OpenCL - 内核方法返回意外结果

Question

我是 OpenCL 的初学者。我尝试运行一个非常简单的内核代码，将向量的每个值加 1。一切运行良好，不返回错误代码（我在每一步后检查了返回值）。源代码：

cl_device_id device_id = NULL;
cl_context context = NULL;
cl_command_queue command_queue = NULL;
cl_mem memobj , resobj = NULL;

cl_program program = NULL;
cl_kernel kernel = NULL;
cl_platform_id platform_id = NULL;
cl_uint ret_num_devices;
cl_uint ret_num_platforms;
cl_int ret;
size_t work_units_per_kernels;
int input[10] =  {1,2,3,4,5,6,7,8,9,10};
int output[10];
int length = 10 ;


FILE *fp;
char fileName[] = "/home/tuan/OpenCLPlayaround/hello.cl";
char *source_str;
size_t source_size;

/* Load the source code containing the kernel*/
fp = fopen(fileName, "r");
if (!fp) {
fprintf(stderr, "Failed to load kernel.\n");
exit(1);
}
source_str = (char*)malloc(0x100000);
source_size = fread(source_str,1,0x100000, fp);
fclose(fp);

ret = clGetPlatformIDs(1, &platform_id, &ret_num_platforms);
std::cout<<ret<<" code"<<std::endl;

ret = clGetDeviceIDs(platform_id, CL_DEVICE_TYPE_DEFAULT, 1, &device_id, &ret_num_devices);
std::cout<<ret<<" code"<<std::endl;

context = clCreateContext(NULL, 1, &device_id, NULL, NULL, &ret);
std::cout<<ret<<" code"<<std::endl;

command_queue = clCreateCommandQueue(context, device_id, 0, &ret);
//Check Concept of memory
memobj = clCreateBuffer(context, CL_MEM_READ_WRITE | CL_MEM_COPY_HOST_PTR,length * sizeof(int), input, &ret);


resobj = clCreateBuffer(context, CL_MEM_READ_WRITE | CL_MEM_COPY_HOST_PTR, length * sizeof(int), output, &ret);
std::cout<<ret<<" code"<<std::endl;


program = clCreateProgramWithSource(context,1,(const char**)&source_str, (const size_t*)&source_size, &ret);

ret = clBuildProgram(program, 1, &device_id, NULL, NULL, NULL);

kernel = clCreateKernel(program, "hello", &ret);

ret = clSetKernelArg(kernel,0, sizeof(memobj),(void *)&memobj);

ret = clSetKernelArg(kernel,1, sizeof(resobj),(void *)&resobj);

ret = clEnqueueTask(command_queue, kernel, 0, NULL,NULL);

ret = clEnqueueReadBuffer(command_queue, resobj, CL_TRUE, 0, length* sizeof(int),output, 0, NULL, NULL);

for (int i = 0 ; i <10 ; i++) {
    std::cout<<output[i]<<" "<<std::endl;
}

return 0;

结果有点奇怪，而应该是 {2,3,4,5,6,7,8,9,10,11} ：

2 
-16777216 
65535 
1 
-1242789408 
32767 
4201449 
0 
2 
0 

还有我的内核：

__kernel void hello(__global int* a, __global int* b)
{
    int sam = 0;
    int gid = get_global_id(0);
    b[gid] = sam + a[gid] +1 ;
}

有人可以解释为什么吗？它让我头疼了好几个小时！

Answer 1

clEnqueueTask 等效于调用clEnqueueNDRangeKernel 并将work_dim = 1、global_work_offset = NULL、global_work_size[0] 设置为1，并将local_work_size[0] 设置为1。

所以使用clEnqueueNDRangeKernel。