Unable to print out correct CUDA matrix addition result - memory

Question

I am trying to add two matrices using blocks and threads in parallel but I am not getting the correct resultant matrix to print out. I think I am having trouble communicating to the device from the host and vice versa.
this is the output I'm getting. The resultant matrix is messed up.
matrix a:
18 27 48 28 6 16 40 15 30 41
30 15 25 24 8 0 7 18 7 23
0 15 47 13 26 16 6 17 39 30
6 25 11 22 44 34 37 38 31 15
8 16 17 0 29 6 13 3 30 46
24 21 30 48 15 23 47 41 26 21
25 45 14 15 27 20 44 14 7 39
28 49 3 40 35 40 45 0 37 5
31 17 24 1 48 6 6 2 46 42
10 43 20 17 14 23 2 21 26 8
matrix b:
44 40 26 14 2 49 6 20 46 36
0 9 5 46 13 26 29 7 46 13
39 41 30 28 4 6 34 32 43 47
30 15 46 19 46 45 49 35 1 8
31 16 36 7 31 38 34 25 26 24
30 11 11 13 13 25 40 14 42 7
30 40 49 44 13 1 44 26 13 22
13 25 41 31 13 10 36 9 18 14
28 25 33 31 41 27 5 11 44 31
32 5 44 7 22 45 42 26 24 24
The resultants matrix:
400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496
400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496
400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496 400848496
here is my code:
#include <stdio.h>
#include <cuda.h>
#include <stdlib.h>
#include <time.h>
#define N 10
#define BLOCK_DIM 5
__global__ void matrixAdd (int *a, int *b, int *c);
int main() {
int a[N][N], b[N][N], c[N][N];
int *dev_a, *dev_b, *dev_c;
int size = N * N * sizeof(int);
time_t t;
//initialize rand function
srand((unsigned) time(&t));
// initialize a and b with values
for(int i=0; i<N; i++ )
{
for(int j=0; j<N; j++)
{
a[i][j]= rand() % 50;
b[i][j]= rand() % 50;
}
}
// print matrix a
printf("matrix a: ");
for(int i=0; i<N; i++)
{
printf("\n\t ");
for(int j=0; j<N; j++)
{
printf("%d ", a[i][j]);
}
printf(" ");
}
printf("\n ");
//print matrix b
printf("matrix b: ");
for(int i=0; i<N; i++)
{
printf("\n\t ");
for(int j=0; j<N; j++)
{
printf("%d ", b[i][j]);
}
printf(" ");
}
printf("\n ");
//allocate memory
cudaMalloc((void**)&dev_a, size);
cudaMalloc((void**)&dev_b, size);
cudaMalloc((void**)&dev_c, size);
cudaMemcpy(dev_a, a, size, cudaMemcpyHostToDevice);
cudaMemcpy(dev_b, b, size, cudaMemcpyHostToDevice);
dim3 dimBlock(BLOCK_DIM, BLOCK_DIM);
dim3 dimGrid((int)ceil(N/dimBlock.x),(int)ceil(N/dimBlock.y));
matrixAdd<<<dimGrid,dimBlock>>>(dev_a,dev_b,dev_c);
/// __syncthreads(); function call from host not allowed
cudaMemcpy(c, dev_c, size, cudaMemcpyDeviceToHost);
// print matrix c
printf("The resultants matrix: \n");
// print matrix c
for(int i=0; i<N; i++)
{
printf("\n\t ");
for(int j=0; j<N; j++)
{
printf("%d ", c);
}
printf(" ");
}
cudaFree(dev_a);
cudaFree(dev_b);
cudaFree(dev_c);
return 0;
}
__global__ void matrixAdd (int *a, int *b, int *c) {
int col = blockIdx.x * blockDim.x + threadIdx.x;
int row = blockIdx.y * blockDim.y + threadIdx.y;
int index = col + row * N;
if (col < N && row < N)
{
c[index] = a[index] + b[index];
}
__syncthreads();
}

Answer 1

You have a typo in your final output printf statement. This:
printf("%d ", c);
should be this:
printf("%d ", c[i][j]);
(which is consistent with your previous printf statements)
FWIW, the final __syncthreads() statement in your kernel serves no purpose.