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
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
|
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -loop-idiom -S %s | FileCheck %s
define void @reuse_cast_1(float** %ptr, i1 %c) {
; CHECK-LABEL: @reuse_cast_1(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[LOOP_0:%.*]]
; CHECK: loop.0:
; CHECK-NEXT: [[TMP:%.*]] = load float*, float** [[PTR:%.*]], align 8
; CHECK-NEXT: [[TMP2:%.*]] = bitcast float* [[TMP]] to i8*
; CHECK-NEXT: br i1 [[C:%.*]], label [[LOOP_2_PREHEADER:%.*]], label [[LOOP_1_PREHEADER:%.*]]
; CHECK: loop.1.preheader:
; CHECK-NEXT: call void @llvm.memset.p0i8.i64(i8* align 4 [[TMP2]], i8 0, i64 400, i1 false)
; CHECK-NEXT: br label [[LOOP_1:%.*]]
; CHECK: loop.2.preheader:
; CHECK-NEXT: br label [[LOOP_2:%.*]]
; CHECK: loop.1:
; CHECK-NEXT: [[IV_1:%.*]] = phi i64 [ [[IV_1_NEXT:%.*]], [[LOOP_1]] ], [ 0, [[LOOP_1_PREHEADER]] ]
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr inbounds float, float* [[TMP]], i64 [[IV_1]]
; CHECK-NEXT: [[IV_1_NEXT]] = add nuw nsw i64 [[IV_1]], 1
; CHECK-NEXT: [[C_1:%.*]] = icmp ne i64 [[IV_1_NEXT]], 100
; CHECK-NEXT: br i1 [[C_1]], label [[LOOP_1]], label [[LOOP_0_LATCH_LOOPEXIT1:%.*]]
; CHECK: loop.2:
; CHECK-NEXT: [[IV_2:%.*]] = phi i64 [ [[IV_2_NEXT:%.*]], [[LOOP_2]] ], [ 0, [[LOOP_2_PREHEADER]] ]
; CHECK-NEXT: [[TMP10:%.*]] = getelementptr inbounds float, float* [[TMP]], i64 [[IV_2]]
; CHECK-NEXT: store float 0.000000e+00, float* [[TMP10]], align 4
; CHECK-NEXT: [[TMP11:%.*]] = load float*, float** [[PTR]], align 8
; CHECK-NEXT: [[TMP12:%.*]] = getelementptr inbounds float, float* [[TMP11]], i64 [[IV_2]]
; CHECK-NEXT: [[TMP13:%.*]] = load float, float* [[TMP12]], align 4
; CHECK-NEXT: [[IV_2_NEXT]] = add nuw nsw i64 [[IV_2]], 1
; CHECK-NEXT: [[C_2:%.*]] = icmp ne i64 [[IV_2_NEXT]], 100
; CHECK-NEXT: br i1 [[C_2]], label [[LOOP_2]], label [[LOOP_0_LATCH_LOOPEXIT:%.*]]
; CHECK: loop.0.latch.loopexit:
; CHECK-NEXT: br label [[LOOP_0_LATCH:%.*]]
; CHECK: loop.0.latch.loopexit1:
; CHECK-NEXT: br label [[LOOP_0_LATCH]]
; CHECK: loop.0.latch:
; CHECK-NEXT: [[C_0:%.*]] = call i1 @cond()
; CHECK-NEXT: br i1 [[C_0]], label [[LOOP_0]], label [[END:%.*]]
; CHECK: end:
; CHECK-NEXT: ret void
;
entry:
br label %loop.0
loop.0: ; preds = %loop.0.latch, %entry
%tmp = load float*, float** %ptr, align 8
br i1 %c, label %loop.2, label %loop.1
loop.1: ; preds = %loop.1, %loop.0
%iv.1 = phi i64 [ %iv.1.next, %loop.1 ], [ 0, %loop.0 ]
%tmp4 = getelementptr inbounds float, float* %tmp, i64 %iv.1
store float 0.000000e+00, float* %tmp4, align 4
%iv.1.next = add nuw nsw i64 %iv.1, 1
%c.1 = icmp ne i64 %iv.1.next, 100
br i1 %c.1, label %loop.1, label %loop.0.latch
loop.2: ; preds = %loop.2, %loop.0
%iv.2 = phi i64 [ %iv.2.next, %loop.2 ], [ 0, %loop.0 ]
%tmp10 = getelementptr inbounds float, float* %tmp, i64 %iv.2
store float 0.000000e+00, float* %tmp10, align 4
%tmp11 = load float*, float** %ptr, align 8
%tmp12 = getelementptr inbounds float, float* %tmp11, i64 %iv.2
%tmp13 = load float, float* %tmp12, align 4
%iv.2.next = add nuw nsw i64 %iv.2, 1
%c.2 = icmp ne i64 %iv.2.next, 100
br i1 %c.2, label %loop.2, label %loop.0.latch
loop.0.latch: ; preds = %loop.2, %loop.1
%c.0 = call i1 @cond()
br i1 %c.0, label %loop.0, label %end
end: ; preds = %loop.0.latch
ret void
}
declare i1 @cond()
declare void @use.i8(i8*)
declare void @use.i1(i1)
define void @reuse_cast_2(i32 %x, i32* %ptr.1.start) {
; CHECK-LABEL: @reuse_cast_2(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[PTR_1_START2:%.*]] = bitcast i32* [[PTR_1_START:%.*]] to i8*
; CHECK-NEXT: [[STACK:%.*]] = alloca [2 x i32], align 4
; CHECK-NEXT: [[STACK1:%.*]] = bitcast [2 x i32]* [[STACK]] to i8*
; CHECK-NEXT: [[C_0:%.*]] = icmp sgt i32 [[X:%.*]], 0
; CHECK-NEXT: [[CAST_TO_REUSE:%.*]] = bitcast [2 x i32]* [[STACK]] to i8*
; CHECK-NEXT: [[PTR_2_START:%.*]] = getelementptr inbounds [2 x i32], [2 x i32]* [[STACK]], i64 0, i64 0
; CHECK-NEXT: call void @use.i8(i8* [[CAST_TO_REUSE]])
; CHECK-NEXT: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[STACK1]], i8* align 4 [[PTR_1_START2]], i64 8, i1 false)
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[IV:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[PTR_1:%.*]] = phi i32* [ [[PTR_1_START]], [[ENTRY]] ], [ [[PTR_1_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[PTR_2:%.*]] = phi i32* [ [[PTR_2_START]], [[ENTRY]] ], [ [[PTR_2_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[PTR_1_NEXT]] = getelementptr inbounds i32, i32* [[PTR_1]], i64 1
; CHECK-NEXT: [[LV:%.*]] = load i32, i32* [[PTR_1]], align 4
; CHECK-NEXT: [[PTR_2_NEXT]] = getelementptr inbounds i32, i32* [[PTR_2]], i64 1
; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i32 [[IV]], 1
; CHECK-NEXT: [[C_1:%.*]] = icmp eq i32 [[IV]], 0
; CHECK-NEXT: br i1 [[C_1]], label [[LOOP]], label [[EXIT:%.*]]
; CHECK: exit:
; CHECK-NEXT: call void @use.i1(i1 [[C_0]])
; CHECK-NEXT: ret void
;
entry:
%stack = alloca [2 x i32], align 4
%c.0 = icmp sgt i32 %x, 0
%cast.to.reuse = bitcast [2 x i32]* %stack to i8*
%ptr.2.start = getelementptr inbounds [2 x i32], [2 x i32]* %stack, i64 0, i64 0
call void @use.i8(i8* %cast.to.reuse)
br label %loop
loop: ; preds = %loop, %entry
%iv = phi i32 [ 0, %entry ], [ %iv.next, %loop ]
%ptr.1 = phi i32* [ %ptr.1.start, %entry ], [ %ptr.1.next, %loop ]
%ptr.2 = phi i32* [ %ptr.2.start, %entry ], [ %ptr.2.next, %loop ]
%ptr.1.next = getelementptr inbounds i32, i32* %ptr.1, i64 1
%lv = load i32, i32* %ptr.1, align 4
%ptr.2.next = getelementptr inbounds i32, i32* %ptr.2, i64 1
store i32 %lv, i32* %ptr.2, align 4
%iv.next = add nuw nsw i32 %iv, 1
%c.1 = icmp eq i32 %iv, 0
br i1 %c.1, label %loop, label %exit
exit: ; preds = %loop
call void @use.i1(i1 %c.0)
ret void
}
|
