数组之和_opencl实现示例

东方耀 · 发表于 2021-7-28 10:04:03

数组之和_opencl实现示例

CMAKE_MINIMUM_REQUIRED(VERSION 2.8)
PROJECT(main_cl)
#FIND_PACKAGE(CUDA REQUIRED)
FIND_PACKAGE(OpenCL REQUIRED)
ADD_EXECUTABLE(main_cl 数组之和_opencl实现.cpp)
#TARGET_LINK_LIBRARIES(main_cl)
#target_link_libraries(main_cl -lboost_system -lboost_thread)
target_link_libraries(main_cl OpenCL)
#Looking for CL_VERSION_1_2 - found
#Found OpenCL: /usr/lib/x86_64-linux-gnu/libOpenCL.so (found version "1.2")

复制代码

__kernel void hello_kernel(__global const float *a,
__global const float *b,
__global float *result)
{
int gid = get_global_id(0);
result[gid] = a[gid] + b[gid];
}

复制代码

#include <iostream>
#include <fstream>
#include <sstream>
#include <CL/cl.h>
const int ARRAY_SIZE = 1000;
//https://blog.csdn.net/jaccen2012/article/details/78810546
//一、选择OpenCL平台并创建一个上下文
cl_context CreateContext()
{
cl_int errNum;
cl_uint numPlatforms;
cl_platform_id firstPlatformId;
cl_context context = NULL;
//选择可用的平台中的第一个
errNum = clGetPlatformIDs(1, &firstPlatformId, &numPlatforms);
if (errNum != CL_SUCCESS || numPlatforms <= 0)
{
std::cerr << "Failed to find any OpenCL platforms." << std::endl;
return NULL;
}
std::cerr << "发现的openCL平台数目=" << numPlatforms << std::endl;
//创建一个OpenCL上下文环境
cl_context_properties contextProperties[] =
{
CL_CONTEXT_PLATFORM,
(cl_context_properties)firstPlatformId,
0
};
context = clCreateContextFromType(contextProperties, CL_DEVICE_TYPE_GPU,
NULL, NULL, &errNum);
return context;
}
//二、创建设备并创建命令队列
cl_command_queue CreateCommandQueue(cl_context context, cl_device_id *device)
{
cl_int errNum;
cl_device_id *devices;
cl_command_queue commandQueue = NULL;
size_t deviceBufferSize = -1;
// 获取设备缓冲区大小
errNum = clGetContextInfo(context, CL_CONTEXT_DEVICES, 0, NULL, &deviceBufferSize);
if (deviceBufferSize <= 0)
{
std::cerr << "No devices available.";
return NULL;
}
std::cout << "设备缓冲区大小=" << deviceBufferSize << std::endl;
// 为设备分配缓存空间
devices = new cl_device_id[deviceBufferSize / sizeof(cl_device_id)];
errNum = clGetContextInfo(context, CL_CONTEXT_DEVICES, deviceBufferSize, devices, NULL);
//选取可用设备中的第一个
commandQueue = clCreateCommandQueue(context, devices[0], 0, NULL);
*device = devices[0];
delete[] devices;
return commandQueue;
}
// 三、创建和构建程序对象
cl_program CreateProgram(cl_context context, cl_device_id device, const char* fileName)
{
cl_int errNum;
cl_program program;
std::cout << "文件名(这里要用绝对路径)：" << fileName << std::endl;
// 并行计算：两个数组求和
std::ifstream kernelFile(fileName, std::ios::in);
//如果打开操作失败，open函数的返回值为0(假)，
//如果是用调用构造函数的方式打开文件的，则流对象的值为0。可以据此测试打开是否成功
if (!kernelFile.is_open())
{
std::cerr << "打开文件失败了: " << fileName << std::endl;
return NULL;
}
std::ostringstream oss; // 输出的字符串流
oss << kernelFile.rdbuf(); // 这个api干啥的？
std::string srcStdStr = oss.str();
const char *srcStr = srcStdStr.c_str();
// 用源码创建并行的程序
program = clCreateProgramWithSource(context, 1,
(const char**)&srcStr,
NULL, NULL);
errNum = clBuildProgram(program, 0, NULL, NULL, NULL, NULL);
return program;
}
//创建和构建程序对象
bool CreateMemObjects(cl_context context, cl_mem memObjects[3],
float *a, float *b)
{
memObjects[0] = clCreateBuffer(context, CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR,
sizeof(float) * ARRAY_SIZE, a, NULL);
memObjects[1] = clCreateBuffer(context, CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR,
sizeof(float) * ARRAY_SIZE, b, NULL);
memObjects[2] = clCreateBuffer(context, CL_MEM_READ_WRITE,
sizeof(float) * ARRAY_SIZE, NULL, NULL);
return true;
}
// 释放OpenCL资源
void Cleanup(cl_context context, cl_command_queue commandQueue,
cl_program program, cl_kernel kernel, cl_mem memObjects[3])
{
for (int i = 0; i < 3; i++)
{
if (memObjects[i] != 0)
clReleaseMemObject(memObjects[i]);
}
if (commandQueue != 0)
clReleaseCommandQueue(commandQueue);
if (kernel != 0)
clReleaseKernel(kernel);
if (program != 0)
clReleaseProgram(program);
if (context != 0)
clReleaseContext(context);
}
//数组之和_opencl实现示例
int main(int argc, char** argv)
{
cl_context context = 0;
cl_command_queue commandQueue = 0;
cl_program program = 0;
cl_device_id device = 0;
cl_kernel kernel = 0;
cl_mem memObjects[3] = { 0, 0, 0 };
cl_int errNum;
// 一、选择OpenCL平台并创建一个上下文
context = CreateContext();
// 二、创建设备并创建命令队列
commandQueue = CreateCommandQueue(context, &device);
//创建和构建程序对象
program = CreateProgram(context, device, "/home/jiang/jjj_opencl_works/HelloWorld.cl");
// 四、创建OpenCL内核并分配内存空间
kernel = clCreateKernel(program, "hello_kernel", NULL);
//创建要处理的数据
float result[ARRAY_SIZE];
float a[ARRAY_SIZE];
float b[ARRAY_SIZE];
for (int i = 0; i < ARRAY_SIZE; i++)
{
a[i] = (float)i;
b[i] = (float)(ARRAY_SIZE - i);
}
//创建内存对象
if (!CreateMemObjects(context, memObjects, a, b))
{
Cleanup(context, commandQueue, program, kernel, memObjects);
return 1;
}
// 五、设置内核数据并执行内核
errNum = clSetKernelArg(kernel, 0, sizeof(cl_mem), &memObjects[0]);
errNum |= clSetKernelArg(kernel, 1, sizeof(cl_mem), &memObjects[1]);
errNum |= clSetKernelArg(kernel, 2, sizeof(cl_mem), &memObjects[2]);
size_t globalWorkSize[1] = { ARRAY_SIZE };
size_t localWorkSize[1] = { 1 };
errNum = clEnqueueNDRangeKernel(commandQueue, kernel, 1, NULL,
globalWorkSize, localWorkSize,
0, NULL, NULL);
// 六、读取执行结果并释放OpenCL资源
errNum = clEnqueueReadBuffer(commandQueue, memObjects[2], CL_TRUE,
0, ARRAY_SIZE * sizeof(float), result,
0, NULL, NULL);
std::cout << "程序执行计算结果：" << std::endl;
for (int i = 0; i < ARRAY_SIZE; i++)
{
std::cout << result[i] << " ";
}
std::cout << std::endl;
std::cout << "Executed program succesfully." << std::endl;
getchar();
Cleanup(context, commandQueue, program, kernel, memObjects);
return 0;
}

复制代码

		自动登录	找回密码
密码			立即注册

[C/C++] 数组之和_opencl实现示例