// moveArrays.cu

 //

 // demonstrates CUDA interface to data allocation on device (GPU)

 // and data movement between host (CPU) and device.

 #include <stdio.h>

 #include <assert.h>

 #include <cuda.h>

 #include "cuda_runtime.h"

 #include "device_launch_parameters.h"

 int main(void)

 {

     float *a_h, *b_h;     //指向主机的指针

     float *a_d, *b_d;     //指向设备的指针

     int N = ;

     int i;

     //在主机端申请内存

     a_h = (float *)malloc(sizeof(float)*N);

     b_h = (float *)malloc(sizeof(float)*N);

     //在设备端申请存储

     cudaMalloc((void **)&a_d, sizeof(float)*N);

     cudaMalloc((void **)&b_d, sizeof(float)*N);

     //主机端数据初始化

     for (i = ; i < N; i++) {

         a_h[i] = .f + i;

         b_h[i] = .f;

     }

     //将数据从主机端传送到设备端：a_h-->a_d

     cudaMemcpy(a_d, a_h, sizeof(float)*N, cudaMemcpyHostToDevice);

     //在设备内传输数据

     cudaMemcpy(b_d, a_d, sizeof(float)*N, cudaMemcpyDeviceToDevice);

     //将数据从设备端传输到主机：b_d-->b_h

     cudaMemcpy(b_h, b_d, sizeof(float)*N, cudaMemcpyDeviceToHost);

     //核对结果

     for (i = ; i < N; i++)

         assert(a_h[i] == b_h[i]);

     //释放主机端存储

     free(a_h);

     free(b_h);

     //释放设备端存储

     cudaFree(a_d);

     cudaFree(b_d);

 }