Computes FFT's of complex, double precision data where n is an integer power of 2. This appears to be slower than the Radix2 method, but the code is smaller and simpler, and it requires no extra storage.
@author Bruce R. Miller bruce.miller@nist.gov,
@author Derived from GSL (Gnu Scientific Library),
@author GSL's FFT Code by Brian Gough bjg@vvv.lanl.gov