1 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+sse2 | FileCheck %s --check-prefix=SSE --check-prefix=SSE2
2 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+ssse3 | FileCheck %s --check-prefix=SSE --check-prefix=SSSE3
3 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+sse4.1 | FileCheck %s --check-prefix=SSE --check-prefix=SSE41
4 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+avx | FileCheck %s --check-prefix=AVX --check-prefix=AVX1
5 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+avx2 | FileCheck %s --check-prefix=AVX --check-prefix=AVX2
7 define <8 x i32> @zext_8i16_to_8i32(<8 x i16> %A) nounwind uwtable readnone ssp {
8 ; SSE2-LABEL: zext_8i16_to_8i32:
9 ; SSE2: # BB#0: # %entry
10 ; SSE2-NEXT: movdqa %xmm0, %xmm1
11 ; SSE2-NEXT: pxor %xmm2, %xmm2
12 ; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
13 ; SSE2-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
14 ; SSE2-NEXT: pand .LCPI0_0(%rip), %xmm1
17 ; SSSE3-LABEL: zext_8i16_to_8i32:
18 ; SSSE3: # BB#0: # %entry
19 ; SSSE3-NEXT: movdqa %xmm0, %xmm1
20 ; SSSE3-NEXT: pxor %xmm2, %xmm2
21 ; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
22 ; SSSE3-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
23 ; SSSE3-NEXT: pand .LCPI0_0(%rip), %xmm1
26 ; SSE41-LABEL: zext_8i16_to_8i32:
27 ; SSE41: # BB#0: # %entry
28 ; SSE41-NEXT: movdqa %xmm0, %xmm1
29 ; SSE41-NEXT: pmovzxwd {{.*#+}} xmm0 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero
30 ; SSE41-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
31 ; SSE41-NEXT: pand .LCPI0_0(%rip), %xmm1
34 ; AVX1-LABEL: zext_8i16_to_8i32:
35 ; AVX1: # BB#0: # %entry
36 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
37 ; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm1 = xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
38 ; AVX1-NEXT: vpmovzxwd {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
39 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
42 ; AVX2-LABEL: zext_8i16_to_8i32:
43 ; AVX2: # BB#0: # %entry
44 ; AVX2-NEXT: vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
47 %B = zext <8 x i16> %A to <8 x i32>
51 define <4 x i64> @zext_4i32_to_4i64(<4 x i32> %A) nounwind uwtable readnone ssp {
52 ; SSE2-LABEL: zext_4i32_to_4i64:
53 ; SSE2: # BB#0: # %entry
54 ; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,1,1,3]
55 ; SSE2-NEXT: movdqa {{.*#+}} xmm3 = [4294967295,4294967295]
56 ; SSE2-NEXT: pand %xmm3, %xmm2
57 ; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,2,3,3]
58 ; SSE2-NEXT: pand %xmm3, %xmm1
59 ; SSE2-NEXT: movdqa %xmm2, %xmm0
62 ; SSSE3-LABEL: zext_4i32_to_4i64:
63 ; SSSE3: # BB#0: # %entry
64 ; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,1,1,3]
65 ; SSSE3-NEXT: movdqa {{.*#+}} xmm3 = [4294967295,4294967295]
66 ; SSSE3-NEXT: pand %xmm3, %xmm2
67 ; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,2,3,3]
68 ; SSSE3-NEXT: pand %xmm3, %xmm1
69 ; SSSE3-NEXT: movdqa %xmm2, %xmm0
72 ; SSE41-LABEL: zext_4i32_to_4i64:
73 ; SSE41: # BB#0: # %entry
74 ; SSE41-NEXT: pmovzxdq {{.*#+}} xmm2 = xmm0[0],zero,xmm0[1],zero
75 ; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [4294967295,4294967295]
76 ; SSE41-NEXT: pand %xmm3, %xmm2
77 ; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,2,3,3]
78 ; SSE41-NEXT: pand %xmm3, %xmm1
79 ; SSE41-NEXT: movdqa %xmm2, %xmm0
82 ; AVX1-LABEL: zext_4i32_to_4i64:
83 ; AVX1: # BB#0: # %entry
84 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
85 ; AVX1-NEXT: vpunpckhdq {{.*#+}} xmm1 = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
86 ; AVX1-NEXT: vpmovzxdq {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero
87 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
90 ; AVX2-LABEL: zext_4i32_to_4i64:
91 ; AVX2: # BB#0: # %entry
92 ; AVX2-NEXT: vpmovzxdq {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
95 %B = zext <4 x i32> %A to <4 x i64>
99 define <8 x i32> @zext_8i8_to_8i32(<8 x i8> %z) {
100 ; SSE2-LABEL: zext_8i8_to_8i32:
101 ; SSE2: # BB#0: # %entry
102 ; SSE2-NEXT: movdqa %xmm0, %xmm2
103 ; SSE2-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3]
104 ; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [255,255,255,255]
105 ; SSE2-NEXT: pand %xmm1, %xmm2
106 ; SSE2-NEXT: punpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
107 ; SSE2-NEXT: pand %xmm0, %xmm1
108 ; SSE2-NEXT: movdqa %xmm2, %xmm0
111 ; SSSE3-LABEL: zext_8i8_to_8i32:
112 ; SSSE3: # BB#0: # %entry
113 ; SSSE3-NEXT: movdqa %xmm0, %xmm2
114 ; SSSE3-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3]
115 ; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [255,255,255,255]
116 ; SSSE3-NEXT: pand %xmm1, %xmm2
117 ; SSSE3-NEXT: punpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
118 ; SSSE3-NEXT: pand %xmm0, %xmm1
119 ; SSSE3-NEXT: movdqa %xmm2, %xmm0
122 ; SSE41-LABEL: zext_8i8_to_8i32:
123 ; SSE41: # BB#0: # %entry
124 ; SSE41-NEXT: pmovzxwd {{.*#+}} xmm2 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
125 ; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [255,255,255,255]
126 ; SSE41-NEXT: pand %xmm1, %xmm2
127 ; SSE41-NEXT: punpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
128 ; SSE41-NEXT: pand %xmm0, %xmm1
129 ; SSE41-NEXT: movdqa %xmm2, %xmm0
132 ; AVX1-LABEL: zext_8i8_to_8i32:
133 ; AVX1: # BB#0: # %entry
134 ; AVX1-NEXT: vpmovzxwd {{.*#+}} xmm1 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
135 ; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
136 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
137 ; AVX1-NEXT: vandps {{.*}}(%rip), %ymm0, %ymm0
140 ; AVX2-LABEL: zext_8i8_to_8i32:
141 ; AVX2: # BB#0: # %entry
142 ; AVX2-NEXT: vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
143 ; AVX2-NEXT: vpbroadcastd {{.*}}(%rip), %ymm1
144 ; AVX2-NEXT: vpand %ymm1, %ymm0, %ymm0
147 %t = zext <8 x i8> %z to <8 x i32>
152 define <16 x i16> @zext_16i8_to_16i16(<16 x i8> %z) {
153 ; SSE2-LABEL: zext_16i8_to_16i16:
154 ; SSE2: # BB#0: # %entry
155 ; SSE2-NEXT: movdqa %xmm0, %xmm1
156 ; SSE2-NEXT: pxor %xmm2, %xmm2
157 ; 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]
158 ; SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
159 ; SSE2-NEXT: pand .LCPI3_0(%rip), %xmm1
162 ; SSSE3-LABEL: zext_16i8_to_16i16:
163 ; SSSE3: # BB#0: # %entry
164 ; SSSE3-NEXT: movdqa %xmm0, %xmm1
165 ; SSSE3-NEXT: pxor %xmm2, %xmm2
166 ; 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]
167 ; SSSE3-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
168 ; SSSE3-NEXT: pand .LCPI3_0(%rip), %xmm1
171 ; SSE41-LABEL: zext_16i8_to_16i16:
172 ; SSE41: # BB#0: # %entry
173 ; SSE41-NEXT: movdqa %xmm0, %xmm1
174 ; SSE41-NEXT: pmovzxbw %xmm1, %xmm0 {{.*#+}} xmm0 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
175 ; SSE41-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
176 ; SSE41-NEXT: pand .LCPI3_0(%rip), %xmm1
179 ; AVX1-LABEL: zext_16i8_to_16i16:
180 ; AVX1: # BB#0: # %entry
181 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
182 ; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm1 = 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]
183 ; AVX1-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
184 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
187 ; AVX2-LABEL: zext_16i8_to_16i16:
188 ; AVX2: # BB#0: # %entry
189 ; AVX2-NEXT: vpmovzxbw {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
192 %t = zext <16 x i8> %z to <16 x i16>
196 define <16 x i16> @load_zext_16i8_to_16i16(<16 x i8> *%ptr) {
197 ; SSE2-LABEL: load_zext_16i8_to_16i16:
198 ; SSE2: # BB#0: # %entry
199 ; SSE2-NEXT: movdqa (%rdi), %xmm1
200 ; SSE2-NEXT: pxor %xmm2, %xmm2
201 ; SSE2-NEXT: movdqa %xmm1, %xmm0
202 ; 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]
203 ; SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
204 ; SSE2-NEXT: pand .LCPI4_0(%rip), %xmm1
207 ; SSSE3-LABEL: load_zext_16i8_to_16i16:
208 ; SSSE3: # BB#0: # %entry
209 ; SSSE3-NEXT: movdqa (%rdi), %xmm1
210 ; SSSE3-NEXT: pxor %xmm2, %xmm2
211 ; SSSE3-NEXT: movdqa %xmm1, %xmm0
212 ; 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]
213 ; SSSE3-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
214 ; SSSE3-NEXT: pand .LCPI4_0(%rip), %xmm1
217 ; SSE41-LABEL: load_zext_16i8_to_16i16:
218 ; SSE41: # BB#0: # %entry
219 ; SSE41-NEXT: pmovzxbw {{.*#+}} xmm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero,mem[4],zero,mem[5],zero,mem[6],zero,mem[7],zero
220 ; SSE41-NEXT: pmovzxbw {{.*#+}} xmm1 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero,mem[4],zero,mem[5],zero,mem[6],zero,mem[7],zero
223 ; AVX1-LABEL: load_zext_16i8_to_16i16:
224 ; AVX1: # BB#0: # %entry
225 ; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero,mem[4],zero,mem[5],zero,mem[6],zero,mem[7],zero
226 ; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm1 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero,mem[4],zero,mem[5],zero,mem[6],zero,mem[7],zero
227 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
230 ; AVX2-LABEL: load_zext_16i8_to_16i16:
231 ; AVX2: # BB#0: # %entry
232 ; AVX2-NEXT: vpmovzxbw {{.*#+}} ymm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero,mem[4],zero,mem[5],zero,mem[6],zero,mem[7],zero,mem[8],zero,mem[9],zero,mem[10],zero,mem[11],zero,mem[12],zero,mem[13],zero,mem[14],zero,mem[15],zero
235 %X = load <16 x i8>, <16 x i8>* %ptr
236 %Y = zext <16 x i8> %X to <16 x i16>
240 define <8 x i32> @load_zext_8i16_to_8i32(<8 x i16> *%ptr) {
241 ; SSE2-LABEL: load_zext_8i16_to_8i32:
242 ; SSE2: # BB#0: # %entry
243 ; SSE2-NEXT: movdqa (%rdi), %xmm1
244 ; SSE2-NEXT: pxor %xmm2, %xmm2
245 ; SSE2-NEXT: movdqa %xmm1, %xmm0
246 ; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
247 ; SSE2-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
248 ; SSE2-NEXT: pand .LCPI5_0(%rip), %xmm1
251 ; SSSE3-LABEL: load_zext_8i16_to_8i32:
252 ; SSSE3: # BB#0: # %entry
253 ; SSSE3-NEXT: movdqa (%rdi), %xmm1
254 ; SSSE3-NEXT: pxor %xmm2, %xmm2
255 ; SSSE3-NEXT: movdqa %xmm1, %xmm0
256 ; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
257 ; SSSE3-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
258 ; SSSE3-NEXT: pand .LCPI5_0(%rip), %xmm1
261 ; SSE41-LABEL: load_zext_8i16_to_8i32:
262 ; SSE41: # BB#0: # %entry
263 ; SSE41-NEXT: pmovzxwd {{.*#+}} xmm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero
264 ; SSE41-NEXT: pmovzxwd {{.*#+}} xmm1 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero
267 ; AVX1-LABEL: load_zext_8i16_to_8i32:
268 ; AVX1: # BB#0: # %entry
269 ; AVX1-NEXT: vpmovzxwd {{.*#+}} xmm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero
270 ; AVX1-NEXT: vpmovzxwd {{.*#+}} xmm1 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero
271 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
274 ; AVX2-LABEL: load_zext_8i16_to_8i32:
275 ; AVX2: # BB#0: # %entry
276 ; AVX2-NEXT: vpmovzxwd {{.*#+}} ymm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero,mem[4],zero,mem[5],zero,mem[6],zero,mem[7],zero
279 %X = load <8 x i16>, <8 x i16>* %ptr
280 %Y = zext <8 x i16> %X to <8 x i32>
284 define <4 x i64> @load_zext_4i32_to_4i64(<4 x i32> *%ptr) {
285 ; SSE2-LABEL: load_zext_4i32_to_4i64:
286 ; SSE2: # BB#0: # %entry
287 ; SSE2-NEXT: movdqa (%rdi), %xmm1
288 ; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,1,1,3]
289 ; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [4294967295,4294967295]
290 ; SSE2-NEXT: pand %xmm2, %xmm0
291 ; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,2,3,3]
292 ; SSE2-NEXT: pand %xmm2, %xmm1
295 ; SSSE3-LABEL: load_zext_4i32_to_4i64:
296 ; SSSE3: # BB#0: # %entry
297 ; SSSE3-NEXT: movdqa (%rdi), %xmm1
298 ; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,1,1,3]
299 ; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [4294967295,4294967295]
300 ; SSSE3-NEXT: pand %xmm2, %xmm0
301 ; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,2,3,3]
302 ; SSSE3-NEXT: pand %xmm2, %xmm1
305 ; SSE41-LABEL: load_zext_4i32_to_4i64:
306 ; SSE41: # BB#0: # %entry
307 ; SSE41-NEXT: pmovzxdq {{.*#+}} xmm0 = mem[0],zero,mem[1],zero
308 ; SSE41-NEXT: pmovzxdq {{.*#+}} xmm1 = mem[0],zero,mem[1],zero
311 ; AVX1-LABEL: load_zext_4i32_to_4i64:
312 ; AVX1: # BB#0: # %entry
313 ; AVX1-NEXT: vpmovzxdq {{.*#+}} xmm0 = mem[0],zero,mem[1],zero
314 ; AVX1-NEXT: vpmovzxdq {{.*#+}} xmm1 = mem[0],zero,mem[1],zero
315 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
318 ; AVX2-LABEL: load_zext_4i32_to_4i64:
319 ; AVX2: # BB#0: # %entry
320 ; AVX2-NEXT: vpmovzxdq {{.*#+}} ymm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero
323 %X = load <4 x i32>, <4 x i32>* %ptr
324 %Y = zext <4 x i32> %X to <4 x i64>
328 define <8 x i32> @shuf_zext_8i16_to_8i32(<8 x i16> %A) nounwind uwtable readnone ssp {
329 ; SSE2-LABEL: shuf_zext_8i16_to_8i32:
330 ; SSE2: # BB#0: # %entry
331 ; SSE2-NEXT: movdqa %xmm0, %xmm1
332 ; SSE2-NEXT: pxor %xmm2, %xmm2
333 ; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
334 ; SSE2-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
337 ; SSSE3-LABEL: shuf_zext_8i16_to_8i32:
338 ; SSSE3: # BB#0: # %entry
339 ; SSSE3-NEXT: movdqa %xmm0, %xmm1
340 ; SSSE3-NEXT: pxor %xmm2, %xmm2
341 ; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
342 ; SSSE3-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
345 ; SSE41-LABEL: shuf_zext_8i16_to_8i32:
346 ; SSE41: # BB#0: # %entry
347 ; SSE41-NEXT: movdqa %xmm0, %xmm1
348 ; SSE41-NEXT: pxor %xmm2, %xmm2
349 ; SSE41-NEXT: pmovzxwd {{.*#+}} xmm0 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero
350 ; SSE41-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
353 ; AVX1-LABEL: shuf_zext_8i16_to_8i32:
354 ; AVX1: # BB#0: # %entry
355 ; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
356 ; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm1 = xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
357 ; AVX1-NEXT: vpmovzxwd {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
358 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
361 ; AVX2-LABEL: shuf_zext_8i16_to_8i32:
362 ; AVX2: # BB#0: # %entry
363 ; AVX2-NEXT: vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
366 %B = shufflevector <8 x i16> %A, <8 x i16> zeroinitializer, <16 x i32> <i32 0, i32 8, i32 1, i32 8, i32 2, i32 8, i32 3, i32 8, i32 4, i32 8, i32 5, i32 8, i32 6, i32 8, i32 7, i32 8>
367 %Z = bitcast <16 x i16> %B to <8 x i32>
371 define <4 x i64> @shuf_zext_4i32_to_4i64(<4 x i32> %A) nounwind uwtable readnone ssp {
372 ; SSE2-LABEL: shuf_zext_4i32_to_4i64:
373 ; SSE2: # BB#0: # %entry
374 ; SSE2-NEXT: movdqa %xmm0, %xmm1
375 ; SSE2-NEXT: pxor %xmm2, %xmm2
376 ; SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
377 ; SSE2-NEXT: punpckhdq {{.*#+}} xmm1 = xmm1[2],xmm2[2],xmm1[3],xmm2[3]
380 ; SSSE3-LABEL: shuf_zext_4i32_to_4i64:
381 ; SSSE3: # BB#0: # %entry
382 ; SSSE3-NEXT: movdqa %xmm0, %xmm1
383 ; SSSE3-NEXT: pxor %xmm2, %xmm2
384 ; SSSE3-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
385 ; SSSE3-NEXT: punpckhdq {{.*#+}} xmm1 = xmm1[2],xmm2[2],xmm1[3],xmm2[3]
388 ; SSE41-LABEL: shuf_zext_4i32_to_4i64:
389 ; SSE41: # BB#0: # %entry
390 ; SSE41-NEXT: movdqa %xmm0, %xmm1
391 ; SSE41-NEXT: pxor %xmm2, %xmm2
392 ; SSE41-NEXT: pmovzxdq {{.*#+}} xmm0 = xmm1[0],zero,xmm1[1],zero
393 ; SSE41-NEXT: punpckhdq {{.*#+}} xmm1 = xmm1[2],xmm2[2],xmm1[3],xmm2[3]
396 ; AVX1-LABEL: shuf_zext_4i32_to_4i64:
397 ; AVX1: # BB#0: # %entry
398 ; AVX1-NEXT: vinsertps {{.*#+}} xmm1 = xmm0[0],zero,xmm0[1],zero
399 ; AVX1-NEXT: vxorpd %xmm2, %xmm2, %xmm2
400 ; AVX1-NEXT: vblendpd {{.*#+}} xmm0 = xmm2[0],xmm0[1]
401 ; AVX1-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[2,0,3,0]
402 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
405 ; AVX2-LABEL: shuf_zext_4i32_to_4i64:
406 ; AVX2: # BB#0: # %entry
407 ; AVX2-NEXT: vpmovzxdq {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
410 %B = shufflevector <4 x i32> %A, <4 x i32> zeroinitializer, <8 x i32> <i32 0, i32 4, i32 1, i32 4, i32 2, i32 4, i32 3, i32 4>
411 %Z = bitcast <8 x i32> %B to <4 x i64>
415 define <8 x i32> @shuf_zext_8i8_to_8i32(<8 x i8> %A) {
416 ; SSE2-LABEL: shuf_zext_8i8_to_8i32:
417 ; SSE2: # BB#0: # %entry
418 ; SSE2-NEXT: pand .LCPI9_0(%rip), %xmm0
419 ; SSE2-NEXT: packuswb %xmm0, %xmm0
420 ; SSE2-NEXT: pxor %xmm1, %xmm1
421 ; SSE2-NEXT: movdqa %xmm0, %xmm2
422 ; 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]
423 ; SSE2-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
424 ; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
425 ; SSE2-NEXT: punpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
426 ; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [0,255,255,255,0,255,255,255,0,255,255,255,0,255,255,255]
427 ; SSE2-NEXT: pandn %xmm0, %xmm1
428 ; SSE2-NEXT: movdqa %xmm2, %xmm0
431 ; SSSE3-LABEL: shuf_zext_8i8_to_8i32:
432 ; SSSE3: # BB#0: # %entry
433 ; SSSE3-NEXT: movdqa %xmm0, %xmm1
434 ; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = xmm1[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
435 ; SSSE3-NEXT: pxor %xmm2, %xmm2
436 ; SSSE3-NEXT: movdqa %xmm1, %xmm0
437 ; 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]
438 ; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
439 ; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = xmm1[4],zero,zero,zero,xmm1[5],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[7],zero,zero,zero
442 ; SSE41-LABEL: shuf_zext_8i8_to_8i32:
443 ; SSE41: # BB#0: # %entry
444 ; SSE41-NEXT: movdqa %xmm0, %xmm1
445 ; SSE41-NEXT: pshufb {{.*#+}} xmm1 = xmm1[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
446 ; SSE41-NEXT: pmovzxbd {{.*#+}} xmm0 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[3],zero,zero,zero
447 ; SSE41-NEXT: pshufb {{.*#+}} xmm1 = xmm1[4],zero,zero,zero,xmm1[5],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[7],zero,zero,zero
450 ; AVX1-LABEL: shuf_zext_8i8_to_8i32:
451 ; AVX1: # BB#0: # %entry
452 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
453 ; AVX1-NEXT: vpmovzxbd {{.*#+}} xmm1 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
454 ; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero
455 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
458 ; AVX2-LABEL: shuf_zext_8i8_to_8i32:
459 ; AVX2: # BB#0: # %entry
460 ; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
461 ; AVX2-NEXT: vpmovzxbd {{.*#+}} ymm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero
464 %B = shufflevector <8 x i8> %A, <8 x i8> zeroinitializer, <32 x i32> <i32 0, i32 8, i32 8, i32 8, i32 1, i32 8, i32 8, i32 8, i32 2, i32 8, i32 8, i32 8, i32 3, i32 8, i32 8, i32 8, i32 4, i32 8, i32 8, i32 8, i32 5, i32 8, i32 8, i32 8, i32 6, i32 8, i32 8, i32 8, i32 7, i32 8, i32 8, i32 8>
465 %Z = bitcast <32 x i8> %B to <8 x i32>