#include <iostream>
using namespace std;

#include <cuda.h>
#include <stdio.h>

// Declarations of wrapper routines (prefixed with gpu_)
void gpu_VectAdd(float *A, float *B, float *C);


// Declarations of GPU routines (prefixed with kernel_)
__global__ void kernel_VectAdd(float *A, float *B, float *C);


//-----------------------
// gpu_VectAdd
//-----------------------
void gpu_VectAdd(float *A, float *B, float *C)
{
	float *Agpu, *Bgpu, *Cgpu;
	
	int size = 3*sizeof(float);
	cudaMalloc(&Agpu, size);
	cudaMalloc(&Bgpu, size);
	cudaMalloc(&Cgpu, size);
	
	cudaMemcpy(Agpu, A, size, cudaMemcpyHostToDevice);
	cudaMemcpy(Bgpu, B, size, cudaMemcpyHostToDevice);
		
	kernel_VectAdd<<<1,3>>>(Agpu, Bgpu, Cgpu);

	cudaMemcpy(C, Cgpu, size, cudaMemcpyDeviceToHost);
}


//-----------------------
// kernel_VectAdd
//-----------------------
__global__ 
void kernel_VectAdd(float *A, float *B, float *C)
{

	int i = threadIdx.x;
	C[i] = A[i] + B[i];
}