| blob | 997ad6be84b9d1c217d476d4a9d128b2d5770df2 |
1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc -O3 --x86-asm-syntax=intel -mtriple=x86_64 -mattr=avx < %s | FileCheck %s
4 define void @extracter0([4 x <4 x i1>] %matrix) nounwind {
5 ; CHECK-LABEL: extracter0:
6 ; CHECK: # %bb.0:
7 ; CHECK-NEXT: push rbp
8 ; CHECK-NEXT: push r14
9 ; CHECK-NEXT: push rbx
10 ; CHECK-NEXT: vpslld xmm0, xmm0, 31
11 ; CHECK-NEXT: vmovmskps edi, xmm0
12 ; CHECK-NEXT: mov ebx, edi
13 ; CHECK-NEXT: shr bl, 3
14 ; CHECK-NEXT: mov ebp, edi
15 ; CHECK-NEXT: and bpl, 4
16 ; CHECK-NEXT: shr bpl, 2
17 ; CHECK-NEXT: mov r14d, edi
18 ; CHECK-NEXT: and r14b, 2
19 ; CHECK-NEXT: shr r14b
20 ; CHECK-NEXT: call print_i1@PLT
21 ; CHECK-NEXT: movzx edi, r14b
22 ; CHECK-NEXT: call print_i1@PLT
23 ; CHECK-NEXT: movzx edi, bpl
24 ; CHECK-NEXT: call print_i1@PLT
25 ; CHECK-NEXT: movzx edi, bl
26 ; CHECK-NEXT: call print_i1@PLT
27 ; CHECK-NEXT: pop rbx
28 ; CHECK-NEXT: pop r14
29 ; CHECK-NEXT: pop rbp
30 ; CHECK-NEXT: ret
31 %1 = extractvalue [4 x <4 x i1>] %matrix, 0
32 %2 = extractelement <4 x i1> %1, i64 0
33 %3 = extractelement <4 x i1> %1, i64 1
34 %4 = extractelement <4 x i1> %1, i64 2
35 %5 = extractelement <4 x i1> %1, i64 3
36 call void @print_i1(i1 %2)
37 call void @print_i1(i1 %3)
38 call void @print_i1(i1 %4)
39 call void @print_i1(i1 %5)
40 ret void
41 }
43 define void @extracter1([4 x <4 x i1>] %matrix) nounwind {
44 ; CHECK-LABEL: extracter1:
45 ; CHECK: # %bb.0:
46 ; CHECK-NEXT: push rbp
47 ; CHECK-NEXT: push r15
48 ; CHECK-NEXT: push r14
49 ; CHECK-NEXT: push r13
50 ; CHECK-NEXT: push r12
51 ; CHECK-NEXT: push rbx
52 ; CHECK-NEXT: push rax
53 ; CHECK-NEXT: vpslld xmm1, xmm1, 31
54 ; CHECK-NEXT: vmovmskps ebx, xmm1
55 ; CHECK-NEXT: mov eax, ebx
56 ; CHECK-NEXT: shr al, 3
57 ; CHECK-NEXT: mov byte ptr [rsp + 7], al # 1-byte Spill
58 ; CHECK-NEXT: mov r14d, ebx
59 ; CHECK-NEXT: and r14b, 4
60 ; CHECK-NEXT: shr r14b, 2
61 ; CHECK-NEXT: mov r15d, ebx
62 ; CHECK-NEXT: and r15b, 2
63 ; CHECK-NEXT: shr r15b
64 ; CHECK-NEXT: vpslld xmm0, xmm0, 31
65 ; CHECK-NEXT: vmovmskps edi, xmm0
66 ; CHECK-NEXT: mov r12d, edi
67 ; CHECK-NEXT: shr r12b, 3
68 ; CHECK-NEXT: mov r13d, edi
69 ; CHECK-NEXT: and r13b, 4
70 ; CHECK-NEXT: shr r13b, 2
71 ; CHECK-NEXT: mov ebp, edi
72 ; CHECK-NEXT: and bpl, 2
73 ; CHECK-NEXT: shr bpl
74 ; CHECK-NEXT: call print_i1@PLT
75 ; CHECK-NEXT: movzx edi, bpl
76 ; CHECK-NEXT: call print_i1@PLT
77 ; CHECK-NEXT: movzx edi, r13b
78 ; CHECK-NEXT: call print_i1@PLT
79 ; CHECK-NEXT: movzx edi, r12b
80 ; CHECK-NEXT: call print_i1@PLT
81 ; CHECK-NEXT: mov edi, ebx
82 ; CHECK-NEXT: call print_i1@PLT
83 ; CHECK-NEXT: movzx edi, r15b
84 ; CHECK-NEXT: call print_i1@PLT
85 ; CHECK-NEXT: movzx edi, r14b
86 ; CHECK-NEXT: call print_i1@PLT
87 ; CHECK-NEXT: movzx edi, byte ptr [rsp + 7] # 1-byte Folded Reload
88 ; CHECK-NEXT: call print_i1@PLT
89 ; CHECK-NEXT: add rsp, 8
90 ; CHECK-NEXT: pop rbx
91 ; CHECK-NEXT: pop r12
92 ; CHECK-NEXT: pop r13
93 ; CHECK-NEXT: pop r14
94 ; CHECK-NEXT: pop r15
95 ; CHECK-NEXT: pop rbp
96 ; CHECK-NEXT: ret
97 %1 = extractvalue [4 x <4 x i1>] %matrix, 0
98 %2 = extractelement <4 x i1> %1, i64 0
99 %3 = extractelement <4 x i1> %1, i64 1
100 %4 = extractelement <4 x i1> %1, i64 2
101 %5 = extractelement <4 x i1> %1, i64 3
102 call void @print_i1(i1 %2)
103 call void @print_i1(i1 %3)
104 call void @print_i1(i1 %4)
105 call void @print_i1(i1 %5)
106 %6 = extractvalue [4 x <4 x i1>] %matrix, 1
107 %7 = extractelement <4 x i1> %6, i64 0
108 %8 = extractelement <4 x i1> %6, i64 1
109 %9 = extractelement <4 x i1> %6, i64 2
110 %10 = extractelement <4 x i1> %6, i64 3
111 call void @print_i1(i1 %7)
112 call void @print_i1(i1 %8)
113 call void @print_i1(i1 %9)
114 call void @print_i1(i1 %10)
115 ret void
116 }
118 declare void @print_i1(i1)