#include <stdio.h>
#include <iostream>
const int N = 1024;


__global__
void hello(int *b)
{
        b[threadIdx.x]=24;
}

int main()
{

        int b[N];
		for(int i=0;i<N;i++)
		{
			b[i]=1;
		}

        int *bd;

        const int isize = N*sizeof(int);


        cudaMalloc( (void**)&bd, isize );

        cudaMemcpy( bd, b, isize, cudaMemcpyHostToDevice );

        hello<<<N,N>>>(bd);
        cudaMemcpy( b, bd, isize, cudaMemcpyDeviceToHost );
        cudaFree( bd );
        std::cout<<b[1]<<std::endl;

        return EXIT_SUCCESS;
}