1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s | FileCheck %s
target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64"
target triple = "x86_64-pc-linux"
%0 = type <{ i64, i64, %1, %1, [21 x %2] }>
%1 = type <{ i64, i64, i64 }>
%2 = type <{ i32, i32, ptr addrspace(2) }>
%3 = type { ptr, ptr, ptr, ptr, i32 }
%4 = type <{ ptr, ptr, i32, i32, [4 x i64], [4 x i64], [4 x i64], [4 x i64], [4 x i64] }>
%5 = type <{ ptr, ptr, ptr }>
define void @foo(ptr nocapture %_stubArgs) nounwind {
; CHECK-LABEL: foo:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: movq 48(%rdi), %rax
; CHECK-NEXT: movl 64(%rdi), %ecx
; CHECK-NEXT: movl $200, %esi
; CHECK-NEXT: addl 68(%rdi), %esi
; CHECK-NEXT: imull $46, %ecx, %edx
; CHECK-NEXT: addq %rsi, %rdx
; CHECK-NEXT: shlq $4, %rdx
; CHECK-NEXT: imull $47, %ecx, %ecx
; CHECK-NEXT: addq %rsi, %rcx
; CHECK-NEXT: shlq $4, %rcx
; CHECK-NEXT: cmpl $0, (%rdi)
; CHECK-NEXT: jne .LBB0_1
; CHECK-NEXT: # %bb.2: # %entry
; CHECK-NEXT: xorps %xmm0, %xmm0
; CHECK-NEXT: jmp .LBB0_3
; CHECK-NEXT: .LBB0_1:
; CHECK-NEXT: movaps (%rax,%rdx), %xmm0
; CHECK-NEXT: .LBB0_3: # %entry
; CHECK-NEXT: leaq -{{[0-9]+}}(%rsp), %rsp
; CHECK-NEXT: movaps (%rax,%rcx), %xmm1
; CHECK-NEXT: movaps %xmm0, -{{[0-9]+}}(%rsp)
; CHECK-NEXT: jne .LBB0_5
; CHECK-NEXT: # %bb.4: # %entry
; CHECK-NEXT: xorps %xmm1, %xmm1
; CHECK-NEXT: .LBB0_5: # %entry
; CHECK-NEXT: movaps %xmm1, -{{[0-9]+}}(%rsp)
; CHECK-NEXT: addq $152, %rsp
; CHECK-NEXT: retq
entry:
%i0 = alloca ptr, align 8
%i2 = alloca ptr, align 8
%b.i = alloca [16 x <2 x double>], align 16
%tmp1 = load i32, ptr %_stubArgs, align 4
%ptr8 = getelementptr i8, ptr %_stubArgs, i64 16
%ptr20 = getelementptr i8, ptr %_stubArgs, i64 48
%tmp21 = load ptr addrspace(1), ptr %ptr20, align 8
%ptr28 = getelementptr i8, ptr %_stubArgs, i64 64
%tmp29 = load i32, ptr %ptr28, align 4
%ptr32 = getelementptr i8, ptr %_stubArgs, i64 68
%tmp33 = load i32, ptr %ptr32, align 4
%tmp17.i = mul i32 10, 20
%tmp19.i = add i32 %tmp17.i, %tmp33
%conv21.i = zext i32 %tmp19.i to i64
%tmp6.i = and i32 42, -32
%tmp42.i = add i32 %tmp6.i, 17
%tmp44.i = insertelement <2 x i32> undef, i32 %tmp42.i, i32 1
%tmp96676677.i = or i32 17, -4
%arrayidx4506.i = getelementptr [16 x <2 x double>], ptr %b.i, i64 0, i64 4
%tmp52.i = insertelement <2 x i32> %tmp44.i, i32 0, i32 0
%tmp78.i = extractelement <2 x i32> %tmp44.i, i32 1
%tmp97.i = add i32 %tmp78.i, %tmp96676677.i
%tmp99.i = insertelement <2 x i32> %tmp52.i, i32 %tmp97.i, i32 1
%tmp154.i = extractelement <2 x i32> %tmp99.i, i32 1
%tmp156.i = extractelement <2 x i32> %tmp52.i, i32 0
%tmp158.i = urem i32 %tmp156.i, %tmp1
%i38 = mul i32 %tmp154.i, %tmp29
%i39 = add i32 %tmp158.i, %i38
%conv160.i = zext i32 %i39 to i64
%tmp22.sum652.i = add i64 %conv160.i, %conv21.i
%arrayidx161.i = getelementptr <2 x double>, ptr addrspace(1) %tmp21, i64 %tmp22.sum652.i
%tmp162.i = load <2 x double>, ptr addrspace(1) %arrayidx161.i, align 16
%tmp222.i = add i32 %tmp154.i, 1
%i43 = mul i32 %tmp222.i, %tmp29
%i44 = add i32 %tmp158.i, %i43
%conv228.i = zext i32 %i44 to i64
%tmp22.sum656.i = add i64 %conv228.i, %conv21.i
%arrayidx229.i = getelementptr <2 x double>, ptr addrspace(1) %tmp21, i64 %tmp22.sum656.i
%tmp230.i = load <2 x double>, ptr addrspace(1) %arrayidx229.i, align 16
%cmp432.i = icmp ult i32 %tmp156.i, %tmp1
; %shl.i should not be sinked below the compare.
%cond.i = select i1 %cmp432.i, <2 x double> %tmp162.i, <2 x double> zeroinitializer
store <2 x double> %cond.i, ptr %b.i, align 16
%cond448.i = select i1 %cmp432.i, <2 x double> %tmp230.i, <2 x double> zeroinitializer
store <2 x double> %cond448.i, ptr %arrayidx4506.i, align 16
ret void
}
|
