1 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --function-signature --include-generated-funcs
2 // RUN: %clang_cc1 -fopenmp-enable-irbuilder -verify -fopenmp -fopenmp-version=45 -x c++ -triple x86_64-unknown-unknown -emit-llvm %s -o - | FileCheck %s
3 // expected-no-diagnostics
4
5 #ifndef HEADER
6 #define HEADER
7
workshareloop_unsigned(float * a,float * b,float * c,float * d)8 extern "C" void workshareloop_unsigned(float *a, float *b, float *c, float *d) {
9 #pragma omp for
10 for (unsigned i = 33; i < 32000000; i += 7) {
11 a[i] = b[i] * c[i] * d[i];
12 }
13 }
14
15 #endif // HEADER
16 // CHECK-LABEL: define {{[^@]+}}@workshareloop_unsigned
17 // CHECK-SAME: (float* [[A:%.*]], float* [[B:%.*]], float* [[C:%.*]], float* [[D:%.*]]) #[[ATTR0:[0-9]+]] {
18 // CHECK-NEXT: entry:
19 // CHECK-NEXT: [[A_ADDR:%.*]] = alloca float*, align 8
20 // CHECK-NEXT: [[B_ADDR:%.*]] = alloca float*, align 8
21 // CHECK-NEXT: [[C_ADDR:%.*]] = alloca float*, align 8
22 // CHECK-NEXT: [[D_ADDR:%.*]] = alloca float*, align 8
23 // CHECK-NEXT: [[I:%.*]] = alloca i32, align 4
24 // CHECK-NEXT: [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON:%.*]], align 8
25 // CHECK-NEXT: [[AGG_CAPTURED1:%.*]] = alloca [[STRUCT_ANON_0:%.*]], align 4
26 // CHECK-NEXT: [[DOTCOUNT_ADDR:%.*]] = alloca i32, align 4
27 // CHECK-NEXT: [[P_LASTITER:%.*]] = alloca i32, align 4
28 // CHECK-NEXT: [[P_LOWERBOUND:%.*]] = alloca i32, align 4
29 // CHECK-NEXT: [[P_UPPERBOUND:%.*]] = alloca i32, align 4
30 // CHECK-NEXT: [[P_STRIDE:%.*]] = alloca i32, align 4
31 // CHECK-NEXT: store float* [[A]], float** [[A_ADDR]], align 8
32 // CHECK-NEXT: store float* [[B]], float** [[B_ADDR]], align 8
33 // CHECK-NEXT: store float* [[C]], float** [[C_ADDR]], align 8
34 // CHECK-NEXT: store float* [[D]], float** [[D_ADDR]], align 8
35 // CHECK-NEXT: store i32 33, i32* [[I]], align 4
36 // CHECK-NEXT: [[TMP0:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[AGG_CAPTURED]], i32 0, i32 0
37 // CHECK-NEXT: store i32* [[I]], i32** [[TMP0]], align 8
38 // CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds [[STRUCT_ANON_0]], %struct.anon.0* [[AGG_CAPTURED1]], i32 0, i32 0
39 // CHECK-NEXT: [[TMP2:%.*]] = load i32, i32* [[I]], align 4
40 // CHECK-NEXT: store i32 [[TMP2]], i32* [[TMP1]], align 4
41 // CHECK-NEXT: call void @__captured_stmt(i32* [[DOTCOUNT_ADDR]], %struct.anon* [[AGG_CAPTURED]])
42 // CHECK-NEXT: [[DOTCOUNT:%.*]] = load i32, i32* [[DOTCOUNT_ADDR]], align 4
43 // CHECK-NEXT: br label [[OMP_LOOP_PREHEADER:%.*]]
44 // CHECK: omp_loop.preheader:
45 // CHECK-NEXT: store i32 0, i32* [[P_LOWERBOUND]], align 4
46 // CHECK-NEXT: [[TMP3:%.*]] = sub i32 [[DOTCOUNT]], 1
47 // CHECK-NEXT: store i32 [[TMP3]], i32* [[P_UPPERBOUND]], align 4
48 // CHECK-NEXT: store i32 1, i32* [[P_STRIDE]], align 4
49 // CHECK-NEXT: [[OMP_GLOBAL_THREAD_NUM:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1:[0-9]+]])
50 // CHECK-NEXT: call void @__kmpc_for_static_init_4u(%struct.ident_t* @[[GLOB1]], i32 [[OMP_GLOBAL_THREAD_NUM]], i32 34, i32* [[P_LASTITER]], i32* [[P_LOWERBOUND]], i32* [[P_UPPERBOUND]], i32* [[P_STRIDE]], i32 1, i32 1)
51 // CHECK-NEXT: [[TMP4:%.*]] = load i32, i32* [[P_LOWERBOUND]], align 4
52 // CHECK-NEXT: [[TMP5:%.*]] = load i32, i32* [[P_UPPERBOUND]], align 4
53 // CHECK-NEXT: [[TMP6:%.*]] = sub i32 [[TMP5]], [[TMP4]]
54 // CHECK-NEXT: [[TMP7:%.*]] = add i32 [[TMP6]], 1
55 // CHECK-NEXT: br label [[OMP_LOOP_HEADER:%.*]]
56 // CHECK: omp_loop.header:
57 // CHECK-NEXT: [[OMP_LOOP_IV:%.*]] = phi i32 [ 0, [[OMP_LOOP_PREHEADER]] ], [ [[OMP_LOOP_NEXT:%.*]], [[OMP_LOOP_INC:%.*]] ]
58 // CHECK-NEXT: br label [[OMP_LOOP_COND:%.*]]
59 // CHECK: omp_loop.cond:
60 // CHECK-NEXT: [[OMP_LOOP_CMP:%.*]] = icmp ult i32 [[OMP_LOOP_IV]], [[TMP7]]
61 // CHECK-NEXT: br i1 [[OMP_LOOP_CMP]], label [[OMP_LOOP_BODY:%.*]], label [[OMP_LOOP_EXIT:%.*]]
62 // CHECK: omp_loop.body:
63 // CHECK-NEXT: [[TMP8:%.*]] = add i32 [[OMP_LOOP_IV]], [[TMP4]]
64 // CHECK-NEXT: call void @__captured_stmt.1(i32* [[I]], i32 [[TMP8]], %struct.anon.0* [[AGG_CAPTURED1]])
65 // CHECK-NEXT: [[TMP9:%.*]] = load float*, float** [[B_ADDR]], align 8
66 // CHECK-NEXT: [[TMP10:%.*]] = load i32, i32* [[I]], align 4
67 // CHECK-NEXT: [[IDXPROM:%.*]] = zext i32 [[TMP10]] to i64
68 // CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, float* [[TMP9]], i64 [[IDXPROM]]
69 // CHECK-NEXT: [[TMP11:%.*]] = load float, float* [[ARRAYIDX]], align 4
70 // CHECK-NEXT: [[TMP12:%.*]] = load float*, float** [[C_ADDR]], align 8
71 // CHECK-NEXT: [[TMP13:%.*]] = load i32, i32* [[I]], align 4
72 // CHECK-NEXT: [[IDXPROM2:%.*]] = zext i32 [[TMP13]] to i64
73 // CHECK-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds float, float* [[TMP12]], i64 [[IDXPROM2]]
74 // CHECK-NEXT: [[TMP14:%.*]] = load float, float* [[ARRAYIDX3]], align 4
75 // CHECK-NEXT: [[MUL:%.*]] = fmul float [[TMP11]], [[TMP14]]
76 // CHECK-NEXT: [[TMP15:%.*]] = load float*, float** [[D_ADDR]], align 8
77 // CHECK-NEXT: [[TMP16:%.*]] = load i32, i32* [[I]], align 4
78 // CHECK-NEXT: [[IDXPROM4:%.*]] = zext i32 [[TMP16]] to i64
79 // CHECK-NEXT: [[ARRAYIDX5:%.*]] = getelementptr inbounds float, float* [[TMP15]], i64 [[IDXPROM4]]
80 // CHECK-NEXT: [[TMP17:%.*]] = load float, float* [[ARRAYIDX5]], align 4
81 // CHECK-NEXT: [[MUL6:%.*]] = fmul float [[MUL]], [[TMP17]]
82 // CHECK-NEXT: [[TMP18:%.*]] = load float*, float** [[A_ADDR]], align 8
83 // CHECK-NEXT: [[TMP19:%.*]] = load i32, i32* [[I]], align 4
84 // CHECK-NEXT: [[IDXPROM7:%.*]] = zext i32 [[TMP19]] to i64
85 // CHECK-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds float, float* [[TMP18]], i64 [[IDXPROM7]]
86 // CHECK-NEXT: store float [[MUL6]], float* [[ARRAYIDX8]], align 4
87 // CHECK-NEXT: br label [[OMP_LOOP_INC]]
88 // CHECK: omp_loop.inc:
89 // CHECK-NEXT: [[OMP_LOOP_NEXT]] = add nuw i32 [[OMP_LOOP_IV]], 1
90 // CHECK-NEXT: br label [[OMP_LOOP_HEADER]]
91 // CHECK: omp_loop.exit:
92 // CHECK-NEXT: call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB1]], i32 [[OMP_GLOBAL_THREAD_NUM]])
93 // CHECK-NEXT: [[OMP_GLOBAL_THREAD_NUM9:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
94 // CHECK-NEXT: call void @__kmpc_barrier(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[OMP_GLOBAL_THREAD_NUM9]])
95 // CHECK-NEXT: br label [[OMP_LOOP_AFTER:%.*]]
96 // CHECK: omp_loop.after:
97 // CHECK-NEXT: ret void
98 //
99 //
100 // CHECK-LABEL: define {{[^@]+}}@__captured_stmt
101 // CHECK-SAME: (i32* nonnull align 4 dereferenceable(4) [[DISTANCE:%.*]], %struct.anon* noalias [[__CONTEXT:%.*]]) #[[ATTR1:[0-9]+]] {
102 // CHECK-NEXT: entry:
103 // CHECK-NEXT: [[DISTANCE_ADDR:%.*]] = alloca i32*, align 8
104 // CHECK-NEXT: [[__CONTEXT_ADDR:%.*]] = alloca %struct.anon*, align 8
105 // CHECK-NEXT: [[DOTSTART:%.*]] = alloca i32, align 4
106 // CHECK-NEXT: [[DOTSTOP:%.*]] = alloca i32, align 4
107 // CHECK-NEXT: [[DOTSTEP:%.*]] = alloca i32, align 4
108 // CHECK-NEXT: store i32* [[DISTANCE]], i32** [[DISTANCE_ADDR]], align 8
109 // CHECK-NEXT: store %struct.anon* [[__CONTEXT]], %struct.anon** [[__CONTEXT_ADDR]], align 8
110 // CHECK-NEXT: [[TMP0:%.*]] = load %struct.anon*, %struct.anon** [[__CONTEXT_ADDR]], align 8
111 // CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds [[STRUCT_ANON:%.*]], %struct.anon* [[TMP0]], i32 0, i32 0
112 // CHECK-NEXT: [[TMP2:%.*]] = load i32*, i32** [[TMP1]], align 8
113 // CHECK-NEXT: [[TMP3:%.*]] = load i32, i32* [[TMP2]], align 4
114 // CHECK-NEXT: store i32 [[TMP3]], i32* [[DOTSTART]], align 4
115 // CHECK-NEXT: store i32 32000000, i32* [[DOTSTOP]], align 4
116 // CHECK-NEXT: store i32 7, i32* [[DOTSTEP]], align 4
117 // CHECK-NEXT: [[TMP4:%.*]] = load i32, i32* [[DOTSTART]], align 4
118 // CHECK-NEXT: [[TMP5:%.*]] = load i32, i32* [[DOTSTOP]], align 4
119 // CHECK-NEXT: [[CMP:%.*]] = icmp ult i32 [[TMP4]], [[TMP5]]
120 // CHECK-NEXT: br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
121 // CHECK: cond.true:
122 // CHECK-NEXT: [[TMP6:%.*]] = load i32, i32* [[DOTSTOP]], align 4
123 // CHECK-NEXT: [[TMP7:%.*]] = load i32, i32* [[DOTSTART]], align 4
124 // CHECK-NEXT: [[SUB:%.*]] = sub i32 [[TMP6]], [[TMP7]]
125 // CHECK-NEXT: [[TMP8:%.*]] = load i32, i32* [[DOTSTEP]], align 4
126 // CHECK-NEXT: [[DIV:%.*]] = udiv i32 [[SUB]], [[TMP8]]
127 // CHECK-NEXT: br label [[COND_END:%.*]]
128 // CHECK: cond.false:
129 // CHECK-NEXT: br label [[COND_END]]
130 // CHECK: cond.end:
131 // CHECK-NEXT: [[COND:%.*]] = phi i32 [ [[DIV]], [[COND_TRUE]] ], [ 0, [[COND_FALSE]] ]
132 // CHECK-NEXT: [[TMP9:%.*]] = load i32*, i32** [[DISTANCE_ADDR]], align 8
133 // CHECK-NEXT: store i32 [[COND]], i32* [[TMP9]], align 4
134 // CHECK-NEXT: ret void
135 //
136 //
137 // CHECK-LABEL: define {{[^@]+}}@__captured_stmt.1
138 // CHECK-SAME: (i32* nonnull align 4 dereferenceable(4) [[LOOPVAR:%.*]], i32 [[LOGICAL:%.*]], %struct.anon.0* noalias [[__CONTEXT:%.*]]) #[[ATTR1]] {
139 // CHECK-NEXT: entry:
140 // CHECK-NEXT: [[LOOPVAR_ADDR:%.*]] = alloca i32*, align 8
141 // CHECK-NEXT: [[LOGICAL_ADDR:%.*]] = alloca i32, align 4
142 // CHECK-NEXT: [[__CONTEXT_ADDR:%.*]] = alloca %struct.anon.0*, align 8
143 // CHECK-NEXT: store i32* [[LOOPVAR]], i32** [[LOOPVAR_ADDR]], align 8
144 // CHECK-NEXT: store i32 [[LOGICAL]], i32* [[LOGICAL_ADDR]], align 4
145 // CHECK-NEXT: store %struct.anon.0* [[__CONTEXT]], %struct.anon.0** [[__CONTEXT_ADDR]], align 8
146 // CHECK-NEXT: [[TMP0:%.*]] = load %struct.anon.0*, %struct.anon.0** [[__CONTEXT_ADDR]], align 8
147 // CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds [[STRUCT_ANON_0:%.*]], %struct.anon.0* [[TMP0]], i32 0, i32 0
148 // CHECK-NEXT: [[TMP2:%.*]] = load i32, i32* [[TMP1]], align 4
149 // CHECK-NEXT: [[TMP3:%.*]] = load i32, i32* [[LOGICAL_ADDR]], align 4
150 // CHECK-NEXT: [[MUL:%.*]] = mul i32 7, [[TMP3]]
151 // CHECK-NEXT: [[ADD:%.*]] = add i32 [[TMP2]], [[MUL]]
152 // CHECK-NEXT: [[TMP4:%.*]] = load i32*, i32** [[LOOPVAR_ADDR]], align 8
153 // CHECK-NEXT: store i32 [[ADD]], i32* [[TMP4]], align 4
154 // CHECK-NEXT: ret void
155 //
156