#include <CL/cl.h>
#include <iostream>
#include <string>
#include <fstream>
#pragma comment(lib, "OpenCL.lib")
const  char * loadfile(const char * fileName)
{
    std::ifstream fs(fileName, std::ios::binary);
    fs.seekg(0, std::ios::end);
    
    int size = fs.tellg();
    char * data = http://www.mamicode.com/new char[size + 1];
    fs.seekg(0);
    fs.read(data, size);
    fs.close();
    data[size] = 0;
    return data;
}


int main()
{
    cl_platform_id platform;
    clGetPlatformIDs(1, &platform, NULL);

    cl_device_id   device;
    clGetDeviceIDs(platform, CL_DEVICE_TYPE_GPU, 1, &device, NULL);
        
    cl_context context = clCreateContext(NULL, 1, &device, NULL, NULL, NULL);
    cl_command_queue queue = clCreateCommandQueue(context, device, 0, NULL);

    const  char * clSourceFile = loadfile("H:/QtTool/build/TritonRayTracing/kernel.txt");
    cl_program program = clCreateProgramWithSource(context, 1, &clSourceFile, NULL, NULL);
    cl_int result = clBuildProgram(program, 1, &device, NULL, NULL, NULL);
    if (result)
    {
        std::cout << "Error buring compilation" << std::endl;
    }
    cl_kernel kernel = clCreateKernel(program, "main", NULL);
    cl_mem output = clCreateBuffer(context, CL_MEM_WRITE_ONLY, 10 * sizeof(cl_int), NULL, 0);
    cl_mem buffer1 = clCreateBuffer(context, CL_MEM_READ_WRITE, 10 * sizeof(cl_int), NULL, 0);
    cl_mem buffer2 = clCreateBuffer(context, CL_MEM_READ_WRITE, 10 * sizeof(cl_int), NULL, 0);
    clSetKernelArg(kernel, 0, sizeof(output), (void *)&output);
    clSetKernelArg(kernel, 1, sizeof(buffer1), (void *)&buffer1);
    clSetKernelArg(kernel, 2, sizeof(buffer2), (void *)&buffer2);
    cl_int * buffer1Ptr = (cl_int *)clEnqueueMapBuffer(queue,
                                                        buffer1,
                                                        CL_TRUE,
                                                        CL_MAP_WRITE,
                                                        0,
                                                        10 * sizeof(cl_int),
                                                        0, NULL, NULL, NULL);
    cl_int * buffer2Ptr = (cl_int *)clEnqueueMapBuffer(queue,
                                                       buffer2,
                                                       CL_TRUE,
                                                       CL_MAP_WRITE,
                                                       0, 
                                                       10 * sizeof(cl_int),
                                                       0, NULL, NULL, NULL);
    for (int i = 0; i < 10; ++i)
    {
        buffer1Ptr[i] = i;
        buffer2Ptr[i] = i;
    }
    clEnqueueUnmapMemObject(queue, buffer1, buffer1Ptr, 0, 0, 0);
    clEnqueueUnmapMemObject(queue, buffer2, buffer2Ptr, 0, 0, 0);
    size_t global_work_size = 10;
    clEnqueueNDRangeKernel(queue,
                           kernel,
                           1, 
                           NULL,
                           &global_work_size,
                           NULL, 0, NULL, NULL);
    cl_int * resultBufferPtr = (cl_int *)clEnqueueMapBuffer(queue,
                                                            output,
                                                            CL_TRUE,
                                                            CL_MAP_READ,
                                                            0,
                                                            10 * sizeof(cl_int),
                                                            0, NULL, NULL, NULL);
    for (int i = 0; i < 10; i++)
    {
        std::cout << "ptr[" << i << "] = " << resultBufferPtr[i] << std::endl;
    }
    return 0;    
}