OpenCL向量加法程序 [英] OpenCL vector addition program

问题描述

这不是重复的它不同于 OpenCL Vector 添加程序.

我的代码输出一堆零，无法理解为什么会发生这种情况.

 #include #include #include//内核字符串的最大源大小#define MAX_SOURCE_SIZE (0x100000)int main(void){//创建两个输入向量国际我;INT LIST_SIZE;printf("请输入多少个元素:");scanf("%d",&LIST_SIZE);int *A = (int*) malloc (sizeof (int) * LIST_SIZE);//初始化输入向量for(i = 0; i < LIST_SIZE; i++){A[i] = i;//如果 LIST_SIZE 非常大}int *B = (int*)malloc(sizeof(int)*LIST_SIZE);//初始化输入向量for(i = 0; i < LIST_SIZE; i++){B[i] = i+10;}int *C = (int*)malloc(sizeof(int)*LIST_SIZE);//将内核源代码加载到数组 source_str文件 *fp;字符 *source_str;size_t source_size;fp = fopen("vectorCLKernel.cl", "r");如果(！fp){fprintf(stderr, "加载内核失败.\n");获取字符()；退出(1)；}source_str = (char*)malloc(MAX_SOURCE_SIZE);source_size = fread( source_str, 1, MAX_SOURCE_SIZE, fp);fclose( fp );//获取平台和设备信息cl_platform_id platform_id = NULL;cl_device_id device_id = NULL;cl_uint ret_num_devices;cl_uint ret_num_platforms;cl_int ret = clGetPlatformIDs(1, &platform_id, &ret_num_platforms);ret = clGetDeviceIDs(platform_id, CL_DEVICE_TYPE_GPU, 1,&device_id,&ret_num_devices);//创建一个 OpenCL 上下文cl_context context = clCreateContext( NULL, 1, &device_id, NULL, NULL,&ret);//创建命令队列cl_command_queue command_queue = clCreateCommandQueue(context,device_id,NULL, &ret);//在设备上为每个向量 A、B 和 C 创建内存缓冲区cl_mem a_mem_obj = clCreateBuffer(context,CL_MEM_READ_ONLY,LIST_SIZE * sizeof(int), A, &ret);cl_mem b_mem_obj = clCreateBuffer(context,CL_MEM_READ_ONLY,LIST_SIZE * sizeof(int), B, &ret);cl_mem c_mem_obj = clCreateBuffer(context,CL_MEM_READ_WRITE,LIST_SIZE * sizeof(int), C, &ret);//将列表 A 和 B 复制到各自的内存缓冲区ret = clEnqueueWriteBuffer(command_queue, a_mem_obj, CL_TRUE,0,LIST_SIZE * sizeof(int), A, 0, NULL, NULL);ret = clEnqueueWriteBuffer(command_queue, b_mem_obj, CL_TRUE,0,LIST_SIZE * sizeof(int), B, 0, NULL, NULL);//从内核源创建一个程序printf("%s\n",source_str);printf("%d\n",source_size);cl_program 程序 = clCreateProgramWithSource(context, 1, (constchar**)&source_str, (const size_t *)&source_size, &ret);ret = clBuildProgram(program, 1, &device_id, NULL, NULL, NULL);//创建 OpenCL 内核对象cl_kernel kernel = clCreateKernel(program, "vector_add", &ret);//设置内核参数ret = clSetKernelArg(kernel, 0, sizeof(cl_mem), (void *)&a_mem_obj);ret = clSetKernelArg(kernel, 1, sizeof(cl_mem), (void *)&b_mem_obj);ret = clSetKernelArg(kernel, 2, sizeof(cl_mem), (void *)&c_mem_obj);//在数组上执行 OpenCL 内核size_t global_item_size = LIST_SIZE;size_t local_item_size = 1;//在设备上执行内核cl_event 事件；ret = clEnqueueNDRangeKernel(command_queue, kernel, 1, NULL,&global_item_size, &local_item_size, 0, NULL, NULL);ret = clFinish(command_queue);//将设备上的内存缓冲区C读到局部变量Cret = clEnqueueReadBuffer(command_queue, c_mem_obj, CL_TRUE, 0,LIST_SIZE* sizeof(int), C, 0, NULL, NULL);//将结果显示在屏幕上for(i = 0; i < LIST_SIZE; i++)printf("%d + %d = %d\n", A[i], B[i], C[i]);//清理ret = clFlush(command_queue);ret = clReleaseKernel(内核);ret = clReleaseProgram(program);ret = clReleaseMemObject(a_mem_obj);ret = clReleaseMemObject(b_mem_obj);ret = clReleaseMemObject(c_mem_obj);ret = clReleaseCommandQueue(command_queue);ret = clReleaseContext(context);免费(A);免费(B)；免费(C)；获取字符()；返回0；}

我的核函数是vectorCLKernel.cl

 __kernel void vector_add(__global int *A, __global int *B, __global int *C){int i = get_global_id(0);C[i] = A[i] + B[i];}

我已经正确编译了代码，但不明白为什么我的输出是零

0 + 10 = 0

1 + 11 = 0

2 + 12 = 0

3 + 13 = 0

4 + 14 = 0

解决方案

首先，我建议你养成创建 & 的习惯.使用一些 C 宏来检查每个 OpenCL API 调用的错误.它会让你的生活更轻松.

第二，问题是您对 clCreateBuffer 的调用无效(所有三个).注意:

clCreateBuffer(context,CL_MEM_READ_ONLY,LIST_SIZE * sizeof(int), A, &ret);

您提供了一个 host_ptr 参数.clCreateBuffer 的文档说:

CL_INVALID_HOST_PTR 如果 host_ptr 为 NULL 并且在标志中设置了 CL_MEM_USE_HOST_PTR 或 CL_MEM_COPY_HOST_PTR，或者如果 host_ptr 不为 NULL 但 CL_MEM_COPY_HOST_PTR 或 CL_MEM_USE_PTR 中未设置标志.

...只需将 NULL 放在那里而不是 A/B/C.

This is not a duplicate It is different from OpenCL Vector add program.

My code outputs a bunch of zeroes cannot understand why that happens.

    #include <stdio.h>
    #include <CL/cl.h>
    #include<stdlib.h>
    //Max source size of the kernel string
    #define MAX_SOURCE_SIZE (0x100000)
    int main(void)
    {
    // Create the two input vectors
    int i;
    int LIST_SIZE;
    printf("Enter how many elements:");
    scanf("%d",&LIST_SIZE);
    int *A = (int*) malloc (sizeof (int) * LIST_SIZE);
    //Initialize the input vectors
    for(i = 0; i < LIST_SIZE; i++)
    {
    A[i] = i; //if LIST_SIZE is very large
    }
    int *B = (int*)malloc(sizeof(int)*LIST_SIZE);
    //Initialize the input vectors
    for(i = 0; i < LIST_SIZE; i++)
    {
    B[i] = i+10;
    }

    int *C = (int*)malloc(sizeof(int)*LIST_SIZE);
    // Load the kernel source code into the array source_str
    FILE *fp;
    char *source_str;
    size_t source_size;
    fp = fopen("vectorCLKernel.cl", "r");
    if (!fp)
    {
    fprintf(stderr, "Failed to load kernel.\n");
    getchar();
    exit(1);
    }
    source_str = (char*)malloc(MAX_SOURCE_SIZE);
    source_size = fread( source_str, 1, MAX_SOURCE_SIZE, fp);
    fclose( fp );
    // Get platform and device information
    cl_platform_id platform_id = NULL;
    cl_device_id device_id = NULL;
    cl_uint ret_num_devices;
    cl_uint ret_num_platforms;
    cl_int ret = clGetPlatformIDs(1, &platform_id, &ret_num_platforms);
    ret = clGetDeviceIDs( platform_id, CL_DEVICE_TYPE_GPU, 1,&device_id,&ret_num_devices);
    // Create an OpenCL context
    cl_context context = clCreateContext( NULL, 1, &device_id, NULL, NULL,&ret);
    // Create a command queue
    cl_command_queue command_queue = clCreateCommandQueue(context,device_id,NULL, &ret);
    // Create memory buffers on the device for each vector A, B and C
    cl_mem a_mem_obj = clCreateBuffer(context,CL_MEM_READ_ONLY,LIST_SIZE * sizeof(int), A, &ret);
    cl_mem b_mem_obj = clCreateBuffer(context,CL_MEM_READ_ONLY,LIST_SIZE * sizeof(int), B, &ret);
    cl_mem c_mem_obj = clCreateBuffer(context,CL_MEM_READ_WRITE,LIST_SIZE * sizeof(int), C, &ret);

    // Copy the lists A and B to their respective memory buffers
    ret = clEnqueueWriteBuffer(command_queue, a_mem_obj, CL_TRUE,0,LIST_SIZE * sizeof(int), A, 0, NULL, NULL);
    ret = clEnqueueWriteBuffer(command_queue, b_mem_obj, CL_TRUE,0,LIST_SIZE * sizeof(int), B, 0, NULL, NULL);
    // Create a program from the kernel source
    printf("%s\n",source_str);
    printf("%d\n",source_size);
    cl_program program = clCreateProgramWithSource(context, 1, (const
    char**)&source_str, (const size_t *)&source_size, &ret);
    ret = clBuildProgram(program, 1, &device_id, NULL, NULL, NULL);
    // Create the OpenCL kernel object
    cl_kernel kernel = clCreateKernel(program, "vector_add", &ret);
    // Set the arguments of the kernel
    ret = clSetKernelArg(kernel, 0, sizeof(cl_mem), (void *)&a_mem_obj);
    ret = clSetKernelArg(kernel, 1, sizeof(cl_mem), (void *)&b_mem_obj);
    ret = clSetKernelArg(kernel, 2, sizeof(cl_mem), (void *)&c_mem_obj);
    // Execute the OpenCL kernel on the array
    size_t global_item_size = LIST_SIZE;
    size_t local_item_size = 1;
    //Execute the kernel on the device
    cl_event event;
    ret = clEnqueueNDRangeKernel(command_queue, kernel, 1, NULL,
    &global_item_size, &local_item_size, 0, NULL, NULL);
    ret = clFinish(command_queue);
    // Read the memory buffer C on the device to the local variable C
    ret = clEnqueueReadBuffer(command_queue, c_mem_obj, CL_TRUE, 0,LIST_SIZE
    * sizeof(int), C, 0, NULL, NULL);
    // Display the result to the screen
    for(i = 0; i < LIST_SIZE; i++)
    printf("%d + %d = %d\n", A[i], B[i], C[i]);
    // Clean up
    ret = clFlush(command_queue);
    ret = clReleaseKernel(kernel);
    ret = clReleaseProgram(program);
    ret = clReleaseMemObject(a_mem_obj);
    ret = clReleaseMemObject(b_mem_obj);
    ret = clReleaseMemObject(c_mem_obj);
    ret = clReleaseCommandQueue(command_queue);
    ret = clReleaseContext(context);
    free(A);
    free(B);
    free(C);
    getchar();
    return 0;
    }

My kernel function is vectorCLKernel.cl

   __kernel void vector_add(__global int *A, __global int *B, __global int *C)
    {
    int i = get_global_id(0);
    C[i] = A[i] + B[i];
    }

I have properly compiled the code but cannot understand why i am getting zeroes as output

0 + 10 = 0

1 + 11 = 0

2 + 12 = 0

3 + 13 = 0

4 + 14 = 0

解决方案

First of all, I suggest you get in the habit of creating & using some C macro that checks the error from each and every OpenCL API call. It will make your life a lot easier.

2nd, the problem is that your calls to clCreateBuffer are invalid (all three). Notice:

clCreateBuffer(context,CL_MEM_READ_ONLY,LIST_SIZE * sizeof(int), A, &ret);

You're providing a host_ptr argument. Documentation to clCreateBuffer says:

CL_INVALID_HOST_PTR if host_ptr is NULL and CL_MEM_USE_HOST_PTR or CL_MEM_COPY_HOST_PTR are set in flags or if host_ptr is not NULL but CL_MEM_COPY_HOST_PTR or CL_MEM_USE_HOST_PTR are not set in flags.

... just put NULL in there instead of A/B/C.

这篇关于OpenCL向量加法程序的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！