Confused about cufftPlan

Accelerated Computing CUDA CUDA Programming and Performance
1

I am just trying out CUDA and am trying to use the FFT capabilities. I am confused about the cufftPlan1d function… what is the purpose of the batch variable?

If I have C code like this:

for(i=0; i <nffts; i++) {

pfacc(1, fft_length, data[i]) // 1D C2C FFT in C

}

What would an efficient implementation of this look like in CUDA? What would I set batch to?

Thanks!