1) a.cpp for loading/saving the input/output data and doing FFT/IFFT
     2) b.cu for multiplying some constant (my own code)

....


cufftHandle forward_plan, inverse_plan;

int rank = 2;
int nCols = samples;
int batch = frames;
int nRows = 128;
int n[2] = {nRows, nCols};

int idist = nRows*nCols;
int odist = nRows*(nCols/2+1);

int inembed[] = {nRows, nCols};
int onembed[] = {nRows, nCols/2+1};

int istride = 1;
int ostride = 1;

cufftPlanMany(&forward_plan, rank, n,
                    inembed, istride, idist,
                    onembed, ostride, odist,
                    CUFFT_R2C, batch);

cufftReal *d_in;
cufftComplex *d_freq;
cufftComplex *outputData = (cufftComplex *)malloc(nRows*(nCols/2+1)*batch*sizeof(cufftComplex));

cudaMalloc(&d_in, sizeof(cufftReal)*nRows*nCols*batch);
cudaMAlloc(&d_freq, sizeof(cufftComplex)*nRows*(nCols/2+1)*batch);

cudaMemcpy(d_in, (cufftReal *)data, sizeof(cufftReal)*nRows*nCols*batch, cudaMemcpyHostToDevice);


/* FFT */
cufftExecR2C(forward_plan, d_in, d_freq);

cudaDeviceSynchronize();

/* multiply with H(f) */
blockSize = 128;
gridSize = ceil(128/(samples*batch));
mulH<<<gridSize,blockSize>>>(d_freq, samples);

cudaDeviceSynchronize();

  /* IFFT */
cufftPlanMany(&inverse_plan, rank, n, onembed, ostride, odist,
                    inembed, istride, idist, CUFFT_C2C, batch);

cufftExeC2C(inverse_plan, d_freq, d_freq, CUFFT_INVERSE);

cudaDeviceSynchronize();


cudaMemcpy(outputData, d_in, sizeof(short)*nRows*nCols*batch,
                                            cudaMemcpyDeviceToHost);

 .......