1 // Test host codegen. 2 // 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 CHECK --check-prefix CHECK-64 3 // RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s 4 // RUN: %clang_cc1 -fopenmp -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 --check-prefix HCHECK 5 // 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 CHECK --check-prefix CHECK-32 --check-prefix HCHECK 6 // RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s 7 // RUN: %clang_cc1 -fopenmp -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 --check-prefix HCHECK 8 9 // Test target codegen - host bc file has to be created first. (no significant differences with host version of target region) 10 // 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 11 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s 12 // RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s 13 // RUN: %clang_cc1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s 14 // 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 15 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s 16 // RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s 17 // RUN: %clang_cc1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s 18 19 // expected-no-diagnostics 20 #ifndef HEADER 21 #define HEADER 22 23 // CHECK-DAG: %ident_t = type { i32, i32, i32, i32, i8* } 24 // CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00" 25 // CHECK-DAG: [[DEF_LOC_0:@.+]] = private unnamed_addr constant %ident_t { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) } 26 // CHECK-DAG: [[DEF_LOC_DISTRIBUTE_0:@.+]] = private unnamed_addr constant %ident_t { i32 0, i32 2050, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) } 27 28 // CHECK-LABEL: define {{.*void}} @{{.*}}without_schedule_clause{{.*}}(float* {{.+}}, float* {{.+}}, float* {{.+}}, float* {{.+}}) 29 void without_schedule_clause(float *a, float *b, float *c, float *d) { 30 #pragma omp target 31 #pragma omp teams 32 #pragma omp distribute simd simdlen(8) aligned(a) 33 for (int i = 33; i < 32000000; i += 7) { 34 a[i] = b[i] * c[i] * d[i]; 35 } 36 } 37 38 // CHECK: define {{.*}}void @{{.+}}(i32* noalias [[GBL_TIDP:%.+]], i32* noalias [[BND_TID:%.+]], float** dereferenceable({{[0-9]+}}) [[APTR:%.+]], float** dereferenceable({{[0-9]+}}) [[BPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[CPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[DPTR:%.+]]) 39 // CHECK: [[TID_ADDR:%.+]] = alloca i32* 40 // CHECK: [[IV:%.+iv]] = alloca i32 41 // CHECK: [[LB:%.+lb]] = alloca i32 42 // CHECK: [[UB:%.+ub]] = alloca i32 43 // CHECK: [[ST:%.+stride]] = alloca i32 44 // CHECK: [[LAST:%.+last]] = alloca i32 45 // CHECK-DAG: store i32* [[GBL_TIDP]], i32** [[TID_ADDR]] 46 // CHECK-DAG: call void @llvm.assume( 47 // CHECK-DAG: store i32 0, i32* [[LB]] 48 // CHECK-DAG: store i32 4571423, i32* [[UB]] 49 // CHECK-DAG: store i32 1, i32* [[ST]] 50 // CHECK-DAG: store i32 0, i32* [[LAST]] 51 // CHECK-DAG: [[GBL_TID:%.+]] = load i32*, i32** [[TID_ADDR]] 52 // CHECK-DAG: [[GBL_TIDV:%.+]] = load i32, i32* [[GBL_TID]] 53 // CHECK: call void @__kmpc_for_static_init_{{.+}}(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]], i32 92, i32* %.omp.is_last, i32* %.omp.lb, i32* %.omp.ub, i32* %.omp.stride, i32 1, i32 1) 54 // CHECK-DAG: [[UBV0:%.+]] = load i32, i32* [[UB]] 55 // CHECK-DAG: [[USWITCH:%.+]] = icmp sgt i32 [[UBV0]], 4571423 56 // CHECK: br i1 [[USWITCH]], label %[[BBCT:.+]], label %[[BBCF:.+]] 57 // CHECK-DAG: [[BBCT]]: 58 // CHECK-DAG: br label %[[BBCE:.+]] 59 // CHECK-DAG: [[BBCF]]: 60 // CHECK-DAG: [[UBV1:%.+]] = load i32, i32* [[UB]] 61 // CHECK-DAG: br label %[[BBCE]] 62 // CHECK: [[BBCE]]: 63 // CHECK: [[SELUB:%.+]] = phi i32 [ 4571423, %[[BBCT]] ], [ [[UBV1]], %[[BBCF]] ] 64 // CHECK: store i32 [[SELUB]], i32* [[UB]] 65 // CHECK: [[LBV0:%.+]] = load i32, i32* [[LB]] 66 // CHECK: store i32 [[LBV0]], i32* [[IV]] 67 // CHECK: br label %[[BBINNFOR:.+]] 68 // CHECK: [[BBINNFOR]]: 69 // CHECK: [[IVVAL0:%.+]] = load i32, i32* [[IV]] 70 // CHECK: [[UBV2:%.+]] = load i32, i32* [[UB]] 71 // CHECK: [[IVLEUB:%.+]] = icmp sle i32 [[IVVAL0]], [[UBV2]] 72 // CHECK: br i1 [[IVLEUB]], label %[[BBINNBODY:.+]], label %[[BBINNEND:.+]] 73 // CHECK: [[BBINNBODY]]: 74 // CHECK: {{.+}} = load i32, i32* [[IV]] 75 // ... loop body ... 76 // CHECK: br label %[[BBBODYCONT:.+]] 77 // CHECK: [[BBBODYCONT]]: 78 // CHECK: br label %[[BBINNINC:.+]] 79 // CHECK: [[BBINNINC]]: 80 // CHECK: [[IVVAL1:%.+]] = load i32, i32* [[IV]] 81 // CHECK: [[IVINC:%.+]] = add nsw i32 [[IVVAL1]], 1 82 // CHECK: store i32 [[IVINC]], i32* [[IV]] 83 // CHECK: br label %[[BBINNFOR]] 84 // CHECK: [[BBINNEND]]: 85 // CHECK: br label %[[LPEXIT:.+]] 86 // CHECK: [[LPEXIT]]: 87 // CHECK: call void @__kmpc_for_static_fini(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]]) 88 // CHECK: ret void 89 90 91 // CHECK-LABEL: define {{.*void}} @{{.*}}static_not_chunked{{.*}}(float* {{.+}}, float* {{.+}}, float* {{.+}}, float* {{.+}}) 92 void static_not_chunked(float *a, float *b, float *c, float *d) { 93 #pragma omp target 94 #pragma omp teams 95 #pragma omp distribute simd dist_schedule(static) safelen(32) 96 for (int i = 32000000; i > 33; i += -7) { 97 a[i] = b[i] * c[i] * d[i]; 98 } 99 } 100 101 // CHECK: define {{.*}}void @.omp_outlined.{{.*}}(i32* noalias [[GBL_TIDP:%.+]], i32* noalias [[BND_TID:%.+]], float** dereferenceable({{[0-9]+}}) [[APTR:%.+]], float** dereferenceable({{[0-9]+}}) [[BPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[CPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[DPTR:%.+]]) 102 // CHECK: [[TID_ADDR:%.+]] = alloca i32* 103 // CHECK: [[IV:%.+iv]] = alloca i32 104 // CHECK: [[LB:%.+lb]] = alloca i32 105 // CHECK: [[UB:%.+ub]] = alloca i32 106 // CHECK: [[ST:%.+stride]] = alloca i32 107 // CHECK: [[LAST:%.+last]] = alloca i32 108 // CHECK-DAG: store i32* [[GBL_TIDP]], i32** [[TID_ADDR]] 109 // CHECK-DAG: store i32 0, i32* [[LB]] 110 // CHECK-DAG: store i32 4571423, i32* [[UB]] 111 // CHECK-DAG: store i32 1, i32* [[ST]] 112 // CHECK-DAG: store i32 0, i32* [[LAST]] 113 // CHECK-DAG: [[GBL_TID:%.+]] = load i32*, i32** [[TID_ADDR]] 114 // CHECK-DAG: [[GBL_TIDV:%.+]] = load i32, i32* [[GBL_TID]] 115 // CHECK: call void @__kmpc_for_static_init_{{.+}}(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]], i32 92, i32* %.omp.is_last, i32* %.omp.lb, i32* %.omp.ub, i32* %.omp.stride, i32 1, i32 1) 116 // CHECK-DAG: [[UBV0:%.+]] = load i32, i32* [[UB]] 117 // CHECK-DAG: [[USWITCH:%.+]] = icmp sgt i32 [[UBV0]], 4571423 118 // CHECK: br i1 [[USWITCH]], label %[[BBCT:.+]], label %[[BBCF:.+]] 119 // CHECK-DAG: [[BBCT]]: 120 // CHECK-DAG: br label %[[BBCE:.+]] 121 // CHECK-DAG: [[BBCF]]: 122 // CHECK-DAG: [[UBV1:%.+]] = load i32, i32* [[UB]] 123 // CHECK-DAG: br label %[[BBCE]] 124 // CHECK: [[BBCE]]: 125 // CHECK: [[SELUB:%.+]] = phi i32 [ 4571423, %[[BBCT]] ], [ [[UBV1]], %[[BBCF]] ] 126 // CHECK: store i32 [[SELUB]], i32* [[UB]] 127 // CHECK: [[LBV0:%.+]] = load i32, i32* [[LB]] 128 // CHECK: store i32 [[LBV0]], i32* [[IV]] 129 // CHECK: br label %[[BBINNFOR:.+]] 130 // CHECK: [[BBINNFOR]]: 131 // CHECK: [[IVVAL0:%.+]] = load i32, i32* [[IV]] 132 // CHECK: [[UBV2:%.+]] = load i32, i32* [[UB]] 133 // CHECK: [[IVLEUB:%.+]] = icmp sle i32 [[IVVAL0]], [[UBV2]] 134 // CHECK: br i1 [[IVLEUB]], label %[[BBINNBODY:.+]], label %[[BBINNEND:.+]] 135 // CHECK: [[BBINNBODY]]: 136 // CHECK: {{.+}} = load i32, i32* [[IV]] 137 // ... loop body ... 138 // CHECK: br label %[[BBBODYCONT:.+]] 139 // CHECK: [[BBBODYCONT]]: 140 // CHECK: br label %[[BBINNINC:.+]] 141 // CHECK: [[BBINNINC]]: 142 // CHECK: [[IVVAL1:%.+]] = load i32, i32* [[IV]] 143 // CHECK: [[IVINC:%.+]] = add nsw i32 [[IVVAL1]], 1 144 // CHECK: store i32 [[IVINC]], i32* [[IV]] 145 // CHECK: br label %[[BBINNFOR]] 146 // CHECK: [[BBINNEND]]: 147 // CHECK: br label %[[LPEXIT:.+]] 148 // CHECK: [[LPEXIT]]: 149 // CHECK: call void @__kmpc_for_static_fini(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]]) 150 // CHECK: ret void 151 152 153 // CHECK-LABEL: define {{.*void}} @{{.*}}static_chunked{{.*}}(float* {{.+}}, float* {{.+}}, float* {{.+}}, float* {{.+}}) 154 void static_chunked(float *a, float *b, float *c, float *d) { 155 #pragma omp target 156 #pragma omp teams 157 #pragma omp distribute simd dist_schedule(static, 5) 158 for (unsigned i = 131071; i <= 2147483647; i += 127) { 159 a[i] = b[i] * c[i] * d[i]; 160 } 161 } 162 163 // CHECK: define {{.*}}void @.omp_outlined.{{.*}}(i32* noalias [[GBL_TIDP:%.+]], i32* noalias [[BND_TID:%.+]], float** dereferenceable({{[0-9]+}}) [[APTR:%.+]], float** dereferenceable({{[0-9]+}}) [[BPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[CPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[DPTR:%.+]]) 164 // CHECK: [[TID_ADDR:%.+]] = alloca i32* 165 // CHECK: [[IV:%.+iv]] = alloca i32 166 // CHECK: [[LB:%.+lb]] = alloca i32 167 // CHECK: [[UB:%.+ub]] = alloca i32 168 // CHECK: [[ST:%.+stride]] = alloca i32 169 // CHECK: [[LAST:%.+last]] = alloca i32 170 // CHECK-DAG: store i32* [[GBL_TIDP]], i32** [[TID_ADDR]] 171 // CHECK-DAG: store i32 0, i32* [[LB]] 172 // CHECK-DAG: store i32 16908288, i32* [[UB]] 173 // CHECK-DAG: store i32 1, i32* [[ST]] 174 // CHECK-DAG: store i32 0, i32* [[LAST]] 175 // CHECK-DAG: [[GBL_TID:%.+]] = load i32*, i32** [[TID_ADDR]] 176 // CHECK-DAG: [[GBL_TIDV:%.+]] = load i32, i32* [[GBL_TID]] 177 // CHECK: call void @__kmpc_for_static_init_{{.+}}(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]], i32 91, i32* %.omp.is_last, i32* %.omp.lb, i32* %.omp.ub, i32* %.omp.stride, i32 1, i32 5) 178 // CHECK-DAG: [[UBV0:%.+]] = load i32, i32* [[UB]] 179 // CHECK-DAG: [[USWITCH:%.+]] = icmp ugt i32 [[UBV0]], 16908288 180 // CHECK: br i1 [[USWITCH]], label %[[BBCT:.+]], label %[[BBCF:.+]] 181 // CHECK-DAG: [[BBCT]]: 182 // CHECK-DAG: br label %[[BBCE:.+]] 183 // CHECK-DAG: [[BBCF]]: 184 // CHECK-DAG: [[UBV1:%.+]] = load i32, i32* [[UB]] 185 // CHECK-DAG: br label %[[BBCE]] 186 // CHECK: [[BBCE]]: 187 // CHECK: [[SELUB:%.+]] = phi i32 [ 16908288, %[[BBCT]] ], [ [[UBV1]], %[[BBCF]] ] 188 // CHECK: store i32 [[SELUB]], i32* [[UB]] 189 // CHECK: [[LBV0:%.+]] = load i32, i32* [[LB]] 190 // CHECK: store i32 [[LBV0]], i32* [[IV]] 191 // CHECK: br label %[[BBINNFOR:.+]] 192 // CHECK: [[BBINNFOR]]: 193 // CHECK: [[IVVAL0:%.+]] = load i32, i32* [[IV]] 194 // CHECK: [[UBV2:%.+]] = load i32, i32* [[UB]] 195 // CHECK: [[IVLEUB:%.+]] = icmp ule i32 [[IVVAL0]], [[UBV2]] 196 // CHECK: br i1 [[IVLEUB]], label %[[BBINNBODY:.+]], label %[[BBINNEND:.+]] 197 // CHECK: [[BBINNBODY]]: 198 // CHECK: {{.+}} = load i32, i32* [[IV]] 199 // ... loop body ... 200 // CHECK: br label %[[BBBODYCONT:.+]] 201 // CHECK: [[BBBODYCONT]]: 202 // CHECK: br label %[[BBINNINC:.+]] 203 // CHECK: [[BBINNINC]]: 204 // CHECK: [[IVVAL1:%.+]] = load i32, i32* [[IV]] 205 // CHECK: [[IVINC:%.+]] = add i32 [[IVVAL1]], 1 206 // CHECK: store i32 [[IVINC]], i32* [[IV]] 207 // CHECK: br label %[[BBINNFOR]] 208 // CHECK: [[BBINNEND]]: 209 // CHECK: br label %[[LPEXIT:.+]] 210 // CHECK: [[LPEXIT]]: 211 // CHECK: call void @__kmpc_for_static_fini(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]]) 212 // CHECK: ret void 213 214 // CHECK-LABEL: test_precond 215 void test_precond() { 216 char a = 0; char i; 217 #pragma omp target 218 #pragma omp teams 219 #pragma omp distribute simd linear(i) 220 for(i = a; i < 10; ++i); 221 } 222 223 // a is passed as a parameter to the outlined functions 224 // CHECK: define {{.*}}void @.omp_outlined.{{.*}}(i32* noalias [[GBL_TIDP:%.+]], i32* noalias [[BND_TID:%.+]], i8* dereferenceable({{[0-9]+}}) [[APARM:%.+]]) 225 // CHECK: store i8* [[APARM]], i8** [[APTRADDR:%.+]] 226 // ..many loads of %0.. 227 // CHECK: [[A2:%.+]] = load i8*, i8** [[APTRADDR]] 228 // CHECK: [[AVAL0:%.+]] = load i8, i8* [[A2]] 229 // CHECK: store i8 [[AVAL0]], i8* [[CAP_EXPR:%.+]], 230 // CHECK: [[AVAL1:%.+]] = load i8, i8* [[CAP_EXPR]] 231 // CHECK: load i8, i8* [[CAP_EXPR]] 232 // CHECK: [[AVAL2:%.+]] = load i8, i8* [[CAP_EXPR]] 233 // CHECK: [[ACONV:%.+]] = sext i8 [[AVAL2]] to i32 234 // CHECK: [[ACMP:%.+]] = icmp slt i32 [[ACONV]], 10 235 // CHECK: br i1 [[ACMP]], label %[[PRECOND_THEN:.+]], label %[[PRECOND_END:.+]] 236 // CHECK: [[PRECOND_THEN]] 237 // CHECK: call void @__kmpc_for_static_init_4 238 // CHECK: call void @__kmpc_for_static_fini 239 // CHECK: [[PRECOND_END]] 240 241 // no templates for now, as these require special handling in target regions and/or declare target 242 243 // HCHECK-LABEL: fint 244 // HCHECK: call {{.*}}i32 {{.+}}ftemplate 245 // HCHECK: ret i32 246 247 // HCHECK: load i16, i16* 248 // HCHECK: store i16 % 249 // HCHECK: call i32 @__tgt_target_teams( 250 // HCHECK: call void @__kmpc_for_static_init_4( 251 template <typename T> 252 T ftemplate() { 253 short aa = 0; 254 255 #pragma omp target 256 #pragma omp teams 257 #pragma omp distribute simd dist_schedule(static, aa) 258 for (int i = 0; i < 100; i++) { 259 } 260 return T(); 261 } 262 263 int fint(void) { return ftemplate<int>(); } 264 265 #endif 266 267 // CHECK: !{!"llvm.loop.vectorize.width", i32 8} 268 // CHECK: !{!"llvm.loop.vectorize.enable", i1 true} 269 // CHECK: !{!"llvm.loop.vectorize.width", i32 32} 270