two_array_sum.cu

//#include "cuda_runtime.h"                                                                                                                                             
//#include "device_launch_parameters.h"                                                                                                                                 
#include <stdio.h>                                                                                                                                                      
#include<cuda.h>                                                                                                                                                        
#define n 5                                                                                                                                                             
__global__ void add(int *a, int *b, int *c) {                                                                                                                           

        int i = blockIdx.x * blockDim.x + threadIdx.x;                                                                                                                  
        if (i < n) {
                c[i] = a[i] + b[i];     
        }                                                                                                                                                               
}                                                                                                                                                                       
int main(){                                                                                                                                                             
        int a[n];                                                                                                                                                       
        int b[n];                                                                                                                                                       
        int i;                                                                                                                                                          
        int c[n]= {0};                                                                                                                                                  
                                                                                                                                                                        
        int* dev_a;                                                                                                                                                     
        int* dev_b;                                                                                                                                                     
        int* dev_c;                                                                                                                                                     
        cudaMalloc((void**)&dev_c, n * sizeof(int));                                                                                                                    
        cudaMalloc((void**)&dev_a, n * sizeof(int));                                                                                                                    
        cudaMalloc((void**)&dev_b, n * sizeof(int));                                                                                                                    
        printf("\narray elements (1st):\n");                                                                                                                            
        for(i=0;i<n;i++){                                                                                                                                               
        scanf("%d",&a[i]);                                                                                                                                              
        }                                                                                                                                                               
        printf("\narray elements (2nd):\n");                                                                                                                            
        for(i=0;i<n;i++){                                                                                                                                               
        scanf("%d",&b[i]);                                                                                                                                              
        }                                                                                                                                                               
        cudaMemcpy(dev_a, a, n * sizeof(int), cudaMemcpyHostToDevice);                                                                                                  
        cudaMemcpy(dev_b, b, n* sizeof(int), cudaMemcpyHostToDevice);                                                                                                   
        add<<<n,1>>>(dev_a, dev_b, dev_c);                                                                                                                              
        cudaMemcpy(c, dev_c, n* sizeof(int), cudaMemcpyDeviceToHost);                                                                                                   
        printf("\nsum is\n");                                                                                                                                           
        for(i=0;i<n;i++){                                                                                                                                               
                printf("%d\n",c[i]);                                                                                                                                    
        }                                                                                                                                                               
                                                                                                                                                                        
        cudaFree(dev_c);                                                                                                                                                
        cudaFree(dev_a);                                                                                                                                                
        cudaFree(dev_b);                                                                                                                                                
                                                                                                                                                                        
        cudaDeviceReset();                                                                                                                                              
        return 0;                                                                                                                                                       
}