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
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
|
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -S -passes=ipsccp < %s | FileCheck %s
;
; void bar(int, float, double);
;
; void foo(int N) {
; float p = 3;
; double q = 5;
; N = 7;
;
; #pragma omp parallel for firstprivate(q)
; for (int i = 2; i < N; i++) {
; bar(i, p, q);
; }
; }
;
; Verify the constant value of q is propagated into the outlined function.
;
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
%struct.ident_t = type { i32, i32, i32, i32, ptr }
@.str = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
@0 = private unnamed_addr global %struct.ident_t { i32 0, i32 514, i32 0, i32 0, ptr @.str }, align 8
@1 = private unnamed_addr global %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @.str }, align 8
define dso_local void @foo(i32 %N) {
; CHECK-LABEL: @foo(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[P:%.*]] = alloca float, align 4
; CHECK-NEXT: store i32 [[N:%.*]], ptr [[N_ADDR]], align 4
; CHECK-NEXT: store float 3.000000e+00, ptr [[P]], align 4
; CHECK-NEXT: store i32 7, ptr [[N_ADDR]], align 4
; CHECK-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr nonnull @1, i32 3, ptr @.omp_outlined., ptr nonnull [[N_ADDR]], ptr nonnull [[P]], i64 4617315517961601024)
; CHECK-NEXT: ret void
;
entry:
%N.addr = alloca i32, align 4
%p = alloca float, align 4
store i32 %N, ptr %N.addr, align 4
store float 3.000000e+00, ptr %p, align 4
store i32 7, ptr %N.addr, align 4
call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr nonnull @1, i32 3, ptr @.omp_outlined., ptr nonnull %N.addr, ptr nonnull %p, i64 4617315517961601024)
ret void
}
define internal void @.omp_outlined.(ptr noalias %.global_tid., ptr noalias %.bound_tid., ptr dereferenceable(4) %N, ptr dereferenceable(4) %p, i64 %q) {
; CHECK-LABEL: @.omp_outlined.(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[Q_ADDR:%.*]] = alloca i64, align 8
; CHECK-NEXT: [[DOTOMP_LB:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[DOTOMP_UB:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
; CHECK-NEXT: store i64 [[Q:%.*]], ptr [[Q_ADDR]], align 8
; CHECK-NEXT: [[TMP:%.*]] = load i32, ptr [[N:%.*]], align 4
; CHECK-NEXT: [[SUB3:%.*]] = add nsw i32 [[TMP]], -3
; CHECK-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP]], 2
; CHECK-NEXT: br i1 [[CMP]], label [[OMP_PRECOND_THEN:%.*]], label [[OMP_PRECOND_END:%.*]]
; CHECK: omp.precond.then:
; CHECK-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
; CHECK-NEXT: store i32 [[SUB3]], ptr [[DOTOMP_UB]], align 4
; CHECK-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
; CHECK-NEXT: store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
; CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr [[DOTGLOBAL_TID_:%.*]], align 4
; CHECK-NEXT: call void @__kmpc_for_static_init_4(ptr nonnull @0, i32 [[TMP5]], i32 34, ptr nonnull [[DOTOMP_IS_LAST]], ptr nonnull [[DOTOMP_LB]], ptr nonnull [[DOTOMP_UB]], ptr nonnull [[DOTOMP_STRIDE]], i32 1, i32 1)
; CHECK-NEXT: [[TMP6:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
; CHECK-NEXT: [[CMP6:%.*]] = icmp sgt i32 [[TMP6]], [[SUB3]]
; CHECK-NEXT: br i1 [[CMP6]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
; CHECK: cond.true:
; CHECK-NEXT: br label [[COND_END:%.*]]
; CHECK: cond.false:
; CHECK-NEXT: [[TMP7:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
; CHECK-NEXT: br label [[COND_END]]
; CHECK: cond.end:
; CHECK-NEXT: [[COND:%.*]] = phi i32 [ [[SUB3]], [[COND_TRUE]] ], [ [[TMP7]], [[COND_FALSE]] ]
; CHECK-NEXT: store i32 [[COND]], ptr [[DOTOMP_UB]], align 4
; CHECK-NEXT: [[TMP8:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
; CHECK-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
; CHECK: omp.inner.for.cond:
; CHECK-NEXT: [[DOTOMP_IV_0:%.*]] = phi i32 [ [[TMP8]], [[COND_END]] ], [ [[ADD11:%.*]], [[OMP_INNER_FOR_INC:%.*]] ]
; CHECK-NEXT: [[TMP9:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
; CHECK-NEXT: [[CMP8:%.*]] = icmp sgt i32 [[DOTOMP_IV_0]], [[TMP9]]
; CHECK-NEXT: br i1 [[CMP8]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]], label [[OMP_INNER_FOR_BODY:%.*]]
; CHECK: omp.inner.for.cond.cleanup:
; CHECK-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
; CHECK: omp.inner.for.body:
; CHECK-NEXT: [[ADD10:%.*]] = add nsw i32 [[DOTOMP_IV_0]], 2
; CHECK-NEXT: [[TMP10:%.*]] = load float, ptr [[P:%.*]], align 4
; CHECK-NEXT: [[TMP11:%.*]] = load double, ptr [[Q_ADDR]], align 8
; CHECK-NEXT: call void @bar(i32 [[ADD10]], float [[TMP10]], double [[TMP11]])
; CHECK-NEXT: br label [[OMP_BODY_CONTINUE:%.*]]
; CHECK: omp.body.continue:
; CHECK-NEXT: br label [[OMP_INNER_FOR_INC]]
; CHECK: omp.inner.for.inc:
; CHECK-NEXT: [[ADD11]] = add nsw i32 [[DOTOMP_IV_0]], 1
; CHECK-NEXT: br label [[OMP_INNER_FOR_COND]]
; CHECK: omp.inner.for.end:
; CHECK-NEXT: br label [[OMP_LOOP_EXIT:%.*]]
; CHECK: omp.loop.exit:
; CHECK-NEXT: [[TMP12:%.*]] = load i32, ptr [[DOTGLOBAL_TID_]], align 4
; CHECK-NEXT: call void @__kmpc_for_static_fini(ptr nonnull @0, i32 [[TMP12]])
; CHECK-NEXT: br label [[OMP_PRECOND_END]]
; CHECK: omp.precond.end:
; CHECK-NEXT: ret void
;
entry:
%q.addr = alloca i64, align 8
%.omp.lb = alloca i32, align 4
%.omp.ub = alloca i32, align 4
%.omp.stride = alloca i32, align 4
%.omp.is_last = alloca i32, align 4
store i64 %q, ptr %q.addr, align 8
%tmp = load i32, ptr %N, align 4
%sub3 = add nsw i32 %tmp, -3
%cmp = icmp sgt i32 %tmp, 2
br i1 %cmp, label %omp.precond.then, label %omp.precond.end
omp.precond.then: ; preds = %entry
store i32 0, ptr %.omp.lb, align 4
store i32 %sub3, ptr %.omp.ub, align 4
store i32 1, ptr %.omp.stride, align 4
store i32 0, ptr %.omp.is_last, align 4
%tmp5 = load i32, ptr %.global_tid., align 4
call void @__kmpc_for_static_init_4(ptr nonnull @0, i32 %tmp5, i32 34, ptr nonnull %.omp.is_last, ptr nonnull %.omp.lb, ptr nonnull %.omp.ub, ptr nonnull %.omp.stride, i32 1, i32 1)
%tmp6 = load i32, ptr %.omp.ub, align 4
%cmp6 = icmp sgt i32 %tmp6, %sub3
br i1 %cmp6, label %cond.true, label %cond.false
cond.true: ; preds = %omp.precond.then
br label %cond.end
cond.false: ; preds = %omp.precond.then
%tmp7 = load i32, ptr %.omp.ub, align 4
br label %cond.end
cond.end: ; preds = %cond.false, %cond.true
%cond = phi i32 [ %sub3, %cond.true ], [ %tmp7, %cond.false ]
store i32 %cond, ptr %.omp.ub, align 4
%tmp8 = load i32, ptr %.omp.lb, align 4
br label %omp.inner.for.cond
omp.inner.for.cond: ; preds = %omp.inner.for.inc, %cond.end
%.omp.iv.0 = phi i32 [ %tmp8, %cond.end ], [ %add11, %omp.inner.for.inc ]
%tmp9 = load i32, ptr %.omp.ub, align 4
%cmp8 = icmp sgt i32 %.omp.iv.0, %tmp9
br i1 %cmp8, label %omp.inner.for.cond.cleanup, label %omp.inner.for.body
omp.inner.for.cond.cleanup: ; preds = %omp.inner.for.cond
br label %omp.inner.for.end
omp.inner.for.body: ; preds = %omp.inner.for.cond
%add10 = add nsw i32 %.omp.iv.0, 2
%tmp10 = load float, ptr %p, align 4
%tmp11 = load double, ptr %q.addr, align 8
call void @bar(i32 %add10, float %tmp10, double %tmp11)
br label %omp.body.continue
omp.body.continue: ; preds = %omp.inner.for.body
br label %omp.inner.for.inc
omp.inner.for.inc: ; preds = %omp.body.continue
%add11 = add nsw i32 %.omp.iv.0, 1
br label %omp.inner.for.cond
omp.inner.for.end: ; preds = %omp.inner.for.cond.cleanup
br label %omp.loop.exit
omp.loop.exit: ; preds = %omp.inner.for.end
%tmp12 = load i32, ptr %.global_tid., align 4
call void @__kmpc_for_static_fini(ptr nonnull @0, i32 %tmp12)
br label %omp.precond.end
omp.precond.end: ; preds = %omp.loop.exit, %entry
ret void
}
declare dso_local void @__kmpc_for_static_init_4(ptr, i32, i32, ptr, ptr, ptr, ptr, i32, i32)
declare dso_local void @bar(i32, float, double)
declare dso_local void @__kmpc_for_static_fini(ptr, i32)
declare !callback !0 dso_local void @__kmpc_fork_call(ptr, i32, ptr, ...)
!1 = !{i64 2, i64 -1, i64 -1, i1 true}
!0 = !{!1}
|
