linux opencl 安装

sudo apt update
sudo apt install ocl-icd-opencl-dev

clinfo

#include <CL/cl.h>
#include <stdio.h>
#include <stdlib.h>

#define ARRAY_SIZE 1024

int main() {
    cl_platform_id platform;
    cl_device_id device;
    cl_context context;
    cl_command_queue queue;
    cl_program program;
    cl_kernel kernel;
    cl_mem bufferA, bufferB, bufferC;
    float *A, *B, *C;

    // Initialize data
    A = (float*)malloc(ARRAY_SIZE * sizeof(float));
    B = (float*)malloc(ARRAY_SIZE * sizeof(float));
    C = (float*)malloc(ARRAY_SIZE * sizeof(float));
    for (int i = 0; i < ARRAY_SIZE; i++) {
        A[i] = i;
        B[i] = i * 2;
    }

    // Get platform and device information
    clGetPlatformIDs(1, &platform, NULL);
    clGetDeviceIDs(platform, CL_DEVICE_TYPE_GPU, 1, &device, NULL);

    // Create context and command queue
    context = clCreateContext(NULL, 1, &device, NULL, NULL, NULL);
    queue = clCreateCommandQueue(context, device, 0, NULL);

    // Create buffers
    bufferA = clCreateBuffer(context, CL_MEM_READ_ONLY, ARRAY_SIZE * sizeof(float), NULL, NULL);
    bufferB = clCreateBuffer(context, CL_MEM_READ_ONLY, ARRAY_SIZE * sizeof(float), NULL, NULL);
    bufferC = clCreateBuffer(context, CL_MEM_WRITE_ONLY, ARRAY_SIZE * sizeof(float), NULL, NULL);

    // Copy data to buffers
    clEnqueueWriteBuffer(queue, bufferA, CL_TRUE, 0, ARRAY_SIZE * sizeof(float), A, 0, NULL, NULL);
    clEnqueueWriteBuffer(queue, bufferB, CL_TRUE, 0, ARRAY_SIZE * sizeof(float), B, 0, NULL, NULL);

    // Create and compile program
    const char *source = "__kernel void vector_add(__global const float *A, __global const float *B, __global float *C) { int i = get_global_id(0); C[i] = A[i] + B[i]; }";
    program = clCreateProgramWithSource(context, 1, &source, NULL, NULL);
    clBuildProgram(program, 1, &device, NULL, NULL, NULL);

    // Create kernel
    kernel = clCreateKernel(program, "vector_add", NULL);

    // Set kernel arguments
    clSetKernelArg(kernel, 0, sizeof(cl_mem), (void*)&bufferA);
    clSetKernelArg(kernel, 1, sizeof(cl_mem), (void*)&bufferB);
    clSetKernelArg(kernel, 2, sizeof(cl_mem), (void*)&bufferC);

    // Execute kernel
    size_t global_size = ARRAY_SIZE;
    clEnqueueNDRangeKernel(queue, kernel, 1, NULL, &global_size, NULL, 0, NULL, NULL);

    // Read result
    clEnqueueReadBuffer(queue, bufferC, CL_TRUE, 0, ARRAY_SIZE * sizeof(float), C, 0, NULL, NULL);

    // Verify result
    for (int i = 0; i < ARRAY_SIZE; i++) {
        if (C[i] != A[i] + B[i]) {
            printf("Error at index %d\n", i);
            break;
        }
    }

    // Clean up
    clReleaseMemObject(bufferA);
    clReleaseMemObject(bufferB);
    clReleaseMemObject(bufferC);
    clReleaseKernel(kernel);
    clReleaseProgram(program);
    clReleaseCommandQueue(queue);
    clReleaseContext(context);
    free(A);
    free(B);
    free(C);

    return 0;
}

gcc -o vector_add vector_add.c -lOpenCL
./vector_add