1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2; RUN: llc < %s -disable-peephole -mtriple=i686-apple-darwin -mattr=+avx512f,avx512bw | FileCheck %s --check-prefix=X32
3; RUN: llc < %s -disable-peephole -mtriple=x86_64-apple-darwin -mattr=+avx512f,avx512bw | FileCheck %s --check-prefix=X64
4
5define <32 x i16> @test_llvm_x86_avx512_pmovsxbw(<32 x i8>* %a) {
6; X32-LABEL: test_llvm_x86_avx512_pmovsxbw:
7; X32:       ## %bb.0:
8; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
9; X32-NEXT:    vpmovsxbw (%eax), %zmm0
10; X32-NEXT:    retl
11;
12; X64-LABEL: test_llvm_x86_avx512_pmovsxbw:
13; X64:       ## %bb.0:
14; X64-NEXT:    vpmovsxbw (%rdi), %zmm0
15; X64-NEXT:    retq
16  %1 = load <32 x i8>, <32 x i8>* %a, align 1
17  %2 = sext <32 x i8> %1 to <32 x i16>
18  ret <32 x i16> %2
19}
20
21define <16 x i32> @test_llvm_x86_avx512_pmovsxbd(<16 x i8>* %a) {
22; X32-LABEL: test_llvm_x86_avx512_pmovsxbd:
23; X32:       ## %bb.0:
24; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
25; X32-NEXT:    vpmovsxbd (%eax), %zmm0
26; X32-NEXT:    retl
27;
28; X64-LABEL: test_llvm_x86_avx512_pmovsxbd:
29; X64:       ## %bb.0:
30; X64-NEXT:    vpmovsxbd (%rdi), %zmm0
31; X64-NEXT:    retq
32  %1 = load <16 x i8>, <16 x i8>* %a, align 1
33  %2 = sext <16 x i8> %1 to <16 x i32>
34  ret <16 x i32> %2
35}
36
37define <8 x i64> @test_llvm_x86_avx512_pmovsxbq(<16 x i8>* %a) {
38; X32-LABEL: test_llvm_x86_avx512_pmovsxbq:
39; X32:       ## %bb.0:
40; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
41; X32-NEXT:    vpmovsxbq (%eax), %zmm0
42; X32-NEXT:    retl
43;
44; X64-LABEL: test_llvm_x86_avx512_pmovsxbq:
45; X64:       ## %bb.0:
46; X64-NEXT:    vpmovsxbq (%rdi), %zmm0
47; X64-NEXT:    retq
48  %1 = load <16 x i8>, <16 x i8>* %a, align 1
49  %2 = shufflevector <16 x i8> %1, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
50  %3 = sext <8 x i8> %2 to <8 x i64>
51  ret <8 x i64> %3
52}
53
54define <16 x i32> @test_llvm_x86_avx512_pmovsxwd(<16 x i16>* %a) {
55; X32-LABEL: test_llvm_x86_avx512_pmovsxwd:
56; X32:       ## %bb.0:
57; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
58; X32-NEXT:    vpmovsxwd (%eax), %zmm0
59; X32-NEXT:    retl
60;
61; X64-LABEL: test_llvm_x86_avx512_pmovsxwd:
62; X64:       ## %bb.0:
63; X64-NEXT:    vpmovsxwd (%rdi), %zmm0
64; X64-NEXT:    retq
65  %1 = load <16 x i16>, <16 x i16>* %a, align 1
66  %2 = sext <16 x i16> %1 to <16 x i32>
67  ret <16 x i32> %2
68}
69
70define <8 x i64> @test_llvm_x86_avx512_pmovsxwq(<8 x i16>* %a) {
71; X32-LABEL: test_llvm_x86_avx512_pmovsxwq:
72; X32:       ## %bb.0:
73; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
74; X32-NEXT:    vpmovsxwq (%eax), %zmm0
75; X32-NEXT:    retl
76;
77; X64-LABEL: test_llvm_x86_avx512_pmovsxwq:
78; X64:       ## %bb.0:
79; X64-NEXT:    vpmovsxwq (%rdi), %zmm0
80; X64-NEXT:    retq
81  %1 = load <8 x i16>, <8 x i16>* %a, align 1
82  %2 = sext <8 x i16> %1 to <8 x i64>
83  ret <8 x i64> %2
84}
85
86define <8 x i64> @test_llvm_x86_avx512_pmovsxdq(<8 x i32>* %a) {
87; X32-LABEL: test_llvm_x86_avx512_pmovsxdq:
88; X32:       ## %bb.0:
89; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
90; X32-NEXT:    vpmovsxdq (%eax), %zmm0
91; X32-NEXT:    retl
92;
93; X64-LABEL: test_llvm_x86_avx512_pmovsxdq:
94; X64:       ## %bb.0:
95; X64-NEXT:    vpmovsxdq (%rdi), %zmm0
96; X64-NEXT:    retq
97  %1 = load <8 x i32>, <8 x i32>* %a, align 1
98  %2 = sext <8 x i32> %1 to <8 x i64>
99  ret <8 x i64> %2
100}
101
102define <32 x i16> @test_llvm_x86_avx512_pmovzxbw(<32 x i8>* %a) {
103; X32-LABEL: test_llvm_x86_avx512_pmovzxbw:
104; X32:       ## %bb.0:
105; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
106; X32-NEXT:    vpmovzxbw {{.*#+}} zmm0 = 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,mem[16],zero,mem[17],zero,mem[18],zero,mem[19],zero,mem[20],zero,mem[21],zero,mem[22],zero,mem[23],zero,mem[24],zero,mem[25],zero,mem[26],zero,mem[27],zero,mem[28],zero,mem[29],zero,mem[30],zero,mem[31],zero
107; X32-NEXT:    retl
108;
109; X64-LABEL: test_llvm_x86_avx512_pmovzxbw:
110; X64:       ## %bb.0:
111; X64-NEXT:    vpmovzxbw {{.*#+}} zmm0 = 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,mem[16],zero,mem[17],zero,mem[18],zero,mem[19],zero,mem[20],zero,mem[21],zero,mem[22],zero,mem[23],zero,mem[24],zero,mem[25],zero,mem[26],zero,mem[27],zero,mem[28],zero,mem[29],zero,mem[30],zero,mem[31],zero
112; X64-NEXT:    retq
113  %1 = load <32 x i8>, <32 x i8>* %a, align 1
114  %2 = zext <32 x i8> %1 to <32 x i16>
115  ret <32 x i16> %2
116}
117
118define <16 x i32> @test_llvm_x86_avx512_pmovzxbd(<16 x i8>* %a) {
119; X32-LABEL: test_llvm_x86_avx512_pmovzxbd:
120; X32:       ## %bb.0:
121; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
122; X32-NEXT:    vpmovzxbd {{.*#+}} zmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero,mem[4],zero,zero,zero,mem[5],zero,zero,zero,mem[6],zero,zero,zero,mem[7],zero,zero,zero,mem[8],zero,zero,zero,mem[9],zero,zero,zero,mem[10],zero,zero,zero,mem[11],zero,zero,zero,mem[12],zero,zero,zero,mem[13],zero,zero,zero,mem[14],zero,zero,zero,mem[15],zero,zero,zero
123; X32-NEXT:    retl
124;
125; X64-LABEL: test_llvm_x86_avx512_pmovzxbd:
126; X64:       ## %bb.0:
127; X64-NEXT:    vpmovzxbd {{.*#+}} zmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero,mem[4],zero,zero,zero,mem[5],zero,zero,zero,mem[6],zero,zero,zero,mem[7],zero,zero,zero,mem[8],zero,zero,zero,mem[9],zero,zero,zero,mem[10],zero,zero,zero,mem[11],zero,zero,zero,mem[12],zero,zero,zero,mem[13],zero,zero,zero,mem[14],zero,zero,zero,mem[15],zero,zero,zero
128; X64-NEXT:    retq
129  %1 = load <16 x i8>, <16 x i8>* %a, align 1
130  %2 = zext <16 x i8> %1 to <16 x i32>
131  ret <16 x i32> %2
132}
133
134define <8 x i64> @test_llvm_x86_avx512_pmovzxbq(<16 x i8>* %a) {
135; X32-LABEL: test_llvm_x86_avx512_pmovzxbq:
136; X32:       ## %bb.0:
137; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
138; X32-NEXT:    vpmovzxbq {{.*#+}} zmm0 = mem[0],zero,zero,zero,zero,zero,zero,zero,mem[1],zero,zero,zero,zero,zero,zero,zero,mem[2],zero,zero,zero,zero,zero,zero,zero,mem[3],zero,zero,zero,zero,zero,zero,zero,mem[4],zero,zero,zero,zero,zero,zero,zero,mem[5],zero,zero,zero,zero,zero,zero,zero,mem[6],zero,zero,zero,zero,zero,zero,zero,mem[7],zero,zero,zero,zero,zero,zero,zero
139; X32-NEXT:    retl
140;
141; X64-LABEL: test_llvm_x86_avx512_pmovzxbq:
142; X64:       ## %bb.0:
143; X64-NEXT:    vpmovzxbq {{.*#+}} zmm0 = mem[0],zero,zero,zero,zero,zero,zero,zero,mem[1],zero,zero,zero,zero,zero,zero,zero,mem[2],zero,zero,zero,zero,zero,zero,zero,mem[3],zero,zero,zero,zero,zero,zero,zero,mem[4],zero,zero,zero,zero,zero,zero,zero,mem[5],zero,zero,zero,zero,zero,zero,zero,mem[6],zero,zero,zero,zero,zero,zero,zero,mem[7],zero,zero,zero,zero,zero,zero,zero
144; X64-NEXT:    retq
145  %1 = load <16 x i8>, <16 x i8>* %a, align 1
146  %2 = shufflevector <16 x i8> %1, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
147  %3 = zext <8 x i8> %2 to <8 x i64>
148  ret <8 x i64> %3
149}
150
151define <16 x i32> @test_llvm_x86_avx512_pmovzxwd(<16 x i16>* %a) {
152; X32-LABEL: test_llvm_x86_avx512_pmovzxwd:
153; X32:       ## %bb.0:
154; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
155; X32-NEXT:    vpmovzxwd {{.*#+}} zmm0 = 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
156; X32-NEXT:    retl
157;
158; X64-LABEL: test_llvm_x86_avx512_pmovzxwd:
159; X64:       ## %bb.0:
160; X64-NEXT:    vpmovzxwd {{.*#+}} zmm0 = 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
161; X64-NEXT:    retq
162  %1 = load <16 x i16>, <16 x i16>* %a, align 1
163  %2 = zext <16 x i16> %1 to <16 x i32>
164  ret <16 x i32> %2
165}
166
167define <8 x i64> @test_llvm_x86_avx512_pmovzxwq(<8 x i16>* %a) {
168; X32-LABEL: test_llvm_x86_avx512_pmovzxwq:
169; X32:       ## %bb.0:
170; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
171; X32-NEXT:    vpmovzxwq {{.*#+}} zmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero,mem[4],zero,zero,zero,mem[5],zero,zero,zero,mem[6],zero,zero,zero,mem[7],zero,zero,zero
172; X32-NEXT:    retl
173;
174; X64-LABEL: test_llvm_x86_avx512_pmovzxwq:
175; X64:       ## %bb.0:
176; X64-NEXT:    vpmovzxwq {{.*#+}} zmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero,mem[4],zero,zero,zero,mem[5],zero,zero,zero,mem[6],zero,zero,zero,mem[7],zero,zero,zero
177; X64-NEXT:    retq
178  %1 = load <8 x i16>, <8 x i16>* %a, align 1
179  %2 = zext <8 x i16> %1 to <8 x i64>
180  ret <8 x i64> %2
181}
182
183define <8 x i64> @test_llvm_x86_avx512_pmovzxdq(<8 x i32>* %a) {
184; X32-LABEL: test_llvm_x86_avx512_pmovzxdq:
185; X32:       ## %bb.0:
186; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
187; X32-NEXT:    vpmovzxdq {{.*#+}} zmm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero,mem[4],zero,mem[5],zero,mem[6],zero,mem[7],zero
188; X32-NEXT:    retl
189;
190; X64-LABEL: test_llvm_x86_avx512_pmovzxdq:
191; X64:       ## %bb.0:
192; X64-NEXT:    vpmovzxdq {{.*#+}} zmm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero,mem[4],zero,mem[5],zero,mem[6],zero,mem[7],zero
193; X64-NEXT:    retq
194  %1 = load <8 x i32>, <8 x i32>* %a, align 1
195  %2 = zext <8 x i32> %1 to <8 x i64>
196  ret <8 x i64> %2
197}
198