5 CUDA学习笔记——内存申请，拷贝与释放

申请主机端内存

    int nx = 1<<14;
    int ny = 1<<14;
    int nxy = nx*ny;
    int nBytes = nxy*sizeof(float);
     
    //mlloc host global memory
    float *h_A, *h_B, *hostRef, *gpuRef;
    h_A=(float *)malloc(nBytes);
    h_B=(float *)malloc(nBytes);
    hostRef = (float *)malloc(nBytes);
    gpuRef =  (float *)malloc(nBytes);

申请设备端内存

    //malloc device gloabl memory
    float *d_MatA, *d_MatB, *d_MatC;
    cudaMalloc((void **)&d_MatA, nBytes);
    cudaMalloc((void **)&d_MatB, nBytes);
    cudaMalloc((void **)&d_MatC, nBytes);

将数据拷贝到设备端

    //transform data from host to device
    cudaMemcpy(d_MatA, h_A, nBytes, cudaMemcpyHostToDevice);
    cudaMemcpy(d_MatB, h_B, nBytes, cudaMemcpyHostToDevice);

将数据拷贝到主机端

    //copy kernel result back to host side
    cudaMemcpy(gpuRef,d_MatC,nBytes,cudaMemcpyDeviceToHost);

内存释放

    //free memory
    cudaFree(d_MatA);
    cudaFree(d_MatB);
    cudaFree(d_MatC);
    free(h_A);
    free(h_B);
    free(gpuRef);
    free(hostRef);