1 // RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64 2 // RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s 3 // RUN: %clang_cc1 -DCHECK -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 CHECK --check-prefix CHECK-64 4 // RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 5 // RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s 6 // RUN: %clang_cc1 -DCHECK -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 CHECK --check-prefix CHECK-32 7 8 // RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64 9 // RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s 10 // RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -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 LAMBDA --check-prefix LAMBDA-64 11 12 // expected-no-diagnostics 13 #ifndef HEADER 14 #define HEADER 15 16 struct St { 17 int a, b; 18 St() : a(0), b(0) {} 19 St(const St &st) : a(st.a + st.b), b(0) {} 20 ~St() {} 21 }; 22 23 volatile int g = 1212; 24 volatile int &g1 = g; 25 26 template <class T> 27 struct S { 28 T f; 29 S(T a) : f(a + g) {} 30 S() : f(g) {} 31 S(const S &s, St t = St()) : f(s.f + t.a) {} 32 operator T() { return T(); } 33 ~S() {} 34 }; 35 36 // CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float } 37 // CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} } 38 // CHECK-DAG: [[ST_TY:%.+]] = type { i{{[0-9]+}}, i{{[0-9]+}} } 39 40 template <typename T> 41 T tmain() { 42 S<T> test; 43 T t_var = T(); 44 T vec[] = {1, 2}; 45 S<T> s_arr[] = {1, 2}; 46 S<T> &var = test; 47 #pragma omp target 48 #pragma omp teams distribute parallel for firstprivate(t_var, vec, s_arr, var) 49 for (int i = 0; i < 2; ++i) { 50 vec[i] = t_var; 51 s_arr[i] = var; 52 } 53 return T(); 54 } 55 56 // CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer, 57 S<float> test; 58 // CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333, 59 int t_var = 333; 60 // CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2], 61 int vec[] = {1, 2}; 62 // CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer, 63 S<float> s_arr[] = {1, 2}; 64 // CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer, 65 S<float> var(3); 66 // CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0, 67 68 int main() { 69 static int sivar; 70 #ifdef LAMBDA 71 // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212, 72 // LAMBDA-LABEL: @main 73 // LAMBDA: call void [[OUTER_LAMBDA:@.+]]( 74 [&]() { 75 // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]]( 76 // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0) 77 // LAMBDA: call void @[[LOFFL1:.+]](i{{64|32}} %{{.+}}) 78 // LAMBDA: ret 79 #pragma omp target 80 #pragma omp teams distribute parallel for firstprivate(g, g1, sivar) 81 for (int i = 0; i < 2; ++i) { 82 // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} {{%.+}}, i{{64|32}} {{%.+}}) 83 // LAMBDA: {{%.+}} = alloca i{{[0-9]+}}, 84 // LAMBDA: {{%.+}} = alloca i{{[0-9]+}}, 85 // LAMBDA: {{%.+}} = alloca i{{[0-9]+}}, 86 // LAMBDA: [[G_CAST:%.+]] = alloca i{{[0-9]+}}, 87 // LAMBDA: [[G1_CAST:%.+]] = alloca i{{[0-9]+}}, 88 // LAMBDA: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}}, 89 // LAMBDA-DAG: [[G_CAST_VAL:%.+]] = load{{.+}} [[G_CAST]], 90 // LAMBDA-DAG: [[G1_CAST_VAL:%.+]] = load{{.+}} [[G1_CAST]], 91 // LAMBDA-DAG: [[SIVAR_CAST_VAL:%.+]] = load{{.+}} [[SIVAR_CAST]], 92 // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[G_CAST_VAL]], {{.+}} [[G1_CAST_VAL]], {{.+}} [[SIVAR_CAST_VAL]]) 93 // LAMBDA: ret void 94 95 // LAMBDA: define internal void @[[LOUTL1]]({{.+}}) 96 // Skip global and bound tid vars 97 // LAMBDA: {{.+}} = alloca i32*, 98 // LAMBDA: {{.+}} = alloca i32*, 99 // LAMBDA: [[G_ADDR:%.+]] = alloca i{{[0-9]+}}, 100 // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}}, 101 // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}}, 102 // LAMBDA: [[G1_TMP:%.+]] = alloca i32*, 103 // skip loop vars 104 // LAMBDA-DAG: store {{.+}}, {{.+}} [[G_ADDR]], 105 // LAMBDA-DAG: store {{.+}}, {{.+}} [[G1_ADDR]], 106 // LAMBDA-DAG: store {{.+}}, {{.+}} [[SIVAR_ADDR]], 107 // LAMBDA-DAG: [[G_CONV:%.+]] = bitcast {{.+}} [[G_ADDR]] to 108 // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast {{.+}} [[G1_ADDR]] to 109 // LAMBDA-DAG: [[SIVAR_CONV:%.+]] = bitcast {{.+}} [[SIVAR_ADDR]] to 110 // LAMBDA-DAG: store{{.+}} [[G1_CONV]], {{.+}} [[G1_TMP]], 111 g = 1; 112 g1 = 1; 113 sivar = 2; 114 // LAMBDA: call void @__kmpc_for_static_init_4( 115 // LAMBDA: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[LPAR_OUTL:.+]] to 116 // LAMBDA: call void @__kmpc_for_static_fini( 117 // LAMBDA: ret void 118 119 // LAMBDA: define internal void @[[LPAR_OUTL]]({{.+}}) 120 // Skip global and bound tid vars, and prev lb and ub vars 121 // LAMBDA: {{.+}} = alloca i32*, 122 // LAMBDA: {{.+}} = alloca i32*, 123 // LAMBDA: {{.+}} = alloca i{{[0-9]+}}, 124 // LAMBDA: {{.+}} = alloca i{{[0-9]+}}, 125 // LAMBDA: [[G_ADDR:%.+]] = alloca i{{[0-9]+}}, 126 // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}}, 127 // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}}, 128 // LAMBDA: [[G1_TMP:%.+]] = alloca i32*, 129 // skip loop vars 130 // LAMBDA: alloca i32, 131 // LAMBDA: alloca i32, 132 // LAMBDA: alloca i32, 133 // LAMBDA: alloca i32, 134 // LAMBDA: alloca i32, 135 // LAMBDA: alloca i32, 136 // LAMBDA: [[G_PRIV:%.+]] = alloca i{{[0-9]+}}, 137 // LAMBDA: [[G1_PRIV:%.+]] = alloca i{{[0-9]+}}, 138 // LAMBDA: [[G1_TMP_PRIV:%.+]] = alloca i{{[0-9]+}}*, 139 // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}}, 140 // LAMBDA-DAG: store {{.+}}, {{.+}} [[G_ADDR]], 141 // LAMBDA-DAG: store {{.+}}, {{.+}} [[G1_ADDR]], 142 // LAMBDA-DAG: store {{.+}}, {{.+}} [[SIVAR_ADDR]], 143 // LAMBDA-DAG: [[G_CONV:%.+]] = bitcast {{.+}} [[G_ADDR]] to 144 // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast {{.+}} [[G1_ADDR]] to 145 // LAMBDA-DAG: [[SIVAR_CONV:%.+]] = bitcast {{.+}} [[SIVAR_ADDR]] to 146 // LAMBDA-DAG: store{{.+}} [[G1_CONV]], {{.+}} [[G1_TMP]], 147 148 // use of private vars 149 // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_PRIV]], 150 // LAMBDA-DAG: [[G1:%.+]] = load{{.+}}, {{.+}}* [[G1_TMP_PRIV]] 151 // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1]], 152 // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_PRIV]], 153 // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[G1_TMP]], 154 // LAMBDA: call void [[INNER_LAMBDA:@.+]]( 155 // LAMBDA: call void @__kmpc_for_static_fini( 156 // LAMBDA: ret void 157 [&]() { 158 // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]]) 159 // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]], 160 g = 2; 161 g1 = 2; 162 sivar = 4; 163 // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]] 164 165 // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0 166 // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]] 167 // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]] 168 // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1 169 // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]] 170 // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]] 171 // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2 172 // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]] 173 // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]] 174 }(); 175 } 176 }(); 177 return 0; 178 #else 179 #pragma omp target 180 #pragma omp teams distribute parallel for firstprivate(t_var, vec, s_arr, var, sivar) 181 for (int i = 0; i < 2; ++i) { 182 vec[i] = t_var; 183 s_arr[i] = var; 184 sivar += i; 185 } 186 return tmain<int>(); 187 #endif 188 } 189 190 // CHECK: define {{.*}}i{{[0-9]+}} @main() 191 // CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0) 192 // CHECK: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}}) 193 // CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]() 194 // CHECK: ret 195 196 // CHECK: define{{.*}} void @[[OFFL1]]({{.+}}) 197 // CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}}, 198 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*, 199 // CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*, 200 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*, 201 // CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}}, 202 // CHECK: [[T_VAR_CAST:%.+]] = alloca i{{[0-9]+}}, 203 // CHECK: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}}, 204 205 // CHECK-DAG: [[VEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_PRIV]], 206 // CHECK-DAG: [[T_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_CAST]], 207 // CHECK-DAG: [[S_ARR_TE_PAR:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_PRIV]], 208 // CHECK-DAG: [[VAR_TE_PAR:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_PRIV]], 209 // CHECK-DAG: [[SIVAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_CAST]], 210 211 // CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[VEC_TE_PAR]], i{{[0-9]+}} [[T_VAR_TE_PAR]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_TE_PAR]], [[S_FLOAT_TY]]* [[VAR_TE_PAR]], i{{[0-9]+}} [[SIVAR_TE_PAR]]) 212 // CHECK: ret void 213 214 // CHECK: define internal void @[[OUTL1]]({{.+}}) 215 // Skip global and bound tid vars 216 // CHECK: {{.+}} = alloca i32*, 217 // CHECK: {{.+}} = alloca i32*, 218 // CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*, 219 // CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}, 220 // CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*, 221 // CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*, 222 // CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}}, 223 // Skip temp vars for loop 224 // CHECK: alloca i{{[0-9]+}}, 225 // CHECK: alloca i{{[0-9]+}}, 226 // CHECK: alloca i{{[0-9]+}}, 227 // CHECK: alloca i{{[0-9]+}}, 228 // CHECK: alloca i{{[0-9]+}}, 229 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}], 230 // CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]], 231 // CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]], 232 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]], 233 // CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]], 234 235 // param copy 236 // CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]], 237 // CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]], 238 // CHECK: store [2 x [[S_FLOAT_TY]]]* {{.+}}, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]], 239 // CHECK: store [[S_FLOAT_TY]]* {{.+}}, [[S_FLOAT_TY]]** [[VAR_ADDR]], 240 // CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[SIVAR_ADDR]], 241 242 // T_VAR and SIVAR 243 // CHECK-DAG-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32* 244 // CHECK-DAG-64: [[CONV_SIVAR:%.+]] = bitcast i64* [[SIVAR_ADDR]] to i32* 245 246 // preparation vars 247 // CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]], 248 // CHECK-DAG: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]], 249 // CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]], 250 251 // firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2 252 // CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 253 // CHECK-DAG: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 254 // CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}}) 255 256 // firstprivate(s_arr) 257 // CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]], 258 // CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]] to 259 // CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]], 260 // CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ] 261 // CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}}], [ [[S_ARR_DST:%.+]], {{.+}} ] 262 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]]) 263 // CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]]) 264 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]]) 265 // CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]], 266 // CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]], 267 268 // firstprivate(var) 269 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]]) 270 // CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]]) 271 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]]) 272 273 // CHECK: call void @__kmpc_for_static_init_4( 274 // CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL:.+]] to 275 // CHECK: call void @__kmpc_for_static_fini( 276 // CHECK: ret void 277 278 // CHECK: define internal void @[[PAR_OUTL]]({{.+}}) 279 // Skip global and bound tid vars, and prev lb ub vars 280 // CHECK: {{.+}} = alloca i32*, 281 // CHECK: {{.+}} = alloca i32*, 282 // CHECK: {{.+}} = alloca i{{[0-9]+}}, 283 // CHECK: {{.+}} = alloca i{{[0-9]+}}, 284 // CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*, 285 // CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}, 286 // CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*, 287 // CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*, 288 // CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}}, 289 // Skip temp vars for loop 290 // CHECK: alloca i{{[0-9]+}}, 291 // CHECK: alloca i{{[0-9]+}}, 292 // CHECK: alloca i{{[0-9]+}}, 293 // CHECK: alloca i{{[0-9]+}}, 294 // CHECK: alloca i{{[0-9]+}}, 295 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}], 296 // CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]], 297 // CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]], 298 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]], 299 // CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]], 300 301 // param copy 302 // CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]], 303 // CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]], 304 // CHECK: store [2 x [[S_FLOAT_TY]]]* {{.+}}, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]], 305 // CHECK: store [[S_FLOAT_TY]]* {{.+}}, [[S_FLOAT_TY]]** [[VAR_ADDR]], 306 // CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[SIVAR_ADDR]], 307 308 // T_VAR and SIVAR 309 // CHECK-DAG-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32* 310 // CHECK-DAG-64: [[CONV_SIVAR:%.+]] = bitcast i64* [[SIVAR_ADDR]] to i32* 311 312 // preparation vars 313 // CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]], 314 // CHECK-DAG: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]], 315 // CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]], 316 317 // firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2 318 // CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 319 // CHECK-DAG: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 320 // CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}}) 321 322 // firstprivate(s_arr) 323 // CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]], 324 // CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]] to 325 // CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]], 326 // CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ] 327 // CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}}], [ [[S_ARR_DST:%.+]], {{.+}} ] 328 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]]) 329 // CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]]) 330 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]]) 331 // CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]], 332 // CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]], 333 334 // firstprivate(var) 335 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]]) 336 // CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]]) 337 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]]) 338 339 // CHECK: call void @__kmpc_for_static_init_4( 340 // CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]] 341 // CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]] 342 // CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]] 343 // CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]] 344 // CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]] 345 // CHECK-DAG-32: {{.+}} = {{.+}} [[SIVAR_ADDR]] 346 // CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_SIVAR]] 347 // CHECK: call void @__kmpc_for_static_fini( 348 // CHECK: ret void 349 350 // CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]() 351 // CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0) 352 // CHECK: call void @[[TOFFL1:.+]](i{{64|32}} %{{.+}}) 353 // CHECK: ret 354 355 // CHECK: define {{.*}}void @[[TOFFL1]]({{.+}}) 356 // CHECK: [[TT_VAR_PRIV:%.+]] = alloca i{{[0-9]+}}, 357 // CHECK: [[TVEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*, 358 // CHECK: [[TS_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]]*, 359 // CHECK: [[TVAR_PRIV:%.+]] = alloca [[S_INT_TY]]*, 360 // CHECK: [[TT_VAR_CAST:%.+]] = alloca i{{[0-9]+}}, 361 362 // CHECK-DAG: [[TVEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[TVEC_PRIV]], 363 // CHECK-DAG: [[TT_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[TT_VAR_CAST]], 364 // CHECK-DAG: [[TS_ARR_TE_PAR:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[TS_ARR_PRIV]], 365 // CHECK-DAG: [[TVAR_TE_PAR:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TVAR_PRIV]], 366 367 // CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[TOUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[TVEC_TE_PAR]], i{{[0-9]+}} [[TT_VAR_TE_PAR]], [2 x [[S_INT_TY]]]* [[TS_ARR_TE_PAR]], [[S_INT_TY]]* [[TVAR_TE_PAR]]) 368 // CHECK: ret void 369 370 // CHECK: define internal void @[[TOUTL1]]({{.+}}) 371 // Skip global and bound tid vars 372 // CHECK: {{.+}} = alloca i32*, 373 // CHECK: {{.+}} = alloca i32*, 374 // CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*, 375 // CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}, 376 // CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*, 377 // CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*, 378 // Skip temp vars for loop 379 // CHECK: alloca i{{[0-9]+}}, 380 // CHECK: alloca i{{[0-9]+}}, 381 // CHECK: alloca i{{[0-9]+}}, 382 // CHECK: alloca i{{[0-9]+}}, 383 // CHECK: alloca i{{[0-9]+}}, 384 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}], 385 // CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]], 386 // CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]], 387 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]], 388 // CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]], 389 // CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*, 390 391 // param copy 392 // CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]], 393 // CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]], 394 // CHECK: store [2 x [[S_INT_TY]]]* {{.+}}, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]], 395 // CHECK: store [[S_INT_TY]]* {{.+}}, [[S_INT_TY]]** [[VAR_ADDR]], 396 397 // T_VAR and preparation variables 398 // CHECK: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]], 399 // CHECK-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32* 400 // CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]], 401 402 // firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2 403 // CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 404 // CHECK-DAG: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 405 // CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}}) 406 407 // firstprivate(s_arr) 408 // CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]], 409 // CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]] to 410 // CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]], 411 // CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ] 412 // CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_DST:%.+]], {{.+}} ] 413 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]]) 414 // CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]]) 415 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]]) 416 // CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]], 417 // CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]], 418 419 // firstprivate(var) 420 // CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]], 421 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]]) 422 // CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]]) 423 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]]) 424 // CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV]], [[S_INT_TY]]** [[TMP]], 425 426 // CHECK: call void @__kmpc_for_static_init_4( 427 // CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[TPAR_OUTL:.+]] to 428 // CHECK: call void @__kmpc_for_static_fini( 429 // CHECK: ret void 430 431 // CHECK: define internal void @[[TPAR_OUTL]]({{.+}}) 432 // Skip global and bound tid vars 433 // CHECK: {{.+}} = alloca i32*, 434 // CHECK: {{.+}} = alloca i32*, 435 // CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*, 436 // CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}, 437 // CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*, 438 // CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*, 439 // Skip temp vars for loop 440 // CHECK: alloca i{{[0-9]+}}, 441 // CHECK: alloca i{{[0-9]+}}, 442 // CHECK: alloca i{{[0-9]+}}, 443 // CHECK: alloca i{{[0-9]+}}, 444 // CHECK: alloca i{{[0-9]+}}, 445 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}], 446 // CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]], 447 // CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]], 448 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]], 449 // CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]], 450 // CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*, 451 452 // param copy 453 // CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]], 454 // CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]], 455 // CHECK: store [2 x [[S_INT_TY]]]* {{.+}}, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]], 456 // CHECK: store [[S_INT_TY]]* {{.+}}, [[S_INT_TY]]** [[VAR_ADDR]], 457 458 // T_VAR and preparation variables 459 // CHECK: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]], 460 // CHECK-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32* 461 // CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]], 462 463 // firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2 464 // CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 465 // CHECK-DAG: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 466 // CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}}) 467 468 // firstprivate(s_arr) 469 // CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]], 470 // CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]] to 471 // CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]], 472 // CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ] 473 // CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_DST:%.+]], {{.+}} ] 474 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]]) 475 // CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]]) 476 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]]) 477 // CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]], 478 // CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]], 479 480 // firstprivate(var) 481 // CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]], 482 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]]) 483 // CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]]) 484 // CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]]) 485 // CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV]], [[S_INT_TY]]** [[TMP]], 486 487 // CHECK: call void @__kmpc_for_static_init_4( 488 // CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]] 489 // CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]] 490 // CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]] 491 // CHECK-DAG: {{.+}} = {{.+}} [[TMP]] 492 // CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]] 493 // CHECK: call void @__kmpc_for_static_fini( 494 // CHECK: ret void 495 496 #endif 497