eosadmins-MacBook-Pro:st7 smasuti$ cat simple1.cu
#include <stdio.h>

__global__ void kernelCode(float *result)
{
    int index = threadIdx.x+blockIdx.x*blockDim.x;
    if(index<5000)
    {
	atomicAdd(result,1.0f);
    }
}

int main ()
{
	float *d;
	dim3 dimGrid (5, 1,1) ;
	dim3 dimBlock (1000,1,1) ;
	float abc ;
	cudaError_t cuError = cudaSuccess ;

	cuError = cudaMalloc ((void **) &d, sizeof(float)) ;
	if (cuError != cudaSuccess)
	{
	    	printf ("Error\n") ;
	}
	cudaMemset (d,0,sizeof(float)) ;
	if (cuError != cudaSuccess)
	{
		printf ("Error\n") ;
	}

	kernelCode <<<dimGrid, dimBlock>>> (d) ;

	if (cudaDeviceSynchronize() != cudaSuccess)
	{
	    printf ("Error\n") ;
	}

	cudaMemcpy (&abc, d, sizeof(float), cudaMemcpyDeviceToHost) ;

	printf ("%e\n", abc) ;
	return 0 ;

}
eosadmins-MacBook-Pro:st7 smasuti$ ./simple1
5.000000e+03
eosadmins-MacBook-Pro:st7 smasuti$