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
| ; RUN: opt %loadPolly -basicaa -polly-codegen -S < %s | FileCheck %s
;
; Verify that we remove the misc intrinsics from the optimized SCoP.
;
; CHECK: for.body:
; CHECK: call {}* @llvm.invariant.start
; CHECK: for.body4:
; CHECK: call void @llvm.assume
; CHECK: call i1 @llvm.expect.i1
; CHECK: call void @llvm.donothing
; CHECK: for.end:
; CHECK: call void @llvm.invariant.end
; CHECK-NOT: call void @llvm.{{.*}}
;
; int A[1024];
; void jd() {
; for (int i = 0; i < 1024; i++) {
; int tmp[1024];
; for (int j = i; j < 1024; j++)
; tmp[i] += A[j];
; A[i] = tmp[i];
; }
; }
;
; ModuleID = 'test/Isl/CodeGen/lifetime_intrinsics.ll'
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
@A = common global [1024 x i32] zeroinitializer, align 16
; Function Attrs: nounwind uwtable
define void @jd() #0 {
entry:
%tmp = alloca [1024 x i32], align 16
%tmp3 = bitcast [1024 x i32]* @A to i8*
br label %for.cond
for.cond: ; preds = %for.inc11, %entry
%indvars.iv3 = phi i64 [ %indvars.iv.next4, %for.inc11 ], [ 0, %entry ]
%exitcond5 = icmp ne i64 %indvars.iv3, 1024
br i1 %exitcond5, label %for.body, label %for.end13
for.body: ; preds = %for.cond
%lis = call {}* @llvm.invariant.start(i64 4096, i8* %tmp3) #1
br label %for.cond2
for.cond2: ; preds = %for.inc, %for.body
%indvars.iv1 = phi i64 [ %indvars.iv.next2, %for.inc ], [ %indvars.iv3, %for.body ]
%lftr.wideiv = trunc i64 %indvars.iv1 to i32
%exitcond = icmp ne i32 %lftr.wideiv, 1024
br i1 %exitcond, label %for.body4, label %for.end
for.body4: ; preds = %for.cond2
call void @llvm.assume(i1 %exitcond)
call i1 @llvm.expect.i1(i1 %exitcond, i1 1)
%arrayidx = getelementptr inbounds [1024 x i32], [1024 x i32]* @A, i64 0, i64 %indvars.iv1
%tmp6 = load i32, i32* %arrayidx, align 4
%arrayidx6 = getelementptr inbounds [1024 x i32], [1024 x i32]* %tmp, i64 0, i64 %indvars.iv3
call void @llvm.donothing()
%tmp7 = load i32, i32* %arrayidx6, align 4
%add = add nsw i32 %tmp7, %tmp6
store i32 %add, i32* %arrayidx6, align 4
br label %for.inc
for.inc: ; preds = %for.body4
%indvars.iv.next2 = add nuw nsw i64 %indvars.iv1, 1
br label %for.cond2
for.end: ; preds = %for.cond2
%arrayidx8 = getelementptr inbounds [1024 x i32], [1024 x i32]* %tmp, i64 0, i64 %indvars.iv3
%tmp8 = load i32, i32* %arrayidx8, align 4
%arrayidx10 = getelementptr inbounds [1024 x i32], [1024 x i32]* @A, i64 0, i64 %indvars.iv3
call void @llvm.invariant.end({}* %lis, i64 4096, i8* %tmp3) #1
store i32 %tmp8, i32* %arrayidx10, align 4
br label %for.inc11
for.inc11: ; preds = %for.end
%indvars.iv.next4 = add nuw nsw i64 %indvars.iv3, 1
br label %for.cond
for.end13: ; preds = %for.cond
ret void
}
; Function Attrs: nounwind
declare void @llvm.donothing() #1
; Function Attrs: nounwind
declare void @llvm.assume(i1) #1
; Function Attrs: nounwind
declare i1 @llvm.expect.i1(i1, i1) #1
; Function Attrs: nounwind
declare {}* @llvm.invariant.start(i64, i8* nocapture) #1
; Function Attrs: nounwind
declare void @llvm.invariant.end({}*, i64, i8* nocapture) #1
attributes #0 = { nounwind uwtable "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "stack-protector-buffer-size"="8" "unsafe-fp-math"="true" "use-soft-float"="false" }
attributes #1 = { nounwind }
|