OpenMP* Threading for FFT Computation

Dear Sir/Madam,

As I understood from an example on https://software.intel.com/en-us/mkl-developer-reference-c-examples-of-u..., you calculate with 4 threads 4 different 2D-FFT (50x100 points). My question is - is it possible to calculate NxN FFT where I divide data between 2 processors where each will get NxM FFT and NXK FFT where M+K = N? I worry about the result, is it will be correct by following your example? Also, how do you do it inside and how I can apply it to fftw?

Thanks,
Sem