384 ;
385 ; SSE41-LABEL: shuffle_v16i8_07_06_05_04_03_02_01_00_15_14_13_12_11_10_09_08:
386 ; SSE41: # %bb.0:
387 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[7,6,5,4,3,2,1,0,15,14,13,12,11,10,9,8]
388 ; SSE41-NEXT: retq
389 ;
390 ; AVX-LABEL: shuffle_v16i8_07_06_05_04_03_02_01_00_15_14_13_12_11_10_09_08:
391 ; AVX: # %bb.0:
392 ; AVX-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[7,6,5,4,3,2,1,0,15,14,13,12,11,10,9,8]
393 ; AVX-NEXT: retq
394 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8>
395 ret <16 x i8> %shuffle
396 }
398 define <16 x i8> @shuffle_v16i8_03_02_01_00_07_06_05_04_11_10_09_08_15_14_13_12(<16 x i8> %a, <16 x i8> %b) {
399 ; SSE2-LABEL: shuffle_v16i8_03_02_01_00_07_06_05_04_11_10_09_08_15_14_13_12:
400 ; SSE2: # %bb.0:
401 ; SSE2-NEXT: pxor %xmm1, %xmm1
402 ; SSE2-NEXT: movdqa %xmm0, %xmm2
403 ; SSE2-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8],xmm1[8],xmm2[9],xmm1[9],xmm2[10],xmm1[10],xmm2[11],xmm1[11],xmm2[12],xmm1[12],xmm2[13],xmm1[13],xmm2[14],xmm1[14],xmm2[15],xmm1[15]
404 ; SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[3,2,1,0,4,5,6,7]
405 ; SSE2-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,7,6,5,4]
406 ; SSE2-NEXT: punpcklbw {{.*#+}} 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]
407 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7]
408 ; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,6,5,4]
409 ; SSE2-NEXT: packuswb %xmm2, %xmm0
410 ; SSE2-NEXT: retq
411 ;
412 ; SSSE3-LABEL: shuffle_v16i8_03_02_01_00_07_06_05_04_11_10_09_08_15_14_13_12:
413 ; SSSE3: # %bb.0:
414 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
415 ; SSSE3-NEXT: retq
416 ;
417 ; SSE41-LABEL: shuffle_v16i8_03_02_01_00_07_06_05_04_11_10_09_08_15_14_13_12:
418 ; SSE41: # %bb.0:
419 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
420 ; SSE41-NEXT: retq
421 ;
422 ; AVX-LABEL: shuffle_v16i8_03_02_01_00_07_06_05_04_11_10_09_08_15_14_13_12:
423 ; AVX: # %bb.0:
424 ; AVX-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
425 ; AVX-NEXT: retq
426 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4, i32 11, i32 10, i32 9, i32 8, i32 15, i32 14, i32 13, i32 12>
427 ret <16 x i8> %shuffle
428 }
430 define <16 x i8> @shuffle_v16i8_01_00_03_02_05_04_07_06_09_08_11_10_13_12_15_14(<16 x i8> %a, <16 x i8> %b) {
431 ; SSE2-LABEL: shuffle_v16i8_01_00_03_02_05_04_07_06_09_08_11_10_13_12_15_14:
432 ; SSE2: # %bb.0:
433 ; SSE2-NEXT: movdqa %xmm0, %xmm1
434 ; SSE2-NEXT: psrlw $8, %xmm1
435 ; SSE2-NEXT: psllw $8, %xmm0
436 ; SSE2-NEXT: por %xmm1, %xmm0
437 ; SSE2-NEXT: retq
438 ;
439 ; SSSE3-LABEL: shuffle_v16i8_01_00_03_02_05_04_07_06_09_08_11_10_13_12_15_14:
440 ; SSSE3: # %bb.0:
441 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,0,3,2,5,4,7,6,9,8,11,10,13,12,15,14]
442 ; SSSE3-NEXT: retq
443 ;
444 ; SSE41-LABEL: shuffle_v16i8_01_00_03_02_05_04_07_06_09_08_11_10_13_12_15_14:
445 ; SSE41: # %bb.0:
446 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,0,3,2,5,4,7,6,9,8,11,10,13,12,15,14]
447 ; SSE41-NEXT: retq
448 ;
449 ; AVX1-LABEL: shuffle_v16i8_01_00_03_02_05_04_07_06_09_08_11_10_13_12_15_14:
450 ; AVX1: # %bb.0:
451 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[1,0,3,2,5,4,7,6,9,8,11,10,13,12,15,14]
452 ; AVX1-NEXT: retq
453 ;
454 ; AVX2OR512VL-LABEL: shuffle_v16i8_01_00_03_02_05_04_07_06_09_08_11_10_13_12_15_14:
455 ; AVX2OR512VL: # %bb.0:
456 ; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[1,0,3,2,5,4,7,6,9,8,11,10,13,12,15,14]
457 ; AVX2OR512VL-NEXT: retq
458 ;
459 ; XOP-LABEL: shuffle_v16i8_01_00_03_02_05_04_07_06_09_08_11_10_13_12_15_14:
460 ; XOP: # %bb.0:
461 ; XOP-NEXT: vprotw $8, %xmm0, %xmm0
462 ; XOP-NEXT: retq
463 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 9, i32 8, i32 11, i32 10, i32 13, i32 12, i32 15, i32 14>
464 ret <16 x i8> %shuffle
465 }
467 define <16 x i8> @shuffle_v16i8_03_02_01_00_07_06_05_04_19_18_17_16_23_22_21_20(<16 x i8> %a, <16 x i8> %b) {
468 ; SSE2-LABEL: shuffle_v16i8_03_02_01_00_07_06_05_04_19_18_17_16_23_22_21_20:
469 ; SSE2: # %bb.0:
470 ; SSE2-NEXT: pxor %xmm2, %xmm2
471 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
472 ; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[3,2,1,0,4,5,6,7]
473 ; SSE2-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,7,6,5,4]
474 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
475 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7]
476 ; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,6,5,4]
477 ; SSE2-NEXT: packuswb %xmm1, %xmm0
478 ; SSE2-NEXT: retq
479 ;
480 ; SSSE3-LABEL: shuffle_v16i8_03_02_01_00_07_06_05_04_19_18_17_16_23_22_21_20:
481 ; SSSE3: # %bb.0:
482 ; SSSE3-NEXT: punpcklbw {{.*#+}} 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]
483 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[6,4,2,0,14,12,10,8,7,5,3,1,15,13,11,9]
484 ; SSSE3-NEXT: retq
485 ;
486 ; SSE41-LABEL: shuffle_v16i8_03_02_01_00_07_06_05_04_19_18_17_16_23_22_21_20:
487 ; SSE41: # %bb.0:
488 ; SSE41-NEXT: punpcklbw {{.*#+}} 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]
489 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[6,4,2,0,14,12,10,8,7,5,3,1,15,13,11,9]
490 ; SSE41-NEXT: retq
491 ;
492 ; AVX1-LABEL: shuffle_v16i8_03_02_01_00_07_06_05_04_19_18_17_16_23_22_21_20:
493 ; AVX1: # %bb.0:
494 ; 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]
495 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[6,4,2,0,14,12,10,8,7,5,3,1,15,13,11,9]
496 ; AVX1-NEXT: retq
497 ;
498 ; AVX2-LABEL: shuffle_v16i8_03_02_01_00_07_06_05_04_19_18_17_16_23_22_21_20:
499 ; AVX2: # %bb.0:
500 ; 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]
501 ; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[6,4,2,0,14,12,10,8,7,5,3,1,15,13,11,9]
502 ; AVX2-NEXT: retq
503 ;
504 ; AVX512VLBW-LABEL: shuffle_v16i8_03_02_01_00_07_06_05_04_19_18_17_16_23_22_21_20:
505 ; AVX512VLBW: # %bb.0:
506 ; AVX512VLBW-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]
507 ; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[6,4,2,0,14,12,10,8,7,5,3,1,15,13,11,9]
508 ; AVX512VLBW-NEXT: retq
509 ;
510 ; AVX512VLVBMI-LABEL: shuffle_v16i8_03_02_01_00_07_06_05_04_19_18_17_16_23_22_21_20:
511 ; AVX512VLVBMI: # %bb.0:
512 ; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm2 = [3,2,1,0,7,6,5,4,19,18,17,16,23,22,21,20]
513 ; AVX512VLVBMI-NEXT: vpermt2b %xmm1, %xmm2, %xmm0
514 ; AVX512VLVBMI-NEXT: retq
515 ;
516 ; XOP-LABEL: shuffle_v16i8_03_02_01_00_07_06_05_04_19_18_17_16_23_22_21_20:
517 ; XOP: # %bb.0:
518 ; XOP-NEXT: vpperm {{.*#+}} xmm0 = xmm0[3,2,1,0,7,6,5,4],xmm1[3,2,1,0,7,6,5,4]
519 ; XOP-NEXT: retq
520 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4, i32 19, i32 18, i32 17, i32 16, i32 23, i32 22, i32 21, i32 20>
521 ret <16 x i8> %shuffle
522 }
524 define <16 x i8> @shuffle_v16i8_03_02_01_00_31_30_29_28_11_10_09_08_23_22_21_20(<16 x i8> %a, <16 x i8> %b) {
525 ; SSE2-LABEL: shuffle_v16i8_03_02_01_00_31_30_29_28_11_10_09_08_23_22_21_20:
526 ; SSE2: # %bb.0:
527 ; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,3,2,3]
528 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
529 ; SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
530 ; SSE2-NEXT: pxor %xmm1, %xmm1
531 ; SSE2-NEXT: movdqa %xmm0, %xmm2
532 ; SSE2-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8],xmm1[8],xmm2[9],xmm1[9],xmm2[10],xmm1[10],xmm2[11],xmm1[11],xmm2[12],xmm1[12],xmm2[13],xmm1[13],xmm2[14],xmm1[14],xmm2[15],xmm1[15]
533 ; SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm2[3,2,1,0,4,5,6,7]
534 ; SSE2-NEXT: punpcklbw {{.*#+}} 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]
535 ; SSE2-NEXT: pshufhw {{.*#+}} xmm1 = xmm0[0,1,2,3,7,6,5,4]
536 ; SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm3[0],xmm1[1]
537 ; SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm0[3,2,1,0,4,5,6,7]
538 ; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm2[0,1,2,3,7,6,5,4]
539 ; SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm3[0],xmm0[1]
540 ; SSE2-NEXT: packuswb %xmm1, %xmm0
541 ; SSE2-NEXT: retq
542 ;
543 ; SSSE3-LABEL: shuffle_v16i8_03_02_01_00_31_30_29_28_11_10_09_08_23_22_21_20:
544 ; SSSE3: # %bb.0:
545 ; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = xmm1[15,14,13,12,7,6,5,4,u,u,u,u,u,u,u,u]
546 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[3,2,1,0,11,10,9,8,u,u,u,u,u,u,u,u]
547 ; SSSE3-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
548 ; SSSE3-NEXT: retq
549 ;
550 ; SSE41-LABEL: shuffle_v16i8_03_02_01_00_31_30_29_28_11_10_09_08_23_22_21_20:
551 ; SSE41: # %bb.0:
552 ; SSE41-NEXT: pshufb {{.*#+}} xmm1 = xmm1[15,14,13,12,7,6,5,4,u,u,u,u,u,u,u,u]
553 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[3,2,1,0,11,10,9,8,u,u,u,u,u,u,u,u]
554 ; SSE41-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
555 ; SSE41-NEXT: retq
556 ;
557 ; AVX1-LABEL: shuffle_v16i8_03_02_01_00_31_30_29_28_11_10_09_08_23_22_21_20:
558 ; AVX1: # %bb.0:
559 ; AVX1-NEXT: vpshufb {{.*#+}} xmm1 = xmm1[15,14,13,12,7,6,5,4,u,u,u,u,u,u,u,u]
560 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[3,2,1,0,11,10,9,8,u,u,u,u,u,u,u,u]
561 ; AVX1-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
562 ; AVX1-NEXT: retq
563 ;
564 ; AVX2-LABEL: shuffle_v16i8_03_02_01_00_31_30_29_28_11_10_09_08_23_22_21_20:
565 ; AVX2: # %bb.0:
566 ; AVX2-NEXT: vpshufb {{.*#+}} xmm1 = xmm1[15,14,13,12,7,6,5,4,u,u,u,u,u,u,u,u]
567 ; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[3,2,1,0,11,10,9,8,u,u,u,u,u,u,u,u]
568 ; AVX2-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
569 ; AVX2-NEXT: retq
570 ;
571 ; AVX512VLBW-LABEL: shuffle_v16i8_03_02_01_00_31_30_29_28_11_10_09_08_23_22_21_20:
572 ; AVX512VLBW: # %bb.0:
573 ; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm1 = xmm1[15,14,13,12,7,6,5,4,u,u,u,u,u,u,u,u]
574 ; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[3,2,1,0,11,10,9,8,u,u,u,u,u,u,u,u]
575 ; AVX512VLBW-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
576 ; AVX512VLBW-NEXT: retq
577 ;
578 ; AVX512VLVBMI-LABEL: shuffle_v16i8_03_02_01_00_31_30_29_28_11_10_09_08_23_22_21_20:
579 ; AVX512VLVBMI: # %bb.0:
580 ; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm2 = [3,2,1,0,31,30,29,28,11,10,9,8,23,22,21,20]
581 ; AVX512VLVBMI-NEXT: vpermt2b %xmm1, %xmm2, %xmm0
582 ; AVX512VLVBMI-NEXT: retq
583 ;
584 ; XOP-LABEL: shuffle_v16i8_03_02_01_00_31_30_29_28_11_10_09_08_23_22_21_20:
585 ; XOP: # %bb.0:
586 ; XOP-NEXT: vpperm {{.*#+}} xmm0 = xmm0[3,2,1,0],xmm1[15,14,13,12],xmm0[11,10,9,8],xmm1[7,6,5,4]
587 ; XOP-NEXT: retq
588 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 31, i32 30, i32 29, i32 28, i32 11, i32 10, i32 9, i32 8, i32 23, i32 22, i32 21, i32 20>
589 ret <16 x i8> %shuffle
590 }
592 define <16 x i8> @shuffle_v16i8_00_17_02_19_04_21_06_23_08_25_10_27_12_29_14_31(<16 x i8> %a, <16 x i8> %b) {
593 ; SSE2-LABEL: shuffle_v16i8_00_17_02_19_04_21_06_23_08_25_10_27_12_29_14_31:
594 ; SSE2: # %bb.0:
595 ; SSE2-NEXT: movaps {{.*#+}} xmm2 = [255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0]
596 ; SSE2-NEXT: andps %xmm2, %xmm0
597 ; SSE2-NEXT: andnps %xmm1, %xmm2
598 ; SSE2-NEXT: orps %xmm2, %xmm0
599 ; SSE2-NEXT: retq
600 ;
601 ; SSSE3-LABEL: shuffle_v16i8_00_17_02_19_04_21_06_23_08_25_10_27_12_29_14_31:
602 ; SSSE3: # %bb.0:
603 ; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = xmm1[1,3,5,7,9,11,13,15,u,u,u,u,u,u,u,u]
604 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
605 ; SSSE3-NEXT: punpcklbw {{.*#+}} 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]
606 ; SSSE3-NEXT: retq
607 ;
608 ; SSE41-LABEL: shuffle_v16i8_00_17_02_19_04_21_06_23_08_25_10_27_12_29_14_31:
609 ; SSE41: # %bb.0:
610 ; SSE41-NEXT: movdqa %xmm0, %xmm2
611 ; SSE41-NEXT: movaps {{.*#+}} xmm0 = [255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0]
612 ; SSE41-NEXT: pblendvb %xmm0, %xmm2, %xmm1
613 ; SSE41-NEXT: movdqa %xmm1, %xmm0
614 ; SSE41-NEXT: retq
615 ;
616 ; AVX1-LABEL: shuffle_v16i8_00_17_02_19_04_21_06_23_08_25_10_27_12_29_14_31:
617 ; AVX1: # %bb.0:
618 ; AVX1-NEXT: vbroadcastss {{.*#+}} xmm2 = [255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0]
619 ; AVX1-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
620 ; AVX1-NEXT: retq
621 ;
622 ; AVX2-LABEL: shuffle_v16i8_00_17_02_19_04_21_06_23_08_25_10_27_12_29_14_31:
623 ; AVX2: # %bb.0:
624 ; AVX2-NEXT: vpbroadcastw {{.*#+}} xmm2 = [255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0]
625 ; AVX2-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
626 ; AVX2-NEXT: retq
627 ;
628 ; AVX512VL-LABEL: shuffle_v16i8_00_17_02_19_04_21_06_23_08_25_10_27_12_29_14_31:
629 ; AVX512VL: # %bb.0:
630 ; AVX512VL-NEXT: movw $-21846, %ax # imm = 0xAAAA
631 ; AVX512VL-NEXT: kmovd %eax, %k1
632 ; AVX512VL-NEXT: vmovdqu8 %xmm1, %xmm0 {%k1}
633 ; AVX512VL-NEXT: retq
634 ;
635 ; XOPAVX1-LABEL: shuffle_v16i8_00_17_02_19_04_21_06_23_08_25_10_27_12_29_14_31:
636 ; XOPAVX1: # %bb.0:
637 ; XOPAVX1-NEXT: vbroadcastss {{.*#+}} xmm2 = [255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0]
638 ; XOPAVX1-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
639 ; XOPAVX1-NEXT: retq
640 ;
641 ; XOPAVX2-LABEL: shuffle_v16i8_00_17_02_19_04_21_06_23_08_25_10_27_12_29_14_31:
642 ; XOPAVX2: # %bb.0:
643 ; XOPAVX2-NEXT: vpbroadcastw {{.*#+}} xmm2 = [255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0]
644 ; XOPAVX2-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
645 ; XOPAVX2-NEXT: retq
646 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
647 ret <16 x i8> %shuffle
648 }
650 define <16 x i8> @shuffle_v16i8_00_01_02_19_04_05_06_23_08_09_10_27_12_13_14_31(<16 x i8> %a, <16 x i8> %b) {
651 ; SSE2-LABEL: shuffle_v16i8_00_01_02_19_04_05_06_23_08_09_10_27_12_13_14_31:
652 ; SSE2: # %bb.0:
653 ; SSE2-NEXT: movaps {{.*#+}} xmm2 = [255,255,255,0,255,255,255,0,255,255,255,0,255,255,255,0]
654 ; SSE2-NEXT: andps %xmm2, %xmm0
655 ; SSE2-NEXT: andnps %xmm1, %xmm2
656 ; SSE2-NEXT: orps %xmm2, %xmm0
657 ; SSE2-NEXT: retq
658 ;
659 ; SSSE3-LABEL: shuffle_v16i8_00_01_02_19_04_05_06_23_08_09_10_27_12_13_14_31:
660 ; SSSE3: # %bb.0:
661 ; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = zero,zero,zero,xmm1[3],zero,zero,zero,xmm1[7],zero,zero,zero,xmm1[11],zero,zero,zero,xmm1[15]
662 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,1,2],zero,xmm0[4,5,6],zero,xmm0[8,9,10],zero,xmm0[12,13,14],zero
663 ; SSSE3-NEXT: por %xmm1, %xmm0
664 ; SSSE3-NEXT: retq
665 ;
666 ; SSE41-LABEL: shuffle_v16i8_00_01_02_19_04_05_06_23_08_09_10_27_12_13_14_31:
667 ; SSE41: # %bb.0:
668 ; SSE41-NEXT: movdqa %xmm0, %xmm2
669 ; SSE41-NEXT: movaps {{.*#+}} xmm0 = [255,255,255,0,255,255,255,0,255,255,255,0,255,255,255,0]
670 ; SSE41-NEXT: pblendvb %xmm0, %xmm2, %xmm1
671 ; SSE41-NEXT: movdqa %xmm1, %xmm0
672 ; SSE41-NEXT: retq
673 ;
674 ; AVX1-LABEL: shuffle_v16i8_00_01_02_19_04_05_06_23_08_09_10_27_12_13_14_31:
675 ; AVX1: # %bb.0:
676 ; AVX1-NEXT: vbroadcastss {{.*#+}} xmm2 = [255,255,255,0,255,255,255,0,255,255,255,0,255,255,255,0]
677 ; AVX1-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
678 ; AVX1-NEXT: retq
679 ;
680 ; AVX2-LABEL: shuffle_v16i8_00_01_02_19_04_05_06_23_08_09_10_27_12_13_14_31:
681 ; AVX2: # %bb.0:
682 ; AVX2-NEXT: vpbroadcastd {{.*#+}} xmm2 = [255,255,255,0,255,255,255,0,255,255,255,0,255,255,255,0]
683 ; AVX2-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
684 ; AVX2-NEXT: retq
685 ;
686 ; AVX512VL-LABEL: shuffle_v16i8_00_01_02_19_04_05_06_23_08_09_10_27_12_13_14_31:
687 ; AVX512VL: # %bb.0:
688 ; AVX512VL-NEXT: movw $-30584, %ax # imm = 0x8888
689 ; AVX512VL-NEXT: kmovd %eax, %k1
690 ; AVX512VL-NEXT: vmovdqu8 %xmm1, %xmm0 {%k1}
691 ; AVX512VL-NEXT: retq
692 ;
693 ; XOPAVX1-LABEL: shuffle_v16i8_00_01_02_19_04_05_06_23_08_09_10_27_12_13_14_31:
694 ; XOPAVX1: # %bb.0:
695 ; XOPAVX1-NEXT: vbroadcastss {{.*#+}} xmm2 = [255,255,255,0,255,255,255,0,255,255,255,0,255,255,255,0]
696 ; XOPAVX1-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
697 ; XOPAVX1-NEXT: retq
698 ;
699 ; XOPAVX2-LABEL: shuffle_v16i8_00_01_02_19_04_05_06_23_08_09_10_27_12_13_14_31:
700 ; XOPAVX2: # %bb.0:
701 ; XOPAVX2-NEXT: vpbroadcastd {{.*#+}} xmm2 = [255,255,255,0,255,255,255,0,255,255,255,0,255,255,255,0]
702 ; XOPAVX2-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
703 ; XOPAVX2-NEXT: retq
704 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 19, i32 4, i32 5, i32 6, i32 23, i32 8, i32 9, i32 10, i32 27, i32 12, i32 13, i32 14, i32 31>
705 ret <16 x i8> %shuffle
706 }
708 define <16 x i8> @shuffle_v16i8_00_01_02_zz_04_05_06_zz_08_09_10_zz_12_13_14_zz(<16 x i8> %a) {
709 ; SSE-LABEL: shuffle_v16i8_00_01_02_zz_04_05_06_zz_08_09_10_zz_12_13_14_zz:
710 ; SSE: # %bb.0:
711 ; SSE-NEXT: andps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
712 ; SSE-NEXT: retq
713 ;
714 ; AVX-LABEL: shuffle_v16i8_00_01_02_zz_04_05_06_zz_08_09_10_zz_12_13_14_zz:
715 ; AVX: # %bb.0:
716 ; AVX-NEXT: vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
717 ; AVX-NEXT: retq
718 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 1, i32 2, i32 19, i32 4, i32 5, i32 6, i32 23, i32 8, i32 9, i32 10, i32 27, i32 12, i32 13, i32 14, i32 31>
719 ret <16 x i8> %shuffle
720 }
722 define <16 x i8> @shuffle_v16i8_00_01_02_03_20_05_06_23_08_09_10_11_28_13_14_31(<16 x i8> %a, <16 x i8> %b) {
723 ; SSE2-LABEL: shuffle_v16i8_00_01_02_03_20_05_06_23_08_09_10_11_28_13_14_31:
724 ; SSE2: # %bb.0:
725 ; SSE2-NEXT: movaps {{.*#+}} xmm2 = [255,255,255,255,0,255,255,0,255,255,255,255,0,255,255,0]
726 ; SSE2-NEXT: andps %xmm2, %xmm0
727 ; SSE2-NEXT: andnps %xmm1, %xmm2
728 ; SSE2-NEXT: orps %xmm2, %xmm0
729 ; SSE2-NEXT: retq
730 ;
731 ; SSSE3-LABEL: shuffle_v16i8_00_01_02_03_20_05_06_23_08_09_10_11_28_13_14_31:
732 ; SSSE3: # %bb.0:
733 ; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = zero,zero,zero,zero,xmm1[4],zero,zero,xmm1[7],zero,zero,zero,zero,xmm1[12],zero,zero,xmm1[15]
734 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,1,2,3],zero,xmm0[5,6],zero,xmm0[8,9,10,11],zero,xmm0[13,14],zero
735 ; SSSE3-NEXT: por %xmm1, %xmm0
736 ; SSSE3-NEXT: retq
737 ;
738 ; SSE41-LABEL: shuffle_v16i8_00_01_02_03_20_05_06_23_08_09_10_11_28_13_14_31:
739 ; SSE41: # %bb.0:
740 ; SSE41-NEXT: movdqa %xmm0, %xmm2
741 ; SSE41-NEXT: movaps {{.*#+}} xmm0 = [255,255,255,255,0,255,255,0,255,255,255,255,0,255,255,0]
742 ; SSE41-NEXT: pblendvb %xmm0, %xmm2, %xmm1
743 ; SSE41-NEXT: movdqa %xmm1, %xmm0
744 ; SSE41-NEXT: retq
745 ;
746 ; AVX1-LABEL: shuffle_v16i8_00_01_02_03_20_05_06_23_08_09_10_11_28_13_14_31:
747 ; AVX1: # %bb.0:
748 ; AVX1-NEXT: vmovddup {{.*#+}} xmm2 = [255,255,255,255,0,255,255,0,255,255,255,255,0,255,255,0]
749 ; AVX1-NEXT: # xmm2 = mem[0,0]
750 ; AVX1-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
751 ; AVX1-NEXT: retq
752 ;
753 ; AVX2-LABEL: shuffle_v16i8_00_01_02_03_20_05_06_23_08_09_10_11_28_13_14_31:
754 ; AVX2: # %bb.0:
755 ; AVX2-NEXT: vpbroadcastq {{.*#+}} xmm2 = [255,255,255,255,0,255,255,0,255,255,255,255,0,255,255,0]
756 ; AVX2-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
757 ; AVX2-NEXT: retq
758 ;
759 ; AVX512VL-LABEL: shuffle_v16i8_00_01_02_03_20_05_06_23_08_09_10_11_28_13_14_31:
760 ; AVX512VL: # %bb.0:
761 ; AVX512VL-NEXT: movw $-28528, %ax # imm = 0x9090
762 ; AVX512VL-NEXT: kmovd %eax, %k1
763 ; AVX512VL-NEXT: vmovdqu8 %xmm1, %xmm0 {%k1}
764 ; AVX512VL-NEXT: retq
765 ;
766 ; XOPAVX1-LABEL: shuffle_v16i8_00_01_02_03_20_05_06_23_08_09_10_11_28_13_14_31:
767 ; XOPAVX1: # %bb.0:
768 ; XOPAVX1-NEXT: vmovddup {{.*#+}} xmm2 = [255,255,255,255,0,255,255,0,255,255,255,255,0,255,255,0]
769 ; XOPAVX1-NEXT: # xmm2 = mem[0,0]
770 ; XOPAVX1-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
771 ; XOPAVX1-NEXT: retq
772 ;
773 ; XOPAVX2-LABEL: shuffle_v16i8_00_01_02_03_20_05_06_23_08_09_10_11_28_13_14_31:
774 ; XOPAVX2: # %bb.0:
775 ; XOPAVX2-NEXT: vpbroadcastq {{.*#+}} xmm2 = [255,255,255,255,0,255,255,0,255,255,255,255,0,255,255,0]
776 ; XOPAVX2-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
777 ; XOPAVX2-NEXT: retq
778 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 5, i32 6, i32 23, i32 8, i32 9, i32 10, i32 11, i32 28, i32 13, i32 14, i32 31>
779 ret <16 x i8> %shuffle
780 }
782 define <16 x i8> @shuffle_v16i8_16_17_18_19_04_05_06_07_24_25_10_11_28_13_30_15(<16 x i8> %a, <16 x i8> %b) {
783 ; SSE2-LABEL: shuffle_v16i8_16_17_18_19_04_05_06_07_24_25_10_11_28_13_30_15:
784 ; SSE2: # %bb.0:
785 ; SSE2-NEXT: movaps {{.*#+}} xmm2 = [255,255,255,255,0,0,0,0,255,255,0,0,255,0,255,0]
786 ; SSE2-NEXT: andps %xmm2, %xmm1
787 ; SSE2-NEXT: andnps %xmm0, %xmm2
788 ; SSE2-NEXT: orps %xmm1, %xmm2
789 ; SSE2-NEXT: movaps %xmm2, %xmm0
790 ; SSE2-NEXT: retq
791 ;
792 ; SSSE3-LABEL: shuffle_v16i8_16_17_18_19_04_05_06_07_24_25_10_11_28_13_30_15:
793 ; SSSE3: # %bb.0:
794 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = zero,zero,zero,zero,xmm0[4,5,6,7],zero,zero,xmm0[10,11],zero,xmm0[13],zero,xmm0[15]
795 ; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = xmm1[0,1,2,3],zero,zero,zero,zero,xmm1[8,9],zero,zero,xmm1[12],zero,xmm1[14],zero
796 ; SSSE3-NEXT: por %xmm1, %xmm0
797 ; SSSE3-NEXT: retq
798 ;
799 ; SSE41-LABEL: shuffle_v16i8_16_17_18_19_04_05_06_07_24_25_10_11_28_13_30_15:
800 ; SSE41: # %bb.0:
801 ; SSE41-NEXT: movdqa %xmm0, %xmm2
802 ; SSE41-NEXT: movaps {{.*#+}} xmm0 = [255,255,255,255,0,0,0,0,255,255,0,0,255,0,255,0]
803 ; SSE41-NEXT: pblendvb %xmm0, %xmm1, %xmm2
804 ; SSE41-NEXT: movdqa %xmm2, %xmm0
805 ; SSE41-NEXT: retq
806 ;
807 ; AVX1OR2-LABEL: shuffle_v16i8_16_17_18_19_04_05_06_07_24_25_10_11_28_13_30_15:
808 ; AVX1OR2: # %bb.0:
809 ; AVX1OR2-NEXT: vmovdqa {{.*#+}} xmm2 = [255,255,255,255,0,0,0,0,255,255,0,0,255,0,255,0]
810 ; AVX1OR2-NEXT: vpblendvb %xmm2, %xmm1, %xmm0, %xmm0
811 ; AVX1OR2-NEXT: retq
812 ;
813 ; AVX512VL-LABEL: shuffle_v16i8_16_17_18_19_04_05_06_07_24_25_10_11_28_13_30_15:
814 ; AVX512VL: # %bb.0:
815 ; AVX512VL-NEXT: movw $-21264, %ax # imm = 0xACF0
816 ; AVX512VL-NEXT: kmovd %eax, %k1
817 ; AVX512VL-NEXT: vpblendmb %xmm0, %xmm1, %xmm0 {%k1}
818 ; AVX512VL-NEXT: retq
819 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 24, i32 25, i32 10, i32 11, i32 28, i32 13, i32 30, i32 15>
820 ret <16 x i8> %shuffle
821 }
823 define <16 x i8> @shuffle_v16i8_02_20_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu(<16 x i8> %a, <16 x i8> %b) {
824 ; SSE2-LABEL: shuffle_v16i8_02_20_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu:
825 ; SSE2: # %bb.0:
826 ; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,1,1]
827 ; SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
828 ; SSE2-NEXT: psrlq $16, %xmm0
829 ; SSE2-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
830 ; SSE2-NEXT: packuswb %xmm0, %xmm0
831 ; SSE2-NEXT: retq
832 ;
833 ; SSSE3-LABEL: shuffle_v16i8_02_20_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu:
834 ; SSSE3: # %bb.0:
835 ; SSSE3-NEXT: punpcklbw {{.*#+}} 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]
836 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[4,9,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
837 ; SSSE3-NEXT: retq
838 ;
839 ; SSE41-LABEL: shuffle_v16i8_02_20_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu:
840 ; SSE41: # %bb.0:
841 ; SSE41-NEXT: punpcklbw {{.*#+}} 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]
842 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[4,9,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
843 ; SSE41-NEXT: retq
844 ;
845 ; AVX1-LABEL: shuffle_v16i8_02_20_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu:
846 ; AVX1: # %bb.0:
847 ; 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]
848 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[4,9,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
849 ; AVX1-NEXT: retq
850 ;
851 ; AVX2-LABEL: shuffle_v16i8_02_20_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu:
852 ; AVX2: # %bb.0:
853 ; 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]
854 ; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[4,9,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
855 ; AVX2-NEXT: retq
856 ;
857 ; AVX512VLBW-LABEL: shuffle_v16i8_02_20_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu:
858 ; AVX512VLBW: # %bb.0:
859 ; AVX512VLBW-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]
860 ; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[4,9,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
861 ; AVX512VLBW-NEXT: retq
862 ;
863 ; AVX512VLVBMI-LABEL: shuffle_v16i8_02_20_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu:
864 ; AVX512VLVBMI: # %bb.0:
865 ; AVX512VLVBMI-NEXT: vpbroadcastw {{.*#+}} xmm2 = [2,20,2,20,2,20,2,20,2,20,2,20,2,20,2,20]
866 ; AVX512VLVBMI-NEXT: vpermt2b %xmm1, %xmm2, %xmm0
867 ; AVX512VLVBMI-NEXT: retq
868 ;
869 ; XOP-LABEL: shuffle_v16i8_02_20_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu_uu:
870 ; XOP: # %bb.0:
871 ; XOP-NEXT: vpperm {{.*#+}} xmm0 = xmm0[2],xmm1[4],xmm0[u,u,u,u,u,u,u,u,u,u,u,u,u,u]
872 ; XOP-NEXT: retq
873 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 2, i32 20, 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>
874 ret <16 x i8> %shuffle
875 }
877 ; PR39387
878 define <16 x i8> @shuffle_v16i8_5_6_7_8_9_10_27_28_29_30_31_0_1_2_3_4(<16 x i8> %a, <16 x i8> %b) {
879 ; SSE2-LABEL: shuffle_v16i8_5_6_7_8_9_10_27_28_29_30_31_0_1_2_3_4:
880 ; SSE2: # %bb.0:
881 ; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,255]
882 ; SSE2-NEXT: movdqa %xmm0, %xmm3
883 ; SSE2-NEXT: pand %xmm2, %xmm3
884 ; SSE2-NEXT: pandn %xmm1, %xmm2
885 ; SSE2-NEXT: por %xmm3, %xmm2
886 ; SSE2-NEXT: pxor %xmm1, %xmm1
887 ; SSE2-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8],xmm1[8],xmm2[9],xmm1[9],xmm2[10],xmm1[10],xmm2[11],xmm1[11],xmm2[12],xmm1[12],xmm2[13],xmm1[13],xmm2[14],xmm1[14],xmm2[15],xmm1[15]
888 ; SSE2-NEXT: punpcklbw {{.*#+}} 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]
889 ; SSE2-NEXT: movdqa %xmm0, %xmm1
890 ; SSE2-NEXT: movdqa {{.*#+}} xmm3 = [65535,65535,65535,65535,65535,0,0,65535]
891 ; SSE2-NEXT: pand %xmm3, %xmm0
892 ; SSE2-NEXT: pandn %xmm2, %xmm3
893 ; SSE2-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,xmm2[0,1,2,3,4,5,6,7,8,9]
894 ; SSE2-NEXT: psrldq {{.*#+}} xmm1 = xmm1[10,11,12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
895 ; SSE2-NEXT: por %xmm2, %xmm1
896 ; SSE2-NEXT: por %xmm0, %xmm3
897 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm3[3,1,2,0]
898 ; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,7,5,7]
899 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,3,2,1]
900 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,0,0,3,4,5,6,7]
901 ; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,6,7,4]
902 ; SSE2-NEXT: packuswb %xmm0, %xmm1
903 ; SSE2-NEXT: movdqa %xmm1, %xmm0
904 ; SSE2-NEXT: retq
905 ;
906 ; SSSE3-LABEL: shuffle_v16i8_5_6_7_8_9_10_27_28_29_30_31_0_1_2_3_4:
907 ; SSSE3: # %bb.0:
908 ; SSSE3-NEXT: palignr {{.*#+}} xmm0 = xmm1[11,12,13,14,15],xmm0[0,1,2,3,4,5,6,7,8,9,10]
909 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[10,11,12,13,14,15,0,1,2,3,3,6,6,7,8,9]
910 ; SSSE3-NEXT: retq
911 ;
912 ; SSE41-LABEL: shuffle_v16i8_5_6_7_8_9_10_27_28_29_30_31_0_1_2_3_4:
913 ; SSE41: # %bb.0:
914 ; SSE41-NEXT: palignr {{.*#+}} xmm0 = xmm1[11,12,13,14,15],xmm0[0,1,2,3,4,5,6,7,8,9,10]
915 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[10,11,12,13,14,15,0,1,2,3,3,6,6,7,8,9]
916 ; SSE41-NEXT: retq
917 ;
918 ; AVX1-LABEL: shuffle_v16i8_5_6_7_8_9_10_27_28_29_30_31_0_1_2_3_4:
919 ; AVX1: # %bb.0:
920 ; AVX1-NEXT: vpalignr {{.*#+}} xmm0 = xmm1[11,12,13,14,15],xmm0[0,1,2,3,4,5,6,7,8,9,10]
921 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[10,11,12,13,14,15,0,1,2,3,3,6,6,7,8,9]
922 ; AVX1-NEXT: retq
923 ;
924 ; AVX2-LABEL: shuffle_v16i8_5_6_7_8_9_10_27_28_29_30_31_0_1_2_3_4:
925 ; AVX2: # %bb.0:
926 ; AVX2-NEXT: vpalignr {{.*#+}} xmm0 = xmm1[11,12,13,14,15],xmm0[0,1,2,3,4,5,6,7,8,9,10]
927 ; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[10,11,12,13,14,15,0,1,2,3,3,6,6,7,8,9]
928 ; AVX2-NEXT: retq
929 ;
930 ; AVX512VLBW-LABEL: shuffle_v16i8_5_6_7_8_9_10_27_28_29_30_31_0_1_2_3_4:
931 ; AVX512VLBW: # %bb.0:
932 ; AVX512VLBW-NEXT: vpalignr {{.*#+}} xmm0 = xmm1[11,12,13,14,15],xmm0[0,1,2,3,4,5,6,7,8,9,10]
933 ; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[10,11,12,13,14,15,0,1,2,3,3,6,6,7,8,9]
934 ; AVX512VLBW-NEXT: retq
935 ;
936 ; AVX512VLVBMI-LABEL: shuffle_v16i8_5_6_7_8_9_10_27_28_29_30_31_0_1_2_3_4:
937 ; AVX512VLVBMI: # %bb.0:
938 ; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm2 = [5,6,7,8,9,10,27,28,29,30,30,1,1,2,3,4]
939 ; AVX512VLVBMI-NEXT: vpermt2b %xmm1, %xmm2, %xmm0
940 ; AVX512VLVBMI-NEXT: retq
941 ;
942 ; XOP-LABEL: shuffle_v16i8_5_6_7_8_9_10_27_28_29_30_31_0_1_2_3_4:
943 ; XOP: # %bb.0:
944 ; XOP-NEXT: vpperm {{.*#+}} xmm0 = xmm0[5,6,7,8,9,10],xmm1[11,12,13,14,14],xmm0[1,1,2,3,4]
945 ; XOP-NEXT: retq
946 %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 27, i32 28, i32 29, i32 30, i32 30, i32 1, i32 1, i32 2, i32 3, i32 4>
947 ret <16 x i8> %1
948 }
950 define <16 x i8> @shuffle_v16i8_00_02_04_06_08_10_12_14_16_18_20_22_24_26_28_30(<16 x i8> %a, <16 x i8> %b) {
951 ; SSE-LABEL: shuffle_v16i8_00_02_04_06_08_10_12_14_16_18_20_22_24_26_28_30:
952 ; SSE: # %bb.0:
953 ; SSE-NEXT: movdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
954 ; SSE-NEXT: pand %xmm2, %xmm1
955 ; SSE-NEXT: pand %xmm2, %xmm0
956 ; SSE-NEXT: packuswb %xmm1, %xmm0
957 ; SSE-NEXT: retq
958 ;
959 ; AVX1-LABEL: shuffle_v16i8_00_02_04_06_08_10_12_14_16_18_20_22_24_26_28_30:
960 ; AVX1: # %bb.0:
961 ; AVX1-NEXT: vbroadcastss {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
962 ; AVX1-NEXT: vpand %xmm2, %xmm1, %xmm1
963 ; AVX1-NEXT: vpand %xmm2, %xmm0, %xmm0
964 ; AVX1-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
965 ; AVX1-NEXT: retq
966 ;
967 ; AVX2-LABEL: shuffle_v16i8_00_02_04_06_08_10_12_14_16_18_20_22_24_26_28_30:
968 ; AVX2: # %bb.0:
969 ; AVX2-NEXT: vpbroadcastw {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
970 ; AVX2-NEXT: vpand %xmm2, %xmm1, %xmm1
971 ; AVX2-NEXT: vpand %xmm2, %xmm0, %xmm0
972 ; AVX2-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
973 ; AVX2-NEXT: retq
974 ;
975 ; AVX512VL-LABEL: shuffle_v16i8_00_02_04_06_08_10_12_14_16_18_20_22_24_26_28_30:
976 ; AVX512VL: # %bb.0:
977 ; AVX512VL-NEXT: # kill: def $xmm0 killed $xmm0 def $ymm0
978 ; AVX512VL-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
979 ; AVX512VL-NEXT: vpmovwb %ymm0, %xmm0
980 ; AVX512VL-NEXT: vzeroupper
981 ; AVX512VL-NEXT: retq
982 ;
983 ; XOP-LABEL: shuffle_v16i8_00_02_04_06_08_10_12_14_16_18_20_22_24_26_28_30:
984 ; XOP: # %bb.0:
985 ; XOP-NEXT: vpperm {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14],xmm1[0,2,4,6,8,10,12,14]
986 ; XOP-NEXT: retq
987 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30>
988 ret <16 x i8> %shuffle
989 }
991 define <16 x i8> @shuffle_v16i8_01_03_05_07_09_11_13_15_17_19_21_23_25_27_29_31(<16 x i8> %a, <16 x i8> %b) {
992 ; SSE2-LABEL: shuffle_v16i8_01_03_05_07_09_11_13_15_17_19_21_23_25_27_29_31:
993 ; SSE2: # %bb.0:
994 ; SSE2-NEXT: psrlw $8, %xmm1
995 ; SSE2-NEXT: psrlw $8, %xmm0
996 ; SSE2-NEXT: packuswb %xmm1, %xmm0
997 ; SSE2-NEXT: retq
998 ;
999 ; SSSE3-LABEL: shuffle_v16i8_01_03_05_07_09_11_13_15_17_19_21_23_25_27_29_31:
1000 ; SSSE3: # %bb.0:
1001 ; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = <1,3,5,7,9,11,13,15,u,u,u,u,u,u,u,u>
1002 ; SSSE3-NEXT: pshufb %xmm2, %xmm1
1003 ; SSSE3-NEXT: pshufb %xmm2, %xmm0
1004 ; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
1005 ; SSSE3-NEXT: retq
1006 ;
1007 ; SSE41-LABEL: shuffle_v16i8_01_03_05_07_09_11_13_15_17_19_21_23_25_27_29_31:
1008 ; SSE41: # %bb.0:
1009 ; SSE41-NEXT: movdqa {{.*#+}} xmm2 = <1,3,5,7,9,11,13,15,u,u,u,u,u,u,u,u>
1010 ; SSE41-NEXT: pshufb %xmm2, %xmm1
1011 ; SSE41-NEXT: pshufb %xmm2, %xmm0
1012 ; SSE41-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
1013 ; SSE41-NEXT: retq
1014 ;
1015 ; AVX1-LABEL: shuffle_v16i8_01_03_05_07_09_11_13_15_17_19_21_23_25_27_29_31:
1016 ; AVX1: # %bb.0:
1017 ; AVX1-NEXT: vmovddup {{.*#+}} xmm2 = [1,3,5,7,9,11,13,15,1,3,5,7,9,11,13,15]
1018 ; AVX1-NEXT: # xmm2 = mem[0,0]
1019 ; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
1020 ; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
1021 ; AVX1-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
1022 ; AVX1-NEXT: retq
1023 ;
1024 ; AVX2-LABEL: shuffle_v16i8_01_03_05_07_09_11_13_15_17_19_21_23_25_27_29_31:
1025 ; AVX2: # %bb.0:
1026 ; AVX2-NEXT: vpbroadcastq {{.*#+}} xmm2 = [1,3,5,7,9,11,13,15,1,3,5,7,9,11,13,15]
1027 ; AVX2-NEXT: vpshufb %xmm2, %xmm1, %xmm1
1028 ; AVX2-NEXT: vpshufb %xmm2, %xmm0, %xmm0
1029 ; AVX2-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
1030 ; AVX2-NEXT: retq
1031 ;
1032 ; AVX512VLBW-LABEL: shuffle_v16i8_01_03_05_07_09_11_13_15_17_19_21_23_25_27_29_31:
1033 ; AVX512VLBW: # %bb.0:
1034 ; AVX512VLBW-NEXT: vpbroadcastq {{.*#+}} xmm2 = [1,3,5,7,9,11,13,15,1,3,5,7,9,11,13,15]
1035 ; AVX512VLBW-NEXT: vpshufb %xmm2, %xmm1, %xmm1
1036 ; AVX512VLBW-NEXT: vpshufb %xmm2, %xmm0, %xmm0
1037 ; AVX512VLBW-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
1038 ; AVX512VLBW-NEXT: retq
1039 ;
1040 ; AVX512VLVBMI-LABEL: shuffle_v16i8_01_03_05_07_09_11_13_15_17_19_21_23_25_27_29_31:
1041 ; AVX512VLVBMI: # %bb.0:
1042 ; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm2 = [1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31]
1043 ; AVX512VLVBMI-NEXT: vpermt2b %xmm1, %xmm2, %xmm0
1044 ; AVX512VLVBMI-NEXT: retq
1045 ;
1046 ; XOP-LABEL: shuffle_v16i8_01_03_05_07_09_11_13_15_17_19_21_23_25_27_29_31:
1047 ; XOP: # %bb.0:
1048 ; XOP-NEXT: vpperm {{.*#+}} xmm0 = xmm0[1,3,5,7,9,11,13,15],xmm1[1,3,5,7,9,11,13,15]
1049 ; XOP-NEXT: retq
1050 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15, i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31>
1051 ret <16 x i8> %shuffle
1052 }
1054 ; PR27780 - https://bugs.llvm.org/show_bug.cgi?id=27780
1056 define <16 x i8> @load_fold_pblendvb(ptr %px, <16 x i8> %y) {
1057 ; SSE2-LABEL: load_fold_pblendvb:
1058 ; SSE2: # %bb.0:
1059 ; SSE2-NEXT: movaps {{.*#+}} xmm1 = [255,255,0,255,0,0,0,255,255,255,0,255,0,0,0,255]
1060 ; SSE2-NEXT: andps %xmm1, %xmm0
1061 ; SSE2-NEXT: andnps (%rdi), %xmm1
1062 ; SSE2-NEXT: orps %xmm1, %xmm0
1063 ; SSE2-NEXT: retq
1064 ;
1065 ; SSSE3-LABEL: load_fold_pblendvb:
1066 ; SSSE3: # %bb.0:
1067 ; SSSE3-NEXT: movdqa (%rdi), %xmm1
1068 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,1],zero,xmm0[3],zero,zero,zero,xmm0[7,8,9],zero,xmm0[11],zero,zero,zero,xmm0[15]
1069 ; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = zero,zero,xmm1[2],zero,xmm1[4,5,6],zero,zero,zero,xmm1[10],zero,xmm1[12,13,14],zero
1070 ; SSSE3-NEXT: por %xmm1, %xmm0
1071 ; SSSE3-NEXT: retq
1072 ;
1073 ; SSE41-LABEL: load_fold_pblendvb:
1074 ; SSE41: # %bb.0:
1075 ; SSE41-NEXT: movdqa %xmm0, %xmm1
1076 ; SSE41-NEXT: movaps {{.*#+}} xmm0 = [0,0,255,0,255,255,255,0,0,0,255,0,255,255,255,0]
1077 ; SSE41-NEXT: pblendvb %xmm0, (%rdi), %xmm1
1078 ; SSE41-NEXT: movdqa %xmm1, %xmm0
1079 ; SSE41-NEXT: retq
1080 ;
1081 ; AVX1-LABEL: load_fold_pblendvb:
1082 ; AVX1: # %bb.0:
1083 ; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = [0,0,255,0,255,255,255,0,0,0,255,0,255,255,255,0]
1084 ; AVX1-NEXT: # xmm1 = mem[0,0]
1085 ; AVX1-NEXT: vpblendvb %xmm1, (%rdi), %xmm0, %xmm0
1086 ; AVX1-NEXT: retq
1087 ;
1088 ; AVX2-LABEL: load_fold_pblendvb:
1089 ; AVX2: # %bb.0:
1090 ; AVX2-NEXT: vpbroadcastq {{.*#+}} xmm1 = [0,0,255,0,255,255,255,0,0,0,255,0,255,255,255,0]
1091 ; AVX2-NEXT: vpblendvb %xmm1, (%rdi), %xmm0, %xmm0
1092 ; AVX2-NEXT: retq
1093 ;
1094 ; AVX512VL-LABEL: load_fold_pblendvb:
1095 ; AVX512VL: # %bb.0:
1096 ; AVX512VL-NEXT: movw $29812, %ax # imm = 0x7474
1097 ; AVX512VL-NEXT: kmovd %eax, %k1
1098 ; AVX512VL-NEXT: vmovdqu8 (%rdi), %xmm0 {%k1}
1099 ; AVX512VL-NEXT: retq
1100 ;
1101 ; XOPAVX1-LABEL: load_fold_pblendvb:
1102 ; XOPAVX1: # %bb.0:
1103 ; XOPAVX1-NEXT: vmovddup {{.*#+}} xmm1 = [0,0,255,0,255,255,255,0,0,0,255,0,255,255,255,0]
1104 ; XOPAVX1-NEXT: # xmm1 = mem[0,0]
1105 ; XOPAVX1-NEXT: vpblendvb %xmm1, (%rdi), %xmm0, %xmm0
1106 ; XOPAVX1-NEXT: retq
1107 ;
1108 ; XOPAVX2-LABEL: load_fold_pblendvb:
1109 ; XOPAVX2: # %bb.0:
1110 ; XOPAVX2-NEXT: vpbroadcastq {{.*#+}} xmm1 = [0,0,255,0,255,255,255,0,0,0,255,0,255,255,255,0]
1111 ; XOPAVX2-NEXT: vpblendvb %xmm1, (%rdi), %xmm0, %xmm0
1112 ; XOPAVX2-NEXT: retq
1113 %x = load <16 x i8>, ptr %px, align 16
1114 %select = shufflevector <16 x i8> %x, <16 x i8> %y, <16 x i32> <i32 16, i32 17, i32 2, i32 19, i32 4, i32 5, i32 6, i32 23, i32 24, i32 25, i32 10, i32 27, i32 12, i32 13, i32 14, i32 31>
1115 ret <16 x i8> %select
1116 }
1118 define <16 x i8> @load_fold_pblendvb_commute(ptr %px, <16 x i8> %y) {
1119 ; SSE2-LABEL: load_fold_pblendvb_commute:
1120 ; SSE2: # %bb.0:
1121 ; SSE2-NEXT: movaps {{.*#+}} xmm1 = [255,255,0,255,0,0,0,255,255,255,0,255,0,0,0,255]
1122 ; SSE2-NEXT: movaps %xmm1, %xmm2
1123 ; SSE2-NEXT: andnps %xmm0, %xmm2
1124 ; SSE2-NEXT: andps (%rdi), %xmm1
1125 ; SSE2-NEXT: orps %xmm2, %xmm1
1126 ; SSE2-NEXT: movaps %xmm1, %xmm0
1127 ; SSE2-NEXT: retq
1128 ;
1129 ; SSSE3-LABEL: load_fold_pblendvb_commute:
1130 ; SSSE3: # %bb.0:
1131 ; SSSE3-NEXT: movdqa (%rdi), %xmm1
1132 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = zero,zero,xmm0[2],zero,xmm0[4,5,6],zero,zero,zero,xmm0[10],zero,xmm0[12,13,14],zero
1133 ; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = xmm1[0,1],zero,xmm1[3],zero,zero,zero,xmm1[7,8,9],zero,xmm1[11],zero,zero,zero,xmm1[15]
1134 ; SSSE3-NEXT: por %xmm1, %xmm0
1135 ; SSSE3-NEXT: retq
1136 ;
1137 ; SSE41-LABEL: load_fold_pblendvb_commute:
1138 ; SSE41: # %bb.0:
1139 ; SSE41-NEXT: movdqa %xmm0, %xmm1
1140 ; SSE41-NEXT: movaps {{.*#+}} xmm0 = [255,255,0,255,0,0,0,255,255,255,0,255,0,0,0,255]
1141 ; SSE41-NEXT: pblendvb %xmm0, (%rdi), %xmm1
1142 ; SSE41-NEXT: movdqa %xmm1, %xmm0
1143 ; SSE41-NEXT: retq
1144 ;
1145 ; AVX1-LABEL: load_fold_pblendvb_commute:
1146 ; AVX1: # %bb.0:
1147 ; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = [255,255,0,255,0,0,0,255,255,255,0,255,0,0,0,255]
1148 ; AVX1-NEXT: # xmm1 = mem[0,0]
1149 ; AVX1-NEXT: vpblendvb %xmm1, (%rdi), %xmm0, %xmm0
1150 ; AVX1-NEXT: retq
1151 ;
1152 ; AVX2-LABEL: load_fold_pblendvb_commute:
1153 ; AVX2: # %bb.0:
1154 ; AVX2-NEXT: vpbroadcastq {{.*#+}} xmm1 = [255,255,0,255,0,0,0,255,255,255,0,255,0,0,0,255]
1155 ; AVX2-NEXT: vpblendvb %xmm1, (%rdi), %xmm0, %xmm0
1156 ; AVX2-NEXT: retq
1157 ;
1158 ; AVX512VL-LABEL: load_fold_pblendvb_commute:
1159 ; AVX512VL: # %bb.0:
1160 ; AVX512VL-NEXT: vmovdqa (%rdi), %xmm1
1161 ; AVX512VL-NEXT: movw $29812, %ax # imm = 0x7474
1162 ; AVX512VL-NEXT: kmovd %eax, %k1
1163 ; AVX512VL-NEXT: vmovdqu8 %xmm0, %xmm1 {%k1}
1164 ; AVX512VL-NEXT: vmovdqa %xmm1, %xmm0
1165 ; AVX512VL-NEXT: retq
1166 ;
1167 ; XOPAVX1-LABEL: load_fold_pblendvb_commute:
1168 ; XOPAVX1: # %bb.0:
1169 ; XOPAVX1-NEXT: vmovddup {{.*#+}} xmm1 = [255,255,0,255,0,0,0,255,255,255,0,255,0,0,0,255]
1170 ; XOPAVX1-NEXT: # xmm1 = mem[0,0]
1171 ; XOPAVX1-NEXT: vpblendvb %xmm1, (%rdi), %xmm0, %xmm0
1172 ; XOPAVX1-NEXT: retq
1173 ;
1174 ; XOPAVX2-LABEL: load_fold_pblendvb_commute:
1175 ; XOPAVX2: # %bb.0:
1176 ; XOPAVX2-NEXT: vpbroadcastq {{.*#+}} xmm1 = [255,255,0,255,0,0,0,255,255,255,0,255,0,0,0,255]
1177 ; XOPAVX2-NEXT: vpblendvb %xmm1, (%rdi), %xmm0, %xmm0
1178 ; XOPAVX2-NEXT: retq
1179 %x = load <16 x i8>, ptr %px, align 16
1180 %select = shufflevector <16 x i8> %y, <16 x i8> %x, <16 x i32> <i32 16, i32 17, i32 2, i32 19, i32 4, i32 5, i32 6, i32 23, i32 24, i32 25, i32 10, i32 27, i32 12, i32 13, i32 14, i32 31>
1181 ret <16 x i8> %select
1182 }
1184 define <16 x i8> @trunc_v4i32_shuffle(<16 x i8> %a) {
1185 ; SSE2-LABEL: trunc_v4i32_shuffle:
1186 ; SSE2: # %bb.0:
1187 ; SSE2-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
1188 ; SSE2-NEXT: packuswb %xmm0, %xmm0
1189 ; SSE2-NEXT: packuswb %xmm0, %xmm0
1190 ; SSE2-NEXT: retq
1191 ;
1192 ; SSSE3-LABEL: trunc_v4i32_shuffle:
1193 ; SSSE3: # %bb.0:
1194 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u]
1195 ; SSSE3-NEXT: retq
1196 ;
1197 ; SSE41-LABEL: trunc_v4i32_shuffle:
1198 ; SSE41: # %bb.0:
1199 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u]
1200 ; SSE41-NEXT: retq
1201 ;
1202 ; AVX1OR2-LABEL: trunc_v4i32_shuffle:
1203 ; AVX1OR2: # %bb.0:
1204 ; AVX1OR2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u]
1205 ; AVX1OR2-NEXT: retq
1206 ;
1207 ; AVX512VL-LABEL: trunc_v4i32_shuffle:
1208 ; AVX512VL: # %bb.0:
1209 ; AVX512VL-NEXT: vpmovdb %xmm0, %xmm0
1210 ; AVX512VL-NEXT: retq
1211 %shuffle = shufflevector <16 x i8> %a, <16 x i8> undef, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
1212 ret <16 x i8> %shuffle
1213 }
1215 define <16 x i8> @stress_test0(<16 x i8> %s.0.1, <16 x i8> %s.0.2, <16 x i8> %s.0.3, <16 x i8> %s.0.4, <16 x i8> %s.0.5, <16 x i8> %s.0.6, <16 x i8> %s.0.7, <16 x i8> %s.0.8, <16 x i8> %s.0.9) {
1216 ; We don't have anything useful to check here. This generates 100s of
1217 ; instructions. Instead, just make sure we survived codegen.
1218 ; ALL-LABEL: stress_test0:
1219 ; ALL: retq
1220 entry:
1221 %s.1.4 = shufflevector <16 x i8> %s.0.4, <16 x i8> %s.0.5, <16 x i32> <i32 1, i32 22, i32 21, i32 28, i32 3, i32 16, i32 6, i32 1, i32 19, i32 29, i32 12, i32 31, i32 2, i32 3, i32 3, i32 6>
1222 %s.1.5 = shufflevector <16 x i8> %s.0.5, <16 x i8> %s.0.6, <16 x i32> <i32 31, i32 20, i32 12, i32 19, i32 2, i32 15, i32 12, i32 31, i32 2, i32 28, i32 2, i32 30, i32 7, i32 8, i32 17, i32 28>
1223 %s.1.8 = shufflevector <16 x i8> %s.0.8, <16 x i8> %s.0.9, <16 x i32> <i32 14, i32 10, i32 17, i32 5, i32 17, i32 9, i32 17, i32 21, i32 31, i32 24, i32 16, i32 6, i32 20, i32 28, i32 23, i32 8>
1224 %s.2.2 = shufflevector <16 x i8> %s.0.3, <16 x i8> %s.0.4, <16 x i32> <i32 20, i32 9, i32 21, i32 11, i32 11, i32 4, i32 3, i32 18, i32 3, i32 30, i32 4, i32 31, i32 11, i32 24, i32 13, i32 29>
1225 %s.3.2 = shufflevector <16 x i8> %s.2.2, <16 x i8> %s.1.4, <16 x i32> <i32 15, i32 13, i32 5, i32 11, i32 7, i32 17, i32 14, i32 22, i32 22, i32 16, i32 7, i32 24, i32 16, i32 22, i32 7, i32 29>
1226 %s.5.4 = shufflevector <16 x i8> %s.1.5, <16 x i8> %s.1.8, <16 x i32> <i32 3, i32 13, i32 19, i32 7, i32 23, i32 11, i32 1, i32 9, i32 16, i32 25, i32 2, i32 7, i32 0, i32 21, i32 23, i32 17>
1227 %s.6.1 = shufflevector <16 x i8> %s.3.2, <16 x i8> %s.3.2, <16 x i32> <i32 11, i32 2, i32 28, i32 31, i32 27, i32 3, i32 9, i32 27, i32 25, i32 25, i32 14, i32 7, i32 12, i32 28, i32 12, i32 23>
1228 %s.7.1 = shufflevector <16 x i8> %s.6.1, <16 x i8> %s.3.2, <16 x i32> <i32 15, i32 29, i32 14, i32 0, i32 29, i32 15, i32 26, i32 30, i32 6, i32 7, i32 2, i32 8, i32 12, i32 10, i32 29, i32 17>
1229 %s.7.2 = shufflevector <16 x i8> %s.3.2, <16 x i8> %s.5.4, <16 x i32> <i32 3, i32 29, i32 3, i32 19, i32 undef, i32 20, i32 undef, i32 3, i32 27, i32 undef, i32 undef, i32 11, i32 undef, i32 undef, i32 undef, i32 undef>
1230 %s.16.0 = shufflevector <16 x i8> %s.7.1, <16 x i8> %s.7.2, <16 x i32> <i32 13, i32 1, i32 16, i32 16, i32 6, i32 7, i32 29, i32 18, i32 19, i32 28, i32 undef, i32 undef, i32 31, i32 1, i32 undef, i32 10>
1231 ret <16 x i8> %s.16.0
1232 }
1234 define <16 x i8> @undef_test1(<16 x i8> %s.0.5, <16 x i8> %s.0.8, <16 x i8> %s.0.9) noinline nounwind {
1235 ; There is nothing interesting to check about these instructions other than
1236 ; that they survive codegen. However, we actually do better and delete all of
1237 ; them because the result is 'undef'.
1238 ;
1239 ; ALL-LABEL: undef_test1:
1240 ; ALL: # %bb.0: # %entry
1241 ; ALL-NEXT: retq
1242 entry:
1243 %s.1.8 = shufflevector <16 x i8> %s.0.8, <16 x i8> undef, <16 x i32> <i32 9, i32 9, i32 undef, i32 undef, i32 undef, i32 2, i32 undef, i32 6, i32 undef, i32 6, i32 undef, i32 14, i32 14, i32 undef, i32 undef, i32 0>
1244 %s.2.4 = shufflevector <16 x i8> undef, <16 x i8> %s.0.5, <16 x i32> <i32 21, i32 undef, i32 undef, i32 19, i32 undef, i32 undef, i32 29, i32 24, i32 21, i32 23, i32 21, i32 17, i32 19, i32 undef, i32 20, i32 22>
1245 %s.2.5 = shufflevector <16 x i8> %s.0.5, <16 x i8> undef, <16 x i32> <i32 3, i32 8, i32 undef, i32 7, i32 undef, i32 10, i32 8, i32 0, i32 15, i32 undef, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 9>
1246 %s.2.9 = shufflevector <16 x i8> %s.0.9, <16 x i8> undef, <16 x i32> <i32 7, i32 undef, i32 14, i32 7, i32 8, i32 undef, i32 7, i32 8, i32 5, i32 15, i32 undef, i32 1, i32 11, i32 undef, i32 undef, i32 11>
1247 %s.3.4 = shufflevector <16 x i8> %s.2.4, <16 x i8> %s.0.5, <16 x i32> <i32 5, i32 0, i32 21, i32 6, i32 15, i32 27, i32 22, i32 21, i32 4, i32 22, i32 19, i32 26, i32 9, i32 26, i32 8, i32 29>
1248 %s.3.9 = shufflevector <16 x i8> %s.2.9, <16 x i8> undef, <16 x i32> <i32 8, i32 6, i32 8, i32 1, i32 undef, i32 4, i32 undef, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 6, i32 undef>
1249 %s.4.7 = shufflevector <16 x i8> %s.1.8, <16 x i8> %s.2.9, <16 x i32> <i32 9, i32 0, i32 22, i32 20, i32 24, i32 7, i32 21, i32 17, i32 20, i32 12, i32 19, i32 23, i32 2, i32 9, i32 17, i32 10>
1250 %s.4.8 = shufflevector <16 x i8> %s.2.9, <16 x i8> %s.3.9, <16 x i32> <i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 6, i32 10, i32 undef, i32 0, i32 5, i32 undef, i32 9, i32 undef>
1251 %s.5.7 = shufflevector <16 x i8> %s.4.7, <16 x i8> %s.4.8, <16 x i32> <i32 16, 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>
1252 %s.8.4 = shufflevector <16 x i8> %s.3.4, <16 x i8> %s.5.7, <16 x i32> <i32 undef, i32 undef, i32 undef, i32 28, i32 undef, i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
1253 %s.9.4 = shufflevector <16 x i8> %s.8.4, <16 x i8> undef, <16 x i32> <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 10, i32 5>
1254 %s.10.4 = shufflevector <16 x i8> %s.9.4, <16 x i8> undef, <16 x i32> <i32 undef, i32 7, 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>
1255 %s.12.4 = shufflevector <16 x i8> %s.10.4, <16 x i8> undef, <16 x i32> <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 13, i32 undef, i32 undef, i32 undef>
1257 ret <16 x i8> %s.12.4
1258 }
1260 define <16 x i8> @PR20540(<8 x i8> %a) {
1261 ; SSE-LABEL: PR20540:
1262 ; SSE: # %bb.0:
1263 ; SSE-NEXT: movq {{.*#+}} xmm0 = xmm0[0],zero
1264 ; SSE-NEXT: retq
1265 ;
1266 ; AVX-LABEL: PR20540:
1267 ; AVX: # %bb.0:
1268 ; AVX-NEXT: vmovq {{.*#+}} xmm0 = xmm0[0],zero
1269 ; AVX-NEXT: retq
1270 %shuffle = shufflevector <8 x i8> %a, <8 x i8> zeroinitializer, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8>
1271 ret <16 x i8> %shuffle
1272 }
1274 define <16 x i8> @shuffle_v16i8_16_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz(i8 %i) {
1275 ; SSE-LABEL: shuffle_v16i8_16_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
1276 ; SSE: # %bb.0:
1277 ; SSE-NEXT: movzbl %dil, %eax
1278 ; SSE-NEXT: movd %eax, %xmm0
1279 ; SSE-NEXT: retq
1280 ;
1281 ; AVX-LABEL: shuffle_v16i8_16_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
1282 ; AVX: # %bb.0:
1283 ; AVX-NEXT: movzbl %dil, %eax
1284 ; AVX-NEXT: vmovd %eax, %xmm0
1285 ; AVX-NEXT: retq
1286 %a = insertelement <16 x i8> undef, i8 %i, i32 0
1287 %shuffle = shufflevector <16 x i8> zeroinitializer, <16 x i8> %a, <16 x i32> <i32 16, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
1288 ret <16 x i8> %shuffle
1289 }
1291 define <16 x i8> @shuffle_v16i8_zz_zz_zz_zz_zz_16_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz(i8 %i) {
1292 ; SSE2-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_16_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
1293 ; SSE2: # %bb.0:
1294 ; SSE2-NEXT: shll $8, %edi
1295 ; SSE2-NEXT: pxor %xmm0, %xmm0
1296 ; SSE2-NEXT: pinsrw $2, %edi, %xmm0
1297 ; SSE2-NEXT: retq
1298 ;
1299 ; SSSE3-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_16_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
1300 ; SSSE3: # %bb.0:
1301 ; SSSE3-NEXT: shll $8, %edi
1302 ; SSSE3-NEXT: pxor %xmm0, %xmm0
1303 ; SSSE3-NEXT: pinsrw $2, %edi, %xmm0
1304 ; SSSE3-NEXT: retq
1305 ;
1306 ; SSE41-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_16_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
1307 ; SSE41: # %bb.0:
1308 ; SSE41-NEXT: pxor %xmm0, %xmm0
1309 ; SSE41-NEXT: pinsrb $5, %edi, %xmm0
1310 ; SSE41-NEXT: retq
1311 ;
1312 ; AVX-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_16_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
1313 ; AVX: # %bb.0:
1314 ; AVX-NEXT: vpxor %xmm0, %xmm0, %xmm0
1315 ; AVX-NEXT: vpinsrb $5, %edi, %xmm0, %xmm0
1316 ; AVX-NEXT: retq
1317 %a = insertelement <16 x i8> undef, i8 %i, i32 0
1318 %shuffle = shufflevector <16 x i8> zeroinitializer, <16 x i8> %a, <16 x i32> <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>
1319 ret <16 x i8> %shuffle
1320 }
1322 define <16 x i8> @shuffle_v16i8_zz_uu_uu_zz_uu_uu_zz_zz_zz_zz_zz_zz_zz_zz_zz_16(i8 %i) {
1323 ; SSE2-LABEL: shuffle_v16i8_zz_uu_uu_zz_uu_uu_zz_zz_zz_zz_zz_zz_zz_zz_zz_16:
1324 ; SSE2: # %bb.0:
1325 ; SSE2-NEXT: shll $8, %edi
1326 ; SSE2-NEXT: pxor %xmm0, %xmm0
1327 ; SSE2-NEXT: pinsrw $7, %edi, %xmm0
1328 ; SSE2-NEXT: retq
1329 ;
1330 ; SSSE3-LABEL: shuffle_v16i8_zz_uu_uu_zz_uu_uu_zz_zz_zz_zz_zz_zz_zz_zz_zz_16:
1331 ; SSSE3: # %bb.0:
1332 ; SSSE3-NEXT: shll $8, %edi
1333 ; SSSE3-NEXT: pxor %xmm0, %xmm0
1334 ; SSSE3-NEXT: pinsrw $7, %edi, %xmm0
1335 ; SSSE3-NEXT: retq
1336 ;
1337 ; SSE41-LABEL: shuffle_v16i8_zz_uu_uu_zz_uu_uu_zz_zz_zz_zz_zz_zz_zz_zz_zz_16:
1338 ; SSE41: # %bb.0:
1339 ; SSE41-NEXT: pxor %xmm0, %xmm0
1340 ; SSE41-NEXT: pinsrb $15, %edi, %xmm0
1341 ; SSE41-NEXT: retq
1342 ;
1343 ; AVX-LABEL: shuffle_v16i8_zz_uu_uu_zz_uu_uu_zz_zz_zz_zz_zz_zz_zz_zz_zz_16:
1344 ; AVX: # %bb.0:
1345 ; AVX-NEXT: vpxor %xmm0, %xmm0, %xmm0
1346 ; AVX-NEXT: vpinsrb $15, %edi, %xmm0, %xmm0
1347 ; AVX-NEXT: retq
1348 %a = insertelement <16 x i8> undef, i8 %i, i32 0
1349 %shuffle = shufflevector <16 x i8> zeroinitializer, <16 x i8> %a, <16 x i32> <i32 0, i32 undef, i32 undef, i32 3, i32 undef, i32 undef, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 16>
1350 ret <16 x i8> %shuffle
1351 }
1353 define <16 x i8> @shuffle_v16i8_zz_zz_19_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz(i8 %i) {
1354 ; SSE2-LABEL: shuffle_v16i8_zz_zz_19_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
1355 ; SSE2: # %bb.0:
1356 ; SSE2-NEXT: movzbl %dil, %eax
1357 ; SSE2-NEXT: pxor %xmm0, %xmm0
1358 ; SSE2-NEXT: pinsrw $1, %eax, %xmm0
1359 ; SSE2-NEXT: retq
1360 ;
1361 ; SSSE3-LABEL: shuffle_v16i8_zz_zz_19_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
1362 ; SSSE3: # %bb.0:
1363 ; SSSE3-NEXT: movzbl %dil, %eax
1364 ; SSSE3-NEXT: pxor %xmm0, %xmm0
1365 ; SSSE3-NEXT: pinsrw $1, %eax, %xmm0
1366 ; SSSE3-NEXT: retq
1367 ;
1368 ; SSE41-LABEL: shuffle_v16i8_zz_zz_19_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
1369 ; SSE41: # %bb.0:
1370 ; SSE41-NEXT: pxor %xmm0, %xmm0
1371 ; SSE41-NEXT: pinsrb $2, %edi, %xmm0
1372 ; SSE41-NEXT: retq
1373 ;
1374 ; AVX-LABEL: shuffle_v16i8_zz_zz_19_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
1375 ; AVX: # %bb.0:
1376 ; AVX-NEXT: vpxor %xmm0, %xmm0, %xmm0
1377 ; AVX-NEXT: vpinsrb $2, %edi, %xmm0, %xmm0
1378 ; AVX-NEXT: retq
1379 %a = insertelement <16 x i8> undef, i8 %i, i32 3
1380 %shuffle = shufflevector <16 x i8> zeroinitializer, <16 x i8> %a, <16 x i32> <i32 0, i32 1, i32 19, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
1381 ret <16 x i8> %shuffle
1382 }
1384 define <16 x i8> @shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_16_uu_18_uu(<16 x i8> %a) {
1385 ; SSE-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_16_uu_18_uu:
1386 ; SSE: # %bb.0:
1387 ; SSE-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3]
1388 ; SSE-NEXT: retq
1389 ;
1390 ; AVX-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_16_uu_18_uu:
1391 ; AVX: # %bb.0:
1392 ; AVX-NEXT: vpslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3]
1393 ; AVX-NEXT: retq
1394 %shuffle = shufflevector <16 x i8> zeroinitializer, <16 x i8> %a, <16 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 16, i32 undef, i32 18, i32 undef>
1395 ret <16 x i8> %shuffle
1396 }
1398 define <16 x i8> @shuffle_v16i8_28_uu_30_31_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz(<16 x i8> %a) {
1399 ; SSE-LABEL: shuffle_v16i8_28_uu_30_31_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
1400 ; SSE: # %bb.0:
1401 ; SSE-NEXT: psrldq {{.*#+}} xmm0 = xmm0[12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
1402 ; SSE-NEXT: retq
1403 ;
1404 ; AVX-LABEL: shuffle_v16i8_28_uu_30_31_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
1405 ; AVX: # %bb.0:
1406 ; AVX-NEXT: vpsrldq {{.*#+}} xmm0 = xmm0[12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
1407 ; AVX-NEXT: retq
1408 %shuffle = shufflevector <16 x i8> zeroinitializer, <16 x i8> %a, <16 x i32> <i32 28, i32 undef, i32 30, i32 31, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 09, i32 0, i32 0, i32 0, i32 0, i32 0>
1409 ret <16 x i8> %shuffle
1410 }
1412 define <16 x i8> @shuffle_v16i8_31_00_01_02_03_04_05_06_07_08_09_10_11_12_13_14(<16 x i8> %a, <16 x i8> %b) {
1413 ; SSE2-LABEL: shuffle_v16i8_31_00_01_02_03_04_05_06_07_08_09_10_11_12_13_14:
1414 ; SSE2: # %bb.0:
1415 ; SSE2-NEXT: psrldq {{.*#+}} xmm1 = xmm1[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
1416 ; SSE2-NEXT: pslldq {{.*#+}} xmm0 = zero,xmm0[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14]
1417 ; SSE2-NEXT: por %xmm1, %xmm0
1418 ; SSE2-NEXT: retq
1419 ;
1420 ; SSSE3-LABEL: shuffle_v16i8_31_00_01_02_03_04_05_06_07_08_09_10_11_12_13_14:
1421 ; SSSE3: # %bb.0:
1422 ; SSSE3-NEXT: palignr {{.*#+}} xmm0 = xmm1[15],xmm0[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14]
1423 ; SSSE3-NEXT: retq
1424 ;
1425 ; SSE41-LABEL: shuffle_v16i8_31_00_01_02_03_04_05_06_07_08_09_10_11_12_13_14:
1426 ; SSE41: # %bb.0:
1427 ; SSE41-NEXT: palignr {{.*#+}} xmm0 = xmm1[15],xmm0[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14]
1428 ; SSE41-NEXT: retq
1429 ;
1430 ; AVX-LABEL: shuffle_v16i8_31_00_01_02_03_04_05_06_07_08_09_10_11_12_13_14:
1431 ; AVX: # %bb.0:
1432 ; AVX-NEXT: vpalignr {{.*#+}} xmm0 = xmm1[15],xmm0[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14]
1433 ; AVX-NEXT: retq
1434 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 31, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14>
1435 ret <16 x i8> %shuffle
1436 }
1438 define <16 x i8> @shuffle_v16i8_15_00_01_02_03_04_05_06_07_08_09_10_11_12_13_14(<16 x i8> %a, <16 x i8> %b) {
1439 ; SSE2-LABEL: shuffle_v16i8_15_00_01_02_03_04_05_06_07_08_09_10_11_12_13_14:
1440 ; SSE2: # %bb.0:
1441 ; SSE2-NEXT: movdqa %xmm0, %xmm1
1442 ; SSE2-NEXT: psrldq {{.*#+}} xmm1 = xmm1[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
1443 ; SSE2-NEXT: pslldq {{.*#+}} xmm0 = zero,xmm0[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14]
1444 ; SSE2-NEXT: por %xmm1, %xmm0
1445 ; SSE2-NEXT: retq
1446 ;
1447 ; SSSE3-LABEL: shuffle_v16i8_15_00_01_02_03_04_05_06_07_08_09_10_11_12_13_14:
1448 ; SSSE3: # %bb.0:
1449 ; SSSE3-NEXT: palignr {{.*#+}} xmm0 = xmm0[15,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14]
1450 ; SSSE3-NEXT: retq
1451 ;
1452 ; SSE41-LABEL: shuffle_v16i8_15_00_01_02_03_04_05_06_07_08_09_10_11_12_13_14:
1453 ; SSE41: # %bb.0:
1454 ; SSE41-NEXT: palignr {{.*#+}} xmm0 = xmm0[15,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14]
1455 ; SSE41-NEXT: retq
1456 ;
1457 ; AVX-LABEL: shuffle_v16i8_15_00_01_02_03_04_05_06_07_08_09_10_11_12_13_14:
1458 ; AVX: # %bb.0:
1459 ; AVX-NEXT: vpalignr {{.*#+}} xmm0 = xmm0[15,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14]
1460 ; AVX-NEXT: retq
1461 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 15, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14>
1462 ret <16 x i8> %shuffle
1463 }
1465 define <16 x i8> @shuffle_v16i8_17_18_19_20_21_22_23_24_25_26_27_28_29_30_31_00(<16 x i8> %a, <16 x i8> %b) {
1466 ; SSE2-LABEL: shuffle_v16i8_17_18_19_20_21_22_23_24_25_26_27_28_29_30_31_00:
1467 ; SSE2: # %bb.0:
1468 ; SSE2-NEXT: psrldq {{.*#+}} xmm1 = xmm1[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],zero
1469 ; SSE2-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0]
1470 ; SSE2-NEXT: por %xmm1, %xmm0
1471 ; SSE2-NEXT: retq
1472 ;
1473 ; SSSE3-LABEL: shuffle_v16i8_17_18_19_20_21_22_23_24_25_26_27_28_29_30_31_00:
1474 ; SSSE3: # %bb.0:
1475 ; SSSE3-NEXT: palignr {{.*#+}} xmm0 = xmm1[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],xmm0[0]
1476 ; SSSE3-NEXT: retq
1477 ;
1478 ; SSE41-LABEL: shuffle_v16i8_17_18_19_20_21_22_23_24_25_26_27_28_29_30_31_00:
1479 ; SSE41: # %bb.0:
1480 ; SSE41-NEXT: palignr {{.*#+}} xmm0 = xmm1[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],xmm0[0]
1481 ; SSE41-NEXT: retq
1482 ;
1483 ; AVX-LABEL: shuffle_v16i8_17_18_19_20_21_22_23_24_25_26_27_28_29_30_31_00:
1484 ; AVX: # %bb.0:
1485 ; AVX-NEXT: vpalignr {{.*#+}} xmm0 = xmm1[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],xmm0[0]
1486 ; AVX-NEXT: retq
1487 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 0>
1488 ret <16 x i8> %shuffle
1489 }
1491 define <16 x i8> @shuffle_v16i8_01_02_03_04_05_06_07_08_09_10_11_12_13_14_15_16(<16 x i8> %a, <16 x i8> %b) {
1492 ; SSE2-LABEL: shuffle_v16i8_01_02_03_04_05_06_07_08_09_10_11_12_13_14_15_16:
1493 ; SSE2: # %bb.0:
1494 ; SSE2-NEXT: psrldq {{.*#+}} xmm0 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],zero
1495 ; SSE2-NEXT: pslldq {{.*#+}} xmm1 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm1[0]
1496 ; SSE2-NEXT: por %xmm1, %xmm0
1497 ; SSE2-NEXT: retq
1498 ;
1499 ; SSSE3-LABEL: shuffle_v16i8_01_02_03_04_05_06_07_08_09_10_11_12_13_14_15_16:
1500 ; SSSE3: # %bb.0:
1501 ; SSSE3-NEXT: palignr {{.*#+}} xmm1 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],xmm1[0]
1502 ; SSSE3-NEXT: movdqa %xmm1, %xmm0
1503 ; SSSE3-NEXT: retq
1504 ;
1505 ; SSE41-LABEL: shuffle_v16i8_01_02_03_04_05_06_07_08_09_10_11_12_13_14_15_16:
1506 ; SSE41: # %bb.0:
1507 ; SSE41-NEXT: palignr {{.*#+}} xmm1 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],xmm1[0]
1508 ; SSE41-NEXT: movdqa %xmm1, %xmm0
1509 ; SSE41-NEXT: retq
1510 ;
1511 ; AVX-LABEL: shuffle_v16i8_01_02_03_04_05_06_07_08_09_10_11_12_13_14_15_16:
1512 ; AVX: # %bb.0:
1513 ; AVX-NEXT: vpalignr {{.*#+}} xmm0 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],xmm1[0]
1514 ; AVX-NEXT: retq
1515 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16>
1516 ret <16 x i8> %shuffle
1517 }
1519 define <16 x i8> @shuffle_v16i8_01_02_03_04_05_06_07_08_09_10_11_12_13_14_15_00(<16 x i8> %a, <16 x i8> %b) {
1520 ; SSE2-LABEL: shuffle_v16i8_01_02_03_04_05_06_07_08_09_10_11_12_13_14_15_00:
1521 ; SSE2: # %bb.0:
1522 ; SSE2-NEXT: movdqa %xmm0, %xmm1
1523 ; SSE2-NEXT: psrldq {{.*#+}} xmm1 = xmm1[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],zero
1524 ; SSE2-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0]
1525 ; SSE2-NEXT: por %xmm1, %xmm0
1526 ; SSE2-NEXT: retq
1527 ;
1528 ; SSSE3-LABEL: shuffle_v16i8_01_02_03_04_05_06_07_08_09_10_11_12_13_14_15_00:
1529 ; SSSE3: # %bb.0:
1530 ; SSSE3-NEXT: palignr {{.*#+}} xmm0 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,0]
1531 ; SSSE3-NEXT: retq
1532 ;
1533 ; SSE41-LABEL: shuffle_v16i8_01_02_03_04_05_06_07_08_09_10_11_12_13_14_15_00:
1534 ; SSE41: # %bb.0:
1535 ; SSE41-NEXT: palignr {{.*#+}} xmm0 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,0]
1536 ; SSE41-NEXT: retq
1537 ;
1538 ; AVX-LABEL: shuffle_v16i8_01_02_03_04_05_06_07_08_09_10_11_12_13_14_15_00:
1539 ; AVX: # %bb.0:
1540 ; AVX-NEXT: vpalignr {{.*#+}} xmm0 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,0]
1541 ; AVX-NEXT: retq
1542 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 0>
1543 ret <16 x i8> %shuffle
1544 }
1546 define <16 x i8> @shuffle_v16i8_15_16_17_18_19_20_21_22_23_24_25_26_27_28_29_30(<16 x i8> %a, <16 x i8> %b) {
1547 ; SSE2-LABEL: shuffle_v16i8_15_16_17_18_19_20_21_22_23_24_25_26_27_28_29_30:
1548 ; SSE2: # %bb.0:
1549 ; SSE2-NEXT: psrldq {{.*#+}} xmm0 = xmm0[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
1550 ; SSE2-NEXT: pslldq {{.*#+}} xmm1 = zero,xmm1[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14]
1551 ; SSE2-NEXT: por %xmm1, %xmm0
1552 ; SSE2-NEXT: retq
1553 ;
1554 ; SSSE3-LABEL: shuffle_v16i8_15_16_17_18_19_20_21_22_23_24_25_26_27_28_29_30:
1555 ; SSSE3: # %bb.0:
1556 ; SSSE3-NEXT: palignr {{.*#+}} xmm1 = xmm0[15],xmm1[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14]
1557 ; SSSE3-NEXT: movdqa %xmm1, %xmm0
1558 ; SSSE3-NEXT: retq
1559 ;
1560 ; SSE41-LABEL: shuffle_v16i8_15_16_17_18_19_20_21_22_23_24_25_26_27_28_29_30:
1561 ; SSE41: # %bb.0:
1562 ; SSE41-NEXT: palignr {{.*#+}} xmm1 = xmm0[15],xmm1[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14]
1563 ; SSE41-NEXT: movdqa %xmm1, %xmm0
1564 ; SSE41-NEXT: retq
1565 ;
1566 ; AVX-LABEL: shuffle_v16i8_15_16_17_18_19_20_21_22_23_24_25_26_27_28_29_30:
1567 ; AVX: # %bb.0:
1568 ; AVX-NEXT: vpalignr {{.*#+}} xmm0 = xmm0[15],xmm1[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14]
1569 ; AVX-NEXT: retq
1570 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30>
1571 ret <16 x i8> %shuffle
1572 }
1574 ; PR31151
1575 define <16 x i8> @shuffle_v16i8_00_16_01_17_04_20_05_21_02_18_03_19_06_22_07_23(<16 x i8> %val1, <16 x i8> %val2) {
1576 ; SSE-LABEL: shuffle_v16i8_00_16_01_17_04_20_05_21_02_18_03_19_06_22_07_23:
1577 ; SSE: # %bb.0:
1578 ; SSE-NEXT: punpcklbw {{.*#+}} 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]
1579 ; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
1580 ; SSE-NEXT: retq
1581 ;
1582 ; AVX-LABEL: shuffle_v16i8_00_16_01_17_04_20_05_21_02_18_03_19_06_22_07_23:
1583 ; AVX: # %bb.0:
1584 ; AVX-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]
1585 ; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,1,3]
1586 ; AVX-NEXT: retq
1587 %shuffle = shufflevector <16 x i8> %val1, <16 x i8> %val2, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 4, i32 20, i32 5, i32 21, i32 2, i32 18, i32 3, i32 19, i32 6, i32 22, i32 7, i32 23>
1588 ret <16 x i8> %shuffle
1589 }
1591 define <16 x i8> @shuffle_v16i8_00_uu_uu_uu_uu_uu_uu_uu_01_uu_uu_uu_uu_uu_uu_uu(<16 x i8> %a) {
1592 ; SSE2-LABEL: shuffle_v16i8_00_uu_uu_uu_uu_uu_uu_uu_01_uu_uu_uu_uu_uu_uu_uu:
1593 ; SSE2: # %bb.0:
1594 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
1595 ; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
1596 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,1,3]
1597 ; SSE2-NEXT: retq
1598 ;
1599 ; SSSE3-LABEL: shuffle_v16i8_00_uu_uu_uu_uu_uu_uu_uu_01_uu_uu_uu_uu_uu_uu_uu:
1600 ; SSSE3: # %bb.0:
1601 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,u,u,u,u,u,u,u,1,u,u,u,u,u,u,u]
1602 ; SSSE3-NEXT: retq
1603 ;
1604 ; SSE41-LABEL: shuffle_v16i8_00_uu_uu_uu_uu_uu_uu_uu_01_uu_uu_uu_uu_uu_uu_uu:
1605 ; SSE41: # %bb.0:
1606 ; SSE41-NEXT: pmovzxbq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero
1607 ; SSE41-NEXT: retq
1608 ;
1609 ; AVX-LABEL: shuffle_v16i8_00_uu_uu_uu_uu_uu_uu_uu_01_uu_uu_uu_uu_uu_uu_uu:
1610 ; AVX: # %bb.0:
1611 ; AVX-NEXT: vpmovzxbq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero
1612 ; AVX-NEXT: retq
1613 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
1614 ret <16 x i8> %shuffle
1615 }
1617 define <16 x i8> @shuffle_v16i8_00_zz_zz_zz_zz_zz_zz_zz_01_zz_zz_zz_zz_zz_zz_zz(<16 x i8> %a) {
1618 ; SSE2-LABEL: shuffle_v16i8_00_zz_zz_zz_zz_zz_zz_zz_01_zz_zz_zz_zz_zz_zz_zz:
1619 ; SSE2: # %bb.0:
1620 ; SSE2-NEXT: pxor %xmm1, %xmm1
1621 ; SSE2-NEXT: punpcklbw {{.*#+}} 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]
1622 ; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
1623 ; SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
1624 ; SSE2-NEXT: retq
1625 ;
1626 ; SSSE3-LABEL: shuffle_v16i8_00_zz_zz_zz_zz_zz_zz_zz_01_zz_zz_zz_zz_zz_zz_zz:
1627 ; SSSE3: # %bb.0:
1628 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero
1629 ; SSSE3-NEXT: retq
1630 ;
1631 ; SSE41-LABEL: shuffle_v16i8_00_zz_zz_zz_zz_zz_zz_zz_01_zz_zz_zz_zz_zz_zz_zz:
1632 ; SSE41: # %bb.0:
1633 ; SSE41-NEXT: pmovzxbq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero
1634 ; SSE41-NEXT: retq
1635 ;
1636 ; AVX-LABEL: shuffle_v16i8_00_zz_zz_zz_zz_zz_zz_zz_01_zz_zz_zz_zz_zz_zz_zz:
1637 ; AVX: # %bb.0:
1638 ; AVX-NEXT: vpmovzxbq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero
1639 ; AVX-NEXT: retq
1640 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 1, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
1641 ret <16 x i8> %shuffle
1642 }
1644 define <16 x i8> @shuffle_v16i8_00_uu_uu_uu_01_uu_uu_uu_02_uu_uu_uu_03_uu_uu_uu(<16 x i8> %a) {
1645 ; SSE2-LABEL: shuffle_v16i8_00_uu_uu_uu_01_uu_uu_uu_02_uu_uu_uu_03_uu_uu_uu:
1646 ; SSE2: # %bb.0:
1647 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
1648 ; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
1649 ; SSE2-NEXT: retq
1650 ;
1651 ; SSSE3-LABEL: shuffle_v16i8_00_uu_uu_uu_01_uu_uu_uu_02_uu_uu_uu_03_uu_uu_uu:
1652 ; SSSE3: # %bb.0:
1653 ; SSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
1654 ; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
1655 ; SSSE3-NEXT: retq
1656 ;
1657 ; SSE41-LABEL: shuffle_v16i8_00_uu_uu_uu_01_uu_uu_uu_02_uu_uu_uu_03_uu_uu_uu:
1658 ; SSE41: # %bb.0:
1659 ; SSE41-NEXT: pmovzxbd {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
1660 ; SSE41-NEXT: retq
1661 ;
1662 ; AVX-LABEL: shuffle_v16i8_00_uu_uu_uu_01_uu_uu_uu_02_uu_uu_uu_03_uu_uu_uu:
1663 ; AVX: # %bb.0:
1664 ; AVX-NEXT: vpmovzxbd {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
1665 ; AVX-NEXT: retq
1666 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 1, i32 undef, i32 undef, i32 undef, i32 2, i32 undef, i32 undef, i32 undef, i32 3, i32 undef, i32 undef, i32 undef>
1667 ret <16 x i8> %shuffle
1668 }
1670 define <16 x i8> @shuffle_v16i8_00_zz_zz_zz_01_zz_zz_zz_02_zz_zz_zz_03_zz_zz_zz(<16 x i8> %a) {
1671 ; SSE2-LABEL: shuffle_v16i8_00_zz_zz_zz_01_zz_zz_zz_02_zz_zz_zz_03_zz_zz_zz:
1672 ; SSE2: # %bb.0:
1673 ; SSE2-NEXT: pxor %xmm1, %xmm1
1674 ; SSE2-NEXT: punpcklbw {{.*#+}} 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]
1675 ; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
1676 ; SSE2-NEXT: retq
1677 ;
1678 ; SSSE3-LABEL: shuffle_v16i8_00_zz_zz_zz_01_zz_zz_zz_02_zz_zz_zz_03_zz_zz_zz:
1679 ; SSSE3: # %bb.0:
1680 ; SSSE3-NEXT: pxor %xmm1, %xmm1
1681 ; SSSE3-NEXT: punpcklbw {{.*#+}} 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]
1682 ; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
1683 ; SSSE3-NEXT: retq
1684 ;
1685 ; SSE41-LABEL: shuffle_v16i8_00_zz_zz_zz_01_zz_zz_zz_02_zz_zz_zz_03_zz_zz_zz:
1686 ; SSE41: # %bb.0:
1687 ; SSE41-NEXT: pmovzxbd {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
1688 ; SSE41-NEXT: retq
1689 ;
1690 ; AVX-LABEL: shuffle_v16i8_00_zz_zz_zz_01_zz_zz_zz_02_zz_zz_zz_03_zz_zz_zz:
1691 ; AVX: # %bb.0:
1692 ; AVX-NEXT: vpmovzxbd {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
1693 ; AVX-NEXT: retq
1694 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 17, i32 18, i32 19, i32 1, i32 21, i32 22, i32 23, i32 2, i32 25, i32 26, i32 27, i32 3, i32 29, i32 30, i32 31>
1695 ret <16 x i8> %shuffle
1696 }
1698 define <16 x i8> @shuffle_v16i8_00_uu_01_uu_02_uu_03_uu_04_uu_05_uu_06_uu_07_uu(<16 x i8> %a) {
1699 ; SSE2-LABEL: shuffle_v16i8_00_uu_01_uu_02_uu_03_uu_04_uu_05_uu_06_uu_07_uu:
1700 ; SSE2: # %bb.0:
1701 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
1702 ; SSE2-NEXT: retq
1703 ;
1704 ; SSSE3-LABEL: shuffle_v16i8_00_uu_01_uu_02_uu_03_uu_04_uu_05_uu_06_uu_07_uu:
1705 ; SSSE3: # %bb.0:
1706 ; SSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
1707 ; SSSE3-NEXT: retq
1708 ;
1709 ; SSE41-LABEL: shuffle_v16i8_00_uu_01_uu_02_uu_03_uu_04_uu_05_uu_06_uu_07_uu:
1710 ; SSE41: # %bb.0:
1711 ; SSE41-NEXT: pmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
1712 ; SSE41-NEXT: retq
1713 ;
1714 ; AVX-LABEL: shuffle_v16i8_00_uu_01_uu_02_uu_03_uu_04_uu_05_uu_06_uu_07_uu:
1715 ; AVX: # %bb.0:
1716 ; AVX-NEXT: vpmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
1717 ; AVX-NEXT: retq
1718 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 undef, i32 1, i32 undef, i32 2, i32 undef, i32 3, i32 undef, i32 4, i32 undef, i32 5, i32 undef, i32 6, i32 undef, i32 7, i32 undef>
1719 ret <16 x i8> %shuffle
1720 }
1722 define <16 x i8> @shuffle_v16i8_00_zz_01_zz_02_zz_03_zz_04_zz_05_zz_06_zz_07_zz(<16 x i8> %a) {
1723 ; SSE2-LABEL: shuffle_v16i8_00_zz_01_zz_02_zz_03_zz_04_zz_05_zz_06_zz_07_zz:
1724 ; SSE2: # %bb.0:
1725 ; SSE2-NEXT: pxor %xmm1, %xmm1
1726 ; SSE2-NEXT: punpcklbw {{.*#+}} 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]
1727 ; SSE2-NEXT: retq
1728 ;
1729 ; SSSE3-LABEL: shuffle_v16i8_00_zz_01_zz_02_zz_03_zz_04_zz_05_zz_06_zz_07_zz:
1730 ; SSSE3: # %bb.0:
1731 ; SSSE3-NEXT: pxor %xmm1, %xmm1
1732 ; SSSE3-NEXT: punpcklbw {{.*#+}} 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]
1733 ; SSSE3-NEXT: retq
1734 ;
1735 ; SSE41-LABEL: shuffle_v16i8_00_zz_01_zz_02_zz_03_zz_04_zz_05_zz_06_zz_07_zz:
1736 ; SSE41: # %bb.0:
1737 ; SSE41-NEXT: pmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
1738 ; SSE41-NEXT: retq
1739 ;
1740 ; AVX-LABEL: shuffle_v16i8_00_zz_01_zz_02_zz_03_zz_04_zz_05_zz_06_zz_07_zz:
1741 ; AVX: # %bb.0:
1742 ; AVX-NEXT: vpmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
1743 ; AVX-NEXT: retq
1744 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 17, i32 1, i32 19, i32 2, i32 21, i32 3, i32 23, i32 4, i32 25, i32 5, i32 27, i32 6, i32 29, i32 7, i32 31>
1745 ret <16 x i8> %shuffle
1746 }
1748 define <16 x i8> @shuffle_v16i8_uu_10_02_07_22_14_07_02_18_03_01_14_18_09_11_00(<16 x i8> %a, <16 x i8> %b) {
1749 ; SSE2-LABEL: shuffle_v16i8_uu_10_02_07_22_14_07_02_18_03_01_14_18_09_11_00:
1750 ; SSE2: # %bb.0: # %entry
1751 ; SSE2-NEXT: pxor %xmm2, %xmm2
1752 ; SSE2-NEXT: movdqa %xmm0, %xmm3
1753 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm3 = 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]
1754 ; SSE2-NEXT: pshuflw {{.*#+}} xmm4 = xmm3[1,3,2,0,4,5,6,7]
1755 ; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[0,0,2,1]
1756 ; SSE2-NEXT: movdqa {{.*#+}} xmm5 = [65535,65535,65535,0,65535,0,0,65535]
1757 ; SSE2-NEXT: pand %xmm5, %xmm4
1758 ; SSE2-NEXT: punpckhbw {{.*#+}} xmm0 = xmm0[8],xmm2[8],xmm0[9],xmm2[9],xmm0[10],xmm2[10],xmm0[11],xmm2[11],xmm0[12],xmm2[12],xmm0[13],xmm2[13],xmm0[14],xmm2[14],xmm0[15],xmm2[15]
1759 ; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,3,0,1]
1760 ; SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[2,2,2,2,4,5,6,7]
1761 ; SSE2-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,5,7,7]
1762 ; SSE2-NEXT: pandn %xmm2, %xmm5
1763 ; SSE2-NEXT: por %xmm4, %xmm5
1764 ; SSE2-NEXT: psrlq $16, %xmm0
1765 ; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,2,2,3]
1766 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm3[3,1,1,3]
1767 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,1,2,1,4,5,6,7]
1768 ; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,5,7,4]
1769 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,3,2,3]
1770 ; SSE2-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1]
1771 ; SSE2-NEXT: packuswb %xmm5, %xmm2
1772 ; SSE2-NEXT: movdqa {{.*#+}} xmm0 = [255,255,255,255,0,255,255,255,0,255,255,255,0,255,255,255]
1773 ; SSE2-NEXT: pand %xmm0, %xmm2
1774 ; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[3,1,1,3,4,5,6,7]
1775 ; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,0,1,1]
1776 ; SSE2-NEXT: pandn %xmm1, %xmm0
1777 ; SSE2-NEXT: por %xmm2, %xmm0
1778 ; SSE2-NEXT: retq
1779 ;
1780 ; SSSE3-LABEL: shuffle_v16i8_uu_10_02_07_22_14_07_02_18_03_01_14_18_09_11_00:
1781 ; SSSE3: # %bb.0: # %entry
1782 ; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = xmm1[u],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[2],zero,zero,zero
1783 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[u,10,2,7],zero,xmm0[14,7,2],zero,xmm0[3,1,14],zero,xmm0[9,11,0]
1784 ; SSSE3-NEXT: por %xmm1, %xmm0
1785 ; SSSE3-NEXT: retq
1786 ;
1787 ; SSE41-LABEL: shuffle_v16i8_uu_10_02_07_22_14_07_02_18_03_01_14_18_09_11_00:
1788 ; SSE41: # %bb.0: # %entry
1789 ; SSE41-NEXT: pshufb {{.*#+}} xmm1 = xmm1[u],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[2],zero,zero,zero
1790 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[u,10,2,7],zero,xmm0[14,7,2],zero,xmm0[3,1,14],zero,xmm0[9,11,0]
1791 ; SSE41-NEXT: por %xmm1, %xmm0
1792 ; SSE41-NEXT: retq
1793 ;
1794 ; AVX1-LABEL: shuffle_v16i8_uu_10_02_07_22_14_07_02_18_03_01_14_18_09_11_00:
1795 ; AVX1: # %bb.0: # %entry
1796 ; AVX1-NEXT: vpshufb {{.*#+}} xmm1 = xmm1[u],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[2],zero,zero,zero
1797 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[u,10,2,7],zero,xmm0[14,7,2],zero,xmm0[3,1,14],zero,xmm0[9,11,0]
1798 ; AVX1-NEXT: vpor %xmm1, %xmm0, %xmm0
1799 ; AVX1-NEXT: retq
1800 ;
1801 ; AVX2-LABEL: shuffle_v16i8_uu_10_02_07_22_14_07_02_18_03_01_14_18_09_11_00:
1802 ; AVX2: # %bb.0: # %entry
1803 ; AVX2-NEXT: vpshufb {{.*#+}} xmm1 = xmm1[u],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[2],zero,zero,zero
1804 ; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[u,10,2,7],zero,xmm0[14,7,2],zero,xmm0[3,1,14],zero,xmm0[9,11,0]
1805 ; AVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
1806 ; AVX2-NEXT: retq
1807 ;
1808 ; AVX512VLBW-LABEL: shuffle_v16i8_uu_10_02_07_22_14_07_02_18_03_01_14_18_09_11_00:
1809 ; AVX512VLBW: # %bb.0: # %entry
1810 ; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm1 = xmm1[u],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[2],zero,zero,zero
1811 ; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[u,10,2,7],zero,xmm0[14,7,2],zero,xmm0[3,1,14],zero,xmm0[9,11,0]
1812 ; AVX512VLBW-NEXT: vpor %xmm1, %xmm0, %xmm0
1813 ; AVX512VLBW-NEXT: retq
1814 ;
1815 ; AVX512VLVBMI-LABEL: shuffle_v16i8_uu_10_02_07_22_14_07_02_18_03_01_14_18_09_11_00:
1816 ; AVX512VLVBMI: # %bb.0: # %entry
1817 ; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm2 = <u,10,2,7,22,14,7,2,18,3,1,14,18,9,11,0>
1818 ; AVX512VLVBMI-NEXT: vpermt2b %xmm1, %xmm2, %xmm0
1819 ; AVX512VLVBMI-NEXT: retq
1820 ;
1821 ; XOP-LABEL: shuffle_v16i8_uu_10_02_07_22_14_07_02_18_03_01_14_18_09_11_00:
1822 ; XOP: # %bb.0: # %entry
1823 ; XOP-NEXT: vpperm {{.*#+}} xmm0 = xmm0[u,10,2,7],xmm1[6],xmm0[14,7,2],xmm1[2],xmm0[3,1,14],xmm1[2],xmm0[9,11,0]
1824 ; XOP-NEXT: retq
1825 entry:
1826 %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 undef, i32 10, i32 2, i32 7, i32 22, i32 14, i32 7, i32 2, i32 18, i32 3, i32 1, i32 14, i32 18, i32 9, i32 11, i32 0>
1828 ret <16 x i8> %shuffle
1829 }
1831 define <16 x i8> @shuffe_v16i8_shift_00_02_04_06_08_10_12_14_16_18_20_22_24_26_28_30(<8 x i16> %a0, <8 x i16> %a1) {
1832 ; SSE-LABEL: shuffe_v16i8_shift_00_02_04_06_08_10_12_14_16_18_20_22_24_26_28_30:
1833 ; SSE: # %bb.0:
1834 ; SSE-NEXT: psrlw $8, %xmm0
1835 ; SSE-NEXT: psrlw $8, %xmm1
1836 ; SSE-NEXT: packuswb %xmm1, %xmm0
1837 ; SSE-NEXT: retq
1838 ;
1839 ; AVX1-LABEL: shuffe_v16i8_shift_00_02_04_06_08_10_12_14_16_18_20_22_24_26_28_30:
1840 ; AVX1: # %bb.0:
1841 ; AVX1-NEXT: vpsrlw $8, %xmm0, %xmm0
1842 ; AVX1-NEXT: vpsrlw $8, %xmm1, %xmm1
1843 ; AVX1-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
1844 ; AVX1-NEXT: retq
1845 ;
1846 ; AVX2OR512VL-LABEL: shuffe_v16i8_shift_00_02_04_06_08_10_12_14_16_18_20_22_24_26_28_30:
1847 ; AVX2OR512VL: # %bb.0:
1848 ; AVX2OR512VL-NEXT: vpsrlw $8, %xmm0, %xmm0
1849 ; AVX2OR512VL-NEXT: vpsrlw $8, %xmm1, %xmm1
1850 ; AVX2OR512VL-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
1851 ; AVX2OR512VL-NEXT: retq
1852 ;
1853 ; XOP-LABEL: shuffe_v16i8_shift_00_02_04_06_08_10_12_14_16_18_20_22_24_26_28_30:
1854 ; XOP: # %bb.0:
1855 ; XOP-NEXT: vpperm {{.*#+}} xmm0 = xmm0[1,3,5,7,9,11,13,15],xmm1[1,3,5,7,9,11,13,15]
1856 ; XOP-NEXT: retq
1857 %1 = lshr <8 x i16> %a0, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
1858 %2 = lshr <8 x i16> %a1, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
1859 %3 = bitcast <8 x i16> %1 to <16 x i8>
1860 %4 = bitcast <8 x i16> %2 to <16 x i8>
1861 %5 = shufflevector <16 x i8> %3, <16 x i8> %4, <16 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30>
1862 ret <16 x i8> %5
1863 }
1865 define <16 x i8> @stress_test2(<16 x i8> %s.0.0, <16 x i8> %s.0.1, <16 x i8> %s.0.2) {
1866 ; Nothing interesting to test here. Just make sure we didn't crashe.
1867 ; ALL-LABEL: stress_test2:
1868 ; ALL: retq
1869 entry:
1870 %s.1.0 = shufflevector <16 x i8> %s.0.0, <16 x i8> %s.0.1, <16 x i32> <i32 29, i32 30, i32 2, i32 16, i32 26, i32 21, i32 11, i32 26, i32 26, i32 3, i32 4, i32 5, i32 30, i32 28, i32 15, i32 5>
1871 %s.1.1 = shufflevector <16 x i8> %s.0.1, <16 x i8> %s.0.2, <16 x i32> <i32 31, i32 1, i32 24, i32 12, i32 28, i32 5, i32 2, i32 9, i32 29, i32 1, i32 31, i32 5, i32 6, i32 17, i32 15, i32 22>
1872 %s.2.0 = shufflevector <16 x i8> %s.1.0, <16 x i8> %s.1.1, <16 x i32> <i32 22, i32 1, i32 12, i32 3, i32 30, i32 4, i32 30, i32 undef, i32 1, i32 10, i32 14, i32 18, i32 27, i32 13, i32 16, i32 19>
1874 ret <16 x i8> %s.2.0
1875 }
1877 define void @constant_gets_selected(ptr %ptr1, ptr %ptr2) {
1878 ; SSE-LABEL: constant_gets_selected:
1879 ; SSE: # %bb.0: # %entry
1880 ; SSE-NEXT: xorps %xmm0, %xmm0
1881 ; SSE-NEXT: movaps %xmm0, (%rdi)
1882 ; SSE-NEXT: movaps %xmm0, (%rsi)
1883 ; SSE-NEXT: retq
1884 ;
1885 ; AVX-LABEL: constant_gets_selected:
1886 ; AVX: # %bb.0: # %entry
1887 ; AVX-NEXT: vxorps %xmm0, %xmm0, %xmm0
1888 ; AVX-NEXT: vmovaps %xmm0, (%rdi)
1889 ; AVX-NEXT: vmovaps %xmm0, (%rsi)
1890 ; AVX-NEXT: retq
1891 entry:
1892 %weird_zero = bitcast <4 x i32> zeroinitializer to <16 x i8>
1893 %shuffle.i = shufflevector <16 x i8> <i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 0, i8 0, i8 0, i8 0>, <16 x i8> %weird_zero, <16 x i32> <i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27>
1894 %weirder_zero = bitcast <16 x i8> %shuffle.i to <4 x i32>
1895 store <4 x i32> %weirder_zero, ptr %ptr1, align 16
1896 store <4 x i32> zeroinitializer, ptr %ptr2, align 16
1897 ret void
1898 }
1900 ;
1901 ; Shuffle to logical bit shifts
1902 ;
1904 define <16 x i8> @shuffle_v16i8_zz_00_zz_02_zz_04_zz_06_zz_08_zz_10_zz_12_zz_14(<16 x i8> %a, <16 x i8> %b) {
1905 ; SSE-LABEL: shuffle_v16i8_zz_00_zz_02_zz_04_zz_06_zz_08_zz_10_zz_12_zz_14:
1906 ; SSE: # %bb.0:
1907 ; SSE-NEXT: psllw $8, %xmm0
1908 ; SSE-NEXT: retq
1909 ;
1910 ; AVX-LABEL: shuffle_v16i8_zz_00_zz_02_zz_04_zz_06_zz_08_zz_10_zz_12_zz_14:
1911 ; AVX: # %bb.0:
1912 ; AVX-NEXT: vpsllw $8, %xmm0, %xmm0
1913 ; AVX-NEXT: retq
1914 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32><i32 16, i32 0, i32 16, i32 2, i32 16, i32 4, i32 16, i32 6, i32 16, i32 8, i32 16, i32 10, i32 16, i32 12, i32 16, i32 14>
1915 ret <16 x i8> %shuffle
1916 }
1918 define <16 x i8> @shuffle_v16i8_zz_zz_zz_00_zz_zz_zz_04_zz_zz_zz_08_zz_zz_zz_12(<16 x i8> %a, <16 x i8> %b) {
1919 ; SSE-LABEL: shuffle_v16i8_zz_zz_zz_00_zz_zz_zz_04_zz_zz_zz_08_zz_zz_zz_12:
1920 ; SSE: # %bb.0:
1921 ; SSE-NEXT: pslld $24, %xmm0
1922 ; SSE-NEXT: retq
1923 ;
1924 ; AVX-LABEL: shuffle_v16i8_zz_zz_zz_00_zz_zz_zz_04_zz_zz_zz_08_zz_zz_zz_12:
1925 ; AVX: # %bb.0:
1926 ; AVX-NEXT: vpslld $24, %xmm0, %xmm0
1927 ; AVX-NEXT: retq
1928 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32><i32 16, i32 16, i32 16, i32 0, i32 16, i32 16, i32 16, i32 4, i32 16, i32 16, i32 16, i32 8, i32 16, i32 16, i32 16, i32 12>
1929 ret <16 x i8> %shuffle
1930 }
1932 define <16 x i8> @shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_00_zz_zz_zz_zz_zz_zz_zz_08(<16 x i8> %a, <16 x i8> %b) {
1933 ; SSE-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_00_zz_zz_zz_zz_zz_zz_zz_08:
1934 ; SSE: # %bb.0:
1935 ; SSE-NEXT: psllq $56, %xmm0
1936 ; SSE-NEXT: retq
1937 ;
1938 ; AVX-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_00_zz_zz_zz_zz_zz_zz_zz_08:
1939 ; AVX: # %bb.0:
1940 ; AVX-NEXT: vpsllq $56, %xmm0, %xmm0
1941 ; AVX-NEXT: retq
1942 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32><i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 0, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 8>
1943 ret <16 x i8> %shuffle
1944 }
1946 define <16 x i8> @shuffle_v16i8_zz_00_uu_02_03_uu_05_06_zz_08_09_uu_11_12_13_14(<16 x i8> %a, <16 x i8> %b) {
1947 ; SSE-LABEL: shuffle_v16i8_zz_00_uu_02_03_uu_05_06_zz_08_09_uu_11_12_13_14:
1948 ; SSE: # %bb.0:
1949 ; SSE-NEXT: psllq $8, %xmm0
1950 ; SSE-NEXT: retq
1951 ;
1952 ; AVX-LABEL: shuffle_v16i8_zz_00_uu_02_03_uu_05_06_zz_08_09_uu_11_12_13_14:
1953 ; AVX: # %bb.0:
1954 ; AVX-NEXT: vpsllq $8, %xmm0, %xmm0
1955 ; AVX-NEXT: retq
1956 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32><i32 16, i32 0, i32 undef, i32 2, i32 3, i32 undef, i32 5, i32 6, i32 16, i32 8, i32 9, i32 undef, i32 11, i32 12, i32 13, i32 14>
1957 ret <16 x i8> %shuffle
1958 }
1960 define <16 x i8> @shuffle_v16i8_01_uu_uu_uu_uu_zz_uu_zz_uu_zz_11_zz_13_zz_15_zz(<16 x i8> %a, <16 x i8> %b) {
1961 ; SSE-LABEL: shuffle_v16i8_01_uu_uu_uu_uu_zz_uu_zz_uu_zz_11_zz_13_zz_15_zz:
1962 ; SSE: # %bb.0:
1963 ; SSE-NEXT: psrlw $8, %xmm0
1964 ; SSE-NEXT: retq
1965 ;
1966 ; AVX-LABEL: shuffle_v16i8_01_uu_uu_uu_uu_zz_uu_zz_uu_zz_11_zz_13_zz_15_zz:
1967 ; AVX: # %bb.0:
1968 ; AVX-NEXT: vpsrlw $8, %xmm0, %xmm0
1969 ; AVX-NEXT: retq
1970 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32><i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 16, i32 undef, i32 16, i32 undef, i32 16, i32 11, i32 16, i32 13, i32 16, i32 15, i32 16>
1971 ret <16 x i8> %shuffle
1972 }
1974 define <16 x i8> @shuffle_v16i8_02_03_zz_zz_06_07_uu_uu_uu_uu_uu_uu_14_15_zz_zz(<16 x i8> %a, <16 x i8> %b) {
1975 ; SSE-LABEL: shuffle_v16i8_02_03_zz_zz_06_07_uu_uu_uu_uu_uu_uu_14_15_zz_zz:
1976 ; SSE: # %bb.0:
1977 ; SSE-NEXT: psrld $16, %xmm0
1978 ; SSE-NEXT: retq
1979 ;
1980 ; AVX-LABEL: shuffle_v16i8_02_03_zz_zz_06_07_uu_uu_uu_uu_uu_uu_14_15_zz_zz:
1981 ; AVX: # %bb.0:
1982 ; AVX-NEXT: vpsrld $16, %xmm0, %xmm0
1983 ; AVX-NEXT: retq
1984 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32><i32 2, i32 3, i32 16, i32 16, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 14, i32 15, i32 16, i32 16>
1985 ret <16 x i8> %shuffle
1986 }
1988 define <16 x i8> @shuffle_v16i8_07_zz_zz_zz_zz_zz_uu_uu_15_uu_uu_uu_uu_uu_zz_zz(<16 x i8> %a, <16 x i8> %b) {
1989 ; SSE-LABEL: shuffle_v16i8_07_zz_zz_zz_zz_zz_uu_uu_15_uu_uu_uu_uu_uu_zz_zz:
1990 ; SSE: # %bb.0:
1991 ; SSE-NEXT: psrlq $56, %xmm0
1992 ; SSE-NEXT: retq
1993 ;
1994 ; AVX-LABEL: shuffle_v16i8_07_zz_zz_zz_zz_zz_uu_uu_15_uu_uu_uu_uu_uu_zz_zz:
1995 ; AVX: # %bb.0:
1996 ; AVX-NEXT: vpsrlq $56, %xmm0, %xmm0
1997 ; AVX-NEXT: retq
1998 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32><i32 7, i32 16, i32 16, i32 16, i32 16, i32 16, i32 undef, i32 undef, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 16, i32 16>
1999 ret <16 x i8> %shuffle
2000 }
2002 define <16 x i8> @shuffle_v16i8_zz_zz_zz_zz_01_02_03_04_05_06_zz_zz_zz_zz_zz_zz(<16 x i8> %a) {
2003 ; SSE2-LABEL: shuffle_v16i8_zz_zz_zz_zz_01_02_03_04_05_06_zz_zz_zz_zz_zz_zz:
2004 ; SSE2: # %bb.0:
2005 ; SSE2-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5,6]
2006 ; SSE2-NEXT: psrldq {{.*#+}} xmm0 = xmm0[10,11,12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
2007 ; SSE2-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,xmm0[0,1,2,3,4,5,6,7,8,9,10,11]
2008 ; SSE2-NEXT: retq
2009 ;
2010 ; SSSE3-LABEL: shuffle_v16i8_zz_zz_zz_zz_01_02_03_04_05_06_zz_zz_zz_zz_zz_zz:
2011 ; SSSE3: # %bb.0:
2012 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = zero,zero,zero,zero,xmm0[1,2,3,4,5,6],zero,zero,zero,zero,zero,zero
2013 ; SSSE3-NEXT: retq
2014 ;
2015 ; SSE41-LABEL: shuffle_v16i8_zz_zz_zz_zz_01_02_03_04_05_06_zz_zz_zz_zz_zz_zz:
2016 ; SSE41: # %bb.0:
2017 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = zero,zero,zero,zero,xmm0[1,2,3,4,5,6],zero,zero,zero,zero,zero,zero
2018 ; SSE41-NEXT: retq
2019 ;
2020 ; AVX-LABEL: shuffle_v16i8_zz_zz_zz_zz_01_02_03_04_05_06_zz_zz_zz_zz_zz_zz:
2021 ; AVX: # %bb.0:
2022 ; AVX-NEXT: vpshufb {{.*#+}} xmm0 = zero,zero,zero,zero,xmm0[1,2,3,4,5,6],zero,zero,zero,zero,zero,zero
2023 ; AVX-NEXT: retq
2024 %shuffle = shufflevector <16 x i8> %a, <16 x i8> <i8 0, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef>, <16 x i32> <i32 16, i32 16, i32 16, i32 16, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
2025 ret <16 x i8> %shuffle
2026 }
2028 define <16 x i8> @shuffle_v16i8_01_02_03_04_05_06_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz(<16 x i8> %a) {
2029 ; SSE-LABEL: shuffle_v16i8_01_02_03_04_05_06_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
2030 ; SSE: # %bb.0:
2031 ; SSE-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5,6]
2032 ; SSE-NEXT: psrldq {{.*#+}} xmm0 = xmm0[10,11,12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
2033 ; SSE-NEXT: retq
2034 ;
2035 ; AVX1-LABEL: shuffle_v16i8_01_02_03_04_05_06_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
2036 ; AVX1: # %bb.0:
2037 ; AVX1-NEXT: vpslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5,6]
2038 ; AVX1-NEXT: vpsrldq {{.*#+}} xmm0 = xmm0[10,11,12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
2039 ; AVX1-NEXT: retq
2040 ;
2041 ; AVX2-SLOW-LABEL: shuffle_v16i8_01_02_03_04_05_06_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
2042 ; AVX2-SLOW: # %bb.0:
2043 ; AVX2-SLOW-NEXT: vpslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5,6]
2044 ; AVX2-SLOW-NEXT: vpsrldq {{.*#+}} xmm0 = xmm0[10,11,12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
2045 ; AVX2-SLOW-NEXT: retq
2046 ;
2047 ; AVX2-FAST-LABEL: shuffle_v16i8_01_02_03_04_05_06_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
2048 ; AVX2-FAST: # %bb.0:
2049 ; AVX2-FAST-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[1,2,3,4,5,6],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
2050 ; AVX2-FAST-NEXT: retq
2051 ;
2052 ; AVX512VL-LABEL: shuffle_v16i8_01_02_03_04_05_06_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
2053 ; AVX512VL: # %bb.0:
2054 ; AVX512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[1,2,3,4,5,6],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
2055 ; AVX512VL-NEXT: retq
2056 ;
2057 ; XOP-LABEL: shuffle_v16i8_01_02_03_04_05_06_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz:
2058 ; XOP: # %bb.0:
2059 ; XOP-NEXT: vpslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5,6]
2060 ; XOP-NEXT: vpsrldq {{.*#+}} xmm0 = xmm0[10,11,12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
2061 ; XOP-NEXT: retq
2062 %shuffle = shufflevector <16 x i8> %a, <16 x i8> <i8 0, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef>, <16 x i32> <i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
2063 ret <16 x i8> %shuffle
2064 }
2066 define <16 x i8> @shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_01_02_03_04_05_06(<16 x i8> %a) {
2067 ; SSE-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_01_02_03_04_05_06:
2068 ; SSE: # %bb.0:
2069 ; SSE-NEXT: psrlq $8, %xmm0
2070 ; SSE-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5]
2071 ; SSE-NEXT: retq
2072 ;
2073 ; AVX1-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_01_02_03_04_05_06:
2074 ; AVX1: # %bb.0:
2075 ; AVX1-NEXT: vpsrlq $8, %xmm0, %xmm0
2076 ; AVX1-NEXT: vpslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5]
2077 ; AVX1-NEXT: retq
2078 ;
2079 ; AVX2-SLOW-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_01_02_03_04_05_06:
2080 ; AVX2-SLOW: # %bb.0:
2081 ; AVX2-SLOW-NEXT: vpsrlq $8, %xmm0, %xmm0
2082 ; AVX2-SLOW-NEXT: vpslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5]
2083 ; AVX2-SLOW-NEXT: retq
2084 ;
2085 ; AVX2-FAST-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_01_02_03_04_05_06:
2086 ; AVX2-FAST: # %bb.0:
2087 ; AVX2-FAST-NEXT: vpshufb {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[1,2,3,4,5,6]
2088 ; AVX2-FAST-NEXT: retq
2089 ;
2090 ; AVX512VL-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_01_02_03_04_05_06:
2091 ; AVX512VL: # %bb.0:
2092 ; AVX512VL-NEXT: vpshufb {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[1,2,3,4,5,6]
2093 ; AVX512VL-NEXT: retq
2094 ;
2095 ; XOP-LABEL: shuffle_v16i8_zz_zz_zz_zz_zz_zz_zz_zz_zz_zz_01_02_03_04_05_06:
2096 ; XOP: # %bb.0:
2097 ; XOP-NEXT: vpsrlq $8, %xmm0, %xmm0
2098 ; XOP-NEXT: vpslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5]
2099 ; XOP-NEXT: retq
2100 %shuffle = shufflevector <16 x i8> %a, <16 x i8> <i8 0, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef>, <16 x i32> <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6>
2101 ret <16 x i8> %shuffle
2102 }
2104 define <16 x i8> @shuffle_v16i8_03_00_01_02_07_04_05_06_11_08_09_10_15_12_13_14(<16 x i8> %a) {
2105 ; SSE2-LABEL: shuffle_v16i8_03_00_01_02_07_04_05_06_11_08_09_10_15_12_13_14:
2106 ; SSE2: # %bb.0:
2107 ; SSE2-NEXT: movdqa %xmm0, %xmm1
2108 ; SSE2-NEXT: psrld $24, %xmm1
2109 ; SSE2-NEXT: pslld $8, %xmm0
2110 ; SSE2-NEXT: por %xmm1, %xmm0
2111 ; SSE2-NEXT: retq
2112 ;
2113 ; SSSE3-LABEL: shuffle_v16i8_03_00_01_02_07_04_05_06_11_08_09_10_15_12_13_14:
2114 ; SSSE3: # %bb.0:
2115 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[3,0,1,2,7,4,5,6,11,8,9,10,15,12,13,14]
2116 ; SSSE3-NEXT: retq
2117 ;
2118 ; SSE41-LABEL: shuffle_v16i8_03_00_01_02_07_04_05_06_11_08_09_10_15_12_13_14:
2119 ; SSE41: # %bb.0:
2120 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[3,0,1,2,7,4,5,6,11,8,9,10,15,12,13,14]
2121 ; SSE41-NEXT: retq
2122 ;
2123 ; AVX1-LABEL: shuffle_v16i8_03_00_01_02_07_04_05_06_11_08_09_10_15_12_13_14:
2124 ; AVX1: # %bb.0:
2125 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[3,0,1,2,7,4,5,6,11,8,9,10,15,12,13,14]
2126 ; AVX1-NEXT: retq
2127 ;
2128 ; AVX2-LABEL: shuffle_v16i8_03_00_01_02_07_04_05_06_11_08_09_10_15_12_13_14:
2129 ; AVX2: # %bb.0:
2130 ; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[3,0,1,2,7,4,5,6,11,8,9,10,15,12,13,14]
2131 ; AVX2-NEXT: retq
2132 ;
2133 ; AVX512VL-LABEL: shuffle_v16i8_03_00_01_02_07_04_05_06_11_08_09_10_15_12_13_14:
2134 ; AVX512VL: # %bb.0:
2135 ; AVX512VL-NEXT: vprold $8, %xmm0, %xmm0
2136 ; AVX512VL-NEXT: retq
2137 ;
2138 ; XOP-LABEL: shuffle_v16i8_03_00_01_02_07_04_05_06_11_08_09_10_15_12_13_14:
2139 ; XOP: # %bb.0:
2140 ; XOP-NEXT: vprotd $8, %xmm0, %xmm0
2141 ; XOP-NEXT: retq
2142 %shuffle = shufflevector <16 x i8> %a, <16 x i8> undef, <16 x i32> <i32 3, i32 0, i32 1, i32 2, i32 7, i32 4, i32 5, i32 6, i32 11, i32 8, i32 9, i32 10, i32 15, i32 12, i32 13, i32 14>
2143 ret <16 x i8> %shuffle
2144 }
2146 ; PR44379
2147 define <16 x i8> @shuffle_v16i8_02_03_04_05_06_07_00_01_10_11_12_13_14_15_08_09(<16 x i8> %a) {
2148 ; SSE-LABEL: shuffle_v16i8_02_03_04_05_06_07_00_01_10_11_12_13_14_15_08_09:
2149 ; SSE: # %bb.0:
2150 ; SSE-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,2,3,0,4,5,6,7]
2151 ; SSE-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,6,7,4]
2152 ; SSE-NEXT: retq
2153 ;
2154 ; AVX1-LABEL: shuffle_v16i8_02_03_04_05_06_07_00_01_10_11_12_13_14_15_08_09:
2155 ; AVX1: # %bb.0:
2156 ; AVX1-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[1,2,3,0,4,5,6,7]
2157 ; AVX1-NEXT: vpshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,6,7,4]
2158 ; AVX1-NEXT: retq
2159 ;
2160 ; AVX2-SLOW-LABEL: shuffle_v16i8_02_03_04_05_06_07_00_01_10_11_12_13_14_15_08_09:
2161 ; AVX2-SLOW: # %bb.0:
2162 ; AVX2-SLOW-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[1,2,3,0,4,5,6,7]
2163 ; AVX2-SLOW-NEXT: vpshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,6,7,4]
2164 ; AVX2-SLOW-NEXT: retq
2165 ;
2166 ; AVX2-FAST-LABEL: shuffle_v16i8_02_03_04_05_06_07_00_01_10_11_12_13_14_15_08_09:
2167 ; AVX2-FAST: # %bb.0:
2168 ; AVX2-FAST-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[2,3,4,5,6,7,0,1,10,11,12,13,14,15,8,9]
2169 ; AVX2-FAST-NEXT: retq
2170 ;
2171 ; AVX512VL-LABEL: shuffle_v16i8_02_03_04_05_06_07_00_01_10_11_12_13_14_15_08_09:
2172 ; AVX512VL: # %bb.0:
2173 ; AVX512VL-NEXT: vprolq $48, %xmm0, %xmm0
2174 ; AVX512VL-NEXT: retq
2175 ;
2176 ; XOP-LABEL: shuffle_v16i8_02_03_04_05_06_07_00_01_10_11_12_13_14_15_08_09:
2177 ; XOP: # %bb.0:
2178 ; XOP-NEXT: vprotq $48, %xmm0, %xmm0
2179 ; XOP-NEXT: retq
2180 %shuffle = shufflevector <16 x i8> %a, <16 x i8> undef, <16 x i32> <i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 8, i32 9>
2181 ret <16 x i8> %shuffle
2182 }
2184 define <16 x i8> @PR12412(<16 x i8> %inval1, <16 x i8> %inval2) {
2185 ; SSE-LABEL: PR12412:
2186 ; SSE: # %bb.0: # %entry
2187 ; SSE-NEXT: movdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
2188 ; SSE-NEXT: pand %xmm2, %xmm1
2189 ; SSE-NEXT: pand %xmm2, %xmm0
2190 ; SSE-NEXT: packuswb %xmm1, %xmm0
2191 ; SSE-NEXT: retq
2192 ;
2193 ; AVX1-LABEL: PR12412:
2194 ; AVX1: # %bb.0: # %entry
2195 ; AVX1-NEXT: vbroadcastss {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
2196 ; AVX1-NEXT: vpand %xmm2, %xmm1, %xmm1
2197 ; AVX1-NEXT: vpand %xmm2, %xmm0, %xmm0
2198 ; AVX1-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
2199 ; AVX1-NEXT: retq
2200 ;
2201 ; AVX2-LABEL: PR12412:
2202 ; AVX2: # %bb.0: # %entry
2203 ; AVX2-NEXT: vpbroadcastw {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
2204 ; AVX2-NEXT: vpand %xmm2, %xmm1, %xmm1
2205 ; AVX2-NEXT: vpand %xmm2, %xmm0, %xmm0
2206 ; AVX2-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
2207 ; AVX2-NEXT: retq
2208 ;
2209 ; AVX512VL-LABEL: PR12412:
2210 ; AVX512VL: # %bb.0: # %entry
2211 ; AVX512VL-NEXT: # kill: def $xmm0 killed $xmm0 def $ymm0
2212 ; AVX512VL-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
2213 ; AVX512VL-NEXT: vpmovwb %ymm0, %xmm0
2214 ; AVX512VL-NEXT: vzeroupper
2215 ; AVX512VL-NEXT: retq
2216 ;
2217 ; XOP-LABEL: PR12412:
2218 ; XOP: # %bb.0: # %entry
2219 ; XOP-NEXT: vpperm {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14],xmm1[0,2,4,6,8,10,12,14]
2220 ; XOP-NEXT: retq
2221 entry:
2222 %0 = shufflevector <16 x i8> %inval1, <16 x i8> %inval2, <16 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30>
2223 ret <16 x i8> %0
2224 }
2226 define <16 x i8> @shuffle_v16i8_uu_02_03_zz_uu_06_07_zz_uu_10_11_zz_uu_14_15_zz(<16 x i8> %a) {
2227 ; SSE-LABEL: shuffle_v16i8_uu_02_03_zz_uu_06_07_zz_uu_10_11_zz_uu_14_15_zz:
2228 ; SSE: # %bb.0:
2229 ; SSE-NEXT: psrld $8, %xmm0
2230 ; SSE-NEXT: retq
2231 ;
2232 ; AVX-LABEL: shuffle_v16i8_uu_02_03_zz_uu_06_07_zz_uu_10_11_zz_uu_14_15_zz:
2233 ; AVX: # %bb.0:
2234 ; AVX-NEXT: vpsrld $8, %xmm0, %xmm0
2235 ; AVX-NEXT: retq
2236 %shuffle = shufflevector <16 x i8> %a, <16 x i8> zeroinitializer, <16 x i32> <i32 undef, i32 2, i32 3, i32 16, i32 undef, i32 6, i32 7, i32 16, i32 undef, i32 10, i32 11, i32 16, i32 undef, i32 14, i32 15, i32 16>
2237 ret <16 x i8> %shuffle
2238 }
2240 define <16 x i8> @shuffle_v16i8_bitcast_unpack(<16 x i8> %a, <16 x i8> %b) {
2241 ; SSE-LABEL: shuffle_v16i8_bitcast_unpack:
2242 ; SSE: # %bb.0:
2243 ; SSE-NEXT: punpcklbw {{.*#+}} 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]
2244 ; SSE-NEXT: retq
2245 ;
2246 ; AVX-LABEL: shuffle_v16i8_bitcast_unpack:
2247 ; AVX: # %bb.0:
2248 ; AVX-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]
2249 ; AVX-NEXT: retq
2250 %shuffle8 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 7, i32 23, i32 6, i32 22, i32 5, i32 21, i32 4, i32 20, i32 3, i32 19, i32 2, i32 18, i32 1, i32 17, i32 0, i32 16>
2251 %bitcast32 = bitcast <16 x i8> %shuffle8 to <4 x float>
2252 %shuffle32 = shufflevector <4 x float> %bitcast32, <4 x float> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
2253 %bitcast16 = bitcast <4 x float> %shuffle32 to <8 x i16>
2254 %shuffle16 = shufflevector <8 x i16> %bitcast16, <8 x i16> undef, <8 x i32> <i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6>
2255 %bitcast8 = bitcast <8 x i16> %shuffle16 to <16 x i8>
2256 ret <16 x i8> %bitcast8
2257 }
2259 define <16 x i8> @insert_dup_mem_v16i8_i32(ptr %ptr) {
2260 ; SSE2-LABEL: insert_dup_mem_v16i8_i32:
2261 ; SSE2: # %bb.0:
2262 ; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2263 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
2264 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
2265 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
2266 ; SSE2-NEXT: retq
2267 ;
2268 ; SSSE3-LABEL: insert_dup_mem_v16i8_i32:
2269 ; SSSE3: # %bb.0:
2270 ; SSSE3-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2271 ; SSSE3-NEXT: pxor %xmm1, %xmm1
2272 ; SSSE3-NEXT: pshufb %xmm1, %xmm0
2273 ; SSSE3-NEXT: retq
2274 ;
2275 ; SSE41-LABEL: insert_dup_mem_v16i8_i32:
2276 ; SSE41: # %bb.0:
2277 ; SSE41-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2278 ; SSE41-NEXT: pxor %xmm1, %xmm1
2279 ; SSE41-NEXT: pshufb %xmm1, %xmm0
2280 ; SSE41-NEXT: retq
2281 ;
2282 ; AVX1-LABEL: insert_dup_mem_v16i8_i32:
2283 ; AVX1: # %bb.0:
2284 ; AVX1-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2285 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
2286 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm0
2287 ; AVX1-NEXT: retq
2288 ;
2289 ; AVX2OR512VL-LABEL: insert_dup_mem_v16i8_i32:
2290 ; AVX2OR512VL: # %bb.0:
2291 ; AVX2OR512VL-NEXT: vpbroadcastb (%rdi), %xmm0
2292 ; AVX2OR512VL-NEXT: retq
2293 ;
2294 ; XOPAVX1-LABEL: insert_dup_mem_v16i8_i32:
2295 ; XOPAVX1: # %bb.0:
2296 ; XOPAVX1-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2297 ; XOPAVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
2298 ; XOPAVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm0
2299 ; XOPAVX1-NEXT: retq
2300 ;
2301 ; XOPAVX2-LABEL: insert_dup_mem_v16i8_i32:
2302 ; XOPAVX2: # %bb.0:
2303 ; XOPAVX2-NEXT: vpbroadcastb (%rdi), %xmm0
2304 ; XOPAVX2-NEXT: retq
2305 %tmp = load i32, ptr %ptr, align 4
2306 %tmp1 = insertelement <4 x i32> zeroinitializer, i32 %tmp, i32 0
2307 %tmp2 = bitcast <4 x i32> %tmp1 to <16 x i8>
2308 %tmp3 = shufflevector <16 x i8> %tmp2, <16 x i8> undef, <16 x i32> zeroinitializer
2309 ret <16 x i8> %tmp3
2310 }
2312 define <16 x i8> @insert_dup_mem_v16i8_sext_i8(ptr %ptr) {
2313 ; SSE2-LABEL: insert_dup_mem_v16i8_sext_i8:
2314 ; SSE2: # %bb.0:
2315 ; SSE2-NEXT: movzbl (%rdi), %eax
2316 ; SSE2-NEXT: movd %eax, %xmm0
2317 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
2318 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
2319 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
2320 ; SSE2-NEXT: retq
2321 ;
2322 ; SSSE3-LABEL: insert_dup_mem_v16i8_sext_i8:
2323 ; SSSE3: # %bb.0:
2324 ; SSSE3-NEXT: movzbl (%rdi), %eax
2325 ; SSSE3-NEXT: movd %eax, %xmm0
2326 ; SSSE3-NEXT: pxor %xmm1, %xmm1
2327 ; SSSE3-NEXT: pshufb %xmm1, %xmm0
2328 ; SSSE3-NEXT: retq
2329 ;
2330 ; SSE41-LABEL: insert_dup_mem_v16i8_sext_i8:
2331 ; SSE41: # %bb.0:
2332 ; SSE41-NEXT: movzbl (%rdi), %eax
2333 ; SSE41-NEXT: movd %eax, %xmm0
2334 ; SSE41-NEXT: pxor %xmm1, %xmm1
2335 ; SSE41-NEXT: pshufb %xmm1, %xmm0
2336 ; SSE41-NEXT: retq
2337 ;
2338 ; AVX1-LABEL: insert_dup_mem_v16i8_sext_i8:
2339 ; AVX1: # %bb.0:
2340 ; AVX1-NEXT: movzbl (%rdi), %eax
2341 ; AVX1-NEXT: vmovd %eax, %xmm0
2342 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
2343 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm0
2344 ; AVX1-NEXT: retq
2345 ;
2346 ; AVX2OR512VL-LABEL: insert_dup_mem_v16i8_sext_i8:
2347 ; AVX2OR512VL: # %bb.0:
2348 ; AVX2OR512VL-NEXT: vpbroadcastb (%rdi), %xmm0
2349 ; AVX2OR512VL-NEXT: retq
2350 ;
2351 ; XOPAVX1-LABEL: insert_dup_mem_v16i8_sext_i8:
2352 ; XOPAVX1: # %bb.0:
2353 ; XOPAVX1-NEXT: movzbl (%rdi), %eax
2354 ; XOPAVX1-NEXT: vmovd %eax, %xmm0
2355 ; XOPAVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
2356 ; XOPAVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm0
2357 ; XOPAVX1-NEXT: retq
2358 ;
2359 ; XOPAVX2-LABEL: insert_dup_mem_v16i8_sext_i8:
2360 ; XOPAVX2: # %bb.0:
2361 ; XOPAVX2-NEXT: vpbroadcastb (%rdi), %xmm0
2362 ; XOPAVX2-NEXT: retq
2363 %tmp = load i8, ptr %ptr, align 1
2364 %tmp1 = sext i8 %tmp to i32
2365 %tmp2 = insertelement <4 x i32> zeroinitializer, i32 %tmp1, i32 0
2366 %tmp3 = bitcast <4 x i32> %tmp2 to <16 x i8>
2367 %tmp4 = shufflevector <16 x i8> %tmp3, <16 x i8> undef, <16 x i32> zeroinitializer
2368 ret <16 x i8> %tmp4
2369 }
2371 define <16 x i8> @insert_dup_elt1_mem_v16i8_i32(ptr %ptr) {
2372 ; SSE2-LABEL: insert_dup_elt1_mem_v16i8_i32:
2373 ; SSE2: # %bb.0:
2374 ; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2375 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
2376 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,1,1,1,4,5,6,7]
2377 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
2378 ; SSE2-NEXT: retq
2379 ;
2380 ; SSSE3-LABEL: insert_dup_elt1_mem_v16i8_i32:
2381 ; SSSE3: # %bb.0:
2382 ; SSSE3-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2383 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
2384 ; SSSE3-NEXT: retq
2385 ;
2386 ; SSE41-LABEL: insert_dup_elt1_mem_v16i8_i32:
2387 ; SSE41: # %bb.0:
2388 ; SSE41-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2389 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
2390 ; SSE41-NEXT: retq
2391 ;
2392 ; AVX1-LABEL: insert_dup_elt1_mem_v16i8_i32:
2393 ; AVX1: # %bb.0:
2394 ; AVX1-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2395 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
2396 ; AVX1-NEXT: retq
2397 ;
2398 ; AVX2OR512VL-LABEL: insert_dup_elt1_mem_v16i8_i32:
2399 ; AVX2OR512VL: # %bb.0:
2400 ; AVX2OR512VL-NEXT: vpbroadcastb 1(%rdi), %xmm0
2401 ; AVX2OR512VL-NEXT: retq
2402 ;
2403 ; XOPAVX1-LABEL: insert_dup_elt1_mem_v16i8_i32:
2404 ; XOPAVX1: # %bb.0:
2405 ; XOPAVX1-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2406 ; XOPAVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
2407 ; XOPAVX1-NEXT: retq
2408 ;
2409 ; XOPAVX2-LABEL: insert_dup_elt1_mem_v16i8_i32:
2410 ; XOPAVX2: # %bb.0:
2411 ; XOPAVX2-NEXT: vpbroadcastb 1(%rdi), %xmm0
2412 ; XOPAVX2-NEXT: retq
2413 %tmp = load i32, ptr %ptr, align 4
2414 %tmp1 = insertelement <4 x i32> zeroinitializer, i32 %tmp, i32 0
2415 %tmp2 = bitcast <4 x i32> %tmp1 to <16 x i8>
2416 %tmp3 = shufflevector <16 x i8> %tmp2, <16 x i8> undef, <16 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
2417 ret <16 x i8> %tmp3
2418 }
2420 define <16 x i8> @insert_dup_elt2_mem_v16i8_i32(ptr %ptr) {
2421 ; SSE2-LABEL: insert_dup_elt2_mem_v16i8_i32:
2422 ; SSE2: # %bb.0:
2423 ; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2424 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
2425 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,2,2,2,4,5,6,7]
2426 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
2427 ; SSE2-NEXT: retq
2428 ;
2429 ; SSSE3-LABEL: insert_dup_elt2_mem_v16i8_i32:
2430 ; SSSE3: # %bb.0:
2431 ; SSSE3-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2432 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
2433 ; SSSE3-NEXT: retq
2434 ;
2435 ; SSE41-LABEL: insert_dup_elt2_mem_v16i8_i32:
2436 ; SSE41: # %bb.0:
2437 ; SSE41-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2438 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
2439 ; SSE41-NEXT: retq
2440 ;
2441 ; AVX1-LABEL: insert_dup_elt2_mem_v16i8_i32:
2442 ; AVX1: # %bb.0:
2443 ; AVX1-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2444 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
2445 ; AVX1-NEXT: retq
2446 ;
2447 ; AVX2OR512VL-LABEL: insert_dup_elt2_mem_v16i8_i32:
2448 ; AVX2OR512VL: # %bb.0:
2449 ; AVX2OR512VL-NEXT: vpbroadcastb 2(%rdi), %xmm0
2450 ; AVX2OR512VL-NEXT: retq
2451 ;
2452 ; XOPAVX1-LABEL: insert_dup_elt2_mem_v16i8_i32:
2453 ; XOPAVX1: # %bb.0:
2454 ; XOPAVX1-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
2455 ; XOPAVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
2456 ; XOPAVX1-NEXT: retq
2457 ;
2458 ; XOPAVX2-LABEL: insert_dup_elt2_mem_v16i8_i32:
2459 ; XOPAVX2: # %bb.0:
2460 ; XOPAVX2-NEXT: vpbroadcastb 2(%rdi), %xmm0
2461 ; XOPAVX2-NEXT: retq
2462 %tmp = load i32, ptr %ptr, align 4
2463 %tmp1 = insertelement <4 x i32> zeroinitializer, i32 %tmp, i32 0
2464 %tmp2 = bitcast <4 x i32> %tmp1 to <16 x i8>
2465 %tmp3 = shufflevector <16 x i8> %tmp2, <16 x i8> undef, <16 x i32> <i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2>
2466 ret <16 x i8> %tmp3
2467 }
2469 define <16 x i8> @insert_dup_elt1_mem_v16i8_sext_i8(ptr %ptr) {
2470 ; SSE2-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
2471 ; SSE2: # %bb.0:
2472 ; SSE2-NEXT: movsbl (%rdi), %eax
2473 ; SSE2-NEXT: movd %eax, %xmm0
2474 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
2475 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,1,1,1,4,5,6,7]
2476 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
2477 ; SSE2-NEXT: retq
2478 ;
2479 ; SSSE3-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
2480 ; SSSE3: # %bb.0:
2481 ; SSSE3-NEXT: movsbl (%rdi), %eax
2482 ; SSSE3-NEXT: movd %eax, %xmm0
2483 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
2484 ; SSSE3-NEXT: retq
2485 ;
2486 ; SSE41-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
2487 ; SSE41: # %bb.0:
2488 ; SSE41-NEXT: movsbl (%rdi), %eax
2489 ; SSE41-NEXT: movd %eax, %xmm0
2490 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
2491 ; SSE41-NEXT: retq
2492 ;
2493 ; AVX1-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
2494 ; AVX1: # %bb.0:
2495 ; AVX1-NEXT: movsbl (%rdi), %eax
2496 ; AVX1-NEXT: vmovd %eax, %xmm0
2497 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
2498 ; AVX1-NEXT: retq
2499 ;
2500 ; AVX2-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
2501 ; AVX2: # %bb.0:
2502 ; AVX2-NEXT: movsbl (%rdi), %eax
2503 ; AVX2-NEXT: shrl $8, %eax
2504 ; AVX2-NEXT: vmovd %eax, %xmm0
2505 ; AVX2-NEXT: vpbroadcastb %xmm0, %xmm0
2506 ; AVX2-NEXT: retq
2507 ;
2508 ; AVX512VL-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
2509 ; AVX512VL: # %bb.0:
2510 ; AVX512VL-NEXT: movsbl (%rdi), %eax
2511 ; AVX512VL-NEXT: shrl $8, %eax
2512 ; AVX512VL-NEXT: vpbroadcastb %eax, %xmm0
2513 ; AVX512VL-NEXT: retq
2514 ;
2515 ; XOPAVX1-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
2516 ; XOPAVX1: # %bb.0:
2517 ; XOPAVX1-NEXT: movsbl (%rdi), %eax
2518 ; XOPAVX1-NEXT: vmovd %eax, %xmm0
2519 ; XOPAVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
2520 ; XOPAVX1-NEXT: retq
2521 ;
2522 ; XOPAVX2-LABEL: insert_dup_elt1_mem_v16i8_sext_i8:
2523 ; XOPAVX2: # %bb.0:
2524 ; XOPAVX2-NEXT: movsbl (%rdi), %eax
2525 ; XOPAVX2-NEXT: shrl $8, %eax
2526 ; XOPAVX2-NEXT: vmovd %eax, %xmm0
2527 ; XOPAVX2-NEXT: vpbroadcastb %xmm0, %xmm0
2528 ; XOPAVX2-NEXT: retq
2529 %tmp = load i8, ptr %ptr, align 1
2530 %tmp1 = sext i8 %tmp to i32
2531 %tmp2 = insertelement <4 x i32> zeroinitializer, i32 %tmp1, i32 0
2532 %tmp3 = bitcast <4 x i32> %tmp2 to <16 x i8>
2533 %tmp4 = shufflevector <16 x i8> %tmp3, <16 x i8> undef, <16 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
2534 ret <16 x i8> %tmp4
2535 }
2537 define <16 x i8> @insert_dup_elt2_mem_v16i8_sext_i8(ptr %ptr) {
2538 ; SSE2-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
2539 ; SSE2: # %bb.0:
2540 ; SSE2-NEXT: movsbl (%rdi), %eax
2541 ; SSE2-NEXT: movd %eax, %xmm0
2542 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
2543 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[2,2,2,2,4,5,6,7]
2544 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
2545 ; SSE2-NEXT: retq
2546 ;
2547 ; SSSE3-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
2548 ; SSSE3: # %bb.0:
2549 ; SSSE3-NEXT: movsbl (%rdi), %eax
2550 ; SSSE3-NEXT: movd %eax, %xmm0
2551 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
2552 ; SSSE3-NEXT: retq
2553 ;
2554 ; SSE41-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
2555 ; SSE41: # %bb.0:
2556 ; SSE41-NEXT: movsbl (%rdi), %eax
2557 ; SSE41-NEXT: movd %eax, %xmm0
2558 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
2559 ; SSE41-NEXT: retq
2560 ;
2561 ; AVX1-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
2562 ; AVX1: # %bb.0:
2563 ; AVX1-NEXT: movsbl (%rdi), %eax
2564 ; AVX1-NEXT: vmovd %eax, %xmm0
2565 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
2566 ; AVX1-NEXT: retq
2567 ;
2568 ; AVX2-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
2569 ; AVX2: # %bb.0:
2570 ; AVX2-NEXT: movsbl (%rdi), %eax
2571 ; AVX2-NEXT: shrl $16, %eax
2572 ; AVX2-NEXT: vmovd %eax, %xmm0
2573 ; AVX2-NEXT: vpbroadcastb %xmm0, %xmm0
2574 ; AVX2-NEXT: retq
2575 ;
2576 ; AVX512VL-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
2577 ; AVX512VL: # %bb.0:
2578 ; AVX512VL-NEXT: movsbl (%rdi), %eax
2579 ; AVX512VL-NEXT: shrl $16, %eax
2580 ; AVX512VL-NEXT: vpbroadcastb %eax, %xmm0
2581 ; AVX512VL-NEXT: retq
2582 ;
2583 ; XOPAVX1-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
2584 ; XOPAVX1: # %bb.0:
2585 ; XOPAVX1-NEXT: movsbl (%rdi), %eax
2586 ; XOPAVX1-NEXT: vmovd %eax, %xmm0
2587 ; XOPAVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2]
2588 ; XOPAVX1-NEXT: retq
2589 ;
2590 ; XOPAVX2-LABEL: insert_dup_elt2_mem_v16i8_sext_i8:
2591 ; XOPAVX2: # %bb.0:
2592 ; XOPAVX2-NEXT: movsbl (%rdi), %eax
2593 ; XOPAVX2-NEXT: shrl $16, %eax
2594 ; XOPAVX2-NEXT: vmovd %eax, %xmm0
2595 ; XOPAVX2-NEXT: vpbroadcastb %xmm0, %xmm0
2596 ; XOPAVX2-NEXT: retq
2597 %tmp = load i8, ptr %ptr, align 1
2598 %tmp1 = sext i8 %tmp to i32
2599 %tmp2 = insertelement <4 x i32> zeroinitializer, i32 %tmp1, i32 0
2600 %tmp3 = bitcast <4 x i32> %tmp2 to <16 x i8>
2601 %tmp4 = shufflevector <16 x i8> %tmp3, <16 x i8> undef, <16 x i32> <i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2>
2602 ret <16 x i8> %tmp4
2603 }
2605 define <16 x i8> @PR31364(ptr nocapture readonly %a, ptr nocapture readonly %b) {
2606 ; SSE2-LABEL: PR31364:
2607 ; SSE2: # %bb.0:
2608 ; SSE2-NEXT: movzbl (%rdi), %eax
2609 ; SSE2-NEXT: movzbl (%rsi), %ecx
2610 ; SSE2-NEXT: shll $8, %ecx
2611 ; SSE2-NEXT: orl %eax, %ecx
2612 ; SSE2-NEXT: movd %ecx, %xmm1
2613 ; SSE2-NEXT: pxor %xmm0, %xmm0
2614 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
2615 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm1[1,1,1,3,4,5,6,7]
2616 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,0,1]
2617 ; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,0,0,0]
2618 ; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[1,1,1,1,4,5,6,7]
2619 ; SSE2-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,5,4,4,4]
2620 ; SSE2-NEXT: packuswb %xmm1, %xmm0
2621 ; SSE2-NEXT: retq
2622 ;
2623 ; SSSE3-LABEL: PR31364:
2624 ; SSSE3: # %bb.0:
2625 ; SSSE3-NEXT: movzbl (%rdi), %eax
2626 ; SSSE3-NEXT: movzbl (%rsi), %ecx
2627 ; SSSE3-NEXT: shll $8, %ecx
2628 ; SSSE3-NEXT: orl %eax, %ecx
2629 ; SSSE3-NEXT: movd %ecx, %xmm0
2630 ; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1],zero,xmm0[1,1,1,1,1,0,0,0]
2631 ; SSSE3-NEXT: retq
2632 ;
2633 ; SSE41-LABEL: PR31364:
2634 ; SSE41: # %bb.0:
2635 ; SSE41-NEXT: movzbl (%rdi), %eax
2636 ; SSE41-NEXT: movd %eax, %xmm0
2637 ; SSE41-NEXT: pinsrb $1, (%rsi), %xmm0
2638 ; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1],zero,xmm0[1,1,1,1,1,0,0,0]
2639 ; SSE41-NEXT: retq
2640 ;
2641 ; AVX-LABEL: PR31364:
2642 ; AVX: # %bb.0:
2643 ; AVX-NEXT: movzbl (%rdi), %eax
2644 ; AVX-NEXT: vmovd %eax, %xmm0
2645 ; AVX-NEXT: vpinsrb $1, (%rsi), %xmm0, %xmm0
2646 ; AVX-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[1,1,1,1,1,1,1],zero,xmm0[1,1,1,1,1,0,0,0]
2647 ; AVX-NEXT: retq
2648 %v0 = load i8, ptr %a, align 1
2649 %vecins = insertelement <16 x i8> <i8 undef, i8 undef, i8 undef, i8 0, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef>, i8 %v0, i32 0
2650 %v1 = load i8, ptr %b, align 1
2651 %vecins2 = insertelement <16 x i8> %vecins, i8 %v1, i32 1
2652 %result = shufflevector <16 x i8> %vecins2, <16 x i8> undef, <16 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 3, i32 1, i32 1, i32 1, i32 1, i32 1, i32 0, i32 0, i32 0>
2653 ret <16 x i8> %result
2654 }
2656 define <16 x i8> @PR31301(ptr nocapture readonly %x, ptr nocapture readonly %y) {
2657 ; SSE2-LABEL: PR31301:
2658 ; SSE2: # %bb.0: # %entry
2659 ; SSE2-NEXT: movzbl (%rdi), %eax
2660 ; SSE2-NEXT: movd %eax, %xmm0
2661 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
2662 ; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,0,0,0,4,5,6,7]
2663 ; SSE2-NEXT: movzbl (%rsi), %eax
2664 ; SSE2-NEXT: movd %eax, %xmm1
2665 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
2666 ; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,0,0,0,4,5,6,7]
2667 ; SSE2-NEXT: punpcklbw {{.*#+}} 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]
2668 ; SSE2-NEXT: retq
2669 ;
2670 ; SSSE3-LABEL: PR31301:
2671 ; SSSE3: # %bb.0: # %entry
2672 ; SSSE3-NEXT: movzbl (%rdi), %eax
2673 ; SSSE3-NEXT: movd %eax, %xmm0
2674 ; SSSE3-NEXT: pxor %xmm1, %xmm1
2675 ; SSSE3-NEXT: pshufb %xmm1, %xmm0
2676 ; SSSE3-NEXT: movzbl (%rsi), %eax
2677 ; SSSE3-NEXT: movd %eax, %xmm2
2678 ; SSSE3-NEXT: pshufb %xmm1, %xmm2
2679 ; SSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
2680 ; SSSE3-NEXT: retq
2681 ;
2682 ; SSE41-LABEL: PR31301:
2683 ; SSE41: # %bb.0: # %entry
2684 ; SSE41-NEXT: movzbl (%rdi), %eax
2685 ; SSE41-NEXT: movd %eax, %xmm0
2686 ; SSE41-NEXT: pxor %xmm1, %xmm1
2687 ; SSE41-NEXT: pshufb %xmm1, %xmm0
2688 ; SSE41-NEXT: movzbl (%rsi), %eax
2689 ; SSE41-NEXT: movd %eax, %xmm2
2690 ; SSE41-NEXT: pshufb %xmm1, %xmm2
2691 ; SSE41-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
2692 ; SSE41-NEXT: retq
2693 ;
2694 ; AVX1-LABEL: PR31301:
2695 ; AVX1: # %bb.0: # %entry
2696 ; AVX1-NEXT: movzbl (%rdi), %eax
2697 ; AVX1-NEXT: vmovd %eax, %xmm0
2698 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
2699 ; AVX1-NEXT: vpshufb %xmm1, %xmm0, %xmm0
2700 ; AVX1-NEXT: movzbl (%rsi), %eax
2701 ; AVX1-NEXT: vmovd %eax, %xmm2
2702 ; AVX1-NEXT: vpshufb %xmm1, %xmm2, %xmm1
2703 ; 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]
2704 ; AVX1-NEXT: retq
2705 ;
2706 ; AVX2OR512VL-LABEL: PR31301:
2707 ; AVX2OR512VL: # %bb.0: # %entry
2708 ; AVX2OR512VL-NEXT: vpbroadcastb (%rdi), %xmm0
2709 ; AVX2OR512VL-NEXT: vpbroadcastb (%rsi), %xmm1
2710 ; AVX2OR512VL-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]
2711 ; AVX2OR512VL-NEXT: retq
2712 ;
2713 ; XOPAVX1-LABEL: PR31301:
2714 ; XOPAVX1: # %bb.0: # %entry
2715 ; XOPAVX1-NEXT: movzbl (%rdi), %eax
2716 ; XOPAVX1-NEXT: vmovd %eax, %xmm0
2717 ; XOPAVX1-NEXT: movzbl (%rsi), %eax
2718 ; XOPAVX1-NEXT: vmovd %eax, %xmm1
2719 ; XOPAVX1-NEXT: vpperm {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[0],xmm1[0],xmm0[0],xmm1[0],xmm0[0],xmm1[0],xmm0[0],xmm1[0],xmm0[0],xmm1[0],xmm0[0],xmm1[0],xmm0[0],xmm1[0]
2720 ; XOPAVX1-NEXT: retq
2721 ;
2722 ; XOPAVX2-LABEL: PR31301:
2723 ; XOPAVX2: # %bb.0: # %entry
2724 ; XOPAVX2-NEXT: vpbroadcastb (%rdi), %xmm0
2725 ; XOPAVX2-NEXT: vpbroadcastb (%rsi), %xmm1
2726 ; XOPAVX2-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]
2727 ; XOPAVX2-NEXT: retq
2728 entry:
2729 %0 = load i8, ptr %x, align 1
2730 %1 = insertelement <16 x i8> undef, i8 %0, i32 0
2731 %lane = shufflevector <16 x i8> %1, <16 x i8> undef, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
2732 %2 = load i8, ptr %y, align 1
2733 %3 = insertelement <16 x i8> undef, i8 %2, i32 0
2734 %lane3 = shufflevector <16 x i8> %3, <16 x i8> undef, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
2735 %vzip.i = shufflevector <16 x i8> %lane, <16 x i8> %lane3, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
2736 ret <16 x i8> %vzip.i
2737 }