1 // Test target codegen - host bc file has to be created first. 2 // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc 3 // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - -disable-llvm-optzns | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64 --check-prefix SEQ 4 // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - -disable-llvm-optzns -fopenmp-cuda-parallel-target-regions | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64 --check-prefix PAR 5 // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm-bc %s -o %t-x86-host.bc 6 // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - -disable-llvm-optzns | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix SEQ 7 // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -fexceptions -fcxx-exceptions -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - -disable-llvm-optzns | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix SEQ 8 // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - -disable-llvm-optzns -fopenmp-cuda-parallel-target-regions | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix PAR 9 // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -fexceptions -fcxx-exceptions -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - -disable-llvm-optzns -fopenmp-cuda-parallel-target-regions | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix PAR 10 11 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc 12 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - -disable-llvm-optzns | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64 --check-prefix SEQ 13 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - -disable-llvm-optzns -fopenmp-cuda-parallel-target-regions | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64 --check-prefix PAR 14 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm-bc %s -o %t-x86-host.bc 15 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - -disable-llvm-optzns | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix SEQ 16 // RUN: %clang_cc1 -verify -fopenmp -fexceptions -fcxx-exceptions -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - -disable-llvm-optzns | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix SEQ 17 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - -disable-llvm-optzns -fopenmp-cuda-parallel-target-regions | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix PAR 18 // RUN: %clang_cc1 -verify -fopenmp -fexceptions -fcxx-exceptions -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - -disable-llvm-optzns -fopenmp-cuda-parallel-target-regions | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix PAR 19 20 // expected-no-diagnostics 21 #ifndef HEADER 22 #define HEADER 23 24 int a; 25 26 int foo(int *a); 27 28 int main(int argc, char **argv) { 29 int b[10], c[10], d[10]; 30 #pragma omp target teams map(tofrom:a) 31 #pragma omp distribute parallel for firstprivate(b) lastprivate(c) if(a) 32 for (int i= 0; i < argc; ++i) 33 a = foo(&i) + foo(&a) + foo(&b[i]) + foo(&c[i]) + foo(&d[i]); 34 return 0; 35 } 36 37 // SEQ: [[MEM_TY:%.+]] = type { [128 x i8] } 38 // SEQ-DAG: [[SHARED_GLOBAL_RD:@.+]] = weak addrspace(3) global [[MEM_TY]] undef 39 // SEQ-DAG: [[KERNEL_PTR:@.+]] = internal addrspace(3) global i8* undef 40 // SEQ-DAG: [[KERNEL_SIZE:@.+]] = internal unnamed_addr constant i{{64|32}} 40 41 // SEQ-DAG: [[KERNEL_SHARED:@.+]] = internal unnamed_addr constant i16 1 42 // CHECK-DAG: @__omp_offloading_{{.*}}_main_[[LINE:l.+]]_exec_mode = weak constant i8 0 43 44 // CHECK: define weak void @__omp_offloading_{{.*}}_main_[[LINE]]([10 x i32]* nonnull align 4 dereferenceable(40) %{{.+}}, [10 x i32]* nonnull align 4 dereferenceable(40) %{{.+}}, i32* nonnull align 4 dereferenceable(4) %{{.+}}, i{{64|32}} %{{.+}}, [10 x i32]* nonnull align 4 dereferenceable(40) %{{.+}}) 45 // SEQ: [[SHARED:%.+]] = load i16, i16* [[KERNEL_SHARED]], 46 // SEQ: [[SIZE:%.+]] = load i{{64|32}}, i{{64|32}}* [[KERNEL_SIZE]], 47 // SEQ: call void @__kmpc_get_team_static_memory(i16 1, i8* addrspacecast (i8 addrspace(3)* getelementptr inbounds ([[MEM_TY]], [[MEM_TY]] addrspace(3)* [[SHARED_GLOBAL_RD]], i32 0, i32 0, i32 0) to i8*), i{{64|32}} [[SIZE]], i16 [[SHARED]], i8** addrspacecast (i8* addrspace(3)* [[KERNEL_PTR]] to i8**)) 48 // SEQ: [[PTR:%.+]] = load i8*, i8* addrspace(3)* [[KERNEL_PTR]], 49 // SEQ: [[GEP:%.+]] = getelementptr inbounds i8, i8* [[PTR]], i{{64|32}} 0 50 // PAR: [[GEP:%.+]] = call i8* @__kmpc_data_sharing_push_stack(i{{32|64}} 40, i16 1) 51 // CHECK: [[STACK:%.+]] = bitcast i8* [[GEP]] to %struct._globalized_locals_ty* 52 // CHECK: getelementptr inbounds %struct._globalized_locals_ty, %struct._globalized_locals_ty* [[STACK]], i{{32|64}} 0, i{{32|64}} 0 53 // CHECK-NOT: getelementptr inbounds %struct._globalized_locals_ty, %struct._globalized_locals_ty* [[STACK]], 54 // CHECK: call void @__kmpc_for_static_init_4( 55 56 // CHECK: call void [[PARALLEL:@.+]]( 57 58 // CHECK: call void @__kmpc_for_static_fini(%struct.ident_t* @ 59 60 // SEQ: [[SHARED:%.+]] = load i16, i16* [[KERNEL_SHARED]], 61 // SEQ: call void @__kmpc_restore_team_static_memory(i16 1, i16 [[SHARED]]) 62 // PAR: call void @__kmpc_data_sharing_pop_stack(i8* [[GEP]]) 63 64 // CHECK: define internal void [[PARALLEL]]( 65 // CHECK-NOT: call i8* @__kmpc_data_sharing_push_stack( 66 67 // CHECK-NOT: call void @__kmpc_data_sharing_pop_stack( 68 69 #endif 70