|
| 1 | +// REQUIRES: opencl, opencl_icd |
| 2 | + |
| 3 | +// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out %opencl_lib |
| 4 | +// RUN: %CPU_RUN_PLACEHOLDER %t.out |
| 5 | +// RUN: %GPU_RUN_PLACEHOLDER %t.out |
| 6 | +// RUN: %ACC_RUN_PLACEHOLDER %t.out |
| 7 | + |
| 8 | +// Test for OpenCL make_kernel |
| 9 | + |
| 10 | +#include <CL/opencl.h> |
| 11 | +#include <CL/sycl.hpp> |
| 12 | +#include <CL/sycl/backend/opencl.hpp> |
| 13 | + |
| 14 | +using namespace cl::sycl; |
| 15 | + |
| 16 | +class DummyKernel; |
| 17 | + |
| 18 | +constexpr size_t N = 1024; |
| 19 | + |
| 20 | +const char OpenCLIotaKernelCode[] = "__kernel void iota(__global int *a) {\ |
| 21 | + int i = get_global_id(0);\ |
| 22 | + a[i] = i;\ |
| 23 | + }"; |
| 24 | +const size_t OpenCLIotaKernelCodeSize = sizeof(OpenCLIotaKernelCode); |
| 25 | +const char *OpenCLCode[1] = {OpenCLIotaKernelCode}; |
| 26 | + |
| 27 | +int main() { |
| 28 | + queue Queue{}; |
| 29 | + auto Context = Queue.get_info<info::queue::context>(); |
| 30 | + auto Device = Queue.get_info<info::queue::device>(); |
| 31 | + |
| 32 | + cl_context OpenCLContext = get_native<backend::opencl>(Context); |
| 33 | + cl_device_id OpenCLDevice = get_native<backend::opencl>(Device); |
| 34 | + |
| 35 | + cl_int OpenCLError; |
| 36 | + cl_program OpenCLProgram = clCreateProgramWithSource( |
| 37 | + OpenCLContext, 1, OpenCLCode, &OpenCLIotaKernelCodeSize, &OpenCLError); |
| 38 | + assert(OpenCLError == CL_SUCCESS); |
| 39 | + |
| 40 | + OpenCLError = clBuildProgram(OpenCLProgram, 1, &OpenCLDevice, nullptr, |
| 41 | + nullptr, nullptr); |
| 42 | + assert(OpenCLError == CL_SUCCESS); |
| 43 | + |
| 44 | + cl_kernel OpenCLKernel = clCreateKernel(OpenCLProgram, "iota", &OpenCLError); |
| 45 | + assert(OpenCLError == CL_SUCCESS); |
| 46 | + |
| 47 | + kernel Kernel = make_kernel<backend::opencl>(OpenCLKernel, Context); |
| 48 | + |
| 49 | + // The associated kernel bundle should not contain the dummy-kernel. |
| 50 | + assert(!Kernel.get_kernel_bundle().has_kernel(get_kernel_id<DummyKernel>())); |
| 51 | + |
| 52 | + int *IotaResult = malloc_shared<int>(N, Device, Context); |
| 53 | + Queue |
| 54 | + .submit([&](sycl::handler &CGH) { |
| 55 | + CGH.set_arg(0, IotaResult); |
| 56 | + CGH.parallel_for(N, Kernel); |
| 57 | + }) |
| 58 | + .wait(); |
| 59 | + |
| 60 | + for (int i = 0; i < N; ++i) |
| 61 | + assert(IotaResult[i] == i); |
| 62 | + |
| 63 | + free(IotaResult, Context); |
| 64 | + |
| 65 | + Queue |
| 66 | + .submit( |
| 67 | + [&](sycl::handler &CGH) { CGH.single_task<DummyKernel>([=]() {}); }) |
| 68 | + .wait(); |
| 69 | + |
| 70 | + return 0; |
| 71 | +} |
0 commit comments