CUDA版本的OpenCL在windows 7的下编程初步

参考文献：

http://blog.csdn.net/neoxmu/article/details/8866928

我安装的是CUDA5.5,代码如下：

//#include "stdafx.h"
#include "CL\cl.h"
#include <stdlib.h>
#include <stdio.h>

#pragma comment(lib,"OpenCL.lib")

#define CL_VERBOSE
void openclRetTackle(cl_int retValue, char* processInfo){
	if(retValue!=CL_SUCCESS){
#if (defined CL_DEBUG) || (defined CL_VERBOSE)
		printf("%s Error!\n",processInfo);
#endif
		exit(-1);
	}else{
#ifdef CL_VERBOSE
		printf("%s Success!\n",processInfo);
#endif
	}
}

cl_platform_id cpPlatform;
cl_device_id cdDevice;
cl_context cxGPUContext;
cl_command_queue cqCommandQueue;

int openclInit()
{
	cl_int ret;
	//得到平台ID
	openclRetTackle( clGetPlatformIDs(1, &cpPlatform, NULL), "clGetPlatFormIDs");
	//得到GPU设备ID
	openclRetTackle( clGetDeviceIDs(cpPlatform, CL_DEVICE_TYPE_GPU,1,&cdDevice,NULL), "clGetDeviceIDs");
	//获取GPU设备上下文
	cxGPUContext = clCreateContext(0, 1, &cdDevice, NULL, NULL, &ret);
	openclRetTackle( ret , "clCreateContext" );
	//开辟任务队列
	cqCommandQueue = clCreateCommandQueue(cxGPUContext, cdDevice, 0, &ret);
	openclRetTackle( ret , "clCreateCommandQueue");
	return CL_SUCCESS;
}

int run()
{
	openclInit();
	system("pause");
	return 0;
}

<span style="font-family:Microsoft YaHei;font-size:18px;">//#include "stdafx.h"
#include <stdio.h>
#include <vector>
#include <CL/cl.h>
#include <iostream>
#include <fstream>
#include <string>

#pragma comment(lib,"OpenCL.lib")

int print_device()
{
	cl_int err;
	cl_uint num;
	err = clGetPlatformIDs(0, 0, &num);
	if(err != CL_SUCCESS)
	{
		std::cerr << "Unable to get platforms\n";
		return 0;
	}
	std::vector<cl_platform_id> platforms(num);
	err = clGetPlatformIDs(num, &platforms[0], &num);
	if(err != CL_SUCCESS)
	{
		std::cerr << "Unable to get platform ID\n";
		return 0;
	}

	cl_context_properties prop[] = { CL_CONTEXT_PLATFORM, reinterpret_cast<cl_context_properties>(platforms[0]), 0 };
	cl_context context = clCreateContextFromType(prop, CL_DEVICE_TYPE_DEFAULT, NULL, NULL, NULL);
	if(context == 0)
	{
		std::cerr << "Can't create OpenCL context\n";
		return 0;
	}

	size_t cb;
	clGetContextInfo(context, CL_CONTEXT_DEVICES, 0, NULL, &cb);
	std::vector<cl_device_id> devices(cb / sizeof(cl_device_id));
	clGetContextInfo(context, CL_CONTEXT_DEVICES, cb, &devices[0], 0);

	clGetDeviceInfo(devices[0], CL_DEVICE_NAME, 0, NULL, &cb);
	std::string devname;
	devname.resize(cb);
	clGetDeviceInfo(devices[0], CL_DEVICE_NAME, cb, &devname[0], 0);
	std::cout << "Device: " << devname.c_str() << "\n";

	clReleaseContext(context);
	return 0;

}

cl_program load_program(cl_context context, const char* filename)
{
	std::ifstream in(filename, std::ios_base::binary);
	if(!in.good())
	{
		return 0;

	}// get file length
	in.seekg(0, std::ios_base::end);
	size_t length = in.tellg();
	in.seekg(0, std::ios_base::beg);

	// read program source
	std::vector<char> data(length + 1);
	in.read(&data[0], length);
	data[length] = 0;

	// create and build program
	const char* source = &data[0];
	cl_program program = clCreateProgramWithSource(context, 1, &source, 0, 0);
	if(program == 0)
	{
		return 0;
	}
	if(clBuildProgram(program, 0, 0, 0, 0, 0) != CL_SUCCESS)
	{
		return 0;
	}
	return program;
}
int main()
{
	print_device();
	cl_int err;
	cl_uint num;
	err = clGetPlatformIDs(0, 0, &num);
	if(err != CL_SUCCESS)
	{
		std::cerr << "Unable to get platforms\n";
		return 0;
	}

	std::vector<cl_platform_id> platforms(num);
	err = clGetPlatformIDs(num, &platforms[0], &num);
	if(err != CL_SUCCESS)
	{
		std::cerr << "Unable to get platform ID\n";
		return 0;
	}
	cl_context_properties prop[] = { CL_CONTEXT_PLATFORM, reinterpret_cast<cl_context_properties>(platforms[0]), 0 };
	cl_context context = clCreateContextFromType(prop, CL_DEVICE_TYPE_DEFAULT, NULL, NULL, NULL);
	if(context == 0)
	{
		std::cerr << "Can't create OpenCL context\n";
		return 0;
	}

	size_t cb;
	clGetContextInfo(context, CL_CONTEXT_DEVICES, 0, NULL, &cb);
	std::vector<cl_device_id> devices(cb / sizeof(cl_device_id));
	clGetContextInfo(context, CL_CONTEXT_DEVICES, cb, &devices[0], 0);

	clGetDeviceInfo(devices[0], CL_DEVICE_NAME, 0, NULL, &cb);
	std::string devname;
	devname.resize(cb);
	clGetDeviceInfo(devices[0], CL_DEVICE_NAME, cb, &devname[0], 0);
	std::cout << "Device: " << devname.c_str() << "\n";

	cl_command_queue queue = clCreateCommandQueue(context, devices[0], 0, 0);
	if(queue == 0)
	{
		std::cerr << "Can't create command queue\n";
		clReleaseContext(context);
		return 0;
	}

	const int DATA_SIZE = 1048576;
	std::vector<float> a(DATA_SIZE), b(DATA_SIZE), res(DATA_SIZE);
	for(int i = 0; i < DATA_SIZE; i++)
	{
		a[i] = std::rand();
		b[i] = std::rand();
	}

	cl_mem cl_a = clCreateBuffer(context, CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR, sizeof(cl_float) * DATA_SIZE, &a[0], NULL);
	cl_mem cl_b = clCreateBuffer(context, CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR, sizeof(cl_float) * DATA_SIZE, &b[0], NULL);
	cl_mem cl_res = clCreateBuffer(context, CL_MEM_WRITE_ONLY, sizeof(cl_float) * DATA_SIZE, NULL, NULL);
	if(cl_a == 0 || cl_b == 0 || cl_res == 0)
	{
		std::cerr << "Can't create OpenCL buffer\n";
		clReleaseMemObject(cl_a);
		clReleaseMemObject(cl_b);
		clReleaseMemObject(cl_res);
		clReleaseCommandQueue(queue);
		clReleaseContext(context);
		return 0;
	}

	cl_program program = load_program(context, "..\\shader.txt");
	if(program == 0)
	{
		std::cerr << "Can't load or build program\n";
		clReleaseMemObject(cl_a);
		clReleaseMemObject(cl_b);
		clReleaseMemObject(cl_res);
		clReleaseCommandQueue(queue);
		clReleaseContext(context);
		return 0;
	}
	cl_kernel adder = clCreateKernel(program, "adder", 0);
	if(adder == 0)
	{
		std::cerr << "Can't load kernel\n";
		clReleaseProgram(program);
		clReleaseMemObject(cl_a);
		clReleaseMemObject(cl_b);
		clReleaseMemObject(cl_res);
		clReleaseCommandQueue(queue);
		clReleaseContext(context);
		return 0;
	}

	clSetKernelArg(adder, 0, sizeof(cl_mem), &cl_a);

	clSetKernelArg(adder, 1, sizeof(cl_mem), &cl_b);

	clSetKernelArg(adder, 2, sizeof(cl_mem), &cl_res);

	size_t work_size = DATA_SIZE;

	err = clEnqueueNDRangeKernel(queue, adder, 1, 0, &work_size, 0, 0, 0, 0);
	if(err == CL_SUCCESS)
	{

		err = clEnqueueReadBuffer(queue, cl_res, CL_TRUE, 0, sizeof(float) * DATA_SIZE, &res[0], 0, 0, 0);
	}
	if(err == CL_SUCCESS)
	{
		bool correct = true;
		for(int i = 0; i < DATA_SIZE; i++)
		{
			if(a[i] + b[i] != res[i])
			{
				correct = false;
				break;
			}
		}
		if(correct)
		{

			std::cout << "Data is correct\n";
		}
		else
		{

			std::cout << "Data is incorrect\n";

		}
	}

	else
	{
		std::cerr << "Can't run kernel or read back data\n";
	}

	clReleaseKernel(adder);
	clReleaseProgram(program);
	clReleaseMemObject(cl_a);
	clReleaseMemObject(cl_b);
	clReleaseMemObject(cl_res);
	clReleaseCommandQueue(queue);
	clReleaseContext(context);
	return 0;

}</span>

需要使用的数据：

shader.txt

<span style="font-family:Microsoft YaHei;font-size:18px;">__kernel void adder(__global const float* a, __global const float* b, __global float* result)
{
    int idx = get_global_id(0);
    result[idx] = a[idx] + b[idx];
}</span>