1 // Make sure that __global__ functions are emitted along with correct 2 // annotations and are added to @llvm.used to prevent their elimination. 3 // REQUIRES: nvptx-registered-target 4 // 5 // RUN: %clang_cc1 %s -triple nvptx-unknown-unknown -fcuda-is-device -emit-llvm -o - | FileCheck %s 6 7 #include "Inputs/cuda.h" 8 9 // CHECK-LABEL: define void @device_function 10 extern "C" 11 __device__ void device_function() {} 12 13 // CHECK-LABEL: define void @global_function 14 extern "C" 15 __global__ void global_function() { 16 // CHECK: call void @device_function 17 device_function(); 18 } 19 20 // Make sure host-instantiated kernels are preserved on device side. 21 template <typename T> __global__ void templated_kernel(T param) {} 22 // CHECK-DAG: define void @_Z16templated_kernelIiEvT_( 23 24 namespace { 25 __global__ void anonymous_ns_kernel() {} 26 // CHECK-DAG: define void @_ZN12_GLOBAL__N_119anonymous_ns_kernelEv( 27 } 28 29 void host_function() { 30 templated_kernel<<<0, 0>>>(0); 31 anonymous_ns_kernel<<<0,0>>>(); 32 } 33 34 // CHECK: !{{[0-9]+}} = !{void ()* @global_function, !"kernel", i32 1} 35 // CHECK: !{{[0-9]+}} = !{void (i32)* @_Z16templated_kernelIiEvT_, !"kernel", i32 1} 36
