[clang] Implement lifetime analysis for lifetime_capture_by(X) (#115921)
[llvm-project.git] / clang / test / OpenMP / target_simd_codegen.cpp
blobe2ff3d5f53e44299baa1ec11f1e9c4f007d03b2a
1 // Test host codegen.
2 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64 --check-prefix OMP45
3 // RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
4 // RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64 --check-prefix OMP45
5 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix OMP45
6 // RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
7 // RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix OMP45
8 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp -DOMP5 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64 --check-prefix OMP51
9 // RUN: %clang_cc1 -fopenmp -DOMP5 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
10 // RUN: %clang_cc1 -fopenmp -DOMP5 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64 --check-prefix OMP51
11 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp -DOMP5 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix OMP51
12 // RUN: %clang_cc1 -fopenmp -DOMP5 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
13 // RUN: %clang_cc1 -fopenmp -DOMP5 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix OMP51
15 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp-simd -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck --check-prefix SIMD-ONLY0 %s
16 // RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
17 // RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck --check-prefix SIMD-ONLY0 %s
18 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp-simd -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck --check-prefix SIMD-ONLY0 %s
19 // RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
20 // RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck --check-prefix SIMD-ONLY0 %s
21 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp-simd -DOMP5 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck --check-prefix SIMD-ONLY0 %s
22 // RUN: %clang_cc1 -fopenmp-simd -DOMP5 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
23 // RUN: %clang_cc1 -fopenmp-simd -DOMP5 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck --check-prefix SIMD-ONLY0 %s
24 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp-simd -DOMP5 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck --check-prefix SIMD-ONLY0 %s
25 // RUN: %clang_cc1 -fopenmp-simd -DOMP5 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
26 // RUN: %clang_cc1 -fopenmp-simd -DOMP5 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck --check-prefix SIMD-ONLY0 %s
27 // SIMD-ONLY0-NOT: {{__kmpc|__tgt}}
29 // Test target codegen - host bc file has to be created first.
30 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
31 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64 --check-prefix TOMP45
32 // RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
33 // RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64 --check-prefix TOMP45
34 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
35 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32 --check-prefix TOMP45
36 // RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
37 // RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32 --check-prefix TOMP45
38 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp -DOMP5 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
39 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp -DOMP5 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64 --check-prefix TOMP51
40 // RUN: %clang_cc1 -fopenmp -DOMP5 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
41 // RUN: %clang_cc1 -fopenmp -DOMP5 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64 --check-prefix TOMP51
42 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp -DOMP5 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
43 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp -DOMP5 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32 --check-prefix TOMP51
44 // RUN: %clang_cc1 -fopenmp -DOMP5 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
45 // RUN: %clang_cc1 -fopenmp -DOMP5 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32 --check-prefix TOMP51
47 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp-simd -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
48 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp-simd -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck --check-prefix SIMD-ONLY1 %s
49 // RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
50 // RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck --check-prefix SIMD-ONLY1 %s
51 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp-simd -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
52 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp-simd -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck --check-prefix SIMD-ONLY1 %s
53 // RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
54 // RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck --check-prefix SIMD-ONLY1 %s
55 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp-simd -DOMP5 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
56 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp-simd -DOMP5 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck --check-prefix SIMD-ONLY1 %s
57 // RUN: %clang_cc1 -fopenmp-simd -DOMP5 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
58 // RUN: %clang_cc1 -fopenmp-simd -DOMP5 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck --check-prefix SIMD-ONLY1 %s
59 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp-simd -DOMP5 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
60 // RUN: %clang_cc1 -verify -Wno-vla -fopenmp-simd -DOMP5 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck --check-prefix SIMD-ONLY1 %s
61 // RUN: %clang_cc1 -fopenmp-simd -DOMP5 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
62 // RUN: %clang_cc1 -fopenmp-simd -DOMP5 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify -Wno-vla %s -emit-llvm -o - | FileCheck --check-prefix SIMD-ONLY1 %s
63 // SIMD-ONLY1-NOT: {{__kmpc|__tgt}}
65 // expected-no-diagnostics
66 #ifndef HEADER
67 #define HEADER
69 // CHECK-DAG: [[IDENT_T:%.+]] = type { i32, i32, i32, i32, ptr }
70 // CHECK-DAG: [[KMP_TASK_T_WITH_PRIVATES:%.+]] = type { [[KMP_TASK_T:%.+]] }
71 // CHECK-DAG: [[KMP_TASK_T]] = type { ptr, ptr, i32, %{{[^,]+}}, %{{[^,]+}} }
72 // CHECK-DAG: [[TT:%.+]] = type { i64, i8 }
73 // CHECK-DAG: [[S1:%.+]] = type { double }
74 // CHECK-DAG: [[ENTTY:%.+]] = type { ptr, ptr, i[[SZ:32|64]], i32, i32 }
76 // TCHECK: [[ENTTY:%.+]] = type { ptr, ptr, i{{32|64}}, i32, i32 }
78 // We have 8 target regions, but only 7 that actually will generate offloading
79 // code, only 6 will have mapped arguments, and only 4 have all-constant map
80 // sizes.
82 // CHECK-DAG: [[SIZET2:@.+]] = private unnamed_addr constant [3 x i64] [i64 2, i64 4, i64 4]
83 // CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [3 x i64] [i64 800, i64 800, i64 800]
84 // CHECK-DAG: [[SIZET3:@.+]] = private unnamed_addr constant [2 x i64] [i64 4, i64 2]
85 // CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i64] [i64 800, i64 800]
86 // CHECK-DAG: [[SIZET4:@.+]] = private unnamed_addr constant [9 x i64] [i64 4, i64 40, i64 {{8|4}}, i64 0, i64 400, i64 {{8|4}}, i64 {{8|4}}, i64 0, i64 {{16|12}}]
87 // CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [9 x i64] [i64 800, i64 547, i64 800, i64 547, i64 547, i64 800, i64 800, i64 547, i64 547]
88 // CHECK-DAG: [[SIZET5:@.+]] = private unnamed_addr constant [3 x i64] [i64 4, i64 2, i64 40]
89 // CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [3 x i64] [i64 800, i64 800, i64 547]
90 // CHECK-DAG: [[SIZET6:@.+]] = private unnamed_addr constant [4 x i64] [i64 4, i64 2, i64 1, i64 40]
91 // CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [4 x i64] [i64 800, i64 800, i64 800, i64 547]
92 // OMP45-DAG: [[SIZET7:@.+]] = private unnamed_addr constant [5 x i64] [i64 {{8|4}}, i64 4, i64 {{8|4}}, i64 {{8|4}}, i64 0]
93 // OMP51-DAG: [[SIZET7:@.+]] = private unnamed_addr constant [6 x i64] [i64 {{8|4}}, i64 4, i64 {{8|4}}, i64 {{8|4}}, i64 0, i64 1]
94 // OMP45-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [5 x i64] [i64 547, i64 800, i64 800, i64 800, i64 547]
95 // OMP51-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [6 x i64] [i64 547, i64 800, i64 800, i64 800, i64 547, i64 800]
96 // CHECK-DAG: @{{.*}} = weak constant i8 0
97 // CHECK-DAG: @{{.*}} = weak constant i8 0
98 // CHECK-DAG: @{{.*}} = weak constant i8 0
99 // CHECK-DAG: @{{.*}} = weak constant i8 0
100 // CHECK-DAG: @{{.*}} = weak constant i8 0
101 // CHECK-DAG: @{{.*}} = weak constant i8 0
102 // CHECK-DAG: @{{.*}} = weak constant i8 0
104 // TCHECK: @{{.+}} = weak constant [[ENTTY]]
105 // TCHECK: @{{.+}} = weak constant [[ENTTY]]
106 // TCHECK: @{{.+}} = weak constant [[ENTTY]]
107 // TCHECK: @{{.+}} = weak constant [[ENTTY]]
108 // TCHECK: @{{.+}} = weak constant [[ENTTY]]
109 // TCHECK: @{{.+}} = weak constant [[ENTTY]]
110 // TCHECK: @{{.+}} = weak constant [[ENTTY]]
111 // TCHECK-NOT: @{{.+}} = weak constant [[ENTTY]]
113 template<typename tx, typename ty>
114 struct TT{
115 tx X;
116 ty Y;
119 // CHECK-LABEL: get_val
120 long long get_val() { return 0; }
122 // CHECK: define {{.*}}[[FOO:@.+]](
123 int foo(int n) {
124 int a = 0;
125 short aa = 0;
126 float b[10];
127 float bn[n];
128 double c[5][10];
129 double cn[5][n];
130 TT<long long, char> d;
132 // CHECK-32: [[TASK:%.+]] = call ptr @__kmpc_omp_target_task_alloc(ptr @{{[^,]+}}, i32 %{{[^,]+}}, i32 1, i32 20, i32 1, ptr [[OMP_TASK_ENTRY:@[^,]+]], i64 -1)
133 // CHECK-64: [[TASK:%.+]] = call ptr @__kmpc_omp_target_task_alloc(ptr @{{[^,]+}}, i32 %{{[^,]+}}, i32 1, i64 40, i64 1, ptr [[OMP_TASK_ENTRY:@[^,]+]], i64 -1)
134 // CHECK: call i32 @__kmpc_omp_task(ptr @{{[^,]+}}, i32 %{{[^,]+}}, ptr [[TASK]])
135 #pragma omp target simd nowait
136 for (int i = 3; i < 32; i += 5) {
139 // CHECK: call void [[HVT1:@.+]](i[[SZ]] {{[^,]+}}, {{[^)]+}})
140 long long k = get_val();
141 #pragma omp target simd if(target: 0) linear(k : 3)
142 for (int i = 10; i > 1; i--) {
143 a += 1;
146 // CHECK-DAG: [[RET:%.+]] = call i32 @__tgt_target_kernel(ptr @{{.+}}, i64 [[DEVICE:.+]], i32 1, i32 1, ptr @.{{.+}}.region_id, ptr [[ARGS:%.+]])
147 // CHECK-DAG: [[BPARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 2
148 // CHECK-DAG: store ptr [[BP:%.+]], ptr [[BPARG]]
149 // CHECK-DAG: [[PARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 3
150 // CHECK-DAG: store ptr [[P:%.+]], ptr [[PARG]]
151 // CHECK-DAG: [[BP]] = getelementptr inbounds [3 x ptr], ptr [[BPR:%[^,]+]], i32 0, i32 0
152 // CHECK-DAG: [[P]] = getelementptr inbounds [3 x ptr], ptr [[PR:%[^,]+]], i32 0, i32 0
153 // CHECK-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [3 x ptr], ptr [[BPR]], i32 0, i32 0
154 // CHECK-DAG: [[PADDR0:%.+]] = getelementptr inbounds [3 x ptr], ptr [[PR]], i32 0, i32 0
155 // CHECK-DAG: store i[[SZ]] [[VAL0:%.+]], ptr [[BPADDR0]],
156 // CHECK-DAG: store i[[SZ]] [[VAL0]], ptr [[PADDR0]],
157 // CHECK-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [3 x ptr], ptr [[BPR]], i32 0, i32 1
158 // CHECK-DAG: [[PADDR1:%.+]] = getelementptr inbounds [3 x ptr], ptr [[PR]], i32 0, i32 1
159 // CHECK-DAG: store i[[SZ]] [[VAL1:%.+]], ptr [[BPADDR1]],
160 // CHECK-DAG: store i[[SZ]] [[VAL1]], ptr [[PADDR1]],
161 // CHECK-DAG: [[BPADDR2:%.+]] = getelementptr inbounds [3 x ptr], ptr [[BPR]], i32 0, i32 2
162 // CHECK-DAG: [[PADDR2:%.+]] = getelementptr inbounds [3 x ptr], ptr [[PR]], i32 0, i32 2
163 // CHECK-DAG: store i[[SZ]] [[VAL2:%.+]], ptr [[BPADDR2]],
164 // CHECK-DAG: store i[[SZ]] [[VAL2]], ptr [[PADDR2]],
166 // CHECK-NEXT: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
167 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
168 // CHECK: [[FAIL]]
169 // CHECK: call void [[HVT2:@.+]](i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^)]+}})
170 // CHECK-NEXT: br label %[[END]]
171 // CHECK: [[END]]
172 int lin = 12;
173 #pragma omp target simd if(target: 1) linear(lin, a : get_val())
174 for (unsigned long long it = 2000; it >= 600; it-=400) {
175 aa += 1;
178 // CHECK: [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 10
179 // CHECK: br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
180 // CHECK: [[IFTHEN]]
181 // CHECK-DAG: [[RET:%.+]] = call i32 @__tgt_target_kernel(ptr @{{.+}}, i64 [[DEVICE:.+]], i32 1, i32 1, ptr @.{{.+}}.region_id, ptr [[ARGS:%.+]])
182 // CHECK-DAG: [[BPARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 2
183 // CHECK-DAG: store ptr [[BPR:%.+]], ptr [[BPARG]]
184 // CHECK-DAG: [[PARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 3
185 // CHECK-DAG: store ptr [[PR:%.+]], ptr [[PARG]]
186 // CHECK-DAG: [[BPR]] = getelementptr inbounds [2 x ptr], ptr [[BP:%[^,]+]], i32 0, i32 0
187 // CHECK-DAG: [[PR]] = getelementptr inbounds [2 x ptr], ptr [[P:%[^,]+]], i32 0, i32 0
189 // CHECK-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [2 x ptr], ptr [[BP]], i32 0, i32 0
190 // CHECK-DAG: [[PADDR0:%.+]] = getelementptr inbounds [2 x ptr], ptr [[P]], i32 0, i32 0
191 // CHECK-DAG: store i[[SZ]] [[VAL0:%.+]], ptr [[BPADDR0]],
192 // CHECK-DAG: store i[[SZ]] [[VAL0]], ptr [[PADDR0]],
194 // CHECK-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [2 x ptr], ptr [[BP]], i32 0, i32 1
195 // CHECK-DAG: [[PADDR1:%.+]] = getelementptr inbounds [2 x ptr], ptr [[P]], i32 0, i32 1
196 // CHECK-DAG: store i[[SZ]] [[VAL1:%.+]], ptr [[BPADDR1]],
197 // CHECK-DAG: store i[[SZ]] [[VAL1]], ptr [[PADDR1]],
198 // CHECK: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
199 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
200 // CHECK: [[FAIL]]
201 // CHECK: call void [[HVT3:@.+]]({{[^,]+}}, {{[^,]+}})
202 // CHECK-NEXT: br label %[[END]]
203 // CHECK: [[END]]
204 // CHECK-NEXT: br label %[[IFEND:.+]]
205 // CHECK: [[IFELSE]]
206 // CHECK: call void [[HVT3]]({{[^,]+}}, {{[^,]+}})
207 // CHECK-NEXT: br label %[[IFEND]]
208 // CHECK: [[IFEND]]
210 #pragma omp target simd if(target: n>10)
211 for (short it = 6; it <= 20; it-=-4) {
212 a += 1;
213 aa += 1;
216 // We capture 3 VLA sizes in this target region
217 // CHECK-64: [[A_VAL:%.+]] = load i32, ptr %{{.+}},
218 // CHECK-64: store i32 [[A_VAL]], ptr [[A_CADDR:%.+]],
219 // CHECK-64: [[A_CVAL:%.+]] = load i[[SZ]], ptr [[A_CADDR]],
221 // CHECK-32: [[A_VAL:%.+]] = load i32, ptr %{{.+}},
222 // CHECK-32: store i32 [[A_VAL]], ptr [[A_CADDR:%.+]],
223 // CHECK-32: [[A_CVAL:%.+]] = load i[[SZ]], ptr [[A_CADDR]],
225 // CHECK: [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 20
226 // CHECK: br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
227 // CHECK: [[TRY]]
228 // CHECK-64: [[BNSIZE:%.+]] = mul nuw i64 [[VLA0:%.+]], 4
229 // CHECK-32: [[BNSZSIZE:%.+]] = mul nuw i32 [[VLA0:%.+]], 4
230 // CHECK-32: [[BNSIZE:%.+]] = sext i32 [[BNSZSIZE]] to i64
231 // CHECK: [[CNELEMSIZE2:%.+]] = mul nuw i[[SZ]] 5, [[VLA1:%.+]]
232 // CHECK-64: [[CNSIZE:%.+]] = mul nuw i64 [[CNELEMSIZE2]], 8
233 // CHECK-32: [[CNSZSIZE:%.+]] = mul nuw i32 [[CNELEMSIZE2]], 8
234 // CHECK-32: [[CNSIZE:%.+]] = sext i32 [[CNSZSIZE]] to i64
236 // CHECK-DAG: [[RET:%.+]] = call i32 @__tgt_target_kernel(ptr @{{.+}}, i64 [[DEVICE:.+]], i32 1, i32 1, ptr @.{{.+}}.region_id, ptr [[ARGS:%.+]])
237 // CHECK-DAG: [[BPARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 2
238 // CHECK-DAG: store ptr [[BPR:%.+]], ptr [[BPARG]]
239 // CHECK-DAG: [[PARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 3
240 // CHECK-DAG: store ptr [[PR:%.+]], ptr [[PARG]]
241 // CHECK-DAG: [[SARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 4
242 // CHECK-DAG: store ptr [[SR:%.+]], ptr [[SARG]]
243 // CHECK-DAG: [[BPR]] = getelementptr inbounds [9 x ptr], ptr [[BP:%[^,]+]], i32 0, i32 0
244 // CHECK-DAG: [[PR]] = getelementptr inbounds [9 x ptr], ptr [[P:%[^,]+]], i32 0, i32 0
245 // CHECK-DAG: [[SR]] = getelementptr inbounds [9 x i64], ptr [[S:%[^,]+]], i32 0, i32 0
247 // CHECK-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [9 x ptr], ptr [[BP]], i32 0, i32 [[IDX0:[0-9]+]]
248 // CHECK-DAG: [[PADDR0:%.+]] = getelementptr inbounds [9 x ptr], ptr [[P]], i32 0, i32 [[IDX0]]
249 // CHECK-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [9 x ptr], ptr [[BP]], i32 0, i32 [[IDX1:[0-9]+]]
250 // CHECK-DAG: [[PADDR1:%.+]] = getelementptr inbounds [9 x ptr], ptr [[P]], i32 0, i32 [[IDX1]]
251 // CHECK-DAG: [[BPADDR2:%.+]] = getelementptr inbounds [9 x ptr], ptr [[BP]], i32 0, i32 [[IDX2:[0-9]+]]
252 // CHECK-DAG: [[PADDR2:%.+]] = getelementptr inbounds [9 x ptr], ptr [[P]], i32 0, i32 [[IDX2]]
253 // CHECK-DAG: [[SADDR3:%.+]] = getelementptr inbounds [9 x i64], ptr [[S]], i32 0, i32 [[IDX3:[0-9]+]]
254 // CHECK-DAG: [[BPADDR3:%.+]] = getelementptr inbounds [9 x ptr], ptr [[BP]], i32 0, i32 [[IDX3]]
255 // CHECK-DAG: [[PADDR3:%.+]] = getelementptr inbounds [9 x ptr], ptr [[P]], i32 0, i32 [[IDX3]]
256 // CHECK-DAG: [[BPADDR4:%.+]] = getelementptr inbounds [9 x ptr], ptr [[BP]], i32 0, i32 [[IDX4:[0-9]+]]
257 // CHECK-DAG: [[PADDR4:%.+]] = getelementptr inbounds [9 x ptr], ptr [[P]], i32 0, i32 [[IDX4]]
258 // CHECK-DAG: [[BPADDR5:%.+]] = getelementptr inbounds [9 x ptr], ptr [[BP]], i32 0, i32 [[IDX5:[0-9]+]]
259 // CHECK-DAG: [[PADDR5:%.+]] = getelementptr inbounds [9 x ptr], ptr [[P]], i32 0, i32 [[IDX5]]
260 // CHECK-DAG: [[BPADDR6:%.+]] = getelementptr inbounds [9 x ptr], ptr [[BP]], i32 0, i32 [[IDX6:[0-9]+]]
261 // CHECK-DAG: [[PADDR6:%.+]] = getelementptr inbounds [9 x ptr], ptr [[P]], i32 0, i32 [[IDX6]]
262 // CHECK-DAG: [[SADDR7:%.+]] = getelementptr inbounds [9 x i64], ptr [[S]], i32 0, i32 [[IDX7:[0-9]+]]
263 // CHECK-DAG: [[BPADDR7:%.+]] = getelementptr inbounds [9 x ptr], ptr [[BP]], i32 0, i32 [[IDX7]]
264 // CHECK-DAG: [[PADDR7:%.+]] = getelementptr inbounds [9 x ptr], ptr [[P]], i32 0, i32 [[IDX7]]
265 // CHECK-DAG: [[BPADDR8:%.+]] = getelementptr inbounds [9 x ptr], ptr [[BP]], i32 0, i32 [[IDX8:[0-9]+]]
266 // CHECK-DAG: [[PADDR8:%.+]] = getelementptr inbounds [9 x ptr], ptr [[P]], i32 0, i32 [[IDX8]]
268 // The names below are not necessarily consistent with the names used for the
269 // addresses above as some are repeated.
270 // CHECK-DAG: store i[[SZ]] [[A_CVAL]], ptr {{%[^,]+}},
271 // CHECK-DAG: store i[[SZ]] [[A_CVAL]], ptr {{%[^,]+}},
273 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
274 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
276 // CHECK-DAG: store i[[SZ]] %{{.+}}, ptr {{%[^,]+}},
277 // CHECK-DAG: store i[[SZ]] %{{.+}}, ptr {{%[^,]+}},
279 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
280 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
281 // CHECK-DAG: store i64 [[BNSIZE]], ptr {{%[^,]+}}
283 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
284 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
286 // CHECK-DAG: store i[[SZ]] 5, ptr {{%[^,]+}},
287 // CHECK-DAG: store i[[SZ]] 5, ptr {{%[^,]+}},
289 // CHECK-DAG: store i[[SZ]] %{{.+}}, ptr {{%[^,]+}},
290 // CHECK-DAG: store i[[SZ]] %{{.+}}, ptr {{%[^,]+}},
292 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
293 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
294 // CHECK-DAG: store i64 [[CNSIZE]], ptr {{%[^,]+}}
296 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
297 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
299 // CHECK-NEXT: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
300 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
302 // CHECK: [[FAIL]]
303 // CHECK: call void [[HVT4:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
304 // CHECK-NEXT: br label %[[END]]
305 // CHECK: [[END]]
306 #pragma omp target simd if (target \
307 : n > 20)
308 for (unsigned char it = 'z'; it >= 'a'; it+=-1) {
309 a += 1;
310 b[2] += 1.0;
311 bn[3] += 1.0;
312 c[1][2] += 1.0;
313 cn[1][3] += 1.0;
314 d.X += 1;
315 d.Y += 1;
318 return a;
321 // Check that the offloading functions are emitted and that the arguments are
322 // correct and loaded correctly for the target regions in foo().
324 // CHECK: define internal void [[HVT0:@.+]]()
325 // CHECK: !llvm.loop
326 // CHECK: ret void
327 // CHECK-NEXT: }
329 // CHECK: define internal {{.*}}i32 [[OMP_TASK_ENTRY]](i32 {{.*}}%0, ptr noalias noundef %1)
330 // CHECK: [[RET:%.+]] = call i32 @__tgt_target_kernel(ptr @{{.+}}, i64 [[DEVICE:.+]], i32 1, i32 1, ptr @.{{.+}}.region_id, ptr [[ARGS:%.+]])
331 // CHECK-NEXT: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
332 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
333 // CHECK: [[FAIL]]
334 // CHECK: call void [[HVT0]]()
335 // CHECK-NEXT: br label %[[END]]
336 // CHECK: [[END]]
338 // CHECK: define internal void [[HVT1]](i[[SZ]] noundef %{{.+}}, {{.+}})
339 // CHECK: [[AA_ADDR:%.+]] = alloca i[[SZ]], align
340 // CHECK: store i[[SZ]] %{{.+}}, ptr [[AA_ADDR]], align
341 // CHECK-64: [[AA:%.+]] = load i32, ptr [[AA_ADDR]], align
342 // CHECK-32: [[AA:%.+]] = load i32, ptr [[AA_ADDR]], align
343 // CHECK: !llvm.access.group
344 // CHECK: !llvm.loop
345 // CHECK: ret void
346 // CHECK-NEXT: }
348 // CHECK: define internal void [[HVT2]](i[[SZ]] noundef %{{.+}}, i[[SZ]] noundef %{{.+}}, i[[SZ]] noundef %{{.+}})
349 // CHECK: [[AA_ADDR:%.+]] = alloca i[[SZ]], align
350 // CHECK: store i[[SZ]] %{{.+}}, ptr [[AA_ADDR]], align
351 // CHECK: [[AA:%.+]] = load i16, ptr [[AA_ADDR]], align
352 // CHECK: !llvm.loop
353 // CHECK: ret void
354 // CHECK-NEXT: }
356 // CHECK: define internal void [[HVT3]]
357 // CHECK: [[A_ADDR:%.+]] = alloca i[[SZ]], align
358 // CHECK: [[AA_ADDR:%.+]] = alloca i[[SZ]], align
359 // CHECK-DAG: store i[[SZ]] %{{.+}}, ptr [[A_ADDR]], align
360 // CHECK-DAG: store i[[SZ]] %{{.+}}, ptr [[AA_ADDR]], align
361 // CHECK: !llvm.loop
362 // CHECK: ret void
363 // CHECK-NEXT: }
365 // CHECK: define internal void [[HVT4]]
366 // Create local storage for each capture.
367 // CHECK: [[LOCAL_A:%.+]] = alloca i[[SZ]]
368 // CHECK: [[LOCAL_B:%.+]] = alloca ptr
369 // CHECK: [[LOCAL_VLA1:%.+]] = alloca i[[SZ]]
370 // CHECK: [[LOCAL_BN:%.+]] = alloca ptr
371 // CHECK: [[LOCAL_C:%.+]] = alloca ptr
372 // CHECK: [[LOCAL_VLA2:%.+]] = alloca i[[SZ]]
373 // CHECK: [[LOCAL_VLA3:%.+]] = alloca i[[SZ]]
374 // CHECK: [[LOCAL_CN:%.+]] = alloca ptr
375 // CHECK: [[LOCAL_D:%.+]] = alloca ptr
376 // CHECK-DAG: store i[[SZ]] [[ARG_A:%.+]], ptr [[LOCAL_A]]
377 // CHECK-DAG: store ptr [[ARG_B:%.+]], ptr [[LOCAL_B]]
378 // CHECK-DAG: store i[[SZ]] [[ARG_VLA1:%.+]], ptr [[LOCAL_VLA1]]
379 // CHECK-DAG: store ptr [[ARG_BN:%.+]], ptr [[LOCAL_BN]]
380 // CHECK-DAG: store ptr [[ARG_C:%.+]], ptr [[LOCAL_C]]
381 // CHECK-DAG: store i[[SZ]] [[ARG_VLA2:%.+]], ptr [[LOCAL_VLA2]]
382 // CHECK-DAG: store i[[SZ]] [[ARG_VLA3:%.+]], ptr [[LOCAL_VLA3]]
383 // CHECK-DAG: store ptr [[ARG_CN:%.+]], ptr [[LOCAL_CN]]
384 // CHECK-DAG: store ptr [[ARG_D:%.+]], ptr [[LOCAL_D]]
386 // CHECK-DAG: [[REF_B:%.+]] = load ptr, ptr [[LOCAL_B]],
387 // CHECK-DAG: [[VAL_VLA1:%.+]] = load i[[SZ]], ptr [[LOCAL_VLA1]],
388 // CHECK-DAG: [[REF_BN:%.+]] = load ptr, ptr [[LOCAL_BN]],
389 // CHECK-DAG: [[REF_C:%.+]] = load ptr, ptr [[LOCAL_C]],
390 // CHECK-DAG: [[VAL_VLA2:%.+]] = load i[[SZ]], ptr [[LOCAL_VLA2]],
391 // CHECK-DAG: [[VAL_VLA3:%.+]] = load i[[SZ]], ptr [[LOCAL_VLA3]],
392 // CHECK-DAG: [[REF_CN:%.+]] = load ptr, ptr [[LOCAL_CN]],
393 // CHECK-DAG: [[REF_D:%.+]] = load ptr, ptr [[LOCAL_D]],
396 template<typename tx>
397 tx ftemplate(int n) {
398 tx a = 0;
399 short aa = 0;
400 tx b[10];
402 #pragma omp target simd if(target: n>40)
403 for (long long i = -10; i < 10; i += 3) {
404 a += 1;
405 aa += 1;
406 b[2] += 1;
409 return a;
412 static
413 int fstatic(int n) {
414 int a = 0;
415 short aa = 0;
416 char aaa = 0;
417 int b[10];
419 #pragma omp target simd if(target: n>50)
420 for (unsigned i=100; i<10; i+=10) {
421 a += 1;
422 aa += 1;
423 aaa += 1;
424 b[2] += 1;
427 return a;
430 struct S1 {
431 double a;
433 int r1(int n){
434 int b = n+1;
435 short int c[2][n];
437 #ifdef OMP5
438 #pragma omp target simd if(n>60) nontemporal(a) private(a)
439 #else
440 #pragma omp target simd if(n>60) private(a)
441 #endif // OMP5
442 for (unsigned long long it = 2000; it >= 600; it -= 400) {
443 this->a = (double)b + 1.5;
444 c[1][1] = ++a;
447 return c[1][1] + (int)b;
451 // CHECK: define {{.*}}@{{.*}}bar{{.*}}
452 int bar(int n){
453 int a = 0;
455 // CHECK: call {{.*}}i32 [[FOO]](i32 {{.*}})
456 a += foo(n);
458 S1 S;
459 // CHECK: call {{.*}}i32 [[FS1:@.+]](ptr {{.*}}, i32 {{.*}})
460 a += S.r1(n);
462 // CHECK: call {{.*}}i32 [[FSTATIC:@.+]](i32 {{.*}})
463 a += fstatic(n);
465 // CHECK: call {{.*}}i32 [[FTEMPLATE:@.+]](i32 {{.*}})
466 a += ftemplate<int>(n);
468 return a;
472 // CHECK: define {{.*}}[[FS1]]
474 // CHECK: ptr @llvm.stacksave.p0()
475 // CHECK-32: store i32 %{{.+}}, ptr %__vla_expr
476 // OMP51: [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 60
477 // CHECK-64: store i32 %{{.+}}, ptr [[B_ADDR:%.+]],
478 // CHECK-64: [[B_CVAL:%.+]] = load i[[SZ]], ptr [[B_ADDR]],
480 // CHECK-32: store i32 %{{.+}}, ptr [[B_ADDR:%.+]],
481 // CHECK-32: [[B_CVAL:%.+]] = load i[[SZ]], ptr [[B_ADDR]],
483 // OMP45: [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 60
484 // OMP51: [[TOBOOL:%.+]] = trunc i8 %{{.+}} to i1
485 // OMP51: [[FROMBOOL:%.+]] = zext i1 [[TOBOOL]] to i8
486 // OMP51: store i8 [[FROMBOOL]], ptr [[CAP:%.+]],
487 // OMP51: [[SIMD_COND:%.+]] = load i[[SZ]], ptr [[CAP]],
488 // OMP51: [[IF:%.+]] = trunc i8 %{{.+}} to i1
489 // CHECK: br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
490 // CHECK: [[TRY]]
491 // We capture 2 VLA sizes in this target region
492 // CHECK: [[CELEMSIZE2:%.+]] = mul nuw i[[SZ]] 2, [[VLA0:%.+]]
493 // CHECK-64: [[CSIZE:%.+]] = mul nuw i64 [[CELEMSIZE2]], 2
494 // CHECK-32: [[CSZSIZE:%.+]] = mul nuw i32 [[CELEMSIZE2]], 2
495 // CHECK-32: [[CSIZE:%.+]] = sext i32 [[CSZSIZE]] to i64
497 // OMP45-DAG: [[RET:%.+]] = call i32 @__tgt_target_kernel(ptr @{{.+}}, i64 [[DEVICE:.+]], i32 1, i32 1, ptr @.{{.+}}.region_id, ptr [[ARGS:%.+]])
498 // OMP45-DAG: [[BPARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 2
499 // OMP45-DAG: store ptr [[BPR:%.+]], ptr [[BPARG]]
500 // OMP45-DAG: [[PARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 3
501 // OMP45-DAG: store ptr [[PR:%.+]], ptr [[PARG]]
502 // OMP45-DAG: [[BPR]] = getelementptr inbounds [5 x ptr], ptr [[BP:%.+]], i32 0, i32 0
503 // OMP45-DAG: [[PR]] = getelementptr inbounds [5 x ptr], ptr [[P:%.+]], i32 0, i32 0
504 // OMP45-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [5 x ptr], ptr [[BP]], i32 [[IDX0:[0-9]+]]
505 // OMP45-DAG: [[PADDR0:%.+]] = getelementptr inbounds [5 x ptr], ptr [[P]], i32 [[IDX0]]
506 // OMP45-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [5 x ptr], ptr [[BP]], i32 [[IDX1:[0-9]+]]
507 // OMP45-DAG: [[PADDR1:%.+]] = getelementptr inbounds [5 x ptr], ptr [[P]], i32 [[IDX1]]
508 // OMP45-DAG: [[BPADDR2:%.+]] = getelementptr inbounds [5 x ptr], ptr [[BP]], i32 [[IDX2:[0-9]+]]
509 // OMP45-DAG: [[PADDR2:%.+]] = getelementptr inbounds [5 x ptr], ptr [[P]], i32 [[IDX2]]
510 // OMP45-DAG: [[BPADDR3:%.+]] = getelementptr inbounds [5 x ptr], ptr [[BP]], i32 [[IDX3:[0-9]+]]
511 // OMP45-DAG: [[PADDR3:%.+]] = getelementptr inbounds [5 x ptr], ptr [[P]], i32 [[IDX3]]
512 // OMP45-DAG: [[SADDR4:%.+]] = getelementptr inbounds [5 x i64], ptr [[S]], i32 [[IDX4:[0-9]+]]
513 // OMP45-DAG: [[BPADDR4:%.+]] = getelementptr inbounds [5 x ptr], ptr [[BP]], i32 [[IDX4]]
514 // OMP45-DAG: [[PADDR4:%.+]] = getelementptr inbounds [5 x ptr], ptr [[P]], i32 [[IDX4]]
515 // OMP51-DAG: [[RET:%.+]] = call i32 @__tgt_target_kernel(ptr @{{.+}}, i64 [[DEVICE:.+]], i32 1, i32 1, ptr @.{{.+}}.region_id, ptr [[ARGS:%.+]])
516 // OMP51-DAG: [[BPARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 2
517 // OMP51-DAG: store ptr [[BPR:%.+]], ptr [[BPARG]]
518 // OMP51-DAG: [[PARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 3
519 // OMP51-DAG: store ptr [[PR:%.+]], ptr [[PARG]]
520 // OMP51-DAG: [[BPR]] = getelementptr inbounds [6 x ptr], ptr [[BP:%.+]], i32 0, i32 0
521 // OMP51-DAG: [[PR]] = getelementptr inbounds [6 x ptr], ptr [[P:%.+]], i32 0, i32 0
522 // OMP51-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [6 x ptr], ptr [[BP]], i32 [[IDX0:[0-9]+]]
523 // OMP51-DAG: [[PADDR0:%.+]] = getelementptr inbounds [6 x ptr], ptr [[P]], i32 [[IDX0]]
524 // OMP51-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [6 x ptr], ptr [[BP]], i32 [[IDX1:[0-9]+]]
525 // OMP51-DAG: [[PADDR1:%.+]] = getelementptr inbounds [6 x ptr], ptr [[P]], i32 [[IDX1]]
526 // OMP51-DAG: [[BPADDR2:%.+]] = getelementptr inbounds [6 x ptr], ptr [[BP]], i32 [[IDX2:[0-9]+]]
527 // OMP51-DAG: [[PADDR2:%.+]] = getelementptr inbounds [6 x ptr], ptr [[P]], i32 [[IDX2]]
528 // OMP51-DAG: [[BPADDR3:%.+]] = getelementptr inbounds [6 x ptr], ptr [[BP]], i32 [[IDX3:[0-9]+]]
529 // OMP51-DAG: [[PADDR3:%.+]] = getelementptr inbounds [6 x ptr], ptr [[P]], i32 [[IDX3]]
530 // OMP51-DAG: [[SADDR4:%.+]] = getelementptr inbounds [6 x i64], ptr [[S]], i32 [[IDX4:[0-9]+]]
531 // OMP51-DAG: [[BPADDR4:%.+]] = getelementptr inbounds [6 x ptr], ptr [[BP]], i32 [[IDX4]]
532 // OMP51-DAG: [[PADDR4:%.+]] = getelementptr inbounds [6 x ptr], ptr [[P]], i32 [[IDX4]]
533 // OMP51-DAG: [[BPADDR5:%.+]] = getelementptr inbounds [6 x ptr], ptr [[BP]], i32 [[IDX5:[0-9]+]]
534 // OMP51-DAG: [[PADDR5:%.+]] = getelementptr inbounds [6 x ptr], ptr [[P]], i32 [[IDX5]]
536 // The names below are not necessarily consistent with the names used for the
537 // addresses above as some are repeated.
538 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
539 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
541 // CHECK-DAG: store i[[SZ]] [[B_CVAL]], ptr {{%[^,]+}},
542 // CHECK-DAG: store i[[SZ]] [[B_CVAL]], ptr {{%[^,]+}},
544 // CHECK-DAG: store i[[SZ]] 2, ptr {{%[^,]+}},
545 // CHECK-DAG: store i[[SZ]] 2, ptr {{%[^,]+}},
547 // CHECK-DAG: store i[[SZ]] %{{.+}}, ptr {{%[^,]+}},
548 // CHECK-DAG: store i[[SZ]] %{{.+}}, ptr {{%[^,]+}},
550 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
551 // CHECK-DAG: store ptr %{{.+}}, ptr {{%[^,]+}},
552 // CHECK-DAG: store i64 [[CSIZE]], ptr {{%[^,]+}}
554 // OMP51-DAG: store i[[SZ]] [[SIMD_COND]], ptr {{%[^,]+}}
555 // OMP51-DAG: store i[[SZ]] [[SIMD_COND]], ptr {{%[^,]+}}
557 // CHECK-NEXT: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
558 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
560 // CHECK: [[FAIL]]
561 // OMP45: call void [[HVT7:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
562 // OMP51: call void [[HVT7:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
563 // CHECK-NEXT: br label %[[END]]
564 // CHECK: [[END]]
567 // CHECK: define {{.*}}[[FSTATIC]]
569 // CHECK: [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 50
570 // CHECK: br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
571 // CHECK: [[IFTHEN]]
572 // CHECK-DAG: [[RET:%.+]] = call i32 @__tgt_target_kernel(ptr @{{.+}}, i64 [[DEVICE:.+]], i32 1, i32 1, ptr @.{{.+}}.region_id, ptr [[ARGS:%.+]])
573 // CHECK-DAG: [[BPARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 2
574 // CHECK-DAG: store ptr [[BPR:%.+]], ptr [[BPARG]]
575 // CHECK-DAG: [[PARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 3
576 // CHECK-DAG: store ptr [[PR:%.+]], ptr [[PARG]]
577 // CHECK-DAG: [[BPR]] = getelementptr inbounds [4 x ptr], ptr [[BP:%.+]], i32 0, i32 0
578 // CHECK-DAG: [[PR]] = getelementptr inbounds [4 x ptr], ptr [[P:%.+]], i32 0, i32 0
580 // CHECK-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [4 x ptr], ptr [[BP]], i32 0, i32 0
581 // CHECK-DAG: [[PADDR0:%.+]] = getelementptr inbounds [4 x ptr], ptr [[P]], i32 0, i32 0
582 // CHECK-DAG: store i[[SZ]] [[VAL0:%.+]], ptr [[BPADDR0]],
583 // CHECK-DAG: store i[[SZ]] [[VAL0]], ptr [[PADDR0]],
585 // CHECK-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [4 x ptr], ptr [[BP]], i32 0, i32 1
586 // CHECK-DAG: [[PADDR1:%.+]] = getelementptr inbounds [4 x ptr], ptr [[P]], i32 0, i32 1
587 // CHECK-DAG: store i[[SZ]] [[VAL1:%.+]], ptr [[BPADDR1]],
588 // CHECK-DAG: store i[[SZ]] [[VAL1]], ptr [[PADDR1]],
590 // CHECK-DAG: [[BPADDR2:%.+]] = getelementptr inbounds [4 x ptr], ptr [[BP]], i32 0, i32 2
591 // CHECK-DAG: [[PADDR2:%.+]] = getelementptr inbounds [4 x ptr], ptr [[P]], i32 0, i32 2
592 // CHECK-DAG: store i[[SZ]] [[VAL2:%.+]], ptr [[BPADDR2]],
593 // CHECK-DAG: store i[[SZ]] [[VAL2]], ptr [[PADDR2]],
595 // CHECK-DAG: [[BPADDR3:%.+]] = getelementptr inbounds [4 x ptr], ptr [[BP]], i32 0, i32 3
596 // CHECK-DAG: [[PADDR3:%.+]] = getelementptr inbounds [4 x ptr], ptr [[P]], i32 0, i32 3
597 // CHECK-DAG: store ptr [[VAL3:%.+]], ptr [[BPADDR3]],
598 // CHECK-DAG: store ptr [[VAL3]], ptr [[PADDR3]],
600 // CHECK: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
601 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
602 // CHECK: [[FAIL]]
603 // CHECK: call void [[HVT6:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
604 // CHECK-NEXT: br label %[[END]]
605 // CHECK: [[END]]
606 // CHECK-NEXT: br label %[[IFEND:.+]]
607 // CHECK: [[IFELSE]]
608 // CHECK: call void [[HVT6]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
609 // CHECK-NEXT: br label %[[IFEND]]
610 // CHECK: [[IFEND]]
613 // CHECK: define {{.*}}[[FTEMPLATE]]
615 // CHECK: [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 40
616 // CHECK: br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
617 // CHECK: [[IFTHEN]]
618 // CHECK-DAG: [[RET:%.+]] = call i32 @__tgt_target_kernel(ptr @{{.+}}, i64 [[DEVICE:.+]], i32 1, i32 1, ptr @.{{.+}}.region_id, ptr [[ARGS:%.+]])
619 // CHECK-DAG: [[BPARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 2
620 // CHECK-DAG: store ptr [[BPR:%.+]], ptr [[BPARG]]
621 // CHECK-DAG: [[PARG:%.+]] = getelementptr inbounds {{.+}}[[ARGS]], i32 0, i32 3
622 // CHECK-DAG: store ptr [[PR:%.+]], ptr [[PARG]]
623 // CHECK-DAG: [[BPR]] = getelementptr inbounds [3 x ptr], ptr [[BP:%.+]], i32 0, i32 0
624 // CHECK-DAG: [[PR]] = getelementptr inbounds [3 x ptr], ptr [[P:%.+]], i32 0, i32 0
626 // CHECK-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [3 x ptr], ptr [[BP]], i32 0, i32 0
627 // CHECK-DAG: [[PADDR0:%.+]] = getelementptr inbounds [3 x ptr], ptr [[P]], i32 0, i32 0
628 // CHECK-DAG: store i[[SZ]] [[VAL0:%.+]], ptr [[BPADDR0]],
629 // CHECK-DAG: store i[[SZ]] [[VAL0]], ptr [[PADDR0]],
631 // CHECK-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [3 x ptr], ptr [[BP]], i32 0, i32 1
632 // CHECK-DAG: [[PADDR1:%.+]] = getelementptr inbounds [3 x ptr], ptr [[P]], i32 0, i32 1
633 // CHECK-DAG: store i[[SZ]] [[VAL1:%.+]], ptr [[BPADDR1]],
634 // CHECK-DAG: store i[[SZ]] [[VAL1]], ptr [[PADDR1]],
636 // CHECK-DAG: [[BPADDR2:%.+]] = getelementptr inbounds [3 x ptr], ptr [[BP]], i32 0, i32 2
637 // CHECK-DAG: [[PADDR2:%.+]] = getelementptr inbounds [3 x ptr], ptr [[P]], i32 0, i32 2
638 // CHECK-DAG: store ptr [[VAL2:%.+]], ptr [[BPADDR2]],
639 // CHECK-DAG: store ptr [[VAL2]], ptr [[PADDR2]],
641 // CHECK: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
642 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
643 // CHECK: [[FAIL]]
644 // CHECK: call void [[HVT5:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
645 // CHECK-NEXT: br label %[[END]]
646 // CHECK: [[END]]
647 // CHECK-NEXT: br label %[[IFEND:.+]]
648 // CHECK: [[IFELSE]]
649 // CHECK: call void [[HVT:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
650 // CHECK-NEXT: br label %[[IFEND]]
651 // CHECK: [[IFEND]]
653 // Check that the offloading functions are emitted and that the arguments are
654 // correct and loaded correctly for the target regions of the callees of bar().
656 // CHECK: define internal void [[HVT7]]
657 // Create local storage for each capture.
658 // CHECK: [[LOCAL_THIS:%.+]] = alloca ptr
659 // CHECK: [[LOCAL_B:%.+]] = alloca i[[SZ]]
660 // CHECK: [[LOCAL_VLA1:%.+]] = alloca i[[SZ]]
661 // CHECK: [[LOCAL_VLA2:%.+]] = alloca i[[SZ]]
662 // CHECK: [[LOCAL_C:%.+]] = alloca ptr
663 // OMP51: [[LOCAL_SIMD_COND_CASTED:%.+]] = alloca i[[SZ]],
664 // CHECK-DAG: store ptr [[ARG_THIS:%.+]], ptr [[LOCAL_THIS]]
665 // CHECK-DAG: store i[[SZ]] [[ARG_B:%.+]], ptr [[LOCAL_B]]
666 // CHECK-DAG: store i[[SZ]] [[ARG_VLA1:%.+]], ptr [[LOCAL_VLA1]]
667 // CHECK-DAG: store i[[SZ]] [[ARG_VLA2:%.+]], ptr [[LOCAL_VLA2]]
668 // CHECK-DAG: store ptr [[ARG_C:%.+]], ptr [[LOCAL_C]]
669 // Store captures in the context.
670 // CHECK-DAG: [[REF_THIS:%.+]] = load ptr, ptr [[LOCAL_THIS]],
671 // CHECK-DAG: [[VAL_VLA1:%.+]] = load i[[SZ]], ptr [[LOCAL_VLA1]],
672 // CHECK-DAG: [[VAL_VLA2:%.+]] = load i[[SZ]], ptr [[LOCAL_VLA2]],
673 // CHECK-DAG: [[REF_C:%.+]] = load ptr, ptr [[LOCAL_C]],
674 // OMP51-DAG: [[SIMD_COND:%.+]] = load i8, ptr [[LOCAL_SIMD_COND_CASTED]],
675 // OMP51-DAG: trunc i8 [[SIMD_COND]] to i1
676 // OMP45-NOT: !nontemporal
677 // OMP51: store double {{.*}}!nontemporal
678 // OMP51: load double, {{.*}}!nontemporal
679 // OMP51: store double {{.*}}!nontemporal
681 // CHECK: define internal void [[HVT6]]
682 // Create local storage for each capture.
683 // CHECK: [[LOCAL_A:%.+]] = alloca i[[SZ]]
684 // CHECK: [[LOCAL_AA:%.+]] = alloca i[[SZ]]
685 // CHECK: [[LOCAL_AAA:%.+]] = alloca i[[SZ]]
686 // CHECK: [[LOCAL_B:%.+]] = alloca ptr
687 // CHECK-DAG: store i[[SZ]] [[ARG_A:%.+]], ptr [[LOCAL_A]]
688 // CHECK-DAG: store i[[SZ]] [[ARG_AA:%.+]], ptr [[LOCAL_AA]]
689 // CHECK-DAG: store i[[SZ]] [[ARG_AAA:%.+]], ptr [[LOCAL_AAA]]
690 // CHECK-DAG: store ptr [[ARG_B:%.+]], ptr [[LOCAL_B]]
691 // Store captures in the context.
692 // CHECK-DAG: [[REF_B:%.+]] = load ptr, ptr [[LOCAL_B]],
694 // CHECK: define internal void [[HVT5]]
695 // Create local storage for each capture.
696 // CHECK: [[LOCAL_A:%.+]] = alloca i[[SZ]]
697 // CHECK: [[LOCAL_AA:%.+]] = alloca i[[SZ]]
698 // CHECK: [[LOCAL_B:%.+]] = alloca ptr
699 // CHECK-DAG: store i[[SZ]] [[ARG_A:%.+]], ptr [[LOCAL_A]]
700 // CHECK-DAG: store i[[SZ]] [[ARG_AA:%.+]], ptr [[LOCAL_AA]]
701 // CHECK-DAG: store ptr [[ARG_B:%.+]], ptr [[LOCAL_B]]
702 // Store captures in the context.
703 // CHECK-DAG: [[REF_B:%.+]] = load ptr, ptr [[LOCAL_B]],
705 // OMP45-NOT: !{!"llvm.loop.vectorize.enable", i1 false}
706 // TOMP45-NOT: !{!"llvm.loop.vectorize.enable", i1 false}
707 // OMP51: !{!"llvm.loop.vectorize.enable", i1 false}
708 // TOMP51: !{!"llvm.loop.vectorize.enable", i1 false}
710 #endif