1 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --function-signature --include-generated-funcs --replace-value-regex "__omp_offloading_[0-9a-z]+_[0-9a-z]+" "reduction_size[.].+[.]" "pl_cond[.].+[.|,]" --prefix-filecheck-ir-name _ 2 // Test host codegen. 3 // 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=CHECK1 4 // 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 5 // 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=CHECK1 6 // 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=CHECK3 7 // 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 8 // 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=CHECK3 9 10 // RUN: %clang_cc1 -verify -fopenmp-simd -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}" 11 // 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 12 // 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 %s -emit-llvm -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}" 13 // RUN: %clang_cc1 -verify -fopenmp-simd -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}" 14 // 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 15 // 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 %s -emit-llvm -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}" 16 17 // Test target codegen - host bc file has to be created first. 18 // 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 19 // 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=CHECK9 20 // 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 21 // 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=CHECK9 22 // 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 23 // 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=CHECK11 24 // 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 25 // 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=CHECK11 26 27 // RUN: %clang_cc1 -verify -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 28 // RUN: %clang_cc1 -verify -fopenmp-simd -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 --implicit-check-not="{{__kmpc|__tgt}}" 29 // 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-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s 30 // 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-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}" 31 // RUN: %clang_cc1 -verify -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 32 // RUN: %clang_cc1 -verify -fopenmp-simd -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 --implicit-check-not="{{__kmpc|__tgt}}" 33 // 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-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s 34 // 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-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}" 35 36 // Test host codegen. 37 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK1 38 // RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s 39 // RUN: %clang_cc1 -fopenmp -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=CHECK1 40 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK3 41 // RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s 42 // RUN: %clang_cc1 -fopenmp -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=CHECK3 43 44 // RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}" 45 // RUN: %clang_cc1 -fopenmp-simd -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s 46 // RUN: %clang_cc1 -fopenmp-simd -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 --implicit-check-not="{{__kmpc|__tgt}}" 47 // RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}" 48 // RUN: %clang_cc1 -fopenmp-simd -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s 49 // RUN: %clang_cc1 -fopenmp-simd -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 --implicit-check-not="{{__kmpc|__tgt}}" 50 51 // Test target codegen - host bc file has to be created first. 52 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc 53 // RUN: %clang_cc1 -verify -fopenmp -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=CHECK9 54 // RUN: %clang_cc1 -fopenmp -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 55 // RUN: %clang_cc1 -fopenmp -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=CHECK9 56 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc 57 // RUN: %clang_cc1 -verify -fopenmp -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=CHECK11 58 // RUN: %clang_cc1 -fopenmp -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 59 // RUN: %clang_cc1 -fopenmp -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=CHECK11 60 61 // RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc 62 // RUN: %clang_cc1 -verify -fopenmp-simd -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 --implicit-check-not="{{__kmpc|__tgt}}" 63 // RUN: %clang_cc1 -fopenmp-simd -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 64 // RUN: %clang_cc1 -fopenmp-simd -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 --implicit-check-not="{{__kmpc|__tgt}}" 65 // RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc 66 // RUN: %clang_cc1 -verify -fopenmp-simd -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 --implicit-check-not="{{__kmpc|__tgt}}" 67 // RUN: %clang_cc1 -fopenmp-simd -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 68 // RUN: %clang_cc1 -fopenmp-simd -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 --implicit-check-not="{{__kmpc|__tgt}}" 69 70 // expected-no-diagnostics 71 #ifndef HEADER 72 #define HEADER 73 74 75 76 77 // We have 6 target regions 78 79 80 81 // Check target registration is registered as a Ctor. 82 83 84 template<typename tx> 85 tx ftemplate(int n) { 86 tx a = 0; 87 88 #pragma omp target teams num_teams(tx(20)) 89 { 90 } 91 92 short b = 1; 93 #pragma omp target teams num_teams(b) 94 { 95 a += b; 96 } 97 98 return a; 99 } 100 101 static 102 int fstatic(int n) { 103 104 #pragma omp target teams num_teams(n) 105 { 106 } 107 108 #pragma omp target teams num_teams(32+n) 109 { 110 } 111 112 return n+1; 113 } 114 115 struct S1 { 116 double a; 117 118 int r1(int n){ 119 int b = 1; 120 121 #pragma omp target teams num_teams(n-b) 122 { 123 this->a = (double)b + 1.5; 124 } 125 126 #pragma omp target teams num_teams(1024) 127 { 128 this->a = 2.5; 129 } 130 131 return (int)a; 132 } 133 }; 134 135 int bar(int n){ 136 int a = 0; 137 138 S1 S; 139 a += S.r1(n); 140 141 a += fstatic(n); 142 143 a += ftemplate<int>(n); 144 145 return a; 146 } 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 // Check that the offloading functions are emitted and that the parallel function 169 // is appropriately guarded. 170 171 172 173 174 175 176 #endif 177 // CHECK1-LABEL: define {{[^@]+}}@_Z3bari 178 // CHECK1-SAME: (i32 noundef signext [[N:%.*]]) #[[ATTR0:[0-9]+]] { 179 // CHECK1-NEXT: entry: 180 // CHECK1-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 181 // CHECK1-NEXT: [[A:%.*]] = alloca i32, align 4 182 // CHECK1-NEXT: [[S:%.*]] = alloca [[STRUCT_S1:%.*]], align 8 183 // CHECK1-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 184 // CHECK1-NEXT: store i32 0, ptr [[A]], align 4 185 // CHECK1-NEXT: [[TMP0:%.*]] = load i32, ptr [[N_ADDR]], align 4 186 // CHECK1-NEXT: [[CALL:%.*]] = call noundef signext i32 @_ZN2S12r1Ei(ptr noundef nonnull align 8 dereferenceable(8) [[S]], i32 noundef signext [[TMP0]]) 187 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[A]], align 4 188 // CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP1]], [[CALL]] 189 // CHECK1-NEXT: store i32 [[ADD]], ptr [[A]], align 4 190 // CHECK1-NEXT: [[TMP2:%.*]] = load i32, ptr [[N_ADDR]], align 4 191 // CHECK1-NEXT: [[CALL1:%.*]] = call noundef signext i32 @_ZL7fstatici(i32 noundef signext [[TMP2]]) 192 // CHECK1-NEXT: [[TMP3:%.*]] = load i32, ptr [[A]], align 4 193 // CHECK1-NEXT: [[ADD2:%.*]] = add nsw i32 [[TMP3]], [[CALL1]] 194 // CHECK1-NEXT: store i32 [[ADD2]], ptr [[A]], align 4 195 // CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[N_ADDR]], align 4 196 // CHECK1-NEXT: [[CALL3:%.*]] = call noundef signext i32 @_Z9ftemplateIiET_i(i32 noundef signext [[TMP4]]) 197 // CHECK1-NEXT: [[TMP5:%.*]] = load i32, ptr [[A]], align 4 198 // CHECK1-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP5]], [[CALL3]] 199 // CHECK1-NEXT: store i32 [[ADD4]], ptr [[A]], align 4 200 // CHECK1-NEXT: [[TMP6:%.*]] = load i32, ptr [[A]], align 4 201 // CHECK1-NEXT: ret i32 [[TMP6]] 202 // 203 // 204 // CHECK1-LABEL: define {{[^@]+}}@_ZN2S12r1Ei 205 // CHECK1-SAME: (ptr noundef nonnull align 8 dereferenceable(8) [[THIS:%.*]], i32 noundef signext [[N:%.*]]) #[[ATTR0]] comdat align 2 { 206 // CHECK1-NEXT: entry: 207 // CHECK1-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 208 // CHECK1-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 209 // CHECK1-NEXT: [[B:%.*]] = alloca i32, align 4 210 // CHECK1-NEXT: [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4 211 // CHECK1-NEXT: [[B_CASTED:%.*]] = alloca i64, align 8 212 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i64, align 8 213 // CHECK1-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [3 x ptr], align 8 214 // CHECK1-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [3 x ptr], align 8 215 // CHECK1-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [3 x ptr], align 8 216 // CHECK1-NEXT: [[DOTOFFLOAD_BASEPTRS3:%.*]] = alloca [1 x ptr], align 8 217 // CHECK1-NEXT: [[DOTOFFLOAD_PTRS4:%.*]] = alloca [1 x ptr], align 8 218 // CHECK1-NEXT: [[DOTOFFLOAD_MAPPERS5:%.*]] = alloca [1 x ptr], align 8 219 // CHECK1-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 220 // CHECK1-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 221 // CHECK1-NEXT: [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 222 // CHECK1-NEXT: store i32 1, ptr [[B]], align 4 223 // CHECK1-NEXT: [[TMP0:%.*]] = load i32, ptr [[N_ADDR]], align 4 224 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[B]], align 4 225 // CHECK1-NEXT: [[SUB:%.*]] = sub nsw i32 [[TMP0]], [[TMP1]] 226 // CHECK1-NEXT: store i32 [[SUB]], ptr [[DOTCAPTURE_EXPR_]], align 4 227 // CHECK1-NEXT: [[TMP2:%.*]] = load i32, ptr [[B]], align 4 228 // CHECK1-NEXT: store i32 [[TMP2]], ptr [[B_CASTED]], align 4 229 // CHECK1-NEXT: [[TMP3:%.*]] = load i64, ptr [[B_CASTED]], align 8 230 // CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 231 // CHECK1-NEXT: store i32 [[TMP4]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 232 // CHECK1-NEXT: [[TMP5:%.*]] = load i64, ptr [[DOTCAPTURE_EXPR__CASTED]], align 8 233 // CHECK1-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[THIS1]], i32 0, i32 0 234 // CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 235 // CHECK1-NEXT: store ptr [[THIS1]], ptr [[TMP6]], align 8 236 // CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 237 // CHECK1-NEXT: store ptr [[A]], ptr [[TMP7]], align 8 238 // CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 0 239 // CHECK1-NEXT: store ptr null, ptr [[TMP8]], align 8 240 // CHECK1-NEXT: [[TMP9:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 1 241 // CHECK1-NEXT: store i64 [[TMP3]], ptr [[TMP9]], align 8 242 // CHECK1-NEXT: [[TMP10:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 1 243 // CHECK1-NEXT: store i64 [[TMP3]], ptr [[TMP10]], align 8 244 // CHECK1-NEXT: [[TMP11:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 1 245 // CHECK1-NEXT: store ptr null, ptr [[TMP11]], align 8 246 // CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 2 247 // CHECK1-NEXT: store i64 [[TMP5]], ptr [[TMP12]], align 8 248 // CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 2 249 // CHECK1-NEXT: store i64 [[TMP5]], ptr [[TMP13]], align 8 250 // CHECK1-NEXT: [[TMP14:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 2 251 // CHECK1-NEXT: store ptr null, ptr [[TMP14]], align 8 252 // CHECK1-NEXT: [[TMP15:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 253 // CHECK1-NEXT: [[TMP16:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 254 // CHECK1-NEXT: [[TMP17:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 255 // CHECK1-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8 256 // CHECK1-NEXT: [[TMP18:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 257 // CHECK1-NEXT: store i32 1, ptr [[TMP18]], align 4 258 // CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1 259 // CHECK1-NEXT: store i32 3, ptr [[TMP19]], align 4 260 // CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2 261 // CHECK1-NEXT: store ptr [[TMP15]], ptr [[TMP20]], align 8 262 // CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3 263 // CHECK1-NEXT: store ptr [[TMP16]], ptr [[TMP21]], align 8 264 // CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4 265 // CHECK1-NEXT: store ptr @.offload_sizes, ptr [[TMP22]], align 8 266 // CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5 267 // CHECK1-NEXT: store ptr @.offload_maptypes, ptr [[TMP23]], align 8 268 // CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6 269 // CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8 270 // CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7 271 // CHECK1-NEXT: store ptr null, ptr [[TMP25]], align 8 272 // CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8 273 // CHECK1-NEXT: store i64 0, ptr [[TMP26]], align 8 274 // CHECK1-NEXT: [[TMP27:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1:[0-9]+]], i64 -1, i32 [[TMP17]], i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121.region_id, ptr [[KERNEL_ARGS]]) 275 // CHECK1-NEXT: [[TMP28:%.*]] = icmp ne i32 [[TMP27]], 0 276 // CHECK1-NEXT: br i1 [[TMP28]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]] 277 // CHECK1: omp_offload.failed: 278 // CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121(ptr [[THIS1]], i64 [[TMP3]], i64 [[TMP5]]) #[[ATTR2:[0-9]+]] 279 // CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT]] 280 // CHECK1: omp_offload.cont: 281 // CHECK1-NEXT: [[A2:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[THIS1]], i32 0, i32 0 282 // CHECK1-NEXT: [[TMP29:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS3]], i32 0, i32 0 283 // CHECK1-NEXT: store ptr [[THIS1]], ptr [[TMP29]], align 8 284 // CHECK1-NEXT: [[TMP30:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS4]], i32 0, i32 0 285 // CHECK1-NEXT: store ptr [[A2]], ptr [[TMP30]], align 8 286 // CHECK1-NEXT: [[TMP31:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS5]], i64 0, i64 0 287 // CHECK1-NEXT: store ptr null, ptr [[TMP31]], align 8 288 // CHECK1-NEXT: [[TMP32:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS3]], i32 0, i32 0 289 // CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS4]], i32 0, i32 0 290 // CHECK1-NEXT: [[KERNEL_ARGS6:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS]], align 8 291 // CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 0 292 // CHECK1-NEXT: store i32 1, ptr [[TMP34]], align 4 293 // CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 1 294 // CHECK1-NEXT: store i32 1, ptr [[TMP35]], align 4 295 // CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 2 296 // CHECK1-NEXT: store ptr [[TMP32]], ptr [[TMP36]], align 8 297 // CHECK1-NEXT: [[TMP37:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 3 298 // CHECK1-NEXT: store ptr [[TMP33]], ptr [[TMP37]], align 8 299 // CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 4 300 // CHECK1-NEXT: store ptr @.offload_sizes.2, ptr [[TMP38]], align 8 301 // CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 5 302 // CHECK1-NEXT: store ptr @.offload_maptypes.3, ptr [[TMP39]], align 8 303 // CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 6 304 // CHECK1-NEXT: store ptr null, ptr [[TMP40]], align 8 305 // CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 7 306 // CHECK1-NEXT: store ptr null, ptr [[TMP41]], align 8 307 // CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 8 308 // CHECK1-NEXT: store i64 0, ptr [[TMP42]], align 8 309 // CHECK1-NEXT: [[TMP43:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 1024, i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126.region_id, ptr [[KERNEL_ARGS6]]) 310 // CHECK1-NEXT: [[TMP44:%.*]] = icmp ne i32 [[TMP43]], 0 311 // CHECK1-NEXT: br i1 [[TMP44]], label [[OMP_OFFLOAD_FAILED7:%.*]], label [[OMP_OFFLOAD_CONT8:%.*]] 312 // CHECK1: omp_offload.failed7: 313 // CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126(ptr [[THIS1]]) #[[ATTR2]] 314 // CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT8]] 315 // CHECK1: omp_offload.cont8: 316 // CHECK1-NEXT: [[A9:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[THIS1]], i32 0, i32 0 317 // CHECK1-NEXT: [[TMP45:%.*]] = load double, ptr [[A9]], align 8 318 // CHECK1-NEXT: [[CONV:%.*]] = fptosi double [[TMP45]] to i32 319 // CHECK1-NEXT: ret i32 [[CONV]] 320 // 321 // 322 // CHECK1-LABEL: define {{[^@]+}}@_ZL7fstatici 323 // CHECK1-SAME: (i32 noundef signext [[N:%.*]]) #[[ATTR0]] { 324 // CHECK1-NEXT: entry: 325 // CHECK1-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 326 // CHECK1-NEXT: [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4 327 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i64, align 8 328 // CHECK1-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [1 x ptr], align 8 329 // CHECK1-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [1 x ptr], align 8 330 // CHECK1-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [1 x ptr], align 8 331 // CHECK1-NEXT: [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4 332 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__CASTED2:%.*]] = alloca i64, align 8 333 // CHECK1-NEXT: [[DOTOFFLOAD_BASEPTRS3:%.*]] = alloca [1 x ptr], align 8 334 // CHECK1-NEXT: [[DOTOFFLOAD_PTRS4:%.*]] = alloca [1 x ptr], align 8 335 // CHECK1-NEXT: [[DOTOFFLOAD_MAPPERS5:%.*]] = alloca [1 x ptr], align 8 336 // CHECK1-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 337 // CHECK1-NEXT: [[TMP0:%.*]] = load i32, ptr [[N_ADDR]], align 4 338 // CHECK1-NEXT: store i32 [[TMP0]], ptr [[DOTCAPTURE_EXPR_]], align 4 339 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 340 // CHECK1-NEXT: store i32 [[TMP1]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 341 // CHECK1-NEXT: [[TMP2:%.*]] = load i64, ptr [[DOTCAPTURE_EXPR__CASTED]], align 8 342 // CHECK1-NEXT: [[TMP3:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 343 // CHECK1-NEXT: store i64 [[TMP2]], ptr [[TMP3]], align 8 344 // CHECK1-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 345 // CHECK1-NEXT: store i64 [[TMP2]], ptr [[TMP4]], align 8 346 // CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 0 347 // CHECK1-NEXT: store ptr null, ptr [[TMP5]], align 8 348 // CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 349 // CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 350 // CHECK1-NEXT: [[TMP8:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 351 // CHECK1-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8 352 // CHECK1-NEXT: [[TMP9:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 353 // CHECK1-NEXT: store i32 1, ptr [[TMP9]], align 4 354 // CHECK1-NEXT: [[TMP10:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1 355 // CHECK1-NEXT: store i32 1, ptr [[TMP10]], align 4 356 // CHECK1-NEXT: [[TMP11:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2 357 // CHECK1-NEXT: store ptr [[TMP6]], ptr [[TMP11]], align 8 358 // CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3 359 // CHECK1-NEXT: store ptr [[TMP7]], ptr [[TMP12]], align 8 360 // CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4 361 // CHECK1-NEXT: store ptr @.offload_sizes.5, ptr [[TMP13]], align 8 362 // CHECK1-NEXT: [[TMP14:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5 363 // CHECK1-NEXT: store ptr @.offload_maptypes.6, ptr [[TMP14]], align 8 364 // CHECK1-NEXT: [[TMP15:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6 365 // CHECK1-NEXT: store ptr null, ptr [[TMP15]], align 8 366 // CHECK1-NEXT: [[TMP16:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7 367 // CHECK1-NEXT: store ptr null, ptr [[TMP16]], align 8 368 // CHECK1-NEXT: [[TMP17:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8 369 // CHECK1-NEXT: store i64 0, ptr [[TMP17]], align 8 370 // CHECK1-NEXT: [[TMP18:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 [[TMP8]], i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104.region_id, ptr [[KERNEL_ARGS]]) 371 // CHECK1-NEXT: [[TMP19:%.*]] = icmp ne i32 [[TMP18]], 0 372 // CHECK1-NEXT: br i1 [[TMP19]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]] 373 // CHECK1: omp_offload.failed: 374 // CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104(i64 [[TMP2]]) #[[ATTR2]] 375 // CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT]] 376 // CHECK1: omp_offload.cont: 377 // CHECK1-NEXT: [[TMP20:%.*]] = load i32, ptr [[N_ADDR]], align 4 378 // CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 32, [[TMP20]] 379 // CHECK1-NEXT: store i32 [[ADD]], ptr [[DOTCAPTURE_EXPR_1]], align 4 380 // CHECK1-NEXT: [[TMP21:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4 381 // CHECK1-NEXT: store i32 [[TMP21]], ptr [[DOTCAPTURE_EXPR__CASTED2]], align 4 382 // CHECK1-NEXT: [[TMP22:%.*]] = load i64, ptr [[DOTCAPTURE_EXPR__CASTED2]], align 8 383 // CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS3]], i32 0, i32 0 384 // CHECK1-NEXT: store i64 [[TMP22]], ptr [[TMP23]], align 8 385 // CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS4]], i32 0, i32 0 386 // CHECK1-NEXT: store i64 [[TMP22]], ptr [[TMP24]], align 8 387 // CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS5]], i64 0, i64 0 388 // CHECK1-NEXT: store ptr null, ptr [[TMP25]], align 8 389 // CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS3]], i32 0, i32 0 390 // CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS4]], i32 0, i32 0 391 // CHECK1-NEXT: [[TMP28:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4 392 // CHECK1-NEXT: [[KERNEL_ARGS6:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS]], align 8 393 // CHECK1-NEXT: [[TMP29:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 0 394 // CHECK1-NEXT: store i32 1, ptr [[TMP29]], align 4 395 // CHECK1-NEXT: [[TMP30:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 1 396 // CHECK1-NEXT: store i32 1, ptr [[TMP30]], align 4 397 // CHECK1-NEXT: [[TMP31:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 2 398 // CHECK1-NEXT: store ptr [[TMP26]], ptr [[TMP31]], align 8 399 // CHECK1-NEXT: [[TMP32:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 3 400 // CHECK1-NEXT: store ptr [[TMP27]], ptr [[TMP32]], align 8 401 // CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 4 402 // CHECK1-NEXT: store ptr @.offload_sizes.8, ptr [[TMP33]], align 8 403 // CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 5 404 // CHECK1-NEXT: store ptr @.offload_maptypes.9, ptr [[TMP34]], align 8 405 // CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 6 406 // CHECK1-NEXT: store ptr null, ptr [[TMP35]], align 8 407 // CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 7 408 // CHECK1-NEXT: store ptr null, ptr [[TMP36]], align 8 409 // CHECK1-NEXT: [[TMP37:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 8 410 // CHECK1-NEXT: store i64 0, ptr [[TMP37]], align 8 411 // CHECK1-NEXT: [[TMP38:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 [[TMP28]], i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108.region_id, ptr [[KERNEL_ARGS6]]) 412 // CHECK1-NEXT: [[TMP39:%.*]] = icmp ne i32 [[TMP38]], 0 413 // CHECK1-NEXT: br i1 [[TMP39]], label [[OMP_OFFLOAD_FAILED7:%.*]], label [[OMP_OFFLOAD_CONT8:%.*]] 414 // CHECK1: omp_offload.failed7: 415 // CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108(i64 [[TMP22]]) #[[ATTR2]] 416 // CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT8]] 417 // CHECK1: omp_offload.cont8: 418 // CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[N_ADDR]], align 4 419 // CHECK1-NEXT: [[ADD9:%.*]] = add nsw i32 [[TMP40]], 1 420 // CHECK1-NEXT: ret i32 [[ADD9]] 421 // 422 // 423 // CHECK1-LABEL: define {{[^@]+}}@_Z9ftemplateIiET_i 424 // CHECK1-SAME: (i32 noundef signext [[N:%.*]]) #[[ATTR0]] comdat { 425 // CHECK1-NEXT: entry: 426 // CHECK1-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 427 // CHECK1-NEXT: [[A:%.*]] = alloca i32, align 4 428 // CHECK1-NEXT: [[B:%.*]] = alloca i16, align 2 429 // CHECK1-NEXT: [[DOTCAPTURE_EXPR_:%.*]] = alloca i16, align 2 430 // CHECK1-NEXT: [[A_CASTED:%.*]] = alloca i64, align 8 431 // CHECK1-NEXT: [[B_CASTED:%.*]] = alloca i64, align 8 432 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i64, align 8 433 // CHECK1-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [3 x ptr], align 8 434 // CHECK1-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [3 x ptr], align 8 435 // CHECK1-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [3 x ptr], align 8 436 // CHECK1-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 437 // CHECK1-NEXT: store i32 0, ptr [[A]], align 4 438 // CHECK1-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8 439 // CHECK1-NEXT: [[TMP0:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 440 // CHECK1-NEXT: store i32 1, ptr [[TMP0]], align 4 441 // CHECK1-NEXT: [[TMP1:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1 442 // CHECK1-NEXT: store i32 0, ptr [[TMP1]], align 4 443 // CHECK1-NEXT: [[TMP2:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2 444 // CHECK1-NEXT: store ptr null, ptr [[TMP2]], align 8 445 // CHECK1-NEXT: [[TMP3:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3 446 // CHECK1-NEXT: store ptr null, ptr [[TMP3]], align 8 447 // CHECK1-NEXT: [[TMP4:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4 448 // CHECK1-NEXT: store ptr null, ptr [[TMP4]], align 8 449 // CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5 450 // CHECK1-NEXT: store ptr null, ptr [[TMP5]], align 8 451 // CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6 452 // CHECK1-NEXT: store ptr null, ptr [[TMP6]], align 8 453 // CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7 454 // CHECK1-NEXT: store ptr null, ptr [[TMP7]], align 8 455 // CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8 456 // CHECK1-NEXT: store i64 0, ptr [[TMP8]], align 8 457 // CHECK1-NEXT: [[TMP9:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 20, i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88.region_id, ptr [[KERNEL_ARGS]]) 458 // CHECK1-NEXT: [[TMP10:%.*]] = icmp ne i32 [[TMP9]], 0 459 // CHECK1-NEXT: br i1 [[TMP10]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]] 460 // CHECK1: omp_offload.failed: 461 // CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88() #[[ATTR2]] 462 // CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT]] 463 // CHECK1: omp_offload.cont: 464 // CHECK1-NEXT: store i16 1, ptr [[B]], align 2 465 // CHECK1-NEXT: [[TMP11:%.*]] = load i16, ptr [[B]], align 2 466 // CHECK1-NEXT: store i16 [[TMP11]], ptr [[DOTCAPTURE_EXPR_]], align 2 467 // CHECK1-NEXT: [[TMP12:%.*]] = load i32, ptr [[A]], align 4 468 // CHECK1-NEXT: store i32 [[TMP12]], ptr [[A_CASTED]], align 4 469 // CHECK1-NEXT: [[TMP13:%.*]] = load i64, ptr [[A_CASTED]], align 8 470 // CHECK1-NEXT: [[TMP14:%.*]] = load i16, ptr [[B]], align 2 471 // CHECK1-NEXT: store i16 [[TMP14]], ptr [[B_CASTED]], align 2 472 // CHECK1-NEXT: [[TMP15:%.*]] = load i64, ptr [[B_CASTED]], align 8 473 // CHECK1-NEXT: [[TMP16:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR_]], align 2 474 // CHECK1-NEXT: store i16 [[TMP16]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 2 475 // CHECK1-NEXT: [[TMP17:%.*]] = load i64, ptr [[DOTCAPTURE_EXPR__CASTED]], align 8 476 // CHECK1-NEXT: [[TMP18:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 477 // CHECK1-NEXT: store i64 [[TMP13]], ptr [[TMP18]], align 8 478 // CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 479 // CHECK1-NEXT: store i64 [[TMP13]], ptr [[TMP19]], align 8 480 // CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 0 481 // CHECK1-NEXT: store ptr null, ptr [[TMP20]], align 8 482 // CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 1 483 // CHECK1-NEXT: store i64 [[TMP15]], ptr [[TMP21]], align 8 484 // CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 1 485 // CHECK1-NEXT: store i64 [[TMP15]], ptr [[TMP22]], align 8 486 // CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 1 487 // CHECK1-NEXT: store ptr null, ptr [[TMP23]], align 8 488 // CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 2 489 // CHECK1-NEXT: store i64 [[TMP17]], ptr [[TMP24]], align 8 490 // CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 2 491 // CHECK1-NEXT: store i64 [[TMP17]], ptr [[TMP25]], align 8 492 // CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 2 493 // CHECK1-NEXT: store ptr null, ptr [[TMP26]], align 8 494 // CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 495 // CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 496 // CHECK1-NEXT: [[TMP29:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR_]], align 2 497 // CHECK1-NEXT: [[TMP30:%.*]] = sext i16 [[TMP29]] to i32 498 // CHECK1-NEXT: [[KERNEL_ARGS1:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS]], align 8 499 // CHECK1-NEXT: [[TMP31:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 0 500 // CHECK1-NEXT: store i32 1, ptr [[TMP31]], align 4 501 // CHECK1-NEXT: [[TMP32:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 1 502 // CHECK1-NEXT: store i32 3, ptr [[TMP32]], align 4 503 // CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 2 504 // CHECK1-NEXT: store ptr [[TMP27]], ptr [[TMP33]], align 8 505 // CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 3 506 // CHECK1-NEXT: store ptr [[TMP28]], ptr [[TMP34]], align 8 507 // CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 4 508 // CHECK1-NEXT: store ptr @.offload_sizes.12, ptr [[TMP35]], align 8 509 // CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 5 510 // CHECK1-NEXT: store ptr @.offload_maptypes.13, ptr [[TMP36]], align 8 511 // CHECK1-NEXT: [[TMP37:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 6 512 // CHECK1-NEXT: store ptr null, ptr [[TMP37]], align 8 513 // CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 7 514 // CHECK1-NEXT: store ptr null, ptr [[TMP38]], align 8 515 // CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 8 516 // CHECK1-NEXT: store i64 0, ptr [[TMP39]], align 8 517 // CHECK1-NEXT: [[TMP40:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 [[TMP30]], i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93.region_id, ptr [[KERNEL_ARGS1]]) 518 // CHECK1-NEXT: [[TMP41:%.*]] = icmp ne i32 [[TMP40]], 0 519 // CHECK1-NEXT: br i1 [[TMP41]], label [[OMP_OFFLOAD_FAILED2:%.*]], label [[OMP_OFFLOAD_CONT3:%.*]] 520 // CHECK1: omp_offload.failed2: 521 // CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93(i64 [[TMP13]], i64 [[TMP15]], i64 [[TMP17]]) #[[ATTR2]] 522 // CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT3]] 523 // CHECK1: omp_offload.cont3: 524 // CHECK1-NEXT: [[TMP42:%.*]] = load i32, ptr [[A]], align 4 525 // CHECK1-NEXT: ret i32 [[TMP42]] 526 // 527 // 528 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121 529 // CHECK1-SAME: (ptr noundef [[THIS:%.*]], i64 noundef [[B:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1:[0-9]+]] { 530 // CHECK1-NEXT: entry: 531 // CHECK1-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 532 // CHECK1-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 533 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 534 // CHECK1-NEXT: [[B_CASTED:%.*]] = alloca i64, align 8 535 // CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 536 // CHECK1-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 537 // CHECK1-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 538 // CHECK1-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 539 // CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 540 // CHECK1-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 541 // CHECK1-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP2]], i32 0) 542 // CHECK1-NEXT: [[TMP3:%.*]] = load i32, ptr [[B_ADDR]], align 4 543 // CHECK1-NEXT: store i32 [[TMP3]], ptr [[B_CASTED]], align 4 544 // CHECK1-NEXT: [[TMP4:%.*]] = load i64, ptr [[B_CASTED]], align 8 545 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @.omp_outlined., ptr [[TMP1]], i64 [[TMP4]]) 546 // CHECK1-NEXT: ret void 547 // 548 // 549 // CHECK1-LABEL: define {{[^@]+}}@.omp_outlined. 550 // CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]], i64 noundef [[B:%.*]]) #[[ATTR1]] { 551 // CHECK1-NEXT: entry: 552 // CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 553 // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 554 // CHECK1-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 555 // CHECK1-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 556 // CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 557 // CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 558 // CHECK1-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 559 // CHECK1-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 560 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 561 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[B_ADDR]], align 4 562 // CHECK1-NEXT: [[CONV:%.*]] = sitofp i32 [[TMP1]] to double 563 // CHECK1-NEXT: [[ADD:%.*]] = fadd double [[CONV]], 1.500000e+00 564 // CHECK1-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 565 // CHECK1-NEXT: store double [[ADD]], ptr [[A]], align 8 566 // CHECK1-NEXT: ret void 567 // 568 // 569 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126 570 // CHECK1-SAME: (ptr noundef [[THIS:%.*]]) #[[ATTR3:[0-9]+]] { 571 // CHECK1-NEXT: entry: 572 // CHECK1-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 573 // CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 574 // CHECK1-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 575 // CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 576 // CHECK1-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 1024, i32 0) 577 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 1, ptr @.omp_outlined..1, ptr [[TMP1]]) 578 // CHECK1-NEXT: ret void 579 // 580 // 581 // CHECK1-LABEL: define {{[^@]+}}@.omp_outlined..1 582 // CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]]) #[[ATTR1]] { 583 // CHECK1-NEXT: entry: 584 // CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 585 // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 586 // CHECK1-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 587 // CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 588 // CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 589 // CHECK1-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 590 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 591 // CHECK1-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 592 // CHECK1-NEXT: store double 2.500000e+00, ptr [[A]], align 8 593 // CHECK1-NEXT: ret void 594 // 595 // 596 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104 597 // CHECK1-SAME: (i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] { 598 // CHECK1-NEXT: entry: 599 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 600 // CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 601 // CHECK1-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 602 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 603 // CHECK1-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP1]], i32 0) 604 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @.omp_outlined..4) 605 // CHECK1-NEXT: ret void 606 // 607 // 608 // CHECK1-LABEL: define {{[^@]+}}@.omp_outlined..4 609 // CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] { 610 // CHECK1-NEXT: entry: 611 // CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 612 // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 613 // CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 614 // CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 615 // CHECK1-NEXT: ret void 616 // 617 // 618 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108 619 // CHECK1-SAME: (i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] { 620 // CHECK1-NEXT: entry: 621 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 622 // CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 623 // CHECK1-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 624 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 625 // CHECK1-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP1]], i32 0) 626 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @.omp_outlined..7) 627 // CHECK1-NEXT: ret void 628 // 629 // 630 // CHECK1-LABEL: define {{[^@]+}}@.omp_outlined..7 631 // CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] { 632 // CHECK1-NEXT: entry: 633 // CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 634 // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 635 // CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 636 // CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 637 // CHECK1-NEXT: ret void 638 // 639 // 640 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88 641 // CHECK1-SAME: () #[[ATTR4:[0-9]+]] { 642 // CHECK1-NEXT: entry: 643 // CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 644 // CHECK1-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 20, i32 0) 645 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @.omp_outlined..10) 646 // CHECK1-NEXT: ret void 647 // 648 // 649 // CHECK1-LABEL: define {{[^@]+}}@.omp_outlined..10 650 // CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] { 651 // CHECK1-NEXT: entry: 652 // CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 653 // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 654 // CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 655 // CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 656 // CHECK1-NEXT: ret void 657 // 658 // 659 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93 660 // CHECK1-SAME: (i64 noundef [[A:%.*]], i64 noundef [[B:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] { 661 // CHECK1-NEXT: entry: 662 // CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8 663 // CHECK1-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 664 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 665 // CHECK1-NEXT: [[A_CASTED:%.*]] = alloca i64, align 8 666 // CHECK1-NEXT: [[B_CASTED:%.*]] = alloca i64, align 8 667 // CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 668 // CHECK1-NEXT: store i64 [[A]], ptr [[A_ADDR]], align 8 669 // CHECK1-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 670 // CHECK1-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 671 // CHECK1-NEXT: [[TMP1:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR__ADDR]], align 2 672 // CHECK1-NEXT: [[TMP2:%.*]] = sext i16 [[TMP1]] to i32 673 // CHECK1-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP2]], i32 0) 674 // CHECK1-NEXT: [[TMP3:%.*]] = load i32, ptr [[A_ADDR]], align 4 675 // CHECK1-NEXT: store i32 [[TMP3]], ptr [[A_CASTED]], align 4 676 // CHECK1-NEXT: [[TMP4:%.*]] = load i64, ptr [[A_CASTED]], align 8 677 // CHECK1-NEXT: [[TMP5:%.*]] = load i16, ptr [[B_ADDR]], align 2 678 // CHECK1-NEXT: store i16 [[TMP5]], ptr [[B_CASTED]], align 2 679 // CHECK1-NEXT: [[TMP6:%.*]] = load i64, ptr [[B_CASTED]], align 8 680 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @.omp_outlined..11, i64 [[TMP4]], i64 [[TMP6]]) 681 // CHECK1-NEXT: ret void 682 // 683 // 684 // CHECK1-LABEL: define {{[^@]+}}@.omp_outlined..11 685 // CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i64 noundef [[A:%.*]], i64 noundef [[B:%.*]]) #[[ATTR1]] { 686 // CHECK1-NEXT: entry: 687 // CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 688 // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 689 // CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8 690 // CHECK1-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 691 // CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 692 // CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 693 // CHECK1-NEXT: store i64 [[A]], ptr [[A_ADDR]], align 8 694 // CHECK1-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 695 // CHECK1-NEXT: [[TMP0:%.*]] = load i16, ptr [[B_ADDR]], align 2 696 // CHECK1-NEXT: [[CONV:%.*]] = sext i16 [[TMP0]] to i32 697 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[A_ADDR]], align 4 698 // CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP1]], [[CONV]] 699 // CHECK1-NEXT: store i32 [[ADD]], ptr [[A_ADDR]], align 4 700 // CHECK1-NEXT: ret void 701 // 702 // 703 // CHECK1-LABEL: define {{[^@]+}}@.omp_offloading.requires_reg 704 // CHECK1-SAME: () #[[ATTR5:[0-9]+]] { 705 // CHECK1-NEXT: entry: 706 // CHECK1-NEXT: call void @__tgt_register_requires(i64 1) 707 // CHECK1-NEXT: ret void 708 // 709 // 710 // CHECK3-LABEL: define {{[^@]+}}@_Z3bari 711 // CHECK3-SAME: (i32 noundef [[N:%.*]]) #[[ATTR0:[0-9]+]] { 712 // CHECK3-NEXT: entry: 713 // CHECK3-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 714 // CHECK3-NEXT: [[A:%.*]] = alloca i32, align 4 715 // CHECK3-NEXT: [[S:%.*]] = alloca [[STRUCT_S1:%.*]], align 4 716 // CHECK3-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 717 // CHECK3-NEXT: store i32 0, ptr [[A]], align 4 718 // CHECK3-NEXT: [[TMP0:%.*]] = load i32, ptr [[N_ADDR]], align 4 719 // CHECK3-NEXT: [[CALL:%.*]] = call noundef i32 @_ZN2S12r1Ei(ptr noundef nonnull align 4 dereferenceable(8) [[S]], i32 noundef [[TMP0]]) 720 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[A]], align 4 721 // CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP1]], [[CALL]] 722 // CHECK3-NEXT: store i32 [[ADD]], ptr [[A]], align 4 723 // CHECK3-NEXT: [[TMP2:%.*]] = load i32, ptr [[N_ADDR]], align 4 724 // CHECK3-NEXT: [[CALL1:%.*]] = call noundef i32 @_ZL7fstatici(i32 noundef [[TMP2]]) 725 // CHECK3-NEXT: [[TMP3:%.*]] = load i32, ptr [[A]], align 4 726 // CHECK3-NEXT: [[ADD2:%.*]] = add nsw i32 [[TMP3]], [[CALL1]] 727 // CHECK3-NEXT: store i32 [[ADD2]], ptr [[A]], align 4 728 // CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[N_ADDR]], align 4 729 // CHECK3-NEXT: [[CALL3:%.*]] = call noundef i32 @_Z9ftemplateIiET_i(i32 noundef [[TMP4]]) 730 // CHECK3-NEXT: [[TMP5:%.*]] = load i32, ptr [[A]], align 4 731 // CHECK3-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP5]], [[CALL3]] 732 // CHECK3-NEXT: store i32 [[ADD4]], ptr [[A]], align 4 733 // CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[A]], align 4 734 // CHECK3-NEXT: ret i32 [[TMP6]] 735 // 736 // 737 // CHECK3-LABEL: define {{[^@]+}}@_ZN2S12r1Ei 738 // CHECK3-SAME: (ptr noundef nonnull align 4 dereferenceable(8) [[THIS:%.*]], i32 noundef [[N:%.*]]) #[[ATTR0]] comdat align 2 { 739 // CHECK3-NEXT: entry: 740 // CHECK3-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 741 // CHECK3-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 742 // CHECK3-NEXT: [[B:%.*]] = alloca i32, align 4 743 // CHECK3-NEXT: [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4 744 // CHECK3-NEXT: [[B_CASTED:%.*]] = alloca i32, align 4 745 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i32, align 4 746 // CHECK3-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [3 x ptr], align 4 747 // CHECK3-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [3 x ptr], align 4 748 // CHECK3-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [3 x ptr], align 4 749 // CHECK3-NEXT: [[DOTOFFLOAD_BASEPTRS3:%.*]] = alloca [1 x ptr], align 4 750 // CHECK3-NEXT: [[DOTOFFLOAD_PTRS4:%.*]] = alloca [1 x ptr], align 4 751 // CHECK3-NEXT: [[DOTOFFLOAD_MAPPERS5:%.*]] = alloca [1 x ptr], align 4 752 // CHECK3-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 753 // CHECK3-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 754 // CHECK3-NEXT: [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 755 // CHECK3-NEXT: store i32 1, ptr [[B]], align 4 756 // CHECK3-NEXT: [[TMP0:%.*]] = load i32, ptr [[N_ADDR]], align 4 757 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[B]], align 4 758 // CHECK3-NEXT: [[SUB:%.*]] = sub nsw i32 [[TMP0]], [[TMP1]] 759 // CHECK3-NEXT: store i32 [[SUB]], ptr [[DOTCAPTURE_EXPR_]], align 4 760 // CHECK3-NEXT: [[TMP2:%.*]] = load i32, ptr [[B]], align 4 761 // CHECK3-NEXT: store i32 [[TMP2]], ptr [[B_CASTED]], align 4 762 // CHECK3-NEXT: [[TMP3:%.*]] = load i32, ptr [[B_CASTED]], align 4 763 // CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 764 // CHECK3-NEXT: store i32 [[TMP4]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 765 // CHECK3-NEXT: [[TMP5:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 766 // CHECK3-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[THIS1]], i32 0, i32 0 767 // CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 768 // CHECK3-NEXT: store ptr [[THIS1]], ptr [[TMP6]], align 4 769 // CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 770 // CHECK3-NEXT: store ptr [[A]], ptr [[TMP7]], align 4 771 // CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 0 772 // CHECK3-NEXT: store ptr null, ptr [[TMP8]], align 4 773 // CHECK3-NEXT: [[TMP9:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 1 774 // CHECK3-NEXT: store i32 [[TMP3]], ptr [[TMP9]], align 4 775 // CHECK3-NEXT: [[TMP10:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 1 776 // CHECK3-NEXT: store i32 [[TMP3]], ptr [[TMP10]], align 4 777 // CHECK3-NEXT: [[TMP11:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 1 778 // CHECK3-NEXT: store ptr null, ptr [[TMP11]], align 4 779 // CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 2 780 // CHECK3-NEXT: store i32 [[TMP5]], ptr [[TMP12]], align 4 781 // CHECK3-NEXT: [[TMP13:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 2 782 // CHECK3-NEXT: store i32 [[TMP5]], ptr [[TMP13]], align 4 783 // CHECK3-NEXT: [[TMP14:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 2 784 // CHECK3-NEXT: store ptr null, ptr [[TMP14]], align 4 785 // CHECK3-NEXT: [[TMP15:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 786 // CHECK3-NEXT: [[TMP16:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 787 // CHECK3-NEXT: [[TMP17:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 788 // CHECK3-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8 789 // CHECK3-NEXT: [[TMP18:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 790 // CHECK3-NEXT: store i32 1, ptr [[TMP18]], align 4 791 // CHECK3-NEXT: [[TMP19:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1 792 // CHECK3-NEXT: store i32 3, ptr [[TMP19]], align 4 793 // CHECK3-NEXT: [[TMP20:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2 794 // CHECK3-NEXT: store ptr [[TMP15]], ptr [[TMP20]], align 4 795 // CHECK3-NEXT: [[TMP21:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3 796 // CHECK3-NEXT: store ptr [[TMP16]], ptr [[TMP21]], align 4 797 // CHECK3-NEXT: [[TMP22:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4 798 // CHECK3-NEXT: store ptr @.offload_sizes, ptr [[TMP22]], align 4 799 // CHECK3-NEXT: [[TMP23:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5 800 // CHECK3-NEXT: store ptr @.offload_maptypes, ptr [[TMP23]], align 4 801 // CHECK3-NEXT: [[TMP24:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6 802 // CHECK3-NEXT: store ptr null, ptr [[TMP24]], align 4 803 // CHECK3-NEXT: [[TMP25:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7 804 // CHECK3-NEXT: store ptr null, ptr [[TMP25]], align 4 805 // CHECK3-NEXT: [[TMP26:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8 806 // CHECK3-NEXT: store i64 0, ptr [[TMP26]], align 8 807 // CHECK3-NEXT: [[TMP27:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1:[0-9]+]], i64 -1, i32 [[TMP17]], i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121.region_id, ptr [[KERNEL_ARGS]]) 808 // CHECK3-NEXT: [[TMP28:%.*]] = icmp ne i32 [[TMP27]], 0 809 // CHECK3-NEXT: br i1 [[TMP28]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]] 810 // CHECK3: omp_offload.failed: 811 // CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121(ptr [[THIS1]], i32 [[TMP3]], i32 [[TMP5]]) #[[ATTR2:[0-9]+]] 812 // CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT]] 813 // CHECK3: omp_offload.cont: 814 // CHECK3-NEXT: [[A2:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[THIS1]], i32 0, i32 0 815 // CHECK3-NEXT: [[TMP29:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS3]], i32 0, i32 0 816 // CHECK3-NEXT: store ptr [[THIS1]], ptr [[TMP29]], align 4 817 // CHECK3-NEXT: [[TMP30:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS4]], i32 0, i32 0 818 // CHECK3-NEXT: store ptr [[A2]], ptr [[TMP30]], align 4 819 // CHECK3-NEXT: [[TMP31:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS5]], i32 0, i32 0 820 // CHECK3-NEXT: store ptr null, ptr [[TMP31]], align 4 821 // CHECK3-NEXT: [[TMP32:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS3]], i32 0, i32 0 822 // CHECK3-NEXT: [[TMP33:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS4]], i32 0, i32 0 823 // CHECK3-NEXT: [[KERNEL_ARGS6:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS]], align 8 824 // CHECK3-NEXT: [[TMP34:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 0 825 // CHECK3-NEXT: store i32 1, ptr [[TMP34]], align 4 826 // CHECK3-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 1 827 // CHECK3-NEXT: store i32 1, ptr [[TMP35]], align 4 828 // CHECK3-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 2 829 // CHECK3-NEXT: store ptr [[TMP32]], ptr [[TMP36]], align 4 830 // CHECK3-NEXT: [[TMP37:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 3 831 // CHECK3-NEXT: store ptr [[TMP33]], ptr [[TMP37]], align 4 832 // CHECK3-NEXT: [[TMP38:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 4 833 // CHECK3-NEXT: store ptr @.offload_sizes.2, ptr [[TMP38]], align 4 834 // CHECK3-NEXT: [[TMP39:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 5 835 // CHECK3-NEXT: store ptr @.offload_maptypes.3, ptr [[TMP39]], align 4 836 // CHECK3-NEXT: [[TMP40:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 6 837 // CHECK3-NEXT: store ptr null, ptr [[TMP40]], align 4 838 // CHECK3-NEXT: [[TMP41:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 7 839 // CHECK3-NEXT: store ptr null, ptr [[TMP41]], align 4 840 // CHECK3-NEXT: [[TMP42:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 8 841 // CHECK3-NEXT: store i64 0, ptr [[TMP42]], align 8 842 // CHECK3-NEXT: [[TMP43:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 1024, i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126.region_id, ptr [[KERNEL_ARGS6]]) 843 // CHECK3-NEXT: [[TMP44:%.*]] = icmp ne i32 [[TMP43]], 0 844 // CHECK3-NEXT: br i1 [[TMP44]], label [[OMP_OFFLOAD_FAILED7:%.*]], label [[OMP_OFFLOAD_CONT8:%.*]] 845 // CHECK3: omp_offload.failed7: 846 // CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126(ptr [[THIS1]]) #[[ATTR2]] 847 // CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT8]] 848 // CHECK3: omp_offload.cont8: 849 // CHECK3-NEXT: [[A9:%.*]] = getelementptr inbounds [[STRUCT_S1]], ptr [[THIS1]], i32 0, i32 0 850 // CHECK3-NEXT: [[TMP45:%.*]] = load double, ptr [[A9]], align 4 851 // CHECK3-NEXT: [[CONV:%.*]] = fptosi double [[TMP45]] to i32 852 // CHECK3-NEXT: ret i32 [[CONV]] 853 // 854 // 855 // CHECK3-LABEL: define {{[^@]+}}@_ZL7fstatici 856 // CHECK3-SAME: (i32 noundef [[N:%.*]]) #[[ATTR0]] { 857 // CHECK3-NEXT: entry: 858 // CHECK3-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 859 // CHECK3-NEXT: [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4 860 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i32, align 4 861 // CHECK3-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [1 x ptr], align 4 862 // CHECK3-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [1 x ptr], align 4 863 // CHECK3-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [1 x ptr], align 4 864 // CHECK3-NEXT: [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4 865 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__CASTED2:%.*]] = alloca i32, align 4 866 // CHECK3-NEXT: [[DOTOFFLOAD_BASEPTRS3:%.*]] = alloca [1 x ptr], align 4 867 // CHECK3-NEXT: [[DOTOFFLOAD_PTRS4:%.*]] = alloca [1 x ptr], align 4 868 // CHECK3-NEXT: [[DOTOFFLOAD_MAPPERS5:%.*]] = alloca [1 x ptr], align 4 869 // CHECK3-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 870 // CHECK3-NEXT: [[TMP0:%.*]] = load i32, ptr [[N_ADDR]], align 4 871 // CHECK3-NEXT: store i32 [[TMP0]], ptr [[DOTCAPTURE_EXPR_]], align 4 872 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 873 // CHECK3-NEXT: store i32 [[TMP1]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 874 // CHECK3-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 875 // CHECK3-NEXT: [[TMP3:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 876 // CHECK3-NEXT: store i32 [[TMP2]], ptr [[TMP3]], align 4 877 // CHECK3-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 878 // CHECK3-NEXT: store i32 [[TMP2]], ptr [[TMP4]], align 4 879 // CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 0 880 // CHECK3-NEXT: store ptr null, ptr [[TMP5]], align 4 881 // CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 882 // CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 883 // CHECK3-NEXT: [[TMP8:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 884 // CHECK3-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8 885 // CHECK3-NEXT: [[TMP9:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 886 // CHECK3-NEXT: store i32 1, ptr [[TMP9]], align 4 887 // CHECK3-NEXT: [[TMP10:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1 888 // CHECK3-NEXT: store i32 1, ptr [[TMP10]], align 4 889 // CHECK3-NEXT: [[TMP11:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2 890 // CHECK3-NEXT: store ptr [[TMP6]], ptr [[TMP11]], align 4 891 // CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3 892 // CHECK3-NEXT: store ptr [[TMP7]], ptr [[TMP12]], align 4 893 // CHECK3-NEXT: [[TMP13:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4 894 // CHECK3-NEXT: store ptr @.offload_sizes.5, ptr [[TMP13]], align 4 895 // CHECK3-NEXT: [[TMP14:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5 896 // CHECK3-NEXT: store ptr @.offload_maptypes.6, ptr [[TMP14]], align 4 897 // CHECK3-NEXT: [[TMP15:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6 898 // CHECK3-NEXT: store ptr null, ptr [[TMP15]], align 4 899 // CHECK3-NEXT: [[TMP16:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7 900 // CHECK3-NEXT: store ptr null, ptr [[TMP16]], align 4 901 // CHECK3-NEXT: [[TMP17:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8 902 // CHECK3-NEXT: store i64 0, ptr [[TMP17]], align 8 903 // CHECK3-NEXT: [[TMP18:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 [[TMP8]], i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104.region_id, ptr [[KERNEL_ARGS]]) 904 // CHECK3-NEXT: [[TMP19:%.*]] = icmp ne i32 [[TMP18]], 0 905 // CHECK3-NEXT: br i1 [[TMP19]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]] 906 // CHECK3: omp_offload.failed: 907 // CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104(i32 [[TMP2]]) #[[ATTR2]] 908 // CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT]] 909 // CHECK3: omp_offload.cont: 910 // CHECK3-NEXT: [[TMP20:%.*]] = load i32, ptr [[N_ADDR]], align 4 911 // CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 32, [[TMP20]] 912 // CHECK3-NEXT: store i32 [[ADD]], ptr [[DOTCAPTURE_EXPR_1]], align 4 913 // CHECK3-NEXT: [[TMP21:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4 914 // CHECK3-NEXT: store i32 [[TMP21]], ptr [[DOTCAPTURE_EXPR__CASTED2]], align 4 915 // CHECK3-NEXT: [[TMP22:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__CASTED2]], align 4 916 // CHECK3-NEXT: [[TMP23:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS3]], i32 0, i32 0 917 // CHECK3-NEXT: store i32 [[TMP22]], ptr [[TMP23]], align 4 918 // CHECK3-NEXT: [[TMP24:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS4]], i32 0, i32 0 919 // CHECK3-NEXT: store i32 [[TMP22]], ptr [[TMP24]], align 4 920 // CHECK3-NEXT: [[TMP25:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS5]], i32 0, i32 0 921 // CHECK3-NEXT: store ptr null, ptr [[TMP25]], align 4 922 // CHECK3-NEXT: [[TMP26:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS3]], i32 0, i32 0 923 // CHECK3-NEXT: [[TMP27:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS4]], i32 0, i32 0 924 // CHECK3-NEXT: [[TMP28:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4 925 // CHECK3-NEXT: [[KERNEL_ARGS6:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS]], align 8 926 // CHECK3-NEXT: [[TMP29:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 0 927 // CHECK3-NEXT: store i32 1, ptr [[TMP29]], align 4 928 // CHECK3-NEXT: [[TMP30:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 1 929 // CHECK3-NEXT: store i32 1, ptr [[TMP30]], align 4 930 // CHECK3-NEXT: [[TMP31:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 2 931 // CHECK3-NEXT: store ptr [[TMP26]], ptr [[TMP31]], align 4 932 // CHECK3-NEXT: [[TMP32:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 3 933 // CHECK3-NEXT: store ptr [[TMP27]], ptr [[TMP32]], align 4 934 // CHECK3-NEXT: [[TMP33:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 4 935 // CHECK3-NEXT: store ptr @.offload_sizes.8, ptr [[TMP33]], align 4 936 // CHECK3-NEXT: [[TMP34:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 5 937 // CHECK3-NEXT: store ptr @.offload_maptypes.9, ptr [[TMP34]], align 4 938 // CHECK3-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 6 939 // CHECK3-NEXT: store ptr null, ptr [[TMP35]], align 4 940 // CHECK3-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 7 941 // CHECK3-NEXT: store ptr null, ptr [[TMP36]], align 4 942 // CHECK3-NEXT: [[TMP37:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 8 943 // CHECK3-NEXT: store i64 0, ptr [[TMP37]], align 8 944 // CHECK3-NEXT: [[TMP38:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 [[TMP28]], i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108.region_id, ptr [[KERNEL_ARGS6]]) 945 // CHECK3-NEXT: [[TMP39:%.*]] = icmp ne i32 [[TMP38]], 0 946 // CHECK3-NEXT: br i1 [[TMP39]], label [[OMP_OFFLOAD_FAILED7:%.*]], label [[OMP_OFFLOAD_CONT8:%.*]] 947 // CHECK3: omp_offload.failed7: 948 // CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108(i32 [[TMP22]]) #[[ATTR2]] 949 // CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT8]] 950 // CHECK3: omp_offload.cont8: 951 // CHECK3-NEXT: [[TMP40:%.*]] = load i32, ptr [[N_ADDR]], align 4 952 // CHECK3-NEXT: [[ADD9:%.*]] = add nsw i32 [[TMP40]], 1 953 // CHECK3-NEXT: ret i32 [[ADD9]] 954 // 955 // 956 // CHECK3-LABEL: define {{[^@]+}}@_Z9ftemplateIiET_i 957 // CHECK3-SAME: (i32 noundef [[N:%.*]]) #[[ATTR0]] comdat { 958 // CHECK3-NEXT: entry: 959 // CHECK3-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 960 // CHECK3-NEXT: [[A:%.*]] = alloca i32, align 4 961 // CHECK3-NEXT: [[B:%.*]] = alloca i16, align 2 962 // CHECK3-NEXT: [[DOTCAPTURE_EXPR_:%.*]] = alloca i16, align 2 963 // CHECK3-NEXT: [[A_CASTED:%.*]] = alloca i32, align 4 964 // CHECK3-NEXT: [[B_CASTED:%.*]] = alloca i32, align 4 965 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i32, align 4 966 // CHECK3-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [3 x ptr], align 4 967 // CHECK3-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [3 x ptr], align 4 968 // CHECK3-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [3 x ptr], align 4 969 // CHECK3-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 970 // CHECK3-NEXT: store i32 0, ptr [[A]], align 4 971 // CHECK3-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8 972 // CHECK3-NEXT: [[TMP0:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 973 // CHECK3-NEXT: store i32 1, ptr [[TMP0]], align 4 974 // CHECK3-NEXT: [[TMP1:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1 975 // CHECK3-NEXT: store i32 0, ptr [[TMP1]], align 4 976 // CHECK3-NEXT: [[TMP2:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2 977 // CHECK3-NEXT: store ptr null, ptr [[TMP2]], align 4 978 // CHECK3-NEXT: [[TMP3:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3 979 // CHECK3-NEXT: store ptr null, ptr [[TMP3]], align 4 980 // CHECK3-NEXT: [[TMP4:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4 981 // CHECK3-NEXT: store ptr null, ptr [[TMP4]], align 4 982 // CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5 983 // CHECK3-NEXT: store ptr null, ptr [[TMP5]], align 4 984 // CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6 985 // CHECK3-NEXT: store ptr null, ptr [[TMP6]], align 4 986 // CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7 987 // CHECK3-NEXT: store ptr null, ptr [[TMP7]], align 4 988 // CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8 989 // CHECK3-NEXT: store i64 0, ptr [[TMP8]], align 8 990 // CHECK3-NEXT: [[TMP9:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 20, i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88.region_id, ptr [[KERNEL_ARGS]]) 991 // CHECK3-NEXT: [[TMP10:%.*]] = icmp ne i32 [[TMP9]], 0 992 // CHECK3-NEXT: br i1 [[TMP10]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]] 993 // CHECK3: omp_offload.failed: 994 // CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88() #[[ATTR2]] 995 // CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT]] 996 // CHECK3: omp_offload.cont: 997 // CHECK3-NEXT: store i16 1, ptr [[B]], align 2 998 // CHECK3-NEXT: [[TMP11:%.*]] = load i16, ptr [[B]], align 2 999 // CHECK3-NEXT: store i16 [[TMP11]], ptr [[DOTCAPTURE_EXPR_]], align 2 1000 // CHECK3-NEXT: [[TMP12:%.*]] = load i32, ptr [[A]], align 4 1001 // CHECK3-NEXT: store i32 [[TMP12]], ptr [[A_CASTED]], align 4 1002 // CHECK3-NEXT: [[TMP13:%.*]] = load i32, ptr [[A_CASTED]], align 4 1003 // CHECK3-NEXT: [[TMP14:%.*]] = load i16, ptr [[B]], align 2 1004 // CHECK3-NEXT: store i16 [[TMP14]], ptr [[B_CASTED]], align 2 1005 // CHECK3-NEXT: [[TMP15:%.*]] = load i32, ptr [[B_CASTED]], align 4 1006 // CHECK3-NEXT: [[TMP16:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR_]], align 2 1007 // CHECK3-NEXT: store i16 [[TMP16]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 2 1008 // CHECK3-NEXT: [[TMP17:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 1009 // CHECK3-NEXT: [[TMP18:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 1010 // CHECK3-NEXT: store i32 [[TMP13]], ptr [[TMP18]], align 4 1011 // CHECK3-NEXT: [[TMP19:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 1012 // CHECK3-NEXT: store i32 [[TMP13]], ptr [[TMP19]], align 4 1013 // CHECK3-NEXT: [[TMP20:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 0 1014 // CHECK3-NEXT: store ptr null, ptr [[TMP20]], align 4 1015 // CHECK3-NEXT: [[TMP21:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 1 1016 // CHECK3-NEXT: store i32 [[TMP15]], ptr [[TMP21]], align 4 1017 // CHECK3-NEXT: [[TMP22:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 1 1018 // CHECK3-NEXT: store i32 [[TMP15]], ptr [[TMP22]], align 4 1019 // CHECK3-NEXT: [[TMP23:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 1 1020 // CHECK3-NEXT: store ptr null, ptr [[TMP23]], align 4 1021 // CHECK3-NEXT: [[TMP24:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 2 1022 // CHECK3-NEXT: store i32 [[TMP17]], ptr [[TMP24]], align 4 1023 // CHECK3-NEXT: [[TMP25:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 2 1024 // CHECK3-NEXT: store i32 [[TMP17]], ptr [[TMP25]], align 4 1025 // CHECK3-NEXT: [[TMP26:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 2 1026 // CHECK3-NEXT: store ptr null, ptr [[TMP26]], align 4 1027 // CHECK3-NEXT: [[TMP27:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 1028 // CHECK3-NEXT: [[TMP28:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 1029 // CHECK3-NEXT: [[TMP29:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR_]], align 2 1030 // CHECK3-NEXT: [[TMP30:%.*]] = sext i16 [[TMP29]] to i32 1031 // CHECK3-NEXT: [[KERNEL_ARGS1:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS]], align 8 1032 // CHECK3-NEXT: [[TMP31:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 0 1033 // CHECK3-NEXT: store i32 1, ptr [[TMP31]], align 4 1034 // CHECK3-NEXT: [[TMP32:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 1 1035 // CHECK3-NEXT: store i32 3, ptr [[TMP32]], align 4 1036 // CHECK3-NEXT: [[TMP33:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 2 1037 // CHECK3-NEXT: store ptr [[TMP27]], ptr [[TMP33]], align 4 1038 // CHECK3-NEXT: [[TMP34:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 3 1039 // CHECK3-NEXT: store ptr [[TMP28]], ptr [[TMP34]], align 4 1040 // CHECK3-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 4 1041 // CHECK3-NEXT: store ptr @.offload_sizes.12, ptr [[TMP35]], align 4 1042 // CHECK3-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 5 1043 // CHECK3-NEXT: store ptr @.offload_maptypes.13, ptr [[TMP36]], align 4 1044 // CHECK3-NEXT: [[TMP37:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 6 1045 // CHECK3-NEXT: store ptr null, ptr [[TMP37]], align 4 1046 // CHECK3-NEXT: [[TMP38:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 7 1047 // CHECK3-NEXT: store ptr null, ptr [[TMP38]], align 4 1048 // CHECK3-NEXT: [[TMP39:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 8 1049 // CHECK3-NEXT: store i64 0, ptr [[TMP39]], align 8 1050 // CHECK3-NEXT: [[TMP40:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 [[TMP30]], i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93.region_id, ptr [[KERNEL_ARGS1]]) 1051 // CHECK3-NEXT: [[TMP41:%.*]] = icmp ne i32 [[TMP40]], 0 1052 // CHECK3-NEXT: br i1 [[TMP41]], label [[OMP_OFFLOAD_FAILED2:%.*]], label [[OMP_OFFLOAD_CONT3:%.*]] 1053 // CHECK3: omp_offload.failed2: 1054 // CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93(i32 [[TMP13]], i32 [[TMP15]], i32 [[TMP17]]) #[[ATTR2]] 1055 // CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT3]] 1056 // CHECK3: omp_offload.cont3: 1057 // CHECK3-NEXT: [[TMP42:%.*]] = load i32, ptr [[A]], align 4 1058 // CHECK3-NEXT: ret i32 [[TMP42]] 1059 // 1060 // 1061 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121 1062 // CHECK3-SAME: (ptr noundef [[THIS:%.*]], i32 noundef [[B:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1:[0-9]+]] { 1063 // CHECK3-NEXT: entry: 1064 // CHECK3-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1065 // CHECK3-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1066 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1067 // CHECK3-NEXT: [[B_CASTED:%.*]] = alloca i32, align 4 1068 // CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1069 // CHECK3-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1070 // CHECK3-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1071 // CHECK3-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1072 // CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1073 // CHECK3-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1074 // CHECK3-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP2]], i32 0) 1075 // CHECK3-NEXT: [[TMP3:%.*]] = load i32, ptr [[B_ADDR]], align 4 1076 // CHECK3-NEXT: store i32 [[TMP3]], ptr [[B_CASTED]], align 4 1077 // CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[B_CASTED]], align 4 1078 // CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @.omp_outlined., ptr [[TMP1]], i32 [[TMP4]]) 1079 // CHECK3-NEXT: ret void 1080 // 1081 // 1082 // CHECK3-LABEL: define {{[^@]+}}@.omp_outlined. 1083 // CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]], i32 noundef [[B:%.*]]) #[[ATTR1]] { 1084 // CHECK3-NEXT: entry: 1085 // CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1086 // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1087 // CHECK3-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1088 // CHECK3-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1089 // CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1090 // CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1091 // CHECK3-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1092 // CHECK3-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1093 // CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1094 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[B_ADDR]], align 4 1095 // CHECK3-NEXT: [[CONV:%.*]] = sitofp i32 [[TMP1]] to double 1096 // CHECK3-NEXT: [[ADD:%.*]] = fadd double [[CONV]], 1.500000e+00 1097 // CHECK3-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 1098 // CHECK3-NEXT: store double [[ADD]], ptr [[A]], align 4 1099 // CHECK3-NEXT: ret void 1100 // 1101 // 1102 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126 1103 // CHECK3-SAME: (ptr noundef [[THIS:%.*]]) #[[ATTR3:[0-9]+]] { 1104 // CHECK3-NEXT: entry: 1105 // CHECK3-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1106 // CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1107 // CHECK3-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1108 // CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1109 // CHECK3-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 1024, i32 0) 1110 // CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 1, ptr @.omp_outlined..1, ptr [[TMP1]]) 1111 // CHECK3-NEXT: ret void 1112 // 1113 // 1114 // CHECK3-LABEL: define {{[^@]+}}@.omp_outlined..1 1115 // CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]]) #[[ATTR1]] { 1116 // CHECK3-NEXT: entry: 1117 // CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1118 // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1119 // CHECK3-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1120 // CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1121 // CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1122 // CHECK3-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1123 // CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1124 // CHECK3-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 1125 // CHECK3-NEXT: store double 2.500000e+00, ptr [[A]], align 4 1126 // CHECK3-NEXT: ret void 1127 // 1128 // 1129 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104 1130 // CHECK3-SAME: (i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] { 1131 // CHECK3-NEXT: entry: 1132 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1133 // CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1134 // CHECK3-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1135 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1136 // CHECK3-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP1]], i32 0) 1137 // CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @.omp_outlined..4) 1138 // CHECK3-NEXT: ret void 1139 // 1140 // 1141 // CHECK3-LABEL: define {{[^@]+}}@.omp_outlined..4 1142 // CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] { 1143 // CHECK3-NEXT: entry: 1144 // CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1145 // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1146 // CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1147 // CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1148 // CHECK3-NEXT: ret void 1149 // 1150 // 1151 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108 1152 // CHECK3-SAME: (i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] { 1153 // CHECK3-NEXT: entry: 1154 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1155 // CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1156 // CHECK3-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1157 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1158 // CHECK3-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP1]], i32 0) 1159 // CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @.omp_outlined..7) 1160 // CHECK3-NEXT: ret void 1161 // 1162 // 1163 // CHECK3-LABEL: define {{[^@]+}}@.omp_outlined..7 1164 // CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] { 1165 // CHECK3-NEXT: entry: 1166 // CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1167 // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1168 // CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1169 // CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1170 // CHECK3-NEXT: ret void 1171 // 1172 // 1173 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88 1174 // CHECK3-SAME: () #[[ATTR4:[0-9]+]] { 1175 // CHECK3-NEXT: entry: 1176 // CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1177 // CHECK3-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 20, i32 0) 1178 // CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @.omp_outlined..10) 1179 // CHECK3-NEXT: ret void 1180 // 1181 // 1182 // CHECK3-LABEL: define {{[^@]+}}@.omp_outlined..10 1183 // CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] { 1184 // CHECK3-NEXT: entry: 1185 // CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1186 // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1187 // CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1188 // CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1189 // CHECK3-NEXT: ret void 1190 // 1191 // 1192 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93 1193 // CHECK3-SAME: (i32 noundef [[A:%.*]], i32 noundef [[B:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] { 1194 // CHECK3-NEXT: entry: 1195 // CHECK3-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4 1196 // CHECK3-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1197 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1198 // CHECK3-NEXT: [[A_CASTED:%.*]] = alloca i32, align 4 1199 // CHECK3-NEXT: [[B_CASTED:%.*]] = alloca i32, align 4 1200 // CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1201 // CHECK3-NEXT: store i32 [[A]], ptr [[A_ADDR]], align 4 1202 // CHECK3-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1203 // CHECK3-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1204 // CHECK3-NEXT: [[TMP1:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR__ADDR]], align 2 1205 // CHECK3-NEXT: [[TMP2:%.*]] = sext i16 [[TMP1]] to i32 1206 // CHECK3-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP2]], i32 0) 1207 // CHECK3-NEXT: [[TMP3:%.*]] = load i32, ptr [[A_ADDR]], align 4 1208 // CHECK3-NEXT: store i32 [[TMP3]], ptr [[A_CASTED]], align 4 1209 // CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[A_CASTED]], align 4 1210 // CHECK3-NEXT: [[TMP5:%.*]] = load i16, ptr [[B_ADDR]], align 2 1211 // CHECK3-NEXT: store i16 [[TMP5]], ptr [[B_CASTED]], align 2 1212 // CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[B_CASTED]], align 4 1213 // CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @.omp_outlined..11, i32 [[TMP4]], i32 [[TMP6]]) 1214 // CHECK3-NEXT: ret void 1215 // 1216 // 1217 // CHECK3-LABEL: define {{[^@]+}}@.omp_outlined..11 1218 // CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i32 noundef [[A:%.*]], i32 noundef [[B:%.*]]) #[[ATTR1]] { 1219 // CHECK3-NEXT: entry: 1220 // CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1221 // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1222 // CHECK3-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4 1223 // CHECK3-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1224 // CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1225 // CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1226 // CHECK3-NEXT: store i32 [[A]], ptr [[A_ADDR]], align 4 1227 // CHECK3-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1228 // CHECK3-NEXT: [[TMP0:%.*]] = load i16, ptr [[B_ADDR]], align 2 1229 // CHECK3-NEXT: [[CONV:%.*]] = sext i16 [[TMP0]] to i32 1230 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[A_ADDR]], align 4 1231 // CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP1]], [[CONV]] 1232 // CHECK3-NEXT: store i32 [[ADD]], ptr [[A_ADDR]], align 4 1233 // CHECK3-NEXT: ret void 1234 // 1235 // 1236 // CHECK3-LABEL: define {{[^@]+}}@.omp_offloading.requires_reg 1237 // CHECK3-SAME: () #[[ATTR5:[0-9]+]] { 1238 // CHECK3-NEXT: entry: 1239 // CHECK3-NEXT: call void @__tgt_register_requires(i64 1) 1240 // CHECK3-NEXT: ret void 1241 // 1242 // 1243 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104 1244 // CHECK9-SAME: (i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0:[0-9]+]] { 1245 // CHECK9-NEXT: entry: 1246 // CHECK9-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 1247 // CHECK9-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1:[0-9]+]]) 1248 // CHECK9-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 1249 // CHECK9-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1250 // CHECK9-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP1]], i32 0) 1251 // CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @.omp_outlined.) 1252 // CHECK9-NEXT: ret void 1253 // 1254 // 1255 // CHECK9-LABEL: define {{[^@]+}}@.omp_outlined. 1256 // CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { 1257 // CHECK9-NEXT: entry: 1258 // CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 1259 // CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 1260 // CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 1261 // CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 1262 // CHECK9-NEXT: ret void 1263 // 1264 // 1265 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108 1266 // CHECK9-SAME: (i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0]] { 1267 // CHECK9-NEXT: entry: 1268 // CHECK9-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 1269 // CHECK9-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1270 // CHECK9-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 1271 // CHECK9-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1272 // CHECK9-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP1]], i32 0) 1273 // CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @.omp_outlined..1) 1274 // CHECK9-NEXT: ret void 1275 // 1276 // 1277 // CHECK9-LABEL: define {{[^@]+}}@.omp_outlined..1 1278 // CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { 1279 // CHECK9-NEXT: entry: 1280 // CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 1281 // CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 1282 // CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 1283 // CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 1284 // CHECK9-NEXT: ret void 1285 // 1286 // 1287 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121 1288 // CHECK9-SAME: (ptr noundef [[THIS:%.*]], i64 noundef [[B:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0]] { 1289 // CHECK9-NEXT: entry: 1290 // CHECK9-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 1291 // CHECK9-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 1292 // CHECK9-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 1293 // CHECK9-NEXT: [[B_CASTED:%.*]] = alloca i64, align 8 1294 // CHECK9-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1295 // CHECK9-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 1296 // CHECK9-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 1297 // CHECK9-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 1298 // CHECK9-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 1299 // CHECK9-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1300 // CHECK9-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP2]], i32 0) 1301 // CHECK9-NEXT: [[TMP3:%.*]] = load i32, ptr [[B_ADDR]], align 4 1302 // CHECK9-NEXT: store i32 [[TMP3]], ptr [[B_CASTED]], align 4 1303 // CHECK9-NEXT: [[TMP4:%.*]] = load i64, ptr [[B_CASTED]], align 8 1304 // CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @.omp_outlined..2, ptr [[TMP1]], i64 [[TMP4]]) 1305 // CHECK9-NEXT: ret void 1306 // 1307 // 1308 // CHECK9-LABEL: define {{[^@]+}}@.omp_outlined..2 1309 // CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]], i64 noundef [[B:%.*]]) #[[ATTR0]] { 1310 // CHECK9-NEXT: entry: 1311 // CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 1312 // CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 1313 // CHECK9-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 1314 // CHECK9-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 1315 // CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 1316 // CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 1317 // CHECK9-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 1318 // CHECK9-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 1319 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 1320 // CHECK9-NEXT: [[TMP1:%.*]] = load i32, ptr [[B_ADDR]], align 4 1321 // CHECK9-NEXT: [[CONV:%.*]] = sitofp i32 [[TMP1]] to double 1322 // CHECK9-NEXT: [[ADD:%.*]] = fadd double [[CONV]], 1.500000e+00 1323 // CHECK9-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 1324 // CHECK9-NEXT: store double [[ADD]], ptr [[A]], align 8 1325 // CHECK9-NEXT: ret void 1326 // 1327 // 1328 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126 1329 // CHECK9-SAME: (ptr noundef [[THIS:%.*]]) #[[ATTR2:[0-9]+]] { 1330 // CHECK9-NEXT: entry: 1331 // CHECK9-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 1332 // CHECK9-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1333 // CHECK9-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 1334 // CHECK9-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 1335 // CHECK9-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 1024, i32 0) 1336 // CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 1, ptr @.omp_outlined..3, ptr [[TMP1]]) 1337 // CHECK9-NEXT: ret void 1338 // 1339 // 1340 // CHECK9-LABEL: define {{[^@]+}}@.omp_outlined..3 1341 // CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]]) #[[ATTR0]] { 1342 // CHECK9-NEXT: entry: 1343 // CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 1344 // CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 1345 // CHECK9-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 1346 // CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 1347 // CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 1348 // CHECK9-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 1349 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 1350 // CHECK9-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 1351 // CHECK9-NEXT: store double 2.500000e+00, ptr [[A]], align 8 1352 // CHECK9-NEXT: ret void 1353 // 1354 // 1355 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88 1356 // CHECK9-SAME: () #[[ATTR3:[0-9]+]] { 1357 // CHECK9-NEXT: entry: 1358 // CHECK9-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1359 // CHECK9-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 20, i32 0) 1360 // CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @.omp_outlined..4) 1361 // CHECK9-NEXT: ret void 1362 // 1363 // 1364 // CHECK9-LABEL: define {{[^@]+}}@.omp_outlined..4 1365 // CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { 1366 // CHECK9-NEXT: entry: 1367 // CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 1368 // CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 1369 // CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 1370 // CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 1371 // CHECK9-NEXT: ret void 1372 // 1373 // 1374 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93 1375 // CHECK9-SAME: (i64 noundef [[A:%.*]], i64 noundef [[B:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0]] { 1376 // CHECK9-NEXT: entry: 1377 // CHECK9-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8 1378 // CHECK9-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 1379 // CHECK9-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 1380 // CHECK9-NEXT: [[A_CASTED:%.*]] = alloca i64, align 8 1381 // CHECK9-NEXT: [[B_CASTED:%.*]] = alloca i64, align 8 1382 // CHECK9-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1383 // CHECK9-NEXT: store i64 [[A]], ptr [[A_ADDR]], align 8 1384 // CHECK9-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 1385 // CHECK9-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 1386 // CHECK9-NEXT: [[TMP1:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR__ADDR]], align 2 1387 // CHECK9-NEXT: [[TMP2:%.*]] = sext i16 [[TMP1]] to i32 1388 // CHECK9-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP2]], i32 0) 1389 // CHECK9-NEXT: [[TMP3:%.*]] = load i32, ptr [[A_ADDR]], align 4 1390 // CHECK9-NEXT: store i32 [[TMP3]], ptr [[A_CASTED]], align 4 1391 // CHECK9-NEXT: [[TMP4:%.*]] = load i64, ptr [[A_CASTED]], align 8 1392 // CHECK9-NEXT: [[TMP5:%.*]] = load i16, ptr [[B_ADDR]], align 2 1393 // CHECK9-NEXT: store i16 [[TMP5]], ptr [[B_CASTED]], align 2 1394 // CHECK9-NEXT: [[TMP6:%.*]] = load i64, ptr [[B_CASTED]], align 8 1395 // CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @.omp_outlined..5, i64 [[TMP4]], i64 [[TMP6]]) 1396 // CHECK9-NEXT: ret void 1397 // 1398 // 1399 // CHECK9-LABEL: define {{[^@]+}}@.omp_outlined..5 1400 // CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i64 noundef [[A:%.*]], i64 noundef [[B:%.*]]) #[[ATTR0]] { 1401 // CHECK9-NEXT: entry: 1402 // CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 1403 // CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 1404 // CHECK9-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8 1405 // CHECK9-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 1406 // CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 1407 // CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 1408 // CHECK9-NEXT: store i64 [[A]], ptr [[A_ADDR]], align 8 1409 // CHECK9-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 1410 // CHECK9-NEXT: [[TMP0:%.*]] = load i16, ptr [[B_ADDR]], align 2 1411 // CHECK9-NEXT: [[CONV:%.*]] = sext i16 [[TMP0]] to i32 1412 // CHECK9-NEXT: [[TMP1:%.*]] = load i32, ptr [[A_ADDR]], align 4 1413 // CHECK9-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP1]], [[CONV]] 1414 // CHECK9-NEXT: store i32 [[ADD]], ptr [[A_ADDR]], align 4 1415 // CHECK9-NEXT: ret void 1416 // 1417 // 1418 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104 1419 // CHECK11-SAME: (i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0:[0-9]+]] { 1420 // CHECK11-NEXT: entry: 1421 // CHECK11-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1422 // CHECK11-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1:[0-9]+]]) 1423 // CHECK11-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1424 // CHECK11-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1425 // CHECK11-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP1]], i32 0) 1426 // CHECK11-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @.omp_outlined.) 1427 // CHECK11-NEXT: ret void 1428 // 1429 // 1430 // CHECK11-LABEL: define {{[^@]+}}@.omp_outlined. 1431 // CHECK11-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { 1432 // CHECK11-NEXT: entry: 1433 // CHECK11-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1434 // CHECK11-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1435 // CHECK11-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1436 // CHECK11-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1437 // CHECK11-NEXT: ret void 1438 // 1439 // 1440 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108 1441 // CHECK11-SAME: (i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0]] { 1442 // CHECK11-NEXT: entry: 1443 // CHECK11-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1444 // CHECK11-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1445 // CHECK11-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1446 // CHECK11-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1447 // CHECK11-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP1]], i32 0) 1448 // CHECK11-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @.omp_outlined..1) 1449 // CHECK11-NEXT: ret void 1450 // 1451 // 1452 // CHECK11-LABEL: define {{[^@]+}}@.omp_outlined..1 1453 // CHECK11-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { 1454 // CHECK11-NEXT: entry: 1455 // CHECK11-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1456 // CHECK11-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1457 // CHECK11-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1458 // CHECK11-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1459 // CHECK11-NEXT: ret void 1460 // 1461 // 1462 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121 1463 // CHECK11-SAME: (ptr noundef [[THIS:%.*]], i32 noundef [[B:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0]] { 1464 // CHECK11-NEXT: entry: 1465 // CHECK11-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1466 // CHECK11-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1467 // CHECK11-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1468 // CHECK11-NEXT: [[B_CASTED:%.*]] = alloca i32, align 4 1469 // CHECK11-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1470 // CHECK11-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1471 // CHECK11-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1472 // CHECK11-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1473 // CHECK11-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1474 // CHECK11-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1475 // CHECK11-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP2]], i32 0) 1476 // CHECK11-NEXT: [[TMP3:%.*]] = load i32, ptr [[B_ADDR]], align 4 1477 // CHECK11-NEXT: store i32 [[TMP3]], ptr [[B_CASTED]], align 4 1478 // CHECK11-NEXT: [[TMP4:%.*]] = load i32, ptr [[B_CASTED]], align 4 1479 // CHECK11-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @.omp_outlined..2, ptr [[TMP1]], i32 [[TMP4]]) 1480 // CHECK11-NEXT: ret void 1481 // 1482 // 1483 // CHECK11-LABEL: define {{[^@]+}}@.omp_outlined..2 1484 // CHECK11-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]], i32 noundef [[B:%.*]]) #[[ATTR0]] { 1485 // CHECK11-NEXT: entry: 1486 // CHECK11-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1487 // CHECK11-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1488 // CHECK11-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1489 // CHECK11-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1490 // CHECK11-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1491 // CHECK11-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1492 // CHECK11-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1493 // CHECK11-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1494 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1495 // CHECK11-NEXT: [[TMP1:%.*]] = load i32, ptr [[B_ADDR]], align 4 1496 // CHECK11-NEXT: [[CONV:%.*]] = sitofp i32 [[TMP1]] to double 1497 // CHECK11-NEXT: [[ADD:%.*]] = fadd double [[CONV]], 1.500000e+00 1498 // CHECK11-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 1499 // CHECK11-NEXT: store double [[ADD]], ptr [[A]], align 4 1500 // CHECK11-NEXT: ret void 1501 // 1502 // 1503 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126 1504 // CHECK11-SAME: (ptr noundef [[THIS:%.*]]) #[[ATTR2:[0-9]+]] { 1505 // CHECK11-NEXT: entry: 1506 // CHECK11-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1507 // CHECK11-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1508 // CHECK11-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1509 // CHECK11-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1510 // CHECK11-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 1024, i32 0) 1511 // CHECK11-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 1, ptr @.omp_outlined..3, ptr [[TMP1]]) 1512 // CHECK11-NEXT: ret void 1513 // 1514 // 1515 // CHECK11-LABEL: define {{[^@]+}}@.omp_outlined..3 1516 // CHECK11-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]]) #[[ATTR0]] { 1517 // CHECK11-NEXT: entry: 1518 // CHECK11-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1519 // CHECK11-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1520 // CHECK11-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1521 // CHECK11-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1522 // CHECK11-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1523 // CHECK11-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1524 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1525 // CHECK11-NEXT: [[A:%.*]] = getelementptr inbounds [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 1526 // CHECK11-NEXT: store double 2.500000e+00, ptr [[A]], align 4 1527 // CHECK11-NEXT: ret void 1528 // 1529 // 1530 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88 1531 // CHECK11-SAME: () #[[ATTR3:[0-9]+]] { 1532 // CHECK11-NEXT: entry: 1533 // CHECK11-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1534 // CHECK11-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 20, i32 0) 1535 // CHECK11-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @.omp_outlined..4) 1536 // CHECK11-NEXT: ret void 1537 // 1538 // 1539 // CHECK11-LABEL: define {{[^@]+}}@.omp_outlined..4 1540 // CHECK11-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { 1541 // CHECK11-NEXT: entry: 1542 // CHECK11-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1543 // CHECK11-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1544 // CHECK11-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1545 // CHECK11-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1546 // CHECK11-NEXT: ret void 1547 // 1548 // 1549 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93 1550 // CHECK11-SAME: (i32 noundef [[A:%.*]], i32 noundef [[B:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0]] { 1551 // CHECK11-NEXT: entry: 1552 // CHECK11-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4 1553 // CHECK11-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1554 // CHECK11-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1555 // CHECK11-NEXT: [[A_CASTED:%.*]] = alloca i32, align 4 1556 // CHECK11-NEXT: [[B_CASTED:%.*]] = alloca i32, align 4 1557 // CHECK11-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1558 // CHECK11-NEXT: store i32 [[A]], ptr [[A_ADDR]], align 4 1559 // CHECK11-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1560 // CHECK11-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1561 // CHECK11-NEXT: [[TMP1:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR__ADDR]], align 2 1562 // CHECK11-NEXT: [[TMP2:%.*]] = sext i16 [[TMP1]] to i32 1563 // CHECK11-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP2]], i32 0) 1564 // CHECK11-NEXT: [[TMP3:%.*]] = load i32, ptr [[A_ADDR]], align 4 1565 // CHECK11-NEXT: store i32 [[TMP3]], ptr [[A_CASTED]], align 4 1566 // CHECK11-NEXT: [[TMP4:%.*]] = load i32, ptr [[A_CASTED]], align 4 1567 // CHECK11-NEXT: [[TMP5:%.*]] = load i16, ptr [[B_ADDR]], align 2 1568 // CHECK11-NEXT: store i16 [[TMP5]], ptr [[B_CASTED]], align 2 1569 // CHECK11-NEXT: [[TMP6:%.*]] = load i32, ptr [[B_CASTED]], align 4 1570 // CHECK11-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @.omp_outlined..5, i32 [[TMP4]], i32 [[TMP6]]) 1571 // CHECK11-NEXT: ret void 1572 // 1573 // 1574 // CHECK11-LABEL: define {{[^@]+}}@.omp_outlined..5 1575 // CHECK11-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i32 noundef [[A:%.*]], i32 noundef [[B:%.*]]) #[[ATTR0]] { 1576 // CHECK11-NEXT: entry: 1577 // CHECK11-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1578 // CHECK11-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1579 // CHECK11-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4 1580 // CHECK11-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1581 // CHECK11-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1582 // CHECK11-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1583 // CHECK11-NEXT: store i32 [[A]], ptr [[A_ADDR]], align 4 1584 // CHECK11-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1585 // CHECK11-NEXT: [[TMP0:%.*]] = load i16, ptr [[B_ADDR]], align 2 1586 // CHECK11-NEXT: [[CONV:%.*]] = sext i16 [[TMP0]] to i32 1587 // CHECK11-NEXT: [[TMP1:%.*]] = load i32, ptr [[A_ADDR]], align 4 1588 // CHECK11-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP1]], [[CONV]] 1589 // CHECK11-NEXT: store i32 [[ADD]], ptr [[A_ADDR]], align 4 1590 // CHECK11-NEXT: ret void 1591 // 1592