add_two_arrays_cuda.cu

#include "cuda_runtime.h"
#include "device_launch_parameters.h"
#include <stdio.h>
#include <stdlib.h>

#define N 200


void dataToArray(int* a, int n){
   int i;
   for (i = 0; i < n; ++i){
    a[i] = i;   
   }
}

__global__ void getSum(int *n1, int *n2, int *sumData,int size) {
    
    int index = threadIdx.x + blockIdx.x * blockDim.x;
    if(index<size){
        sumData[index] = n1[index] + n2[index];
    }
	  
}

int main() {

	int *n1, *n2, *sum;
	int *dN1, *dN2, *dSum;

	int size = sizeof(int) * N;

	// allocate device memory
	cudaMalloc((void **)&dN1, size);
	cudaMalloc((void **)&dN2, size);
	cudaMalloc((void **)&dSum, size);

  	// allocate host memory
	n1 = (int *)malloc(size);
	n2 = (int *)malloc(size);
	sum = (int *)malloc(size);

	// add data to arrays
	dataToArray(n1,N);
	dataToArray(n2,N);

	//copy from host memory to device memory
	cudaMemcpy(dN1, n1, size, cudaMemcpyHostToDevice);
	cudaMemcpy(dN2, n2, size, cudaMemcpyHostToDevice);


	dim3 Dg(N/5);
	dim3 Db(5);

	getSum << <Dg, Db >> > (dN1, dN2, dSum,N);

	cudaMemcpy(sum, dSum, size, cudaMemcpyDeviceToHost);

	for(int i=0;i<N;i++){
	   printf("n1[%d] + n2[%d] = %d\n", i,i,sum[i]);
	}
	
	free(n1);
	free(n2);
	free(sum);
	cudaFree(dN1);
	cudaFree(dN2);
	cudaFree(dSum);

	return 0;

}