seperate the cuda host code in .cpp file - visual-c++

Question

main.cpp
#include<iostream>
#include "cuda.h"
using namespace std;
void cuda_calculation();
int main()
{
cuda_calculation();
return 0;
}
cu.h
void call(int , int ,float* , int );
cuda.cpp
#include <stdio.h>
#include <cuda.h>
#include "cu.h"
void cuda_calculation()
{
float *a_h, *a_d; // Pointer to host & device arrays
const int N = 10; // Number of elements in arrays
size_t size = N * sizeof(float);
a_h = (float *)malloc(size); // Allocate array on host
cudaMalloc((void **) &a_d, size); // Allocate array on device
// Initialize host array and copy it to CUDA device
for (int i=0; i<N; i++) a_h[i] = (float)i;
cudaMemcpy(a_d, a_h, size, cudaMemcpyHostToDevice);
// Do calculation on device:
int block_size = 4;
int n_blocks = N/block_size + (N%block_size == 0 ? 0:1);
void call(n_blocks, block_size,&a_d, N);
/*square_array <<< n_blocks, block_size >>> (a_d, N);*/
// Retrieve result from device and store it in host array
cudaMemcpy(a_h, a_d, sizeof(float)*N, cudaMemcpyDeviceToHost);
// Print results
for (int i=0; i<N; i++) printf("%d %f\n", i, a_h[i]);
// Cleanup
free(a_h); cudaFree(a_d);
}
cu.cu
#include <stdio.h>
#include "cu.h"
#include <cuda.h>
// Kernel that executes on the CUDA device
__global__ void square_array(float *a, int N)
{
int idx = blockIdx.x * blockDim.x + threadIdx.x;
if (idx<N) a[idx] = a[idx] * a[idx];
}
//}
void call(int a,int b,float* c,int d)
{
square_array <<< 3,4 >>> (c,d);
}
I tried to seperate the kernal code and host code in a cpp file, however the following error prevails:
Error 'cudaMemcpy': identifier not found and the other cuda related identifier is not identified.
how to use the cuda related identifier in cpp file and call the kernal functions

Answer 1

There are some errors: void cuda_calculation(); needs to be visible to main.cpp through a header file (cu.h).
Also make sure to compile your .cu files with nvcc and NOT as a standard C++ file. Use CUDA compilation rules to make this process easy (installed by default as part of CUDA toolkit)

Answer 2

after a long trial ,I came with the proper output,
to include the cuda identifier in the cpp files we not only need to include cuda.h but also we need to include cuda_runtime.h as
cuda.cpp as
#include <stdio.h>
#include <cuda.h>
#include<cuda_runtime.h>
#include "cu.h"
#include "cud.h"
//void call(int , int ,float * , int );
void cuda_calculation()
{
float *a_h, *a_d; // Pointer to host & device arrays
const int N = 10; // Number of elements in arrays
size_t size = N * sizeof(float);
a_h = (float *)malloc(size); // Allocate array on host
cudaMalloc((void **) &a_d, size); // Allocate array on device
// Initialize host array and copy it to CUDA device
for (int i=0; i<N; i++) a_h[i] = (float)i;
cudaMemcpy(a_d, a_h, size, cudaMemcpyHostToDevice);
// Do calculation on device:
int block_size = 4;
int n_blocks = N/block_size + (N%block_size == 0 ? 0:1);
call(n_blocks, block_size,a_d, N);
/*square_array <<< n_blocks, block_size >>> (a_d, N);*/
// Retrieve result from device and store it in host array
cudaMemcpy(a_h, a_d, sizeof(float)*N, cudaMemcpyDeviceToHost);
// Print results
for (int i=0; i<N; i++) printf("%d %f\n", i, a_h[i]);
// Cleanup
free(a_h);
cudaFree(a_d);
}
so the others files are
main.cpp
#include<iostream>
#include "cud.h"
using namespace std;
int main()
{
cuda_calculation();
return 0;
}
cud.h
void cuda_calculation();
cu.h
void call(int , int ,float* , int );
cu.cu
#include <stdio.h>
#include "cu.h"
#include <cuda.h>
// Kernel that executes on the CUDA device
__global__ void square_array(float *a, int N)
{
int idx = blockIdx.x * blockDim.x + threadIdx.x;
if (idx<N) a[idx] = a[idx] * a[idx];
}
//}
void call(int a,int b,float* c,int d)
{
square_array <<< 3,4 >>> (c,d);
}