cuda_opencv之向量相加

实现向量相加。。。实现与 矩阵相加一样~

#include <stdlib.h>
#include <stdio.h>
#include <opencv/cv.hpp>
#include <opencv/highgui.h>

#include <cuda_runtime.h>
#include <device_launch_parameters.h>

using namespace std;
using namespace cv;

#define N 10

__global__ void Add_kernel(const int2* d_a, const int2* d_b,int2*  d_c ,int pp)
{
    int x=threadIdx.x;
    
    if(x < pp)
    {
        d_c[x].x=d_a[x].x + d_b[x].x;
        d_c[x].y=d_a[x].y + d_b[x].y;

    }

}
int main()
{
    Mat a(1,N,CV_32S,Scalar_<int>(0));
    Mat b(1,N,CV_32S,Scalar_<int>(10));

    cout<<a<<endl;
    cout<<endl;
    cout<<b<<endl;
    cout<<endl;

    size_t memSize = a.step * a.rows;

    int2* d_a = NULL;
    int2* d_b = NULL;
    int2* d_c = NULL;

    cudaMalloc((void**)&d_a,memSize);
    cudaMalloc((void**)&d_b,memSize);
    cudaMalloc((void**)&d_c,memSize);

    cudaMemcpy(d_a,a.data,memSize,cudaMemcpyHostToDevice);
    cudaMemcpy(d_b,b.data,memSize,cudaMemcpyHostToDevice);

    Add_kernel<<<1,N>>>(d_a, d_b, d_c, N);
    cudaMemcpy(a.data, d_c,memSize, cudaMemcpyDeviceToHost);
    cout<<endl;
    cout<<a<<endl;
    system("pause");
    return 0;
}