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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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-target-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 thread_limit(tx(20)) 89 { 90 } 91 92 short b = 1; 93 #pragma omp target teams num_teams(b) thread_limit(1024) 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) thread_limit(n*32) 105 { 106 } 107 108 #pragma omp target teams thread_limit(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 thread_limit(n-b) 122 { 123 this->a = (double)b + 1.5; 124 } 125 126 #pragma omp target teams thread_limit(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 { 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: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8 217 // CHECK1-NEXT: [[DOTOFFLOAD_BASEPTRS3:%.*]] = alloca [1 x ptr], align 8 218 // CHECK1-NEXT: [[DOTOFFLOAD_PTRS4:%.*]] = alloca [1 x ptr], align 8 219 // CHECK1-NEXT: [[DOTOFFLOAD_MAPPERS5:%.*]] = alloca [1 x ptr], align 8 220 // CHECK1-NEXT: [[KERNEL_ARGS6:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS]], align 8 221 // CHECK1-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 222 // CHECK1-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 223 // CHECK1-NEXT: [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 224 // CHECK1-NEXT: store i32 1, ptr [[B]], align 4 225 // CHECK1-NEXT: [[TMP0:%.*]] = load i32, ptr [[N_ADDR]], align 4 226 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[B]], align 4 227 // CHECK1-NEXT: [[SUB:%.*]] = sub nsw i32 [[TMP0]], [[TMP1]] 228 // CHECK1-NEXT: store i32 [[SUB]], ptr [[DOTCAPTURE_EXPR_]], align 4 229 // CHECK1-NEXT: [[TMP2:%.*]] = load i32, ptr [[B]], align 4 230 // CHECK1-NEXT: store i32 [[TMP2]], ptr [[B_CASTED]], align 4 231 // CHECK1-NEXT: [[TMP3:%.*]] = load i64, ptr [[B_CASTED]], align 8 232 // CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 233 // CHECK1-NEXT: store i32 [[TMP4]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 234 // CHECK1-NEXT: [[TMP5:%.*]] = load i64, ptr [[DOTCAPTURE_EXPR__CASTED]], align 8 235 // CHECK1-NEXT: [[A:%.*]] = getelementptr inbounds nuw [[STRUCT_S1:%.*]], ptr [[THIS1]], i32 0, i32 0 236 // CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 237 // CHECK1-NEXT: store ptr [[THIS1]], ptr [[TMP6]], align 8 238 // CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 239 // CHECK1-NEXT: store ptr [[A]], ptr [[TMP7]], align 8 240 // CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 0 241 // CHECK1-NEXT: store ptr null, ptr [[TMP8]], align 8 242 // CHECK1-NEXT: [[TMP9:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 1 243 // CHECK1-NEXT: store i64 [[TMP3]], ptr [[TMP9]], align 8 244 // CHECK1-NEXT: [[TMP10:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 1 245 // CHECK1-NEXT: store i64 [[TMP3]], ptr [[TMP10]], align 8 246 // CHECK1-NEXT: [[TMP11:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 1 247 // CHECK1-NEXT: store ptr null, ptr [[TMP11]], align 8 248 // CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 2 249 // CHECK1-NEXT: store i64 [[TMP5]], ptr [[TMP12]], align 8 250 // CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 2 251 // CHECK1-NEXT: store i64 [[TMP5]], ptr [[TMP13]], align 8 252 // CHECK1-NEXT: [[TMP14:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 2 253 // CHECK1-NEXT: store ptr null, ptr [[TMP14]], align 8 254 // CHECK1-NEXT: [[TMP15:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 255 // CHECK1-NEXT: [[TMP16:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 256 // CHECK1-NEXT: [[TMP17:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 257 // CHECK1-NEXT: [[TMP18:%.*]] = insertvalue [3 x i32] zeroinitializer, i32 [[TMP17]], 0 258 // CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 259 // CHECK1-NEXT: store i32 3, ptr [[TMP19]], align 4 260 // CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1 261 // CHECK1-NEXT: store i32 3, ptr [[TMP20]], align 4 262 // CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2 263 // CHECK1-NEXT: store ptr [[TMP15]], ptr [[TMP21]], align 8 264 // CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3 265 // CHECK1-NEXT: store ptr [[TMP16]], ptr [[TMP22]], align 8 266 // CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4 267 // CHECK1-NEXT: store ptr @.offload_sizes, ptr [[TMP23]], align 8 268 // CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5 269 // CHECK1-NEXT: store ptr @.offload_maptypes, ptr [[TMP24]], align 8 270 // CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6 271 // CHECK1-NEXT: store ptr null, ptr [[TMP25]], align 8 272 // CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7 273 // CHECK1-NEXT: store ptr null, ptr [[TMP26]], align 8 274 // CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8 275 // CHECK1-NEXT: store i64 0, ptr [[TMP27]], align 8 276 // CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 9 277 // CHECK1-NEXT: store i64 0, ptr [[TMP28]], align 8 278 // CHECK1-NEXT: [[TMP29:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 10 279 // CHECK1-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP29]], align 4 280 // CHECK1-NEXT: [[TMP30:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 11 281 // CHECK1-NEXT: store [3 x i32] [[TMP18]], ptr [[TMP30]], align 4 282 // CHECK1-NEXT: [[TMP31:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 12 283 // CHECK1-NEXT: store i32 0, ptr [[TMP31]], align 4 284 // CHECK1-NEXT: [[TMP32:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1:[0-9]+]], i64 -1, i32 0, i32 [[TMP17]], ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121.region_id, ptr [[KERNEL_ARGS]]) 285 // CHECK1-NEXT: [[TMP33:%.*]] = icmp ne i32 [[TMP32]], 0 286 // CHECK1-NEXT: br i1 [[TMP33]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]] 287 // CHECK1: omp_offload.failed: 288 // CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121(ptr [[THIS1]], i64 [[TMP3]], i64 [[TMP5]]) #[[ATTR2:[0-9]+]] 289 // CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT]] 290 // CHECK1: omp_offload.cont: 291 // CHECK1-NEXT: [[A2:%.*]] = getelementptr inbounds nuw [[STRUCT_S1]], ptr [[THIS1]], i32 0, i32 0 292 // CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS3]], i32 0, i32 0 293 // CHECK1-NEXT: store ptr [[THIS1]], ptr [[TMP34]], align 8 294 // CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS4]], i32 0, i32 0 295 // CHECK1-NEXT: store ptr [[A2]], ptr [[TMP35]], align 8 296 // CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS5]], i64 0, i64 0 297 // CHECK1-NEXT: store ptr null, ptr [[TMP36]], align 8 298 // CHECK1-NEXT: [[TMP37:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS3]], i32 0, i32 0 299 // CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS4]], i32 0, i32 0 300 // CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 0 301 // CHECK1-NEXT: store i32 3, ptr [[TMP39]], align 4 302 // CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 1 303 // CHECK1-NEXT: store i32 1, ptr [[TMP40]], align 4 304 // CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 2 305 // CHECK1-NEXT: store ptr [[TMP37]], ptr [[TMP41]], align 8 306 // CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 3 307 // CHECK1-NEXT: store ptr [[TMP38]], ptr [[TMP42]], align 8 308 // CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 4 309 // CHECK1-NEXT: store ptr @.offload_sizes.1, ptr [[TMP43]], align 8 310 // CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 5 311 // CHECK1-NEXT: store ptr @.offload_maptypes.2, ptr [[TMP44]], align 8 312 // CHECK1-NEXT: [[TMP45:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 6 313 // CHECK1-NEXT: store ptr null, ptr [[TMP45]], align 8 314 // CHECK1-NEXT: [[TMP46:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 7 315 // CHECK1-NEXT: store ptr null, ptr [[TMP46]], align 8 316 // CHECK1-NEXT: [[TMP47:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 8 317 // CHECK1-NEXT: store i64 0, ptr [[TMP47]], align 8 318 // CHECK1-NEXT: [[TMP48:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 9 319 // CHECK1-NEXT: store i64 0, ptr [[TMP48]], align 8 320 // CHECK1-NEXT: [[TMP49:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 10 321 // CHECK1-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP49]], align 4 322 // CHECK1-NEXT: [[TMP50:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 11 323 // CHECK1-NEXT: store [3 x i32] [i32 1024, i32 0, i32 0], ptr [[TMP50]], align 4 324 // CHECK1-NEXT: [[TMP51:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 12 325 // CHECK1-NEXT: store i32 0, ptr [[TMP51]], align 4 326 // CHECK1-NEXT: [[TMP52:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 0, i32 1024, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126.region_id, ptr [[KERNEL_ARGS6]]) 327 // CHECK1-NEXT: [[TMP53:%.*]] = icmp ne i32 [[TMP52]], 0 328 // CHECK1-NEXT: br i1 [[TMP53]], label [[OMP_OFFLOAD_FAILED7:%.*]], label [[OMP_OFFLOAD_CONT8:%.*]] 329 // CHECK1: omp_offload.failed7: 330 // CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126(ptr [[THIS1]]) #[[ATTR2]] 331 // CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT8]] 332 // CHECK1: omp_offload.cont8: 333 // CHECK1-NEXT: [[A9:%.*]] = getelementptr inbounds nuw [[STRUCT_S1]], ptr [[THIS1]], i32 0, i32 0 334 // CHECK1-NEXT: [[TMP54:%.*]] = load double, ptr [[A9]], align 8 335 // CHECK1-NEXT: [[CONV:%.*]] = fptosi double [[TMP54]] to i32 336 // CHECK1-NEXT: ret i32 [[CONV]] 337 // 338 // 339 // CHECK1-LABEL: define {{[^@]+}}@_ZL7fstatici 340 // CHECK1-SAME: (i32 noundef signext [[N:%.*]]) #[[ATTR0]] { 341 // CHECK1-NEXT: entry: 342 // CHECK1-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 343 // CHECK1-NEXT: [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4 344 // CHECK1-NEXT: [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4 345 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i64, align 8 346 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__CASTED2:%.*]] = alloca i64, align 8 347 // CHECK1-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [2 x ptr], align 8 348 // CHECK1-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [2 x ptr], align 8 349 // CHECK1-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [2 x ptr], align 8 350 // CHECK1-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8 351 // CHECK1-NEXT: [[DOTCAPTURE_EXPR_3:%.*]] = alloca i32, align 4 352 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__CASTED4:%.*]] = alloca i64, align 8 353 // CHECK1-NEXT: [[DOTOFFLOAD_BASEPTRS5:%.*]] = alloca [1 x ptr], align 8 354 // CHECK1-NEXT: [[DOTOFFLOAD_PTRS6:%.*]] = alloca [1 x ptr], align 8 355 // CHECK1-NEXT: [[DOTOFFLOAD_MAPPERS7:%.*]] = alloca [1 x ptr], align 8 356 // CHECK1-NEXT: [[KERNEL_ARGS8:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS]], align 8 357 // CHECK1-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 358 // CHECK1-NEXT: [[TMP0:%.*]] = load i32, ptr [[N_ADDR]], align 4 359 // CHECK1-NEXT: store i32 [[TMP0]], ptr [[DOTCAPTURE_EXPR_]], align 4 360 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[N_ADDR]], align 4 361 // CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP1]], 32 362 // CHECK1-NEXT: store i32 [[MUL]], ptr [[DOTCAPTURE_EXPR_1]], align 4 363 // CHECK1-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 364 // CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 365 // CHECK1-NEXT: [[TMP3:%.*]] = load i64, ptr [[DOTCAPTURE_EXPR__CASTED]], align 8 366 // CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4 367 // CHECK1-NEXT: store i32 [[TMP4]], ptr [[DOTCAPTURE_EXPR__CASTED2]], align 4 368 // CHECK1-NEXT: [[TMP5:%.*]] = load i64, ptr [[DOTCAPTURE_EXPR__CASTED2]], align 8 369 // CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 370 // CHECK1-NEXT: store i64 [[TMP3]], ptr [[TMP6]], align 8 371 // CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 372 // CHECK1-NEXT: store i64 [[TMP3]], ptr [[TMP7]], align 8 373 // CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 0 374 // CHECK1-NEXT: store ptr null, ptr [[TMP8]], align 8 375 // CHECK1-NEXT: [[TMP9:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 1 376 // CHECK1-NEXT: store i64 [[TMP5]], ptr [[TMP9]], align 8 377 // CHECK1-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 1 378 // CHECK1-NEXT: store i64 [[TMP5]], ptr [[TMP10]], align 8 379 // CHECK1-NEXT: [[TMP11:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 1 380 // CHECK1-NEXT: store ptr null, ptr [[TMP11]], align 8 381 // CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 382 // CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 383 // CHECK1-NEXT: [[TMP14:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 384 // CHECK1-NEXT: [[TMP15:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4 385 // CHECK1-NEXT: [[TMP16:%.*]] = insertvalue [3 x i32] zeroinitializer, i32 [[TMP14]], 0 386 // CHECK1-NEXT: [[TMP17:%.*]] = insertvalue [3 x i32] zeroinitializer, i32 [[TMP15]], 0 387 // CHECK1-NEXT: [[TMP18:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 388 // CHECK1-NEXT: store i32 3, ptr [[TMP18]], align 4 389 // CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1 390 // CHECK1-NEXT: store i32 2, ptr [[TMP19]], align 4 391 // CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2 392 // CHECK1-NEXT: store ptr [[TMP12]], ptr [[TMP20]], align 8 393 // CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3 394 // CHECK1-NEXT: store ptr [[TMP13]], ptr [[TMP21]], align 8 395 // CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4 396 // CHECK1-NEXT: store ptr @.offload_sizes.3, ptr [[TMP22]], align 8 397 // CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5 398 // CHECK1-NEXT: store ptr @.offload_maptypes.4, ptr [[TMP23]], align 8 399 // CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6 400 // CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8 401 // CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7 402 // CHECK1-NEXT: store ptr null, ptr [[TMP25]], align 8 403 // CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8 404 // CHECK1-NEXT: store i64 0, ptr [[TMP26]], align 8 405 // CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 9 406 // CHECK1-NEXT: store i64 0, ptr [[TMP27]], align 8 407 // CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 10 408 // CHECK1-NEXT: store [3 x i32] [[TMP16]], ptr [[TMP28]], align 4 409 // CHECK1-NEXT: [[TMP29:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 11 410 // CHECK1-NEXT: store [3 x i32] [[TMP17]], ptr [[TMP29]], align 4 411 // CHECK1-NEXT: [[TMP30:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 12 412 // CHECK1-NEXT: store i32 0, ptr [[TMP30]], align 4 413 // CHECK1-NEXT: [[TMP31:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 [[TMP14]], i32 [[TMP15]], ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104.region_id, ptr [[KERNEL_ARGS]]) 414 // CHECK1-NEXT: [[TMP32:%.*]] = icmp ne i32 [[TMP31]], 0 415 // CHECK1-NEXT: br i1 [[TMP32]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]] 416 // CHECK1: omp_offload.failed: 417 // CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104(i64 [[TMP3]], i64 [[TMP5]]) #[[ATTR2]] 418 // CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT]] 419 // CHECK1: omp_offload.cont: 420 // CHECK1-NEXT: [[TMP33:%.*]] = load i32, ptr [[N_ADDR]], align 4 421 // CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 32, [[TMP33]] 422 // CHECK1-NEXT: store i32 [[ADD]], ptr [[DOTCAPTURE_EXPR_3]], align 4 423 // CHECK1-NEXT: [[TMP34:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_3]], align 4 424 // CHECK1-NEXT: store i32 [[TMP34]], ptr [[DOTCAPTURE_EXPR__CASTED4]], align 4 425 // CHECK1-NEXT: [[TMP35:%.*]] = load i64, ptr [[DOTCAPTURE_EXPR__CASTED4]], align 8 426 // CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS5]], i32 0, i32 0 427 // CHECK1-NEXT: store i64 [[TMP35]], ptr [[TMP36]], align 8 428 // CHECK1-NEXT: [[TMP37:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS6]], i32 0, i32 0 429 // CHECK1-NEXT: store i64 [[TMP35]], ptr [[TMP37]], align 8 430 // CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS7]], i64 0, i64 0 431 // CHECK1-NEXT: store ptr null, ptr [[TMP38]], align 8 432 // CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS5]], i32 0, i32 0 433 // CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS6]], i32 0, i32 0 434 // CHECK1-NEXT: [[TMP41:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_3]], align 4 435 // CHECK1-NEXT: [[TMP42:%.*]] = insertvalue [3 x i32] zeroinitializer, i32 [[TMP41]], 0 436 // CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 0 437 // CHECK1-NEXT: store i32 3, ptr [[TMP43]], align 4 438 // CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 1 439 // CHECK1-NEXT: store i32 1, ptr [[TMP44]], align 4 440 // CHECK1-NEXT: [[TMP45:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 2 441 // CHECK1-NEXT: store ptr [[TMP39]], ptr [[TMP45]], align 8 442 // CHECK1-NEXT: [[TMP46:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 3 443 // CHECK1-NEXT: store ptr [[TMP40]], ptr [[TMP46]], align 8 444 // CHECK1-NEXT: [[TMP47:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 4 445 // CHECK1-NEXT: store ptr @.offload_sizes.5, ptr [[TMP47]], align 8 446 // CHECK1-NEXT: [[TMP48:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 5 447 // CHECK1-NEXT: store ptr @.offload_maptypes.6, ptr [[TMP48]], align 8 448 // CHECK1-NEXT: [[TMP49:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 6 449 // CHECK1-NEXT: store ptr null, ptr [[TMP49]], align 8 450 // CHECK1-NEXT: [[TMP50:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 7 451 // CHECK1-NEXT: store ptr null, ptr [[TMP50]], align 8 452 // CHECK1-NEXT: [[TMP51:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 8 453 // CHECK1-NEXT: store i64 0, ptr [[TMP51]], align 8 454 // CHECK1-NEXT: [[TMP52:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 9 455 // CHECK1-NEXT: store i64 0, ptr [[TMP52]], align 8 456 // CHECK1-NEXT: [[TMP53:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 10 457 // CHECK1-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP53]], align 4 458 // CHECK1-NEXT: [[TMP54:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 11 459 // CHECK1-NEXT: store [3 x i32] [[TMP42]], ptr [[TMP54]], align 4 460 // CHECK1-NEXT: [[TMP55:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 12 461 // CHECK1-NEXT: store i32 0, ptr [[TMP55]], align 4 462 // CHECK1-NEXT: [[TMP56:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 0, i32 [[TMP41]], ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108.region_id, ptr [[KERNEL_ARGS8]]) 463 // CHECK1-NEXT: [[TMP57:%.*]] = icmp ne i32 [[TMP56]], 0 464 // CHECK1-NEXT: br i1 [[TMP57]], label [[OMP_OFFLOAD_FAILED9:%.*]], label [[OMP_OFFLOAD_CONT10:%.*]] 465 // CHECK1: omp_offload.failed9: 466 // CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108(i64 [[TMP35]]) #[[ATTR2]] 467 // CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT10]] 468 // CHECK1: omp_offload.cont10: 469 // CHECK1-NEXT: [[TMP58:%.*]] = load i32, ptr [[N_ADDR]], align 4 470 // CHECK1-NEXT: [[ADD11:%.*]] = add nsw i32 [[TMP58]], 1 471 // CHECK1-NEXT: ret i32 [[ADD11]] 472 // 473 // 474 // CHECK1-LABEL: define {{[^@]+}}@_Z9ftemplateIiET_i 475 // CHECK1-SAME: (i32 noundef signext [[N:%.*]]) #[[ATTR0]] comdat { 476 // CHECK1-NEXT: entry: 477 // CHECK1-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 478 // CHECK1-NEXT: [[A:%.*]] = alloca i32, align 4 479 // CHECK1-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8 480 // CHECK1-NEXT: [[B:%.*]] = alloca i16, align 2 481 // CHECK1-NEXT: [[DOTCAPTURE_EXPR_:%.*]] = alloca i16, align 2 482 // CHECK1-NEXT: [[A_CASTED:%.*]] = alloca i64, align 8 483 // CHECK1-NEXT: [[B_CASTED:%.*]] = alloca i64, align 8 484 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i64, align 8 485 // CHECK1-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [3 x ptr], align 8 486 // CHECK1-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [3 x ptr], align 8 487 // CHECK1-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [3 x ptr], align 8 488 // CHECK1-NEXT: [[KERNEL_ARGS1:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS]], align 8 489 // CHECK1-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 490 // CHECK1-NEXT: store i32 0, ptr [[A]], align 4 491 // CHECK1-NEXT: [[TMP0:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 492 // CHECK1-NEXT: store i32 3, ptr [[TMP0]], align 4 493 // CHECK1-NEXT: [[TMP1:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1 494 // CHECK1-NEXT: store i32 0, ptr [[TMP1]], align 4 495 // CHECK1-NEXT: [[TMP2:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2 496 // CHECK1-NEXT: store ptr null, ptr [[TMP2]], align 8 497 // CHECK1-NEXT: [[TMP3:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3 498 // CHECK1-NEXT: store ptr null, ptr [[TMP3]], align 8 499 // CHECK1-NEXT: [[TMP4:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4 500 // CHECK1-NEXT: store ptr null, ptr [[TMP4]], align 8 501 // CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5 502 // CHECK1-NEXT: store ptr null, ptr [[TMP5]], align 8 503 // CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6 504 // CHECK1-NEXT: store ptr null, ptr [[TMP6]], align 8 505 // CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7 506 // CHECK1-NEXT: store ptr null, ptr [[TMP7]], align 8 507 // CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8 508 // CHECK1-NEXT: store i64 0, ptr [[TMP8]], align 8 509 // CHECK1-NEXT: [[TMP9:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 9 510 // CHECK1-NEXT: store i64 0, ptr [[TMP9]], align 8 511 // CHECK1-NEXT: [[TMP10:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 10 512 // CHECK1-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP10]], align 4 513 // CHECK1-NEXT: [[TMP11:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 11 514 // CHECK1-NEXT: store [3 x i32] [i32 20, i32 0, i32 0], ptr [[TMP11]], align 4 515 // CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 12 516 // CHECK1-NEXT: store i32 0, ptr [[TMP12]], align 4 517 // CHECK1-NEXT: [[TMP13:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 0, i32 20, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88.region_id, ptr [[KERNEL_ARGS]]) 518 // CHECK1-NEXT: [[TMP14:%.*]] = icmp ne i32 [[TMP13]], 0 519 // CHECK1-NEXT: br i1 [[TMP14]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]] 520 // CHECK1: omp_offload.failed: 521 // CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88() #[[ATTR2]] 522 // CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT]] 523 // CHECK1: omp_offload.cont: 524 // CHECK1-NEXT: store i16 1, ptr [[B]], align 2 525 // CHECK1-NEXT: [[TMP15:%.*]] = load i16, ptr [[B]], align 2 526 // CHECK1-NEXT: store i16 [[TMP15]], ptr [[DOTCAPTURE_EXPR_]], align 2 527 // CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[A]], align 4 528 // CHECK1-NEXT: store i32 [[TMP16]], ptr [[A_CASTED]], align 4 529 // CHECK1-NEXT: [[TMP17:%.*]] = load i64, ptr [[A_CASTED]], align 8 530 // CHECK1-NEXT: [[TMP18:%.*]] = load i16, ptr [[B]], align 2 531 // CHECK1-NEXT: store i16 [[TMP18]], ptr [[B_CASTED]], align 2 532 // CHECK1-NEXT: [[TMP19:%.*]] = load i64, ptr [[B_CASTED]], align 8 533 // CHECK1-NEXT: [[TMP20:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR_]], align 2 534 // CHECK1-NEXT: store i16 [[TMP20]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 2 535 // CHECK1-NEXT: [[TMP21:%.*]] = load i64, ptr [[DOTCAPTURE_EXPR__CASTED]], align 8 536 // CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 537 // CHECK1-NEXT: store i64 [[TMP17]], ptr [[TMP22]], align 8 538 // CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 539 // CHECK1-NEXT: store i64 [[TMP17]], ptr [[TMP23]], align 8 540 // CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 0 541 // CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8 542 // CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 1 543 // CHECK1-NEXT: store i64 [[TMP19]], ptr [[TMP25]], align 8 544 // CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 1 545 // CHECK1-NEXT: store i64 [[TMP19]], ptr [[TMP26]], align 8 546 // CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 1 547 // CHECK1-NEXT: store ptr null, ptr [[TMP27]], align 8 548 // CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 2 549 // CHECK1-NEXT: store i64 [[TMP21]], ptr [[TMP28]], align 8 550 // CHECK1-NEXT: [[TMP29:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 2 551 // CHECK1-NEXT: store i64 [[TMP21]], ptr [[TMP29]], align 8 552 // CHECK1-NEXT: [[TMP30:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 2 553 // CHECK1-NEXT: store ptr null, ptr [[TMP30]], align 8 554 // CHECK1-NEXT: [[TMP31:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 555 // CHECK1-NEXT: [[TMP32:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 556 // CHECK1-NEXT: [[TMP33:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR_]], align 2 557 // CHECK1-NEXT: [[TMP34:%.*]] = sext i16 [[TMP33]] to i32 558 // CHECK1-NEXT: [[TMP35:%.*]] = insertvalue [3 x i32] zeroinitializer, i32 [[TMP34]], 0 559 // CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 0 560 // CHECK1-NEXT: store i32 3, ptr [[TMP36]], align 4 561 // CHECK1-NEXT: [[TMP37:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 1 562 // CHECK1-NEXT: store i32 3, ptr [[TMP37]], align 4 563 // CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 2 564 // CHECK1-NEXT: store ptr [[TMP31]], ptr [[TMP38]], align 8 565 // CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 3 566 // CHECK1-NEXT: store ptr [[TMP32]], ptr [[TMP39]], align 8 567 // CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 4 568 // CHECK1-NEXT: store ptr @.offload_sizes.7, ptr [[TMP40]], align 8 569 // CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 5 570 // CHECK1-NEXT: store ptr @.offload_maptypes.8, ptr [[TMP41]], align 8 571 // CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 6 572 // CHECK1-NEXT: store ptr null, ptr [[TMP42]], align 8 573 // CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 7 574 // CHECK1-NEXT: store ptr null, ptr [[TMP43]], align 8 575 // CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 8 576 // CHECK1-NEXT: store i64 0, ptr [[TMP44]], align 8 577 // CHECK1-NEXT: [[TMP45:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 9 578 // CHECK1-NEXT: store i64 0, ptr [[TMP45]], align 8 579 // CHECK1-NEXT: [[TMP46:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 10 580 // CHECK1-NEXT: store [3 x i32] [[TMP35]], ptr [[TMP46]], align 4 581 // CHECK1-NEXT: [[TMP47:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 11 582 // CHECK1-NEXT: store [3 x i32] [i32 1024, i32 0, i32 0], ptr [[TMP47]], align 4 583 // CHECK1-NEXT: [[TMP48:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 12 584 // CHECK1-NEXT: store i32 0, ptr [[TMP48]], align 4 585 // CHECK1-NEXT: [[TMP49:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 [[TMP34]], i32 1024, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93.region_id, ptr [[KERNEL_ARGS1]]) 586 // CHECK1-NEXT: [[TMP50:%.*]] = icmp ne i32 [[TMP49]], 0 587 // CHECK1-NEXT: br i1 [[TMP50]], label [[OMP_OFFLOAD_FAILED2:%.*]], label [[OMP_OFFLOAD_CONT3:%.*]] 588 // CHECK1: omp_offload.failed2: 589 // CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93(i64 [[TMP17]], i64 [[TMP19]], i64 [[TMP21]]) #[[ATTR2]] 590 // CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT3]] 591 // CHECK1: omp_offload.cont3: 592 // CHECK1-NEXT: [[TMP51:%.*]] = load i32, ptr [[A]], align 4 593 // CHECK1-NEXT: ret i32 [[TMP51]] 594 // 595 // 596 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121 597 // CHECK1-SAME: (ptr noundef [[THIS:%.*]], i64 noundef [[B:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1:[0-9]+]] { 598 // CHECK1-NEXT: entry: 599 // CHECK1-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 600 // CHECK1-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 601 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 602 // CHECK1-NEXT: [[B_CASTED:%.*]] = alloca i64, align 8 603 // CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 604 // CHECK1-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 605 // CHECK1-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 606 // CHECK1-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 607 // CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 608 // CHECK1-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 609 // CHECK1-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 [[TMP2]]) 610 // CHECK1-NEXT: [[TMP3:%.*]] = load i32, ptr [[B_ADDR]], align 4 611 // CHECK1-NEXT: store i32 [[TMP3]], ptr [[B_CASTED]], align 4 612 // CHECK1-NEXT: [[TMP4:%.*]] = load i64, ptr [[B_CASTED]], align 8 613 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121.omp_outlined, ptr [[TMP1]], i64 [[TMP4]]) 614 // CHECK1-NEXT: ret void 615 // 616 // 617 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121.omp_outlined 618 // CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]], i64 noundef [[B:%.*]]) #[[ATTR1]] { 619 // CHECK1-NEXT: entry: 620 // CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 621 // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 622 // CHECK1-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 623 // CHECK1-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 624 // CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 625 // CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 626 // CHECK1-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 627 // CHECK1-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 628 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 629 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[B_ADDR]], align 4 630 // CHECK1-NEXT: [[CONV:%.*]] = sitofp i32 [[TMP1]] to double 631 // CHECK1-NEXT: [[ADD:%.*]] = fadd double [[CONV]], 1.500000e+00 632 // CHECK1-NEXT: [[A:%.*]] = getelementptr inbounds nuw [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 633 // CHECK1-NEXT: store double [[ADD]], ptr [[A]], align 8 634 // CHECK1-NEXT: ret void 635 // 636 // 637 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126 638 // CHECK1-SAME: (ptr noundef [[THIS:%.*]]) #[[ATTR1]] { 639 // CHECK1-NEXT: entry: 640 // CHECK1-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 641 // CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 642 // CHECK1-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 643 // CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 644 // CHECK1-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 1024) 645 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126.omp_outlined, ptr [[TMP1]]) 646 // CHECK1-NEXT: ret void 647 // 648 // 649 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126.omp_outlined 650 // CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]]) #[[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: [[THIS_ADDR:%.*]] = alloca ptr, align 8 655 // CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 656 // CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 657 // CHECK1-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 658 // CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 659 // CHECK1-NEXT: [[A:%.*]] = getelementptr inbounds nuw [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 660 // CHECK1-NEXT: store double 2.500000e+00, ptr [[A]], align 8 661 // CHECK1-NEXT: ret void 662 // 663 // 664 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104 665 // CHECK1-SAME: (i64 noundef [[DOTCAPTURE_EXPR_:%.*]], i64 noundef [[DOTCAPTURE_EXPR_1:%.*]]) #[[ATTR1]] { 666 // CHECK1-NEXT: entry: 667 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 668 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__ADDR2:%.*]] = alloca i64, align 8 669 // CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 670 // CHECK1-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 671 // CHECK1-NEXT: store i64 [[DOTCAPTURE_EXPR_1]], ptr [[DOTCAPTURE_EXPR__ADDR2]], align 8 672 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 673 // CHECK1-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR2]], align 4 674 // CHECK1-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP1]], i32 [[TMP2]]) 675 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104.omp_outlined) 676 // CHECK1-NEXT: ret void 677 // 678 // 679 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104.omp_outlined 680 // CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] { 681 // CHECK1-NEXT: entry: 682 // CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 683 // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 684 // CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 685 // CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 686 // CHECK1-NEXT: ret void 687 // 688 // 689 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108 690 // CHECK1-SAME: (i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] { 691 // CHECK1-NEXT: entry: 692 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 693 // CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 694 // CHECK1-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 695 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 696 // CHECK1-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 [[TMP1]]) 697 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108.omp_outlined) 698 // CHECK1-NEXT: ret void 699 // 700 // 701 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108.omp_outlined 702 // CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] { 703 // CHECK1-NEXT: entry: 704 // CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 705 // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 706 // CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 707 // CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 708 // CHECK1-NEXT: ret void 709 // 710 // 711 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88 712 // CHECK1-SAME: () #[[ATTR1]] { 713 // CHECK1-NEXT: entry: 714 // CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 715 // CHECK1-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 20) 716 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88.omp_outlined) 717 // CHECK1-NEXT: ret void 718 // 719 // 720 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88.omp_outlined 721 // CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] { 722 // CHECK1-NEXT: entry: 723 // CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 724 // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 725 // CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 726 // CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 727 // CHECK1-NEXT: ret void 728 // 729 // 730 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93 731 // CHECK1-SAME: (i64 noundef [[A:%.*]], i64 noundef [[B:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] { 732 // CHECK1-NEXT: entry: 733 // CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8 734 // CHECK1-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 735 // CHECK1-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 736 // CHECK1-NEXT: [[A_CASTED:%.*]] = alloca i64, align 8 737 // CHECK1-NEXT: [[B_CASTED:%.*]] = alloca i64, align 8 738 // CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 739 // CHECK1-NEXT: store i64 [[A]], ptr [[A_ADDR]], align 8 740 // CHECK1-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 741 // CHECK1-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 742 // CHECK1-NEXT: [[TMP1:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR__ADDR]], align 2 743 // CHECK1-NEXT: [[TMP2:%.*]] = sext i16 [[TMP1]] to i32 744 // CHECK1-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP2]], i32 1024) 745 // CHECK1-NEXT: [[TMP3:%.*]] = load i32, ptr [[A_ADDR]], align 4 746 // CHECK1-NEXT: store i32 [[TMP3]], ptr [[A_CASTED]], align 4 747 // CHECK1-NEXT: [[TMP4:%.*]] = load i64, ptr [[A_CASTED]], align 8 748 // CHECK1-NEXT: [[TMP5:%.*]] = load i16, ptr [[B_ADDR]], align 2 749 // CHECK1-NEXT: store i16 [[TMP5]], ptr [[B_CASTED]], align 2 750 // CHECK1-NEXT: [[TMP6:%.*]] = load i64, ptr [[B_CASTED]], align 8 751 // CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93.omp_outlined, i64 [[TMP4]], i64 [[TMP6]]) 752 // CHECK1-NEXT: ret void 753 // 754 // 755 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93.omp_outlined 756 // CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i64 noundef [[A:%.*]], i64 noundef [[B:%.*]]) #[[ATTR1]] { 757 // CHECK1-NEXT: entry: 758 // CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 759 // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 760 // CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8 761 // CHECK1-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 762 // CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 763 // CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 764 // CHECK1-NEXT: store i64 [[A]], ptr [[A_ADDR]], align 8 765 // CHECK1-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 766 // CHECK1-NEXT: [[TMP0:%.*]] = load i16, ptr [[B_ADDR]], align 2 767 // CHECK1-NEXT: [[CONV:%.*]] = sext i16 [[TMP0]] to i32 768 // CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[A_ADDR]], align 4 769 // CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP1]], [[CONV]] 770 // CHECK1-NEXT: store i32 [[ADD]], ptr [[A_ADDR]], align 4 771 // CHECK1-NEXT: ret void 772 // 773 // 774 // CHECK3-LABEL: define {{[^@]+}}@_Z3bari 775 // CHECK3-SAME: (i32 noundef [[N:%.*]]) #[[ATTR0:[0-9]+]] { 776 // CHECK3-NEXT: entry: 777 // CHECK3-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 778 // CHECK3-NEXT: [[A:%.*]] = alloca i32, align 4 779 // CHECK3-NEXT: [[S:%.*]] = alloca [[STRUCT_S1:%.*]], align 4 780 // CHECK3-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 781 // CHECK3-NEXT: store i32 0, ptr [[A]], align 4 782 // CHECK3-NEXT: [[TMP0:%.*]] = load i32, ptr [[N_ADDR]], align 4 783 // CHECK3-NEXT: [[CALL:%.*]] = call noundef i32 @_ZN2S12r1Ei(ptr noundef nonnull align 4 dereferenceable(8) [[S]], i32 noundef [[TMP0]]) 784 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[A]], align 4 785 // CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP1]], [[CALL]] 786 // CHECK3-NEXT: store i32 [[ADD]], ptr [[A]], align 4 787 // CHECK3-NEXT: [[TMP2:%.*]] = load i32, ptr [[N_ADDR]], align 4 788 // CHECK3-NEXT: [[CALL1:%.*]] = call noundef i32 @_ZL7fstatici(i32 noundef [[TMP2]]) 789 // CHECK3-NEXT: [[TMP3:%.*]] = load i32, ptr [[A]], align 4 790 // CHECK3-NEXT: [[ADD2:%.*]] = add nsw i32 [[TMP3]], [[CALL1]] 791 // CHECK3-NEXT: store i32 [[ADD2]], ptr [[A]], align 4 792 // CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[N_ADDR]], align 4 793 // CHECK3-NEXT: [[CALL3:%.*]] = call noundef i32 @_Z9ftemplateIiET_i(i32 noundef [[TMP4]]) 794 // CHECK3-NEXT: [[TMP5:%.*]] = load i32, ptr [[A]], align 4 795 // CHECK3-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP5]], [[CALL3]] 796 // CHECK3-NEXT: store i32 [[ADD4]], ptr [[A]], align 4 797 // CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[A]], align 4 798 // CHECK3-NEXT: ret i32 [[TMP6]] 799 // 800 // 801 // CHECK3-LABEL: define {{[^@]+}}@_ZN2S12r1Ei 802 // CHECK3-SAME: (ptr noundef nonnull align 4 dereferenceable(8) [[THIS:%.*]], i32 noundef [[N:%.*]]) #[[ATTR0]] comdat align 2 { 803 // CHECK3-NEXT: entry: 804 // CHECK3-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 805 // CHECK3-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 806 // CHECK3-NEXT: [[B:%.*]] = alloca i32, align 4 807 // CHECK3-NEXT: [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4 808 // CHECK3-NEXT: [[B_CASTED:%.*]] = alloca i32, align 4 809 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i32, align 4 810 // CHECK3-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [3 x ptr], align 4 811 // CHECK3-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [3 x ptr], align 4 812 // CHECK3-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [3 x ptr], align 4 813 // CHECK3-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8 814 // CHECK3-NEXT: [[DOTOFFLOAD_BASEPTRS3:%.*]] = alloca [1 x ptr], align 4 815 // CHECK3-NEXT: [[DOTOFFLOAD_PTRS4:%.*]] = alloca [1 x ptr], align 4 816 // CHECK3-NEXT: [[DOTOFFLOAD_MAPPERS5:%.*]] = alloca [1 x ptr], align 4 817 // CHECK3-NEXT: [[KERNEL_ARGS6:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS]], align 8 818 // CHECK3-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 819 // CHECK3-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 820 // CHECK3-NEXT: [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 821 // CHECK3-NEXT: store i32 1, ptr [[B]], align 4 822 // CHECK3-NEXT: [[TMP0:%.*]] = load i32, ptr [[N_ADDR]], align 4 823 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[B]], align 4 824 // CHECK3-NEXT: [[SUB:%.*]] = sub nsw i32 [[TMP0]], [[TMP1]] 825 // CHECK3-NEXT: store i32 [[SUB]], ptr [[DOTCAPTURE_EXPR_]], align 4 826 // CHECK3-NEXT: [[TMP2:%.*]] = load i32, ptr [[B]], align 4 827 // CHECK3-NEXT: store i32 [[TMP2]], ptr [[B_CASTED]], align 4 828 // CHECK3-NEXT: [[TMP3:%.*]] = load i32, ptr [[B_CASTED]], align 4 829 // CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 830 // CHECK3-NEXT: store i32 [[TMP4]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 831 // CHECK3-NEXT: [[TMP5:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 832 // CHECK3-NEXT: [[A:%.*]] = getelementptr inbounds nuw [[STRUCT_S1:%.*]], ptr [[THIS1]], i32 0, i32 0 833 // CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 834 // CHECK3-NEXT: store ptr [[THIS1]], ptr [[TMP6]], align 4 835 // CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 836 // CHECK3-NEXT: store ptr [[A]], ptr [[TMP7]], align 4 837 // CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 0 838 // CHECK3-NEXT: store ptr null, ptr [[TMP8]], align 4 839 // CHECK3-NEXT: [[TMP9:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 1 840 // CHECK3-NEXT: store i32 [[TMP3]], ptr [[TMP9]], align 4 841 // CHECK3-NEXT: [[TMP10:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 1 842 // CHECK3-NEXT: store i32 [[TMP3]], ptr [[TMP10]], align 4 843 // CHECK3-NEXT: [[TMP11:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 1 844 // CHECK3-NEXT: store ptr null, ptr [[TMP11]], align 4 845 // CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 2 846 // CHECK3-NEXT: store i32 [[TMP5]], ptr [[TMP12]], align 4 847 // CHECK3-NEXT: [[TMP13:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 2 848 // CHECK3-NEXT: store i32 [[TMP5]], ptr [[TMP13]], align 4 849 // CHECK3-NEXT: [[TMP14:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 2 850 // CHECK3-NEXT: store ptr null, ptr [[TMP14]], align 4 851 // CHECK3-NEXT: [[TMP15:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 852 // CHECK3-NEXT: [[TMP16:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 853 // CHECK3-NEXT: [[TMP17:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 854 // CHECK3-NEXT: [[TMP18:%.*]] = insertvalue [3 x i32] zeroinitializer, i32 [[TMP17]], 0 855 // CHECK3-NEXT: [[TMP19:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 856 // CHECK3-NEXT: store i32 3, ptr [[TMP19]], align 4 857 // CHECK3-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1 858 // CHECK3-NEXT: store i32 3, ptr [[TMP20]], align 4 859 // CHECK3-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2 860 // CHECK3-NEXT: store ptr [[TMP15]], ptr [[TMP21]], align 4 861 // CHECK3-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3 862 // CHECK3-NEXT: store ptr [[TMP16]], ptr [[TMP22]], align 4 863 // CHECK3-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4 864 // CHECK3-NEXT: store ptr @.offload_sizes, ptr [[TMP23]], align 4 865 // CHECK3-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5 866 // CHECK3-NEXT: store ptr @.offload_maptypes, ptr [[TMP24]], align 4 867 // CHECK3-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6 868 // CHECK3-NEXT: store ptr null, ptr [[TMP25]], align 4 869 // CHECK3-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7 870 // CHECK3-NEXT: store ptr null, ptr [[TMP26]], align 4 871 // CHECK3-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8 872 // CHECK3-NEXT: store i64 0, ptr [[TMP27]], align 8 873 // CHECK3-NEXT: [[TMP28:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 9 874 // CHECK3-NEXT: store i64 0, ptr [[TMP28]], align 8 875 // CHECK3-NEXT: [[TMP29:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 10 876 // CHECK3-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP29]], align 4 877 // CHECK3-NEXT: [[TMP30:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 11 878 // CHECK3-NEXT: store [3 x i32] [[TMP18]], ptr [[TMP30]], align 4 879 // CHECK3-NEXT: [[TMP31:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 12 880 // CHECK3-NEXT: store i32 0, ptr [[TMP31]], align 4 881 // CHECK3-NEXT: [[TMP32:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1:[0-9]+]], i64 -1, i32 0, i32 [[TMP17]], ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121.region_id, ptr [[KERNEL_ARGS]]) 882 // CHECK3-NEXT: [[TMP33:%.*]] = icmp ne i32 [[TMP32]], 0 883 // CHECK3-NEXT: br i1 [[TMP33]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]] 884 // CHECK3: omp_offload.failed: 885 // CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121(ptr [[THIS1]], i32 [[TMP3]], i32 [[TMP5]]) #[[ATTR2:[0-9]+]] 886 // CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT]] 887 // CHECK3: omp_offload.cont: 888 // CHECK3-NEXT: [[A2:%.*]] = getelementptr inbounds nuw [[STRUCT_S1]], ptr [[THIS1]], i32 0, i32 0 889 // CHECK3-NEXT: [[TMP34:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS3]], i32 0, i32 0 890 // CHECK3-NEXT: store ptr [[THIS1]], ptr [[TMP34]], align 4 891 // CHECK3-NEXT: [[TMP35:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS4]], i32 0, i32 0 892 // CHECK3-NEXT: store ptr [[A2]], ptr [[TMP35]], align 4 893 // CHECK3-NEXT: [[TMP36:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS5]], i32 0, i32 0 894 // CHECK3-NEXT: store ptr null, ptr [[TMP36]], align 4 895 // CHECK3-NEXT: [[TMP37:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS3]], i32 0, i32 0 896 // CHECK3-NEXT: [[TMP38:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS4]], i32 0, i32 0 897 // CHECK3-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 0 898 // CHECK3-NEXT: store i32 3, ptr [[TMP39]], align 4 899 // CHECK3-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 1 900 // CHECK3-NEXT: store i32 1, ptr [[TMP40]], align 4 901 // CHECK3-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 2 902 // CHECK3-NEXT: store ptr [[TMP37]], ptr [[TMP41]], align 4 903 // CHECK3-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 3 904 // CHECK3-NEXT: store ptr [[TMP38]], ptr [[TMP42]], align 4 905 // CHECK3-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 4 906 // CHECK3-NEXT: store ptr @.offload_sizes.1, ptr [[TMP43]], align 4 907 // CHECK3-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 5 908 // CHECK3-NEXT: store ptr @.offload_maptypes.2, ptr [[TMP44]], align 4 909 // CHECK3-NEXT: [[TMP45:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 6 910 // CHECK3-NEXT: store ptr null, ptr [[TMP45]], align 4 911 // CHECK3-NEXT: [[TMP46:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 7 912 // CHECK3-NEXT: store ptr null, ptr [[TMP46]], align 4 913 // CHECK3-NEXT: [[TMP47:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 8 914 // CHECK3-NEXT: store i64 0, ptr [[TMP47]], align 8 915 // CHECK3-NEXT: [[TMP48:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 9 916 // CHECK3-NEXT: store i64 0, ptr [[TMP48]], align 8 917 // CHECK3-NEXT: [[TMP49:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 10 918 // CHECK3-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP49]], align 4 919 // CHECK3-NEXT: [[TMP50:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 11 920 // CHECK3-NEXT: store [3 x i32] [i32 1024, i32 0, i32 0], ptr [[TMP50]], align 4 921 // CHECK3-NEXT: [[TMP51:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS6]], i32 0, i32 12 922 // CHECK3-NEXT: store i32 0, ptr [[TMP51]], align 4 923 // CHECK3-NEXT: [[TMP52:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 0, i32 1024, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126.region_id, ptr [[KERNEL_ARGS6]]) 924 // CHECK3-NEXT: [[TMP53:%.*]] = icmp ne i32 [[TMP52]], 0 925 // CHECK3-NEXT: br i1 [[TMP53]], label [[OMP_OFFLOAD_FAILED7:%.*]], label [[OMP_OFFLOAD_CONT8:%.*]] 926 // CHECK3: omp_offload.failed7: 927 // CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126(ptr [[THIS1]]) #[[ATTR2]] 928 // CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT8]] 929 // CHECK3: omp_offload.cont8: 930 // CHECK3-NEXT: [[A9:%.*]] = getelementptr inbounds nuw [[STRUCT_S1]], ptr [[THIS1]], i32 0, i32 0 931 // CHECK3-NEXT: [[TMP54:%.*]] = load double, ptr [[A9]], align 4 932 // CHECK3-NEXT: [[CONV:%.*]] = fptosi double [[TMP54]] to i32 933 // CHECK3-NEXT: ret i32 [[CONV]] 934 // 935 // 936 // CHECK3-LABEL: define {{[^@]+}}@_ZL7fstatici 937 // CHECK3-SAME: (i32 noundef [[N:%.*]]) #[[ATTR0]] { 938 // CHECK3-NEXT: entry: 939 // CHECK3-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 940 // CHECK3-NEXT: [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4 941 // CHECK3-NEXT: [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4 942 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i32, align 4 943 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__CASTED2:%.*]] = alloca i32, align 4 944 // CHECK3-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [2 x ptr], align 4 945 // CHECK3-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [2 x ptr], align 4 946 // CHECK3-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [2 x ptr], align 4 947 // CHECK3-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8 948 // CHECK3-NEXT: [[DOTCAPTURE_EXPR_3:%.*]] = alloca i32, align 4 949 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__CASTED4:%.*]] = alloca i32, align 4 950 // CHECK3-NEXT: [[DOTOFFLOAD_BASEPTRS5:%.*]] = alloca [1 x ptr], align 4 951 // CHECK3-NEXT: [[DOTOFFLOAD_PTRS6:%.*]] = alloca [1 x ptr], align 4 952 // CHECK3-NEXT: [[DOTOFFLOAD_MAPPERS7:%.*]] = alloca [1 x ptr], align 4 953 // CHECK3-NEXT: [[KERNEL_ARGS8:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS]], align 8 954 // CHECK3-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 955 // CHECK3-NEXT: [[TMP0:%.*]] = load i32, ptr [[N_ADDR]], align 4 956 // CHECK3-NEXT: store i32 [[TMP0]], ptr [[DOTCAPTURE_EXPR_]], align 4 957 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[N_ADDR]], align 4 958 // CHECK3-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP1]], 32 959 // CHECK3-NEXT: store i32 [[MUL]], ptr [[DOTCAPTURE_EXPR_1]], align 4 960 // CHECK3-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 961 // CHECK3-NEXT: store i32 [[TMP2]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 962 // CHECK3-NEXT: [[TMP3:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 963 // CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4 964 // CHECK3-NEXT: store i32 [[TMP4]], ptr [[DOTCAPTURE_EXPR__CASTED2]], align 4 965 // CHECK3-NEXT: [[TMP5:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__CASTED2]], align 4 966 // CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 967 // CHECK3-NEXT: store i32 [[TMP3]], ptr [[TMP6]], align 4 968 // CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 969 // CHECK3-NEXT: store i32 [[TMP3]], ptr [[TMP7]], align 4 970 // CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 0 971 // CHECK3-NEXT: store ptr null, ptr [[TMP8]], align 4 972 // CHECK3-NEXT: [[TMP9:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 1 973 // CHECK3-NEXT: store i32 [[TMP5]], ptr [[TMP9]], align 4 974 // CHECK3-NEXT: [[TMP10:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 1 975 // CHECK3-NEXT: store i32 [[TMP5]], ptr [[TMP10]], align 4 976 // CHECK3-NEXT: [[TMP11:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 1 977 // CHECK3-NEXT: store ptr null, ptr [[TMP11]], align 4 978 // CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 979 // CHECK3-NEXT: [[TMP13:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 980 // CHECK3-NEXT: [[TMP14:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_]], align 4 981 // CHECK3-NEXT: [[TMP15:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4 982 // CHECK3-NEXT: [[TMP16:%.*]] = insertvalue [3 x i32] zeroinitializer, i32 [[TMP14]], 0 983 // CHECK3-NEXT: [[TMP17:%.*]] = insertvalue [3 x i32] zeroinitializer, i32 [[TMP15]], 0 984 // CHECK3-NEXT: [[TMP18:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 985 // CHECK3-NEXT: store i32 3, ptr [[TMP18]], align 4 986 // CHECK3-NEXT: [[TMP19:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1 987 // CHECK3-NEXT: store i32 2, ptr [[TMP19]], align 4 988 // CHECK3-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2 989 // CHECK3-NEXT: store ptr [[TMP12]], ptr [[TMP20]], align 4 990 // CHECK3-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3 991 // CHECK3-NEXT: store ptr [[TMP13]], ptr [[TMP21]], align 4 992 // CHECK3-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4 993 // CHECK3-NEXT: store ptr @.offload_sizes.3, ptr [[TMP22]], align 4 994 // CHECK3-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5 995 // CHECK3-NEXT: store ptr @.offload_maptypes.4, ptr [[TMP23]], align 4 996 // CHECK3-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6 997 // CHECK3-NEXT: store ptr null, ptr [[TMP24]], align 4 998 // CHECK3-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7 999 // CHECK3-NEXT: store ptr null, ptr [[TMP25]], align 4 1000 // CHECK3-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8 1001 // CHECK3-NEXT: store i64 0, ptr [[TMP26]], align 8 1002 // CHECK3-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 9 1003 // CHECK3-NEXT: store i64 0, ptr [[TMP27]], align 8 1004 // CHECK3-NEXT: [[TMP28:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 10 1005 // CHECK3-NEXT: store [3 x i32] [[TMP16]], ptr [[TMP28]], align 4 1006 // CHECK3-NEXT: [[TMP29:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 11 1007 // CHECK3-NEXT: store [3 x i32] [[TMP17]], ptr [[TMP29]], align 4 1008 // CHECK3-NEXT: [[TMP30:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 12 1009 // CHECK3-NEXT: store i32 0, ptr [[TMP30]], align 4 1010 // CHECK3-NEXT: [[TMP31:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 [[TMP14]], i32 [[TMP15]], ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104.region_id, ptr [[KERNEL_ARGS]]) 1011 // CHECK3-NEXT: [[TMP32:%.*]] = icmp ne i32 [[TMP31]], 0 1012 // CHECK3-NEXT: br i1 [[TMP32]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]] 1013 // CHECK3: omp_offload.failed: 1014 // CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104(i32 [[TMP3]], i32 [[TMP5]]) #[[ATTR2]] 1015 // CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT]] 1016 // CHECK3: omp_offload.cont: 1017 // CHECK3-NEXT: [[TMP33:%.*]] = load i32, ptr [[N_ADDR]], align 4 1018 // CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 32, [[TMP33]] 1019 // CHECK3-NEXT: store i32 [[ADD]], ptr [[DOTCAPTURE_EXPR_3]], align 4 1020 // CHECK3-NEXT: [[TMP34:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_3]], align 4 1021 // CHECK3-NEXT: store i32 [[TMP34]], ptr [[DOTCAPTURE_EXPR__CASTED4]], align 4 1022 // CHECK3-NEXT: [[TMP35:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__CASTED4]], align 4 1023 // CHECK3-NEXT: [[TMP36:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS5]], i32 0, i32 0 1024 // CHECK3-NEXT: store i32 [[TMP35]], ptr [[TMP36]], align 4 1025 // CHECK3-NEXT: [[TMP37:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS6]], i32 0, i32 0 1026 // CHECK3-NEXT: store i32 [[TMP35]], ptr [[TMP37]], align 4 1027 // CHECK3-NEXT: [[TMP38:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS7]], i32 0, i32 0 1028 // CHECK3-NEXT: store ptr null, ptr [[TMP38]], align 4 1029 // CHECK3-NEXT: [[TMP39:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS5]], i32 0, i32 0 1030 // CHECK3-NEXT: [[TMP40:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS6]], i32 0, i32 0 1031 // CHECK3-NEXT: [[TMP41:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_3]], align 4 1032 // CHECK3-NEXT: [[TMP42:%.*]] = insertvalue [3 x i32] zeroinitializer, i32 [[TMP41]], 0 1033 // CHECK3-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 0 1034 // CHECK3-NEXT: store i32 3, ptr [[TMP43]], align 4 1035 // CHECK3-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 1 1036 // CHECK3-NEXT: store i32 1, ptr [[TMP44]], align 4 1037 // CHECK3-NEXT: [[TMP45:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 2 1038 // CHECK3-NEXT: store ptr [[TMP39]], ptr [[TMP45]], align 4 1039 // CHECK3-NEXT: [[TMP46:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 3 1040 // CHECK3-NEXT: store ptr [[TMP40]], ptr [[TMP46]], align 4 1041 // CHECK3-NEXT: [[TMP47:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 4 1042 // CHECK3-NEXT: store ptr @.offload_sizes.5, ptr [[TMP47]], align 4 1043 // CHECK3-NEXT: [[TMP48:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 5 1044 // CHECK3-NEXT: store ptr @.offload_maptypes.6, ptr [[TMP48]], align 4 1045 // CHECK3-NEXT: [[TMP49:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 6 1046 // CHECK3-NEXT: store ptr null, ptr [[TMP49]], align 4 1047 // CHECK3-NEXT: [[TMP50:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 7 1048 // CHECK3-NEXT: store ptr null, ptr [[TMP50]], align 4 1049 // CHECK3-NEXT: [[TMP51:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 8 1050 // CHECK3-NEXT: store i64 0, ptr [[TMP51]], align 8 1051 // CHECK3-NEXT: [[TMP52:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 9 1052 // CHECK3-NEXT: store i64 0, ptr [[TMP52]], align 8 1053 // CHECK3-NEXT: [[TMP53:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 10 1054 // CHECK3-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP53]], align 4 1055 // CHECK3-NEXT: [[TMP54:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 11 1056 // CHECK3-NEXT: store [3 x i32] [[TMP42]], ptr [[TMP54]], align 4 1057 // CHECK3-NEXT: [[TMP55:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS8]], i32 0, i32 12 1058 // CHECK3-NEXT: store i32 0, ptr [[TMP55]], align 4 1059 // CHECK3-NEXT: [[TMP56:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 0, i32 [[TMP41]], ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108.region_id, ptr [[KERNEL_ARGS8]]) 1060 // CHECK3-NEXT: [[TMP57:%.*]] = icmp ne i32 [[TMP56]], 0 1061 // CHECK3-NEXT: br i1 [[TMP57]], label [[OMP_OFFLOAD_FAILED9:%.*]], label [[OMP_OFFLOAD_CONT10:%.*]] 1062 // CHECK3: omp_offload.failed9: 1063 // CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108(i32 [[TMP35]]) #[[ATTR2]] 1064 // CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT10]] 1065 // CHECK3: omp_offload.cont10: 1066 // CHECK3-NEXT: [[TMP58:%.*]] = load i32, ptr [[N_ADDR]], align 4 1067 // CHECK3-NEXT: [[ADD11:%.*]] = add nsw i32 [[TMP58]], 1 1068 // CHECK3-NEXT: ret i32 [[ADD11]] 1069 // 1070 // 1071 // CHECK3-LABEL: define {{[^@]+}}@_Z9ftemplateIiET_i 1072 // CHECK3-SAME: (i32 noundef [[N:%.*]]) #[[ATTR0]] comdat { 1073 // CHECK3-NEXT: entry: 1074 // CHECK3-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4 1075 // CHECK3-NEXT: [[A:%.*]] = alloca i32, align 4 1076 // CHECK3-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8 1077 // CHECK3-NEXT: [[B:%.*]] = alloca i16, align 2 1078 // CHECK3-NEXT: [[DOTCAPTURE_EXPR_:%.*]] = alloca i16, align 2 1079 // CHECK3-NEXT: [[A_CASTED:%.*]] = alloca i32, align 4 1080 // CHECK3-NEXT: [[B_CASTED:%.*]] = alloca i32, align 4 1081 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i32, align 4 1082 // CHECK3-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [3 x ptr], align 4 1083 // CHECK3-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [3 x ptr], align 4 1084 // CHECK3-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [3 x ptr], align 4 1085 // CHECK3-NEXT: [[KERNEL_ARGS1:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS]], align 8 1086 // CHECK3-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4 1087 // CHECK3-NEXT: store i32 0, ptr [[A]], align 4 1088 // CHECK3-NEXT: [[TMP0:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0 1089 // CHECK3-NEXT: store i32 3, ptr [[TMP0]], align 4 1090 // CHECK3-NEXT: [[TMP1:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1 1091 // CHECK3-NEXT: store i32 0, ptr [[TMP1]], align 4 1092 // CHECK3-NEXT: [[TMP2:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2 1093 // CHECK3-NEXT: store ptr null, ptr [[TMP2]], align 4 1094 // CHECK3-NEXT: [[TMP3:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3 1095 // CHECK3-NEXT: store ptr null, ptr [[TMP3]], align 4 1096 // CHECK3-NEXT: [[TMP4:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4 1097 // CHECK3-NEXT: store ptr null, ptr [[TMP4]], align 4 1098 // CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5 1099 // CHECK3-NEXT: store ptr null, ptr [[TMP5]], align 4 1100 // CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6 1101 // CHECK3-NEXT: store ptr null, ptr [[TMP6]], align 4 1102 // CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7 1103 // CHECK3-NEXT: store ptr null, ptr [[TMP7]], align 4 1104 // CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8 1105 // CHECK3-NEXT: store i64 0, ptr [[TMP8]], align 8 1106 // CHECK3-NEXT: [[TMP9:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 9 1107 // CHECK3-NEXT: store i64 0, ptr [[TMP9]], align 8 1108 // CHECK3-NEXT: [[TMP10:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 10 1109 // CHECK3-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP10]], align 4 1110 // CHECK3-NEXT: [[TMP11:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 11 1111 // CHECK3-NEXT: store [3 x i32] [i32 20, i32 0, i32 0], ptr [[TMP11]], align 4 1112 // CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 12 1113 // CHECK3-NEXT: store i32 0, ptr [[TMP12]], align 4 1114 // CHECK3-NEXT: [[TMP13:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 0, i32 20, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88.region_id, ptr [[KERNEL_ARGS]]) 1115 // CHECK3-NEXT: [[TMP14:%.*]] = icmp ne i32 [[TMP13]], 0 1116 // CHECK3-NEXT: br i1 [[TMP14]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]] 1117 // CHECK3: omp_offload.failed: 1118 // CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88() #[[ATTR2]] 1119 // CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT]] 1120 // CHECK3: omp_offload.cont: 1121 // CHECK3-NEXT: store i16 1, ptr [[B]], align 2 1122 // CHECK3-NEXT: [[TMP15:%.*]] = load i16, ptr [[B]], align 2 1123 // CHECK3-NEXT: store i16 [[TMP15]], ptr [[DOTCAPTURE_EXPR_]], align 2 1124 // CHECK3-NEXT: [[TMP16:%.*]] = load i32, ptr [[A]], align 4 1125 // CHECK3-NEXT: store i32 [[TMP16]], ptr [[A_CASTED]], align 4 1126 // CHECK3-NEXT: [[TMP17:%.*]] = load i32, ptr [[A_CASTED]], align 4 1127 // CHECK3-NEXT: [[TMP18:%.*]] = load i16, ptr [[B]], align 2 1128 // CHECK3-NEXT: store i16 [[TMP18]], ptr [[B_CASTED]], align 2 1129 // CHECK3-NEXT: [[TMP19:%.*]] = load i32, ptr [[B_CASTED]], align 4 1130 // CHECK3-NEXT: [[TMP20:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR_]], align 2 1131 // CHECK3-NEXT: store i16 [[TMP20]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 2 1132 // CHECK3-NEXT: [[TMP21:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__CASTED]], align 4 1133 // CHECK3-NEXT: [[TMP22:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 1134 // CHECK3-NEXT: store i32 [[TMP17]], ptr [[TMP22]], align 4 1135 // CHECK3-NEXT: [[TMP23:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 1136 // CHECK3-NEXT: store i32 [[TMP17]], ptr [[TMP23]], align 4 1137 // CHECK3-NEXT: [[TMP24:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 0 1138 // CHECK3-NEXT: store ptr null, ptr [[TMP24]], align 4 1139 // CHECK3-NEXT: [[TMP25:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 1 1140 // CHECK3-NEXT: store i32 [[TMP19]], ptr [[TMP25]], align 4 1141 // CHECK3-NEXT: [[TMP26:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 1 1142 // CHECK3-NEXT: store i32 [[TMP19]], ptr [[TMP26]], align 4 1143 // CHECK3-NEXT: [[TMP27:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 1 1144 // CHECK3-NEXT: store ptr null, ptr [[TMP27]], align 4 1145 // CHECK3-NEXT: [[TMP28:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 2 1146 // CHECK3-NEXT: store i32 [[TMP21]], ptr [[TMP28]], align 4 1147 // CHECK3-NEXT: [[TMP29:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 2 1148 // CHECK3-NEXT: store i32 [[TMP21]], ptr [[TMP29]], align 4 1149 // CHECK3-NEXT: [[TMP30:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 2 1150 // CHECK3-NEXT: store ptr null, ptr [[TMP30]], align 4 1151 // CHECK3-NEXT: [[TMP31:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0 1152 // CHECK3-NEXT: [[TMP32:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0 1153 // CHECK3-NEXT: [[TMP33:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR_]], align 2 1154 // CHECK3-NEXT: [[TMP34:%.*]] = sext i16 [[TMP33]] to i32 1155 // CHECK3-NEXT: [[TMP35:%.*]] = insertvalue [3 x i32] zeroinitializer, i32 [[TMP34]], 0 1156 // CHECK3-NEXT: [[TMP36:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 0 1157 // CHECK3-NEXT: store i32 3, ptr [[TMP36]], align 4 1158 // CHECK3-NEXT: [[TMP37:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 1 1159 // CHECK3-NEXT: store i32 3, ptr [[TMP37]], align 4 1160 // CHECK3-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 2 1161 // CHECK3-NEXT: store ptr [[TMP31]], ptr [[TMP38]], align 4 1162 // CHECK3-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 3 1163 // CHECK3-NEXT: store ptr [[TMP32]], ptr [[TMP39]], align 4 1164 // CHECK3-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 4 1165 // CHECK3-NEXT: store ptr @.offload_sizes.7, ptr [[TMP40]], align 4 1166 // CHECK3-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 5 1167 // CHECK3-NEXT: store ptr @.offload_maptypes.8, ptr [[TMP41]], align 4 1168 // CHECK3-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 6 1169 // CHECK3-NEXT: store ptr null, ptr [[TMP42]], align 4 1170 // CHECK3-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 7 1171 // CHECK3-NEXT: store ptr null, ptr [[TMP43]], align 4 1172 // CHECK3-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 8 1173 // CHECK3-NEXT: store i64 0, ptr [[TMP44]], align 8 1174 // CHECK3-NEXT: [[TMP45:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 9 1175 // CHECK3-NEXT: store i64 0, ptr [[TMP45]], align 8 1176 // CHECK3-NEXT: [[TMP46:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 10 1177 // CHECK3-NEXT: store [3 x i32] [[TMP35]], ptr [[TMP46]], align 4 1178 // CHECK3-NEXT: [[TMP47:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 11 1179 // CHECK3-NEXT: store [3 x i32] [i32 1024, i32 0, i32 0], ptr [[TMP47]], align 4 1180 // CHECK3-NEXT: [[TMP48:%.*]] = getelementptr inbounds nuw [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS1]], i32 0, i32 12 1181 // CHECK3-NEXT: store i32 0, ptr [[TMP48]], align 4 1182 // CHECK3-NEXT: [[TMP49:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB1]], i64 -1, i32 [[TMP34]], i32 1024, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93.region_id, ptr [[KERNEL_ARGS1]]) 1183 // CHECK3-NEXT: [[TMP50:%.*]] = icmp ne i32 [[TMP49]], 0 1184 // CHECK3-NEXT: br i1 [[TMP50]], label [[OMP_OFFLOAD_FAILED2:%.*]], label [[OMP_OFFLOAD_CONT3:%.*]] 1185 // CHECK3: omp_offload.failed2: 1186 // CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93(i32 [[TMP17]], i32 [[TMP19]], i32 [[TMP21]]) #[[ATTR2]] 1187 // CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT3]] 1188 // CHECK3: omp_offload.cont3: 1189 // CHECK3-NEXT: [[TMP51:%.*]] = load i32, ptr [[A]], align 4 1190 // CHECK3-NEXT: ret i32 [[TMP51]] 1191 // 1192 // 1193 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121 1194 // CHECK3-SAME: (ptr noundef [[THIS:%.*]], i32 noundef [[B:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1:[0-9]+]] { 1195 // CHECK3-NEXT: entry: 1196 // CHECK3-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1197 // CHECK3-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1198 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = 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 ptr [[THIS]], ptr [[THIS_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 ptr, ptr [[THIS_ADDR]], align 4 1205 // CHECK3-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1206 // CHECK3-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 [[TMP2]]) 1207 // CHECK3-NEXT: [[TMP3:%.*]] = load i32, ptr [[B_ADDR]], align 4 1208 // CHECK3-NEXT: store i32 [[TMP3]], ptr [[B_CASTED]], align 4 1209 // CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[B_CASTED]], align 4 1210 // CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121.omp_outlined, ptr [[TMP1]], i32 [[TMP4]]) 1211 // CHECK3-NEXT: ret void 1212 // 1213 // 1214 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121.omp_outlined 1215 // CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]], i32 noundef [[B:%.*]]) #[[ATTR1]] { 1216 // CHECK3-NEXT: entry: 1217 // CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1218 // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1219 // CHECK3-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1220 // CHECK3-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1221 // CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1222 // CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1223 // CHECK3-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1224 // CHECK3-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1225 // CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1226 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[B_ADDR]], align 4 1227 // CHECK3-NEXT: [[CONV:%.*]] = sitofp i32 [[TMP1]] to double 1228 // CHECK3-NEXT: [[ADD:%.*]] = fadd double [[CONV]], 1.500000e+00 1229 // CHECK3-NEXT: [[A:%.*]] = getelementptr inbounds nuw [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 1230 // CHECK3-NEXT: store double [[ADD]], ptr [[A]], align 4 1231 // CHECK3-NEXT: ret void 1232 // 1233 // 1234 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126 1235 // CHECK3-SAME: (ptr noundef [[THIS:%.*]]) #[[ATTR1]] { 1236 // CHECK3-NEXT: entry: 1237 // CHECK3-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1238 // CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1239 // CHECK3-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1240 // CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1241 // CHECK3-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 1024) 1242 // CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126.omp_outlined, ptr [[TMP1]]) 1243 // CHECK3-NEXT: ret void 1244 // 1245 // 1246 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126.omp_outlined 1247 // CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]]) #[[ATTR1]] { 1248 // CHECK3-NEXT: entry: 1249 // CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1250 // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1251 // CHECK3-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1252 // CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1253 // CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1254 // CHECK3-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1255 // CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1256 // CHECK3-NEXT: [[A:%.*]] = getelementptr inbounds nuw [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 1257 // CHECK3-NEXT: store double 2.500000e+00, ptr [[A]], align 4 1258 // CHECK3-NEXT: ret void 1259 // 1260 // 1261 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104 1262 // CHECK3-SAME: (i32 noundef [[DOTCAPTURE_EXPR_:%.*]], i32 noundef [[DOTCAPTURE_EXPR_1:%.*]]) #[[ATTR1]] { 1263 // CHECK3-NEXT: entry: 1264 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1265 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__ADDR2:%.*]] = alloca i32, align 4 1266 // CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1267 // CHECK3-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1268 // CHECK3-NEXT: store i32 [[DOTCAPTURE_EXPR_1]], ptr [[DOTCAPTURE_EXPR__ADDR2]], align 4 1269 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1270 // CHECK3-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR2]], align 4 1271 // CHECK3-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP1]], i32 [[TMP2]]) 1272 // CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104.omp_outlined) 1273 // CHECK3-NEXT: ret void 1274 // 1275 // 1276 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104.omp_outlined 1277 // CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] { 1278 // CHECK3-NEXT: entry: 1279 // CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1280 // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1281 // CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1282 // CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1283 // CHECK3-NEXT: ret void 1284 // 1285 // 1286 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108 1287 // CHECK3-SAME: (i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] { 1288 // CHECK3-NEXT: entry: 1289 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1290 // CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1291 // CHECK3-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1292 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1293 // CHECK3-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 [[TMP1]]) 1294 // CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108.omp_outlined) 1295 // CHECK3-NEXT: ret void 1296 // 1297 // 1298 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108.omp_outlined 1299 // CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] { 1300 // CHECK3-NEXT: entry: 1301 // CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1302 // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1303 // CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1304 // CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1305 // CHECK3-NEXT: ret void 1306 // 1307 // 1308 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88 1309 // CHECK3-SAME: () #[[ATTR1]] { 1310 // CHECK3-NEXT: entry: 1311 // CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1312 // CHECK3-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 20) 1313 // CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88.omp_outlined) 1314 // CHECK3-NEXT: ret void 1315 // 1316 // 1317 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88.omp_outlined 1318 // CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] { 1319 // CHECK3-NEXT: entry: 1320 // CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1321 // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1322 // CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1323 // CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1324 // CHECK3-NEXT: ret void 1325 // 1326 // 1327 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93 1328 // CHECK3-SAME: (i32 noundef [[A:%.*]], i32 noundef [[B:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] { 1329 // CHECK3-NEXT: entry: 1330 // CHECK3-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4 1331 // CHECK3-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1332 // CHECK3-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1333 // CHECK3-NEXT: [[A_CASTED:%.*]] = alloca i32, align 4 1334 // CHECK3-NEXT: [[B_CASTED:%.*]] = alloca i32, align 4 1335 // CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1336 // CHECK3-NEXT: store i32 [[A]], ptr [[A_ADDR]], align 4 1337 // CHECK3-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1338 // CHECK3-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1339 // CHECK3-NEXT: [[TMP1:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR__ADDR]], align 2 1340 // CHECK3-NEXT: [[TMP2:%.*]] = sext i16 [[TMP1]] to i32 1341 // CHECK3-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP2]], i32 1024) 1342 // CHECK3-NEXT: [[TMP3:%.*]] = load i32, ptr [[A_ADDR]], align 4 1343 // CHECK3-NEXT: store i32 [[TMP3]], ptr [[A_CASTED]], align 4 1344 // CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[A_CASTED]], align 4 1345 // CHECK3-NEXT: [[TMP5:%.*]] = load i16, ptr [[B_ADDR]], align 2 1346 // CHECK3-NEXT: store i16 [[TMP5]], ptr [[B_CASTED]], align 2 1347 // CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[B_CASTED]], align 4 1348 // CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93.omp_outlined, i32 [[TMP4]], i32 [[TMP6]]) 1349 // CHECK3-NEXT: ret void 1350 // 1351 // 1352 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93.omp_outlined 1353 // CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i32 noundef [[A:%.*]], i32 noundef [[B:%.*]]) #[[ATTR1]] { 1354 // CHECK3-NEXT: entry: 1355 // CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1356 // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1357 // CHECK3-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4 1358 // CHECK3-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1359 // CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1360 // CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1361 // CHECK3-NEXT: store i32 [[A]], ptr [[A_ADDR]], align 4 1362 // CHECK3-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1363 // CHECK3-NEXT: [[TMP0:%.*]] = load i16, ptr [[B_ADDR]], align 2 1364 // CHECK3-NEXT: [[CONV:%.*]] = sext i16 [[TMP0]] to i32 1365 // CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[A_ADDR]], align 4 1366 // CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP1]], [[CONV]] 1367 // CHECK3-NEXT: store i32 [[ADD]], ptr [[A_ADDR]], align 4 1368 // CHECK3-NEXT: ret void 1369 // 1370 // 1371 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104 1372 // CHECK9-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]], i64 noundef [[DOTCAPTURE_EXPR_1:%.*]]) #[[ATTR0:[0-9]+]] { 1373 // CHECK9-NEXT: entry: 1374 // CHECK9-NEXT: [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 8 1375 // CHECK9-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 1376 // CHECK9-NEXT: [[DOTCAPTURE_EXPR__ADDR2:%.*]] = alloca i64, align 8 1377 // CHECK9-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1:[0-9]+]]) 1378 // CHECK9-NEXT: store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR]], align 8 1379 // CHECK9-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 1380 // CHECK9-NEXT: store i64 [[DOTCAPTURE_EXPR_1]], ptr [[DOTCAPTURE_EXPR__ADDR2]], align 8 1381 // CHECK9-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1382 // CHECK9-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR2]], align 4 1383 // CHECK9-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP1]], i32 [[TMP2]]) 1384 // CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104.omp_outlined) 1385 // CHECK9-NEXT: ret void 1386 // 1387 // 1388 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104.omp_outlined 1389 // CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { 1390 // CHECK9-NEXT: entry: 1391 // CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 1392 // CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 1393 // CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 1394 // CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 1395 // CHECK9-NEXT: ret void 1396 // 1397 // 1398 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108 1399 // CHECK9-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0]] { 1400 // CHECK9-NEXT: entry: 1401 // CHECK9-NEXT: [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 8 1402 // CHECK9-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 1403 // CHECK9-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1404 // CHECK9-NEXT: store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR]], align 8 1405 // CHECK9-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 1406 // CHECK9-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1407 // CHECK9-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 [[TMP1]]) 1408 // CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108.omp_outlined) 1409 // CHECK9-NEXT: ret void 1410 // 1411 // 1412 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108.omp_outlined 1413 // CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { 1414 // CHECK9-NEXT: entry: 1415 // CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 1416 // CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 1417 // CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 1418 // CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 1419 // CHECK9-NEXT: ret void 1420 // 1421 // 1422 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121 1423 // CHECK9-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], ptr noundef [[THIS:%.*]], i64 noundef [[B:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0]] { 1424 // CHECK9-NEXT: entry: 1425 // CHECK9-NEXT: [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 8 1426 // CHECK9-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 1427 // CHECK9-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 1428 // CHECK9-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 1429 // CHECK9-NEXT: [[B_CASTED:%.*]] = alloca i64, align 8 1430 // CHECK9-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1431 // CHECK9-NEXT: store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR]], align 8 1432 // CHECK9-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 1433 // CHECK9-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 1434 // CHECK9-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 1435 // CHECK9-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 1436 // CHECK9-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1437 // CHECK9-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 [[TMP2]]) 1438 // CHECK9-NEXT: [[TMP3:%.*]] = load i32, ptr [[B_ADDR]], align 4 1439 // CHECK9-NEXT: store i32 [[TMP3]], ptr [[B_CASTED]], align 4 1440 // CHECK9-NEXT: [[TMP4:%.*]] = load i64, ptr [[B_CASTED]], align 8 1441 // CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121.omp_outlined, ptr [[TMP1]], i64 [[TMP4]]) 1442 // CHECK9-NEXT: ret void 1443 // 1444 // 1445 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121.omp_outlined 1446 // CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]], i64 noundef [[B:%.*]]) #[[ATTR0]] { 1447 // CHECK9-NEXT: entry: 1448 // CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 1449 // CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 1450 // CHECK9-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 1451 // CHECK9-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 1452 // CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 1453 // CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 1454 // CHECK9-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 1455 // CHECK9-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 1456 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 1457 // CHECK9-NEXT: [[TMP1:%.*]] = load i32, ptr [[B_ADDR]], align 4 1458 // CHECK9-NEXT: [[CONV:%.*]] = sitofp i32 [[TMP1]] to double 1459 // CHECK9-NEXT: [[ADD:%.*]] = fadd double [[CONV]], 1.500000e+00 1460 // CHECK9-NEXT: [[A:%.*]] = getelementptr inbounds nuw [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 1461 // CHECK9-NEXT: store double [[ADD]], ptr [[A]], align 8 1462 // CHECK9-NEXT: ret void 1463 // 1464 // 1465 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126 1466 // CHECK9-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], ptr noundef [[THIS:%.*]]) #[[ATTR0]] { 1467 // CHECK9-NEXT: entry: 1468 // CHECK9-NEXT: [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 8 1469 // CHECK9-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 1470 // CHECK9-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1471 // CHECK9-NEXT: store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR]], align 8 1472 // CHECK9-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 1473 // CHECK9-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 1474 // CHECK9-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 1024) 1475 // CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126.omp_outlined, ptr [[TMP1]]) 1476 // CHECK9-NEXT: ret void 1477 // 1478 // 1479 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126.omp_outlined 1480 // CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]]) #[[ATTR0]] { 1481 // CHECK9-NEXT: entry: 1482 // CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 1483 // CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 1484 // CHECK9-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 8 1485 // CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 1486 // CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 1487 // CHECK9-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8 1488 // CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8 1489 // CHECK9-NEXT: [[A:%.*]] = getelementptr inbounds nuw [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 1490 // CHECK9-NEXT: store double 2.500000e+00, ptr [[A]], align 8 1491 // CHECK9-NEXT: ret void 1492 // 1493 // 1494 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88 1495 // CHECK9-SAME: (ptr noalias noundef [[DYN_PTR:%.*]]) #[[ATTR0]] { 1496 // CHECK9-NEXT: entry: 1497 // CHECK9-NEXT: [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 8 1498 // CHECK9-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1499 // CHECK9-NEXT: store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR]], align 8 1500 // CHECK9-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 20) 1501 // CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88.omp_outlined) 1502 // CHECK9-NEXT: ret void 1503 // 1504 // 1505 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88.omp_outlined 1506 // CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { 1507 // CHECK9-NEXT: entry: 1508 // CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 1509 // CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 1510 // CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 1511 // CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 1512 // CHECK9-NEXT: ret void 1513 // 1514 // 1515 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93 1516 // CHECK9-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], i64 noundef [[A:%.*]], i64 noundef [[B:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0]] { 1517 // CHECK9-NEXT: entry: 1518 // CHECK9-NEXT: [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 8 1519 // CHECK9-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8 1520 // CHECK9-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 1521 // CHECK9-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8 1522 // CHECK9-NEXT: [[A_CASTED:%.*]] = alloca i64, align 8 1523 // CHECK9-NEXT: [[B_CASTED:%.*]] = alloca i64, align 8 1524 // CHECK9-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1525 // CHECK9-NEXT: store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR]], align 8 1526 // CHECK9-NEXT: store i64 [[A]], ptr [[A_ADDR]], align 8 1527 // CHECK9-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 1528 // CHECK9-NEXT: store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8 1529 // CHECK9-NEXT: [[TMP1:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR__ADDR]], align 2 1530 // CHECK9-NEXT: [[TMP2:%.*]] = sext i16 [[TMP1]] to i32 1531 // CHECK9-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP2]], i32 1024) 1532 // CHECK9-NEXT: [[TMP3:%.*]] = load i32, ptr [[A_ADDR]], align 4 1533 // CHECK9-NEXT: store i32 [[TMP3]], ptr [[A_CASTED]], align 4 1534 // CHECK9-NEXT: [[TMP4:%.*]] = load i64, ptr [[A_CASTED]], align 8 1535 // CHECK9-NEXT: [[TMP5:%.*]] = load i16, ptr [[B_ADDR]], align 2 1536 // CHECK9-NEXT: store i16 [[TMP5]], ptr [[B_CASTED]], align 2 1537 // CHECK9-NEXT: [[TMP6:%.*]] = load i64, ptr [[B_CASTED]], align 8 1538 // CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93.omp_outlined, i64 [[TMP4]], i64 [[TMP6]]) 1539 // CHECK9-NEXT: ret void 1540 // 1541 // 1542 // CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93.omp_outlined 1543 // CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i64 noundef [[A:%.*]], i64 noundef [[B:%.*]]) #[[ATTR0]] { 1544 // CHECK9-NEXT: entry: 1545 // CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8 1546 // CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8 1547 // CHECK9-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8 1548 // CHECK9-NEXT: [[B_ADDR:%.*]] = alloca i64, align 8 1549 // CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8 1550 // CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8 1551 // CHECK9-NEXT: store i64 [[A]], ptr [[A_ADDR]], align 8 1552 // CHECK9-NEXT: store i64 [[B]], ptr [[B_ADDR]], align 8 1553 // CHECK9-NEXT: [[TMP0:%.*]] = load i16, ptr [[B_ADDR]], align 2 1554 // CHECK9-NEXT: [[CONV:%.*]] = sext i16 [[TMP0]] to i32 1555 // CHECK9-NEXT: [[TMP1:%.*]] = load i32, ptr [[A_ADDR]], align 4 1556 // CHECK9-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP1]], [[CONV]] 1557 // CHECK9-NEXT: store i32 [[ADD]], ptr [[A_ADDR]], align 4 1558 // CHECK9-NEXT: ret void 1559 // 1560 // 1561 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104 1562 // CHECK11-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]], i32 noundef [[DOTCAPTURE_EXPR_1:%.*]]) #[[ATTR0:[0-9]+]] { 1563 // CHECK11-NEXT: entry: 1564 // CHECK11-NEXT: [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 4 1565 // CHECK11-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1566 // CHECK11-NEXT: [[DOTCAPTURE_EXPR__ADDR2:%.*]] = alloca i32, align 4 1567 // CHECK11-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1:[0-9]+]]) 1568 // CHECK11-NEXT: store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR]], align 4 1569 // CHECK11-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1570 // CHECK11-NEXT: store i32 [[DOTCAPTURE_EXPR_1]], ptr [[DOTCAPTURE_EXPR__ADDR2]], align 4 1571 // CHECK11-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1572 // CHECK11-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR2]], align 4 1573 // CHECK11-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP1]], i32 [[TMP2]]) 1574 // CHECK11-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104.omp_outlined) 1575 // CHECK11-NEXT: ret void 1576 // 1577 // 1578 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l104.omp_outlined 1579 // CHECK11-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { 1580 // CHECK11-NEXT: entry: 1581 // CHECK11-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1582 // CHECK11-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1583 // CHECK11-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1584 // CHECK11-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1585 // CHECK11-NEXT: ret void 1586 // 1587 // 1588 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108 1589 // CHECK11-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0]] { 1590 // CHECK11-NEXT: entry: 1591 // CHECK11-NEXT: [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 4 1592 // CHECK11-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1593 // CHECK11-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1594 // CHECK11-NEXT: store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR]], align 4 1595 // CHECK11-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1596 // CHECK11-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1597 // CHECK11-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 [[TMP1]]) 1598 // CHECK11-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108.omp_outlined) 1599 // CHECK11-NEXT: ret void 1600 // 1601 // 1602 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZL7fstatici_l108.omp_outlined 1603 // CHECK11-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { 1604 // CHECK11-NEXT: entry: 1605 // CHECK11-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1606 // CHECK11-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1607 // CHECK11-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1608 // CHECK11-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1609 // CHECK11-NEXT: ret void 1610 // 1611 // 1612 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121 1613 // CHECK11-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], ptr noundef [[THIS:%.*]], i32 noundef [[B:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0]] { 1614 // CHECK11-NEXT: entry: 1615 // CHECK11-NEXT: [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 4 1616 // CHECK11-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1617 // CHECK11-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1618 // CHECK11-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1619 // CHECK11-NEXT: [[B_CASTED:%.*]] = alloca i32, align 4 1620 // CHECK11-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1621 // CHECK11-NEXT: store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR]], align 4 1622 // CHECK11-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1623 // CHECK11-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1624 // CHECK11-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1625 // CHECK11-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1626 // CHECK11-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1627 // CHECK11-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 [[TMP2]]) 1628 // CHECK11-NEXT: [[TMP3:%.*]] = load i32, ptr [[B_ADDR]], align 4 1629 // CHECK11-NEXT: store i32 [[TMP3]], ptr [[B_CASTED]], align 4 1630 // CHECK11-NEXT: [[TMP4:%.*]] = load i32, ptr [[B_CASTED]], align 4 1631 // CHECK11-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121.omp_outlined, ptr [[TMP1]], i32 [[TMP4]]) 1632 // CHECK11-NEXT: ret void 1633 // 1634 // 1635 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l121.omp_outlined 1636 // CHECK11-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]], i32 noundef [[B:%.*]]) #[[ATTR0]] { 1637 // CHECK11-NEXT: entry: 1638 // CHECK11-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1639 // CHECK11-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1640 // CHECK11-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1641 // CHECK11-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1642 // CHECK11-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1643 // CHECK11-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1644 // CHECK11-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1645 // CHECK11-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1646 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1647 // CHECK11-NEXT: [[TMP1:%.*]] = load i32, ptr [[B_ADDR]], align 4 1648 // CHECK11-NEXT: [[CONV:%.*]] = sitofp i32 [[TMP1]] to double 1649 // CHECK11-NEXT: [[ADD:%.*]] = fadd double [[CONV]], 1.500000e+00 1650 // CHECK11-NEXT: [[A:%.*]] = getelementptr inbounds nuw [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 1651 // CHECK11-NEXT: store double [[ADD]], ptr [[A]], align 4 1652 // CHECK11-NEXT: ret void 1653 // 1654 // 1655 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126 1656 // CHECK11-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], ptr noundef [[THIS:%.*]]) #[[ATTR0]] { 1657 // CHECK11-NEXT: entry: 1658 // CHECK11-NEXT: [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 4 1659 // CHECK11-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1660 // CHECK11-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1661 // CHECK11-NEXT: store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR]], align 4 1662 // CHECK11-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1663 // CHECK11-NEXT: [[TMP1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1664 // CHECK11-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 1024) 1665 // CHECK11-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126.omp_outlined, ptr [[TMP1]]) 1666 // CHECK11-NEXT: ret void 1667 // 1668 // 1669 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN2S12r1Ei_l126.omp_outlined 1670 // CHECK11-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef [[THIS:%.*]]) #[[ATTR0]] { 1671 // CHECK11-NEXT: entry: 1672 // CHECK11-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1673 // CHECK11-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1674 // CHECK11-NEXT: [[THIS_ADDR:%.*]] = alloca ptr, align 4 1675 // CHECK11-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1676 // CHECK11-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1677 // CHECK11-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 4 1678 // CHECK11-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 4 1679 // CHECK11-NEXT: [[A:%.*]] = getelementptr inbounds nuw [[STRUCT_S1:%.*]], ptr [[TMP0]], i32 0, i32 0 1680 // CHECK11-NEXT: store double 2.500000e+00, ptr [[A]], align 4 1681 // CHECK11-NEXT: ret void 1682 // 1683 // 1684 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88 1685 // CHECK11-SAME: (ptr noalias noundef [[DYN_PTR:%.*]]) #[[ATTR0]] { 1686 // CHECK11-NEXT: entry: 1687 // CHECK11-NEXT: [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 4 1688 // CHECK11-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1689 // CHECK11-NEXT: store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR]], align 4 1690 // CHECK11-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 0, i32 20) 1691 // CHECK11-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 0, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88.omp_outlined) 1692 // CHECK11-NEXT: ret void 1693 // 1694 // 1695 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l88.omp_outlined 1696 // CHECK11-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { 1697 // CHECK11-NEXT: entry: 1698 // CHECK11-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1699 // CHECK11-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1700 // CHECK11-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1701 // CHECK11-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1702 // CHECK11-NEXT: ret void 1703 // 1704 // 1705 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93 1706 // CHECK11-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], i32 noundef [[A:%.*]], i32 noundef [[B:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0]] { 1707 // CHECK11-NEXT: entry: 1708 // CHECK11-NEXT: [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 4 1709 // CHECK11-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4 1710 // CHECK11-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1711 // CHECK11-NEXT: [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4 1712 // CHECK11-NEXT: [[A_CASTED:%.*]] = alloca i32, align 4 1713 // CHECK11-NEXT: [[B_CASTED:%.*]] = alloca i32, align 4 1714 // CHECK11-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]]) 1715 // CHECK11-NEXT: store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR]], align 4 1716 // CHECK11-NEXT: store i32 [[A]], ptr [[A_ADDR]], align 4 1717 // CHECK11-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1718 // CHECK11-NEXT: store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4 1719 // CHECK11-NEXT: [[TMP1:%.*]] = load i16, ptr [[DOTCAPTURE_EXPR__ADDR]], align 2 1720 // CHECK11-NEXT: [[TMP2:%.*]] = sext i16 [[TMP1]] to i32 1721 // CHECK11-NEXT: call void @__kmpc_push_num_teams(ptr @[[GLOB1]], i32 [[TMP0]], i32 [[TMP2]], i32 1024) 1722 // CHECK11-NEXT: [[TMP3:%.*]] = load i32, ptr [[A_ADDR]], align 4 1723 // CHECK11-NEXT: store i32 [[TMP3]], ptr [[A_CASTED]], align 4 1724 // CHECK11-NEXT: [[TMP4:%.*]] = load i32, ptr [[A_CASTED]], align 4 1725 // CHECK11-NEXT: [[TMP5:%.*]] = load i16, ptr [[B_ADDR]], align 2 1726 // CHECK11-NEXT: store i16 [[TMP5]], ptr [[B_CASTED]], align 2 1727 // CHECK11-NEXT: [[TMP6:%.*]] = load i32, ptr [[B_CASTED]], align 4 1728 // CHECK11-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB1]], i32 2, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93.omp_outlined, i32 [[TMP4]], i32 [[TMP6]]) 1729 // CHECK11-NEXT: ret void 1730 // 1731 // 1732 // CHECK11-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l93.omp_outlined 1733 // CHECK11-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i32 noundef [[A:%.*]], i32 noundef [[B:%.*]]) #[[ATTR0]] { 1734 // CHECK11-NEXT: entry: 1735 // CHECK11-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4 1736 // CHECK11-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4 1737 // CHECK11-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4 1738 // CHECK11-NEXT: [[B_ADDR:%.*]] = alloca i32, align 4 1739 // CHECK11-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4 1740 // CHECK11-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4 1741 // CHECK11-NEXT: store i32 [[A]], ptr [[A_ADDR]], align 4 1742 // CHECK11-NEXT: store i32 [[B]], ptr [[B_ADDR]], align 4 1743 // CHECK11-NEXT: [[TMP0:%.*]] = load i16, ptr [[B_ADDR]], align 2 1744 // CHECK11-NEXT: [[CONV:%.*]] = sext i16 [[TMP0]] to i32 1745 // CHECK11-NEXT: [[TMP1:%.*]] = load i32, ptr [[A_ADDR]], align 4 1746 // CHECK11-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP1]], [[CONV]] 1747 // CHECK11-NEXT: store i32 [[ADD]], ptr [[A_ADDR]], align 4 1748 // CHECK11-NEXT: ret void 1749 // 1750