1 // Test host codegen. 2 // RUN: %clang_cc1 -verify -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 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 %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64 5 // RUN: %clang_cc1 -verify -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 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 %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 8 9 // Test target codegen - host bc file has to be created first. 10 // RUN: %clang_cc1 -verify -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 11 // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64 12 // 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-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s 13 // RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64 14 // RUN: %clang_cc1 -verify -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 15 // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32 16 // 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-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s 17 // RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32 18 19 // expected-no-diagnostics 20 #ifndef HEADER 21 #define HEADER 22 23 // CHECK-DAG: %ident_t = type { i32, i32, i32, i32, i8* } 24 // CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00" 25 // CHECK-DAG: [[DEF_LOC:@.+]] = private unnamed_addr constant %ident_t { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) } 26 27 // CHECK-DAG: [[TT:%.+]] = type { i64, i8 } 28 // CHECK-DAG: [[S1:%.+]] = type { double } 29 // CHECK-DAG: [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 } 30 // CHECK-DAG: [[DEVTY:%.+]] = type { i8*, i8*, [[ENTTY]]*, [[ENTTY]]* } 31 // CHECK-DAG: [[DSCTY:%.+]] = type { i32, [[DEVTY]]*, [[ENTTY]]*, [[ENTTY]]* } 32 33 // TCHECK: [[ENTTY:%.+]] = type { i8*, i8*, i{{32|64}}, i32, i32 } 34 35 // CHECK-DAG: $[[REGFN:\.omp_offloading\..+]] = comdat 36 37 // We have 8 target regions, but only 7 that actually will generate offloading 38 // code, only 6 will have mapped arguments, and only 4 have all-constant map 39 // sizes. 40 41 // CHECK-DAG: [[SIZET:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 2, i[[SZ]] 4, i[[SZ]] 4] 42 // CHECK-DAG: [[MAPT:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 288] 43 // CHECK-DAG: [[SIZET2:@.+]] = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 2] 44 // CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [1 x i64] [i64 288] 45 // CHECK-DAG: [[SIZET3:@.+]] = private unnamed_addr constant [2 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2] 46 // CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i64] [i64 288, i64 288] 47 // CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [9 x i64] [i64 288, i64 547, i64 288, i64 547, i64 547, i64 288, i64 288, i64 547, i64 547] 48 // CHECK-DAG: [[SIZET5:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 40] 49 // CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 547] 50 // CHECK-DAG: [[SIZET6:@.+]] = private unnamed_addr constant [4 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 1, i[[SZ]] 40] 51 // CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [4 x i64] [i64 288, i64 288, i64 288, i64 547] 52 // CHECK-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [5 x i64] [i64 547, i64 288, i64 288, i64 288, i64 547] 53 // CHECK-DAG: @{{.*}} = private constant i8 0 54 // CHECK-DAG: @{{.*}} = private constant i8 0 55 // CHECK-DAG: @{{.*}} = private constant i8 0 56 // CHECK-DAG: @{{.*}} = private constant i8 0 57 // CHECK-DAG: @{{.*}} = private constant i8 0 58 // CHECK-DAG: @{{.*}} = private constant i8 0 59 // CHECK-DAG: @{{.*}} = private constant i8 0 60 61 // TCHECK: @{{.+}} = constant [[ENTTY]] 62 // TCHECK: @{{.+}} = constant [[ENTTY]] 63 // TCHECK: @{{.+}} = constant [[ENTTY]] 64 // TCHECK: @{{.+}} = constant [[ENTTY]] 65 // TCHECK: @{{.+}} = constant [[ENTTY]] 66 // TCHECK: @{{.+}} = constant [[ENTTY]] 67 // TCHECK: @{{.+}} = constant [[ENTTY]] 68 // TCHECK-NOT: @{{.+}} = constant [[ENTTY]] 69 70 // Check if offloading descriptor is created. 71 // CHECK: [[ENTBEGIN:@.+]] = external constant [[ENTTY]] 72 // CHECK: [[ENTEND:@.+]] = external constant [[ENTTY]] 73 // CHECK: [[DEVBEGIN:@.+]] = external constant i8 74 // CHECK: [[DEVEND:@.+]] = external constant i8 75 // CHECK: [[IMAGES:@.+]] = internal unnamed_addr constant [1 x [[DEVTY]]] [{{.+}} { i8* [[DEVBEGIN]], i8* [[DEVEND]], [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }], comdat($[[REGFN]]) 76 // CHECK: [[DESC:@.+]] = internal constant [[DSCTY]] { i32 1, [[DEVTY]]* getelementptr inbounds ([1 x [[DEVTY]]], [1 x [[DEVTY]]]* [[IMAGES]], i32 0, i32 0), [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }, comdat($[[REGFN]]) 77 78 // Check target registration is registered as a Ctor. 79 // CHECK: appending global [1 x { i32, void ()*, i8* }] [{ i32, void ()*, i8* } { i32 0, void ()* bitcast (void (i8*)* @[[REGFN]] to void ()*), i8* bitcast (void (i8*)* @[[REGFN]] to i8*) }] 80 81 82 template<typename tx, typename ty> 83 struct TT{ 84 tx X; 85 ty Y; 86 }; 87 88 int global; 89 90 // CHECK: define {{.*}}[[FOO:@.+]]( 91 int foo(int n) { 92 int a = 0; 93 short aa = 0; 94 float b[10]; 95 float bn[n]; 96 double c[5][10]; 97 double cn[5][n]; 98 TT<long long, char> d; 99 100 // CHECK-DAG: [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT]], i32 0, i32 0), i32 {{[^,]+}}, i32 {{[^)]+}}) 101 // CHECK-DAG: [[BP]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0 102 // CHECK-DAG: [[P]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR:%[^,]+]], i32 0, i32 0 103 // CHECK-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 [[IDX0:[0-9]+]] 104 // CHECK-DAG: [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 [[IDX0]] 105 // CHECK-DAG: [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]* 106 // CHECK-DAG: [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]* 107 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR0]] 108 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR0]] 109 // CHECK-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 [[IDX1:[0-9]+]] 110 // CHECK-DAG: [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 [[IDX1]] 111 // CHECK-DAG: [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]* 112 // CHECK-DAG: [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]* 113 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR1]] 114 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR1]] 115 // CHECK-DAG: [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 [[IDX1:[0-9]+]] 116 // CHECK-DAG: [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 [[IDX1]] 117 // CHECK-DAG: [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to i[[SZ]]* 118 // CHECK-DAG: [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to i[[SZ]]* 119 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR2]] 120 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR2]] 121 // CHECK-NEXT: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0 122 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]] 123 // CHECK: [[FAIL]] 124 // CHECK: call void [[HVT0:@.+]](i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^)]+}}) 125 // CHECK-NEXT: br label %[[END]] 126 // CHECK: [[END]] 127 #pragma omp target teams num_teams(a) thread_limit(a) firstprivate(aa) 128 { 129 } 130 131 // CHECK: call void [[HVT1:@.+]](i[[SZ]] {{[^,]+}}) 132 #pragma omp target teams if(target: 0) 133 { 134 a += 1; 135 } 136 137 // CHECK-DAG: [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET2]], i32 0, i32 0), i64* getelementptr inbounds ([1 x i64], [1 x i64]* [[MAPT2]], i32 0, i32 0), i32 0, i32 0) 138 // CHECK-DAG: [[BP]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0 139 // CHECK-DAG: [[P]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[PR:%[^,]+]], i32 0, i32 0 140 // CHECK-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BPR]], i32 0, i32 [[IDX0:[0-9]+]] 141 // CHECK-DAG: [[PADDR0:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[PR]], i32 0, i32 [[IDX0]] 142 // CHECK-DAG: [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]* 143 // CHECK-DAG: [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]* 144 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR0]] 145 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR0]] 146 147 // CHECK-NEXT: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0 148 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]] 149 // CHECK: [[FAIL]] 150 // CHECK: call void [[HVT2:@.+]](i[[SZ]] {{[^,]+}}) 151 // CHECK-NEXT: br label %[[END]] 152 // CHECK: [[END]] 153 #pragma omp target teams if(target: 1) 154 { 155 aa += 1; 156 } 157 158 // CHECK: [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 10 159 // CHECK: br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]] 160 // CHECK: [[IFTHEN]] 161 // CHECK-DAG: [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i64* getelementptr inbounds ([2 x i64], [2 x i64]* [[MAPT3]], i32 0, i32 0), i32 0, i32 0) 162 // CHECK-DAG: [[BPR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP:%[^,]+]], i32 0, i32 0 163 // CHECK-DAG: [[PR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P:%[^,]+]], i32 0, i32 0 164 165 // CHECK-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 0 166 // CHECK-DAG: [[PADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 0 167 // CHECK-DAG: [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]* 168 // CHECK-DAG: [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]* 169 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR0]] 170 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR0]] 171 172 // CHECK-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 1 173 // CHECK-DAG: [[PADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 1 174 // CHECK-DAG: [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]* 175 // CHECK-DAG: [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]* 176 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR1]] 177 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR1]] 178 // CHECK: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0 179 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]] 180 // CHECK: [[FAIL]] 181 // CHECK: call void [[HVT3:@.+]]({{[^,]+}}, {{[^,]+}}) 182 // CHECK-NEXT: br label %[[END]] 183 // CHECK: [[END]] 184 // CHECK-NEXT: br label %[[IFEND:.+]] 185 // CHECK: [[IFELSE]] 186 // CHECK: call void [[HVT3]]({{[^,]+}}, {{[^,]+}}) 187 // CHECK-NEXT: br label %[[IFEND]] 188 // CHECK: [[IFEND]] 189 #pragma omp target teams if(target: n>10) 190 { 191 a += 1; 192 aa += 1; 193 global += 1; 194 } 195 196 // We capture 3 VLA sizes in this target region 197 // CHECK-64: [[A_VAL:%.+]] = load i32, i32* %{{.+}}, 198 // CHECK-64: [[A_ADDR:%.+]] = bitcast i[[SZ]]* [[A_CADDR:%.+]] to i32* 199 // CHECK-64: store i32 [[A_VAL]], i32* [[A_ADDR]], 200 // CHECK-64: [[A_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CADDR]], 201 202 // CHECK-32: [[A_VAL:%.+]] = load i32, i32* %{{.+}}, 203 // CHECK-32: store i32 [[A_VAL]], i32* [[A_CADDR:%.+]], 204 // CHECK-32: [[A_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CADDR]], 205 206 // CHECK: [[BNSIZE:%.+]] = mul nuw i[[SZ]] [[VLA0:%.+]], 4 207 // CHECK: [[CNELEMSIZE2:%.+]] = mul nuw i[[SZ]] 5, [[VLA1:%.+]] 208 // CHECK: [[CNSIZE:%.+]] = mul nuw i[[SZ]] [[CNELEMSIZE2]], 8 209 210 // CHECK: [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 20 211 // CHECK: br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]] 212 // CHECK: [[TRY]] 213 // CHECK-DAG: [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 9, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([9 x i64], [9 x i64]* [[MAPT4]], i32 0, i32 0), i32 0, i32 0) 214 // CHECK-DAG: [[BPR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP:%[^,]+]], i32 0, i32 0 215 // CHECK-DAG: [[PR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P:%[^,]+]], i32 0, i32 0 216 // CHECK-DAG: [[SR]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S:%[^,]+]], i32 0, i32 0 217 218 // CHECK-DAG: [[SADDR0:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX0:[0-9]+]] 219 // CHECK-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX0]] 220 // CHECK-DAG: [[PADDR0:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX0]] 221 // CHECK-DAG: [[SADDR1:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX1:[0-9]+]] 222 // CHECK-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX1]] 223 // CHECK-DAG: [[PADDR1:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX1]] 224 // CHECK-DAG: [[SADDR2:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX2:[0-9]+]] 225 // CHECK-DAG: [[BPADDR2:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX2]] 226 // CHECK-DAG: [[PADDR2:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX2]] 227 // CHECK-DAG: [[SADDR3:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX3:[0-9]+]] 228 // CHECK-DAG: [[BPADDR3:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX3]] 229 // CHECK-DAG: [[PADDR3:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX3]] 230 // CHECK-DAG: [[SADDR4:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX4:[0-9]+]] 231 // CHECK-DAG: [[BPADDR4:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX4]] 232 // CHECK-DAG: [[PADDR4:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX4]] 233 // CHECK-DAG: [[SADDR5:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX5:[0-9]+]] 234 // CHECK-DAG: [[BPADDR5:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX5]] 235 // CHECK-DAG: [[PADDR5:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX5]] 236 // CHECK-DAG: [[SADDR6:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX6:[0-9]+]] 237 // CHECK-DAG: [[BPADDR6:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX6]] 238 // CHECK-DAG: [[PADDR6:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX6]] 239 // CHECK-DAG: [[SADDR7:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX7:[0-9]+]] 240 // CHECK-DAG: [[BPADDR7:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX7]] 241 // CHECK-DAG: [[PADDR7:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX7]] 242 // CHECK-DAG: [[SADDR8:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX8:[0-9]+]] 243 // CHECK-DAG: [[BPADDR8:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX8]] 244 // CHECK-DAG: [[PADDR8:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX8]] 245 246 // The names below are not necessarily consistent with the names used for the 247 // addresses above as some are repeated. 248 // CHECK-DAG: store i[[SZ]] [[VLA0]], i[[SZ]]* [[CBPADDR0:%.+]], 249 // CHECK-DAG: store i[[SZ]] [[VLA0]], i[[SZ]]* [[CPADDR0:%.+]], 250 // CHECK-DAG: [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 251 // CHECK-DAG: [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 252 // CHECK-DAG: store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}} 253 254 // CHECK-DAG: store i[[SZ]] [[VLA1]], i[[SZ]]* [[CBPADDR1:%.+]], 255 // CHECK-DAG: store i[[SZ]] [[VLA1]], i[[SZ]]* [[CPADDR1:%.+]], 256 // CHECK-DAG: [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 257 // CHECK-DAG: [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 258 // CHECK-DAG: store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}} 259 260 // CHECK-DAG: store i[[SZ]] 5, i[[SZ]]* [[CBPADDR2:%.+]], 261 // CHECK-DAG: store i[[SZ]] 5, i[[SZ]]* [[CPADDR2:%.+]], 262 // CHECK-DAG: [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 263 // CHECK-DAG: [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 264 // CHECK-DAG: store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}} 265 266 // CHECK-DAG: store i[[SZ]] [[A_CVAL]], i[[SZ]]* [[CBPADDR3:%.+]], 267 // CHECK-DAG: store i[[SZ]] [[A_CVAL]], i[[SZ]]* [[CPADDR3:%.+]], 268 // CHECK-DAG: [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 269 // CHECK-DAG: [[CPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 270 // CHECK-DAG: store i[[SZ]] 4, i[[SZ]]* {{%[^,]+}} 271 272 // CHECK-DAG: store [10 x float]* %{{.+}}, [10 x float]** [[CBPADDR4:%.+]], 273 // CHECK-DAG: store [10 x float]* %{{.+}}, [10 x float]** [[CPADDR4:%.+]], 274 // CHECK-DAG: [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x float]** 275 // CHECK-DAG: [[CPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x float]** 276 // CHECK-DAG: store i[[SZ]] 40, i[[SZ]]* {{%[^,]+}} 277 278 // CHECK-DAG: store float* %{{.+}}, float** [[CBPADDR5:%.+]], 279 // CHECK-DAG: store float* %{{.+}}, float** [[CPADDR5:%.+]], 280 // CHECK-DAG: [[CBPADDR5]] = bitcast i8** {{%[^,]+}} to float** 281 // CHECK-DAG: [[CPADDR5]] = bitcast i8** {{%[^,]+}} to float** 282 // CHECK-DAG: store i[[SZ]] [[BNSIZE]], i[[SZ]]* {{%[^,]+}} 283 284 // CHECK-DAG: store [5 x [10 x double]]* %{{.+}}, [5 x [10 x double]]** [[CBPADDR6:%.+]], 285 // CHECK-DAG: store [5 x [10 x double]]* %{{.+}}, [5 x [10 x double]]** [[CPADDR6:%.+]], 286 // CHECK-DAG: [[CBPADDR6]] = bitcast i8** {{%[^,]+}} to [5 x [10 x double]]** 287 // CHECK-DAG: [[CPADDR6]] = bitcast i8** {{%[^,]+}} to [5 x [10 x double]]** 288 // CHECK-DAG: store i[[SZ]] 400, i[[SZ]]* {{%[^,]+}} 289 290 // CHECK-DAG: store double* %{{.+}}, double** [[CBPADDR7:%.+]], 291 // CHECK-DAG: store double* %{{.+}}, double** [[CPADDR7:%.+]], 292 // CHECK-DAG: [[CBPADDR7]] = bitcast i8** {{%[^,]+}} to double** 293 // CHECK-DAG: [[CPADDR7]] = bitcast i8** {{%[^,]+}} to double** 294 // CHECK-DAG: store i[[SZ]] [[CNSIZE]], i[[SZ]]* {{%[^,]+}} 295 296 // CHECK-DAG: store [[TT]]* %{{.+}}, [[TT]]** [[CBPADDR8:%.+]], 297 // CHECK-DAG: store [[TT]]* %{{.+}}, [[TT]]** [[CPADDR8:%.+]], 298 // CHECK-DAG: [[CBPADDR8]] = bitcast i8** {{%[^,]+}} to [[TT]]** 299 // CHECK-DAG: [[CPADDR8]] = bitcast i8** {{%[^,]+}} to [[TT]]** 300 // CHECK-DAG: store i[[SZ]] {{12|16}}, i[[SZ]]* {{%[^,]+}} 301 302 // CHECK-NEXT: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0 303 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]] 304 305 // CHECK: [[FAIL]] 306 // CHECK: call void [[HVT4:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}) 307 // CHECK-NEXT: br label %[[END]] 308 // CHECK: [[END]] 309 #pragma omp target teams if(target: n>20) 310 { 311 a += 1; 312 b[2] += 1.0; 313 bn[3] += 1.0; 314 c[1][2] += 1.0; 315 cn[1][3] += 1.0; 316 d.X += 1; 317 d.Y += 1; 318 } 319 320 return a; 321 } 322 323 // Check that the offloading functions are emitted and that the arguments are 324 // correct and loaded correctly for the target regions in foo(). 325 326 // CHECK: define internal void [[HVT0]](i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^)]+}}) 327 // CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]])* [[OMP_OUTLINED:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] {{[^)]+}}) 328 // 329 // 330 // CHECK: define internal {{.*}}void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] {{[^)]+}}) 331 // CHECK: alloca i16, 332 // CHECK: ret void 333 // CHECK-NEXT: } 334 335 336 // CHECK: define internal void [[HVT1]](i[[SZ]] %{{.+}}) 337 // Create stack storage and store argument in there. 338 // CHECK: [[AA_ADDR:%.+]] = alloca i[[SZ]], align 339 // CHECK: [[AA_CASTED:%.+]] = alloca i[[SZ]], align 340 // CHECK: store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align 341 // CHECK-64: [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i32* 342 // CHECK-64: [[AA:%.+]] = load i32, i32* [[AA_CADDR]], align 343 // CHECK-32: [[AA:%.+]] = load i32, i32* [[AA_ADDR]], align 344 // CHECK-64: [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i32* 345 // CHECK-64: store i32 [[AA]], i32* [[AA_C]], align 346 // CHECK-32: store i32 [[AA]], i32* [[AA_CASTED]], align 347 // CHECK: [[PARAM:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align 348 // CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]])* [[OMP_OUTLINED1:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM]]) 349 // 350 // 351 // CHECK: define internal {{.*}}void [[OMP_OUTLINED1]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}) 352 // CHECK: [[AA_ADDR:%.+]] = alloca i[[SZ]], align 353 // CHECK: store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align 354 // CHECK-64: [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i32* 355 // CHECK-64: [[AA:%.+]] = load i32, i32* [[AA_CADDR]], align 356 // CHECK-32: [[AA:%.+]] = load i32, i32* [[AA_ADDR]], align 357 // CHECK: ret void 358 // CHECK-NEXT: } 359 360 // CHECK: define internal void [[HVT2]](i[[SZ]] %{{.+}}) 361 // Create stack storage and store argument in there. 362 // CHECK: [[AA_ADDR:%.+]] = alloca i[[SZ]], align 363 // CHECK: [[AA_CASTED:%.+]] = alloca i[[SZ]], align 364 // CHECK: store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align 365 // CHECK: [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16* 366 // CHECK: [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align 367 // CHECK: [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i16* 368 // CHECK: store i16 [[AA]], i16* [[AA_C]], align 369 // CHECK: [[PARAM:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align 370 // CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]])* [[OMP_OUTLINED2:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM]]) 371 // 372 // 373 // CHECK: define internal {{.*}}void [[OMP_OUTLINED2]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}) 374 // CHECK: [[AA_ADDR:%.+]] = alloca i[[SZ]], align 375 // CHECK: store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align 376 // CHECK: [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16* 377 // CHECK: [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align 378 // CHECK: ret void 379 // CHECK-NEXT: } 380 381 // CHECK: define internal void [[HVT3]] 382 // Create stack storage and store argument in there. 383 // CHECK: [[A_ADDR:%.+]] = alloca i[[SZ]], align 384 // CHECK: [[AA_ADDR:%.+]] = alloca i[[SZ]], align 385 // CHECK: [[A_CASTED:%.+]] = alloca i[[SZ]], align 386 // CHECK: [[AA_CASTED:%.+]] = alloca i[[SZ]], align 387 // CHECK: [[GLOBAL_CASTED:%.+]] = alloca i[[SZ]], align 388 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[A_ADDR]], align 389 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align 390 // CHECK-64-DAG:[[A_CADDR:%.+]] = bitcast i[[SZ]]* [[A_ADDR]] to i32* 391 // CHECK-DAG: [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16* 392 // CHECK-64-DAG:[[A:%.+]] = load i32, i32* [[A_CADDR]], align 393 // CHECK-32-DAG:[[A:%.+]] = load i32, i32* [[A_ADDR]], align 394 // CHECK-64-DAG:[[A_C:%.+]] = bitcast i[[SZ]]* [[A_CASTED]] to i32* 395 // CHECK-64-DAG:store i32 [[A]], i32* [[A_C]], align 396 // CHECK-32-DAG:store i32 [[A]], i32* [[A_CASTED]], align 397 // CHECK-DAG: [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align 398 // CHECK-DAG: [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i16* 399 // CHECK-DAG: store i16 [[AA]], i16* [[AA_C]], align 400 // CHECK-DAG: [[GLOBAL:%.+]] = load i32, i32* @global, align 401 // CHECK-64-DAG:[[GLOBAL_C:%.+]] = bitcast i[[SZ]]* [[GLOBAL_CASTED]] to i32* 402 // CHECK-64-DAG:store i32 [[GLOBAL]], i32* [[GLOBAL_C]], align 403 // CHECK-32-DAG:store i32 [[GLOBAL]], i32* [[GLOBAL_CASTED]], align 404 // CHECK-DAG: [[PARAM1:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CASTED]], align 405 // CHECK-DAG: [[PARAM2:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align 406 // CHECK-DAG: [[PARAM3:%.+]] = load i[[SZ]], i[[SZ]]* [[GLOBAL_CASTED]], align 407 // CHECK-DAG: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 3, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], i[[SZ]])* [[OMP_OUTLINED3:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM1]], i[[SZ]] [[PARAM2]], i[[SZ]] [[PARAM3]]) 408 // 409 // 410 // CHECK: define internal {{.*}}void [[OMP_OUTLINED3]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}) 411 // CHECK: [[A_ADDR:%.+]] = alloca i[[SZ]], align 412 // CHECK: [[AA_ADDR:%.+]] = alloca i[[SZ]], align 413 // CHECK: [[GLOBAL_ADDR:%.+]] = alloca i[[SZ]], align 414 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[A_ADDR]], align 415 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align 416 // CHECK-DAG: store i[[SZ]] %{{.+}}, i[[SZ]]* [[GLOBAL_ADDR]], align 417 // CHECK-64-DAG:[[A_CADDR:%.+]] = bitcast i[[SZ]]* [[A_ADDR]] to i32* 418 // CHECK-DAG: [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16* 419 // CHECK-64-DAG:[[GLOBAL_CADDR:%.+]] = bitcast i[[SZ]]* [[GLOBAL_ADDR]] to i32* 420 // CHECK-64: store i32 {{.+}}, i32* [[GLOBAL_CADDR]], 421 // CHECK-32: store i32 {{.+}}, i32* [[GLOBAL_ADDR]], 422 // CHECK: ret void 423 // CHECK-NEXT: } 424 425 // CHECK: define internal void [[HVT4]] 426 // Create local storage for each capture. 427 // CHECK: [[LOCAL_A:%.+]] = alloca i[[SZ]] 428 // CHECK: [[LOCAL_B:%.+]] = alloca [10 x float]* 429 // CHECK: [[LOCAL_VLA1:%.+]] = alloca i[[SZ]] 430 // CHECK: [[LOCAL_BN:%.+]] = alloca float* 431 // CHECK: [[LOCAL_C:%.+]] = alloca [5 x [10 x double]]* 432 // CHECK: [[LOCAL_VLA2:%.+]] = alloca i[[SZ]] 433 // CHECK: [[LOCAL_VLA3:%.+]] = alloca i[[SZ]] 434 // CHECK: [[LOCAL_CN:%.+]] = alloca double* 435 // CHECK: [[LOCAL_D:%.+]] = alloca [[TT]]* 436 // CHECK: [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]] 437 // CHECK-DAG: store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]] 438 // CHECK-DAG: store [10 x float]* [[ARG_B:%.+]], [10 x float]** [[LOCAL_B]] 439 // CHECK-DAG: store i[[SZ]] [[ARG_VLA1:%.+]], i[[SZ]]* [[LOCAL_VLA1]] 440 // CHECK-DAG: store float* [[ARG_BN:%.+]], float** [[LOCAL_BN]] 441 // CHECK-DAG: store [5 x [10 x double]]* [[ARG_C:%.+]], [5 x [10 x double]]** [[LOCAL_C]] 442 // CHECK-DAG: store i[[SZ]] [[ARG_VLA2:%.+]], i[[SZ]]* [[LOCAL_VLA2]] 443 // CHECK-DAG: store i[[SZ]] [[ARG_VLA3:%.+]], i[[SZ]]* [[LOCAL_VLA3]] 444 // CHECK-DAG: store double* [[ARG_CN:%.+]], double** [[LOCAL_CN]] 445 // CHECK-DAG: store [[TT]]* [[ARG_D:%.+]], [[TT]]** [[LOCAL_D]] 446 447 // CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32* 448 // CHECK-DAG: [[REF_B:%.+]] = load [10 x float]*, [10 x float]** [[LOCAL_B]], 449 // CHECK-DAG: [[VAL_VLA1:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA1]], 450 // CHECK-DAG: [[REF_BN:%.+]] = load float*, float** [[LOCAL_BN]], 451 // CHECK-DAG: [[REF_C:%.+]] = load [5 x [10 x double]]*, [5 x [10 x double]]** [[LOCAL_C]], 452 // CHECK-DAG: [[VAL_VLA2:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA2]], 453 // CHECK-DAG: [[VAL_VLA3:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA3]], 454 // CHECK-DAG: [[REF_CN:%.+]] = load double*, double** [[LOCAL_CN]], 455 // CHECK-DAG: [[REF_D:%.+]] = load [[TT]]*, [[TT]]** [[LOCAL_D]], 456 457 // CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]] 458 // CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32* 459 // CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align 460 // CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]] 461 // CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align 462 // CHECK-DAG: [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]], 463 464 // CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 9, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], [10 x float]*, i[[SZ]], float*, [5 x [10 x double]]*, i[[SZ]], i[[SZ]], double*, [[TT]]*)* [[OMP_OUTLINED4:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], [10 x float]* [[REF_B]], i[[SZ]] [[VAL_VLA1]], float* [[REF_BN]], [5 x [10 x double]]* [[REF_C]], i[[SZ]] [[VAL_VLA2]], i[[SZ]] [[VAL_VLA3]], double* [[REF_CN]], [[TT]]* [[REF_D]]) 465 // 466 // 467 // CHECK: define internal {{.*}}void [[OMP_OUTLINED4]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, [10 x float]* {{.+}}, i[[SZ]] %{{.+}}, float* {{.+}}, [5 x [10 x double]]* {{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, double* {{.+}}, [[TT]]* {{.+}}) 468 // To reduce complexity, we're only going as far as validating the signature of the outlined parallel function. 469 470 template<typename tx> 471 tx ftemplate(int n) { 472 tx a = 0; 473 short aa = 0; 474 tx b[10]; 475 476 #pragma omp target teams if(target: n>40) 477 { 478 a += 1; 479 aa += 1; 480 b[2] += 1; 481 } 482 483 return a; 484 } 485 486 static 487 int fstatic(int n) { 488 int a = 0; 489 short aa = 0; 490 char aaa = 0; 491 int b[10]; 492 493 #pragma omp target teams if(target: n>50) 494 { 495 a += 1; 496 aa += 1; 497 aaa += 1; 498 b[2] += 1; 499 } 500 501 return a; 502 } 503 504 struct S1 { 505 double a; 506 507 int r1(int n){ 508 int b = n+1; 509 short int c[2][n]; 510 511 #pragma omp target teams if(target: n>60) 512 { 513 this->a = (double)b + 1.5; 514 c[1][1] = ++a; 515 } 516 517 return c[1][1] + (int)b; 518 } 519 }; 520 521 // CHECK: define {{.*}}@{{.*}}bar{{.*}} 522 int bar(int n){ 523 int a = 0; 524 525 // CHECK: call {{.*}}i32 [[FOO]](i32 {{.*}}) 526 a += foo(n); 527 528 S1 S; 529 // CHECK: call {{.*}}i32 [[FS1:@.+]]([[S1]]* {{.*}}, i32 {{.*}}) 530 a += S.r1(n); 531 532 // CHECK: call {{.*}}i32 [[FSTATIC:@.+]](i32 {{.*}}) 533 a += fstatic(n); 534 535 // CHECK: call {{.*}}i32 [[FTEMPLATE:@.+]](i32 {{.*}}) 536 a += ftemplate<int>(n); 537 538 return a; 539 } 540 541 // 542 // CHECK: define {{.*}}[[FS1]] 543 // 544 // CHECK: i8* @llvm.stacksave() 545 // CHECK-64: [[B_ADDR:%.+]] = bitcast i[[SZ]]* [[B_CADDR:%.+]] to i32* 546 // CHECK-64: store i32 %{{.+}}, i32* [[B_ADDR]], 547 // CHECK-64: [[B_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[B_CADDR]], 548 549 // CHECK-32: store i32 %{{.+}}, i32* [[B_ADDR:%.+]], 550 // CHECK-32: [[B_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[B_ADDR]], 551 552 // We capture 2 VLA sizes in this target region 553 // CHECK: [[CELEMSIZE2:%.+]] = mul nuw i[[SZ]] 2, [[VLA0:%.+]] 554 // CHECK: [[CSIZE:%.+]] = mul nuw i[[SZ]] [[CELEMSIZE2]], 2 555 556 // CHECK: [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 60 557 // CHECK: br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]] 558 // CHECK: [[TRY]] 559 // CHECK-DAG: [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[MAPT7]], i32 0, i32 0), i32 0, i32 0) 560 // CHECK-DAG: [[BPR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP:%.+]], i32 0, i32 0 561 // CHECK-DAG: [[PR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P:%.+]], i32 0, i32 0 562 // CHECK-DAG: [[SR]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S:%.+]], i32 0, i32 0 563 // CHECK-DAG: [[SADDR0:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX0:[0-9]+]] 564 // CHECK-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX0]] 565 // CHECK-DAG: [[PADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX0]] 566 // CHECK-DAG: [[SADDR1:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX1:[0-9]+]] 567 // CHECK-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX1]] 568 // CHECK-DAG: [[PADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX1]] 569 // CHECK-DAG: [[SADDR2:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX2:[0-9]+]] 570 // CHECK-DAG: [[BPADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX2]] 571 // CHECK-DAG: [[PADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX2]] 572 // CHECK-DAG: [[SADDR3:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX3:[0-9]+]] 573 // CHECK-DAG: [[BPADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX3]] 574 // CHECK-DAG: [[PADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX3]] 575 576 // The names below are not necessarily consistent with the names used for the 577 // addresses above as some are repeated. 578 // CHECK-DAG: store i[[SZ]] [[VLA0]], i[[SZ]]* [[CBPADDR0:%.+]], 579 // CHECK-DAG: store i[[SZ]] [[VLA0]], i[[SZ]]* [[CPADDR0:%.+]], 580 // CHECK-DAG: [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 581 // CHECK-DAG: [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 582 // CHECK-DAG: store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}} 583 584 // CHECK-DAG: store i[[SZ]] 2, i[[SZ]]* [[CBPADDR1:%.+]], 585 // CHECK-DAG: store i[[SZ]] 2, i[[SZ]]* [[CPADDR1:%.+]], 586 // CHECK-DAG: [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 587 // CHECK-DAG: [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 588 // CHECK-DAG: store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}} 589 590 // CHECK-DAG: store i[[SZ]] [[B_CVAL]], i[[SZ]]* [[CBPADDR2:%.+]], 591 // CHECK-DAG: store i[[SZ]] [[B_CVAL]], i[[SZ]]* [[CPADDR2:%.+]], 592 // CHECK-DAG: [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 593 // CHECK-DAG: [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 594 // CHECK-DAG: store i[[SZ]] 4, i[[SZ]]* {{%[^,]+}} 595 596 // CHECK-DAG: store [[S1]]* %{{.+}}, [[S1]]** [[CBPADDR3:%.+]], 597 // CHECK-DAG: store [[S1]]* %{{.+}}, [[S1]]** [[CPADDR3:%.+]], 598 // CHECK-DAG: [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to [[S1]]** 599 // CHECK-DAG: [[CPADDR3]] = bitcast i8** {{%[^,]+}} to [[S1]]** 600 // CHECK-DAG: store i[[SZ]] 8, i[[SZ]]* {{%[^,]+}} 601 602 // CHECK-DAG: store i16* %{{.+}}, i16** [[CBPADDR4:%.+]], 603 // CHECK-DAG: store i16* %{{.+}}, i16** [[CPADDR4:%.+]], 604 // CHECK-DAG: [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to i16** 605 // CHECK-DAG: [[CPADDR4]] = bitcast i8** {{%[^,]+}} to i16** 606 // CHECK-DAG: store i[[SZ]] [[CSIZE]], i[[SZ]]* {{%[^,]+}} 607 608 // CHECK-NEXT: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0 609 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]] 610 611 // CHECK: [[FAIL]] 612 // CHECK: call void [[HVT7:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}) 613 // CHECK-NEXT: br label %[[END]] 614 // CHECK: [[END]] 615 616 // 617 // CHECK: define {{.*}}[[FSTATIC]] 618 // 619 // CHECK: [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 50 620 // CHECK: br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]] 621 // CHECK: [[IFTHEN]] 622 // CHECK-DAG: [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([4 x i[[SZ]]], [4 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i64* getelementptr inbounds ([4 x i64], [4 x i64]* [[MAPT6]], i32 0, i32 0), i32 0, i32 0) 623 // CHECK-DAG: [[BPR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP:%.+]], i32 0, i32 0 624 // CHECK-DAG: [[PR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P:%.+]], i32 0, i32 0 625 626 // CHECK-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 0 627 // CHECK-DAG: [[PADDR0:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 0 628 // CHECK-DAG: store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0:%.+]], 629 // CHECK-DAG: store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0:%.+]], 630 // CHECK-DAG: [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 631 // CHECK-DAG: [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 632 633 // CHECK-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 1 634 // CHECK-DAG: [[PADDR1:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 1 635 // CHECK-DAG: store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1:%.+]], 636 // CHECK-DAG: store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1:%.+]], 637 // CHECK-DAG: [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 638 // CHECK-DAG: [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 639 640 // CHECK-DAG: [[BPADDR2:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 2 641 // CHECK-DAG: [[PADDR2:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 2 642 // CHECK-DAG: store i[[SZ]] [[VAL2:%.+]], i[[SZ]]* [[CBPADDR2:%.+]], 643 // CHECK-DAG: store i[[SZ]] [[VAL2]], i[[SZ]]* [[CPADDR2:%.+]], 644 // CHECK-DAG: [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 645 // CHECK-DAG: [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 646 647 // CHECK-DAG: [[BPADDR3:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 3 648 // CHECK-DAG: [[PADDR3:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 3 649 // CHECK-DAG: store [10 x i32]* %{{.+}}, [10 x i32]** [[CBPADDR3:%.+]], 650 // CHECK-DAG: store [10 x i32]* %{{.+}}, [10 x i32]** [[CPADDR3:%.+]], 651 // CHECK-DAG: [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to [10 x i32]** 652 // CHECK-DAG: [[CPADDR3]] = bitcast i8** {{%[^,]+}} to [10 x i32]** 653 654 // CHECK: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0 655 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]] 656 // CHECK: [[FAIL]] 657 // CHECK: call void [[HVT6:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}) 658 // CHECK-NEXT: br label %[[END]] 659 // CHECK: [[END]] 660 // CHECK-NEXT: br label %[[IFEND:.+]] 661 // CHECK: [[IFELSE]] 662 // CHECK: call void [[HVT6]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}) 663 // CHECK-NEXT: br label %[[IFEND]] 664 // CHECK: [[IFEND]] 665 666 // 667 // CHECK: define {{.*}}[[FTEMPLATE]] 668 // 669 // CHECK: [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 40 670 // CHECK: br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]] 671 // CHECK: [[IFTHEN]] 672 // CHECK-DAG: [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET5]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT5]], i32 0, i32 0), i32 0, i32 0) 673 // CHECK-DAG: [[BPR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP:%.+]], i32 0, i32 0 674 // CHECK-DAG: [[PR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P:%.+]], i32 0, i32 0 675 676 // CHECK-DAG: [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 0 677 // CHECK-DAG: [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 0 678 // CHECK-DAG: store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0:%.+]], 679 // CHECK-DAG: store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0:%.+]], 680 // CHECK-DAG: [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 681 // CHECK-DAG: [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 682 683 // CHECK-DAG: [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 1 684 // CHECK-DAG: [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 1 685 // CHECK-DAG: store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1:%.+]], 686 // CHECK-DAG: store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1:%.+]], 687 // CHECK-DAG: [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 688 // CHECK-DAG: [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]* 689 690 // CHECK-DAG: [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 2 691 // CHECK-DAG: [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 2 692 // CHECK-DAG: store [10 x i32]* [[VAL2:%.+]], [10 x i32]** [[CBPADDR2:%.+]], 693 // CHECK-DAG: store [10 x i32]* [[VAL2]], [10 x i32]** [[CPADDR2:%.+]], 694 // CHECK-DAG: [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to [10 x i32]** 695 // CHECK-DAG: [[CPADDR2]] = bitcast i8** {{%[^,]+}} to [10 x i32]** 696 697 // CHECK: [[ERROR:%.+]] = icmp ne i32 [[RET]], 0 698 // CHECK-NEXT: br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]] 699 // CHECK: [[FAIL]] 700 // CHECK: call void [[HVT5:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}) 701 // CHECK-NEXT: br label %[[END]] 702 // CHECK: [[END]] 703 // CHECK-NEXT: br label %[[IFEND:.+]] 704 // CHECK: [[IFELSE]] 705 // CHECK: call void [[HVT5]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}) 706 // CHECK-NEXT: br label %[[IFEND]] 707 // CHECK: [[IFEND]] 708 709 710 711 // Check that the offloading functions are emitted and that the arguments are 712 // correct and loaded correctly for the target regions of the callees of bar(). 713 714 // CHECK: define internal void [[HVT7]] 715 // Create local storage for each capture. 716 // CHECK: [[LOCAL_THIS:%.+]] = alloca [[S1]]* 717 // CHECK: [[LOCAL_B:%.+]] = alloca i[[SZ]] 718 // CHECK: [[LOCAL_VLA1:%.+]] = alloca i[[SZ]] 719 // CHECK: [[LOCAL_VLA2:%.+]] = alloca i[[SZ]] 720 // CHECK: [[LOCAL_C:%.+]] = alloca i16* 721 // CHECK: [[LOCAL_B_CASTED:%.+]] = alloca i[[SZ]] 722 // CHECK-DAG: store [[S1]]* [[ARG_THIS:%.+]], [[S1]]** [[LOCAL_THIS]] 723 // CHECK-DAG: store i[[SZ]] [[ARG_B:%.+]], i[[SZ]]* [[LOCAL_B]] 724 // CHECK-DAG: store i[[SZ]] [[ARG_VLA1:%.+]], i[[SZ]]* [[LOCAL_VLA1]] 725 // CHECK-DAG: store i[[SZ]] [[ARG_VLA2:%.+]], i[[SZ]]* [[LOCAL_VLA2]] 726 // CHECK-DAG: store i16* [[ARG_C:%.+]], i16** [[LOCAL_C]] 727 // Store captures in the context. 728 // CHECK-DAG: [[REF_THIS:%.+]] = load [[S1]]*, [[S1]]** [[LOCAL_THIS]], 729 // CHECK-64-DAG:[[CONV_BP:%.+]] = bitcast i[[SZ]]* [[LOCAL_B]] to i32* 730 // CHECK-DAG: [[VAL_VLA1:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA1]], 731 // CHECK-DAG: [[VAL_VLA2:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA2]], 732 // CHECK-DAG: [[REF_C:%.+]] = load i16*, i16** [[LOCAL_C]], 733 734 // CHECK-64-DAG:[[CONV_B:%.+]] = load i32, i32* [[CONV_BP]] 735 // CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_B_CASTED]] to i32* 736 // CHECK-64-DAG:store i32 [[CONV_B]], i32* [[CONV]], align 737 // CHECK-32-DAG:[[LOCAL_BV:%.+]] = load i32, i32* [[LOCAL_B]] 738 // CHECK-32-DAG:store i32 [[LOCAL_BV]], i32* [[LOCAL_B_CASTED]], align 739 // CHECK-DAG: [[REF_B:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_B_CASTED]], 740 741 // CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, [[S1]]*, i[[SZ]], i[[SZ]], i[[SZ]], i16*)* [[OMP_OUTLINED5:@.+]] to void (i32*, i32*, ...)*), [[S1]]* [[REF_THIS]], i[[SZ]] [[REF_B]], i[[SZ]] [[VAL_VLA1]], i[[SZ]] [[VAL_VLA2]], i16* [[REF_C]]) 742 // 743 // 744 // CHECK: define internal {{.*}}void [[OMP_OUTLINED5]](i32* noalias %.global_tid., i32* noalias %.bound_tid., [[S1]]* %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i16* {{.+}}) 745 // To reduce complexity, we're only going as far as validating the signature of the outlined parallel function. 746 747 748 // CHECK: define internal void [[HVT6]] 749 // Create local storage for each capture. 750 // CHECK: [[LOCAL_A:%.+]] = alloca i[[SZ]] 751 // CHECK: [[LOCAL_AA:%.+]] = alloca i[[SZ]] 752 // CHECK: [[LOCAL_AAA:%.+]] = alloca i[[SZ]] 753 // CHECK: [[LOCAL_B:%.+]] = alloca [10 x i32]* 754 // CHECK: [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]] 755 // CHECK: [[LOCAL_AA_CASTED:%.+]] = alloca i[[SZ]] 756 // CHECK: [[LOCAL_AAA_CASTED:%.+]] = alloca i[[SZ]] 757 // CHECK-DAG: store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]] 758 // CHECK-DAG: store i[[SZ]] [[ARG_AA:%.+]], i[[SZ]]* [[LOCAL_AA]] 759 // CHECK-DAG: store i[[SZ]] [[ARG_AAA:%.+]], i[[SZ]]* [[LOCAL_AAA]] 760 // CHECK-DAG: store [10 x i32]* [[ARG_B:%.+]], [10 x i32]** [[LOCAL_B]] 761 // Store captures in the context. 762 // CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32* 763 // CHECK-DAG: [[CONV_AAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA]] to i16* 764 // CHECK-DAG: [[CONV_AAAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AAA]] to i8* 765 // CHECK-DAG: [[REF_B:%.+]] = load [10 x i32]*, [10 x i32]** [[LOCAL_B]], 766 767 // CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]] 768 // CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32* 769 // CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align 770 // CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]] 771 // CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align 772 // CHECK-DAG: [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]], 773 774 // CHECK-DAG: [[CONV_AA:%.+]] = load i16, i16* [[CONV_AAP]] 775 // CHECK-DAG: [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA_CASTED]] to i16* 776 // CHECK-DAG: store i16 [[CONV_AA]], i16* [[CONV]], align 777 // CHECK-DAG: [[REF_AA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AA_CASTED]], 778 779 // CHECK-DAG: [[CONV_AAA:%.+]] = load i8, i8* [[CONV_AAAP]] 780 // CHECK-DAG: [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AAA_CASTED]] to i8* 781 // CHECK-DAG: store i8 [[CONV_AAA]], i8* [[CONV]], align 782 // CHECK-DAG: [[REF_AAA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AAA_CASTED]], 783 784 // CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 4, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], i[[SZ]], [10 x i32]*)* [[OMP_OUTLINED6:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], i[[SZ]] [[REF_AA]], i[[SZ]] [[REF_AAA]], [10 x i32]* [[REF_B]]) 785 // 786 // 787 // CHECK: define internal {{.*}}void [[OMP_OUTLINED6]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, [10 x i32]* {{.+}}) 788 // To reduce complexity, we're only going as far as validating the signature of the outlined parallel function. 789 790 // CHECK: define internal void [[HVT5]] 791 // Create local storage for each capture. 792 // CHECK: [[LOCAL_A:%.+]] = alloca i[[SZ]] 793 // CHECK: [[LOCAL_AA:%.+]] = alloca i[[SZ]] 794 // CHECK: [[LOCAL_B:%.+]] = alloca [10 x i32]* 795 // CHECK: [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]] 796 // CHECK: [[LOCAL_AA_CASTED:%.+]] = alloca i[[SZ]] 797 // CHECK-DAG: store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]] 798 // CHECK-DAG: store i[[SZ]] [[ARG_AA:%.+]], i[[SZ]]* [[LOCAL_AA]] 799 // CHECK-DAG: store [10 x i32]* [[ARG_B:%.+]], [10 x i32]** [[LOCAL_B]] 800 // Store captures in the context. 801 // CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32* 802 // CHECK-DAG: [[CONV_AAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA]] to i16* 803 // CHECK-DAG: [[REF_B:%.+]] = load [10 x i32]*, [10 x i32]** [[LOCAL_B]], 804 805 // CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]] 806 // CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32* 807 // CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align 808 // CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]] 809 // CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align 810 // CHECK-DAG: [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]], 811 812 // CHECK-DAG: [[CONV_AA:%.+]] = load i16, i16* [[CONV_AAP]] 813 // CHECK-DAG: [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA_CASTED]] to i16* 814 // CHECK-DAG: store i16 [[CONV_AA]], i16* [[CONV]], align 815 // CHECK-DAG: [[REF_AA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AA_CASTED]], 816 817 // CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 3, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], [10 x i32]*)* [[OMP_OUTLINED7:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], i[[SZ]] [[REF_AA]], [10 x i32]* [[REF_B]]) 818 // 819 // 820 // CHECK: define internal {{.*}}void [[OMP_OUTLINED7]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, [10 x i32]* {{.+}}) 821 // To reduce complexity, we're only going as far as validating the signature of the outlined parallel function. 822 823 #endif 824