opencl example

example.cc

#include <iostream>
#ifdef __APPLE__
    #include <OpenCL/cl.hpp>
#else
    #include <CL/cl.hpp>
#endif

int main() {
    // get all platforms (drivers), e.g. NVIDIA
    std::vector<cl::Platform> all_platforms;
    cl::Platform::get(&all_platforms);

    std::cout << all_platforms.size() << std::endl;
    if (all_platforms.size()==0) {
        std::cout<<" No platforms found. Check OpenCL installation!\n";
        exit(1);
    }
    cl::Platform default_platform=all_platforms[0];
    // std::cout << "Using platform: "<<default_platform.getInfo<CL_PLATFORM_NAME>()<<"\n";

    // get default device (CPUs, GPUs) of the default platform
    std::vector<cl::Device> all_devices;
    printf("get devices\n");
    default_platform.getDevices(CL_DEVICE_TYPE_ALL, &all_devices);
    printf("get devices done %d\n", all_devices.size());
    if(all_devices.size()==0){
        std::cout<<" No devices found. Check OpenCL installation!\n";
        exit(1);
    }

    // use device[1] because that's a GPU; device[0] is the CPU
    cl::Device default_device=all_devices[0];
    // std::cout<< "Using device: "<<default_device.getInfo<CL_DEVICE_NAME>()<<"\n";
    cl::Context context({default_device});
    cl::CommandQueue queue(context,default_device);

    // a context is like a "runtime link" to the device and platform;
    // i.e. communication is possible
    cl::Buffer buffer_A(context,CL_MEM_READ_WRITE | CL_MEM_ALLOC_HOST_PTR , sizeof(int)*10);
    cl_int error                = CL_SUCCESS;
    // zerocopy
    int* bufferPtr = (int*)queue.enqueueMapBuffer(buffer_A, CL_TRUE, CL_MAP_WRITE, 0, sizeof(int)*10, nullptr, nullptr, &error);

    if (error != CL_SUCCESS) {
        printf("Error to map buffer in copy buffer, error=%d\n", error);
        return 0;
    }
    int b[10] = {1,2,3,4,5,6,7,8,9,10};
    // copy to the mapped buffer
    if(bufferPtr != nullptr){
      memcpy(bufferPtr, b, 10 * sizeof(int));
    }
    int c[10];
    // buffer_A: 1,2,3,4,5,6,7,8,9,10
    queue.enqueueReadBuffer(buffer_A, CL_TRUE, 0, sizeof(int)*10, c);

    printf("buffer_A:\n");
    for (int i = 0; i < 10; i++) {
      printf("%d\n", c[i]);
    }

    int d[10] = {10,2,3,4,15,6,7,8,9,20};
    queue.enqueueWriteBuffer(buffer_A, CL_TRUE, 0, sizeof(int)*10, d);

    // bufferPtr: 10,2,3,4,15,6,7,8,9,20
    printf("map buffer content\n");
    for (int i = 0; i < 10; i++) {
      printf("%d\n", bufferPtr[i]);
    }

    return 0;
}