Prime-factor FFT algorithm

The prime-factor algorithm (PFA), also called the Good–Thomas algorithm (1958/1963), is a fast Fourier transform (FFT) algorithm that re-expresses the discrete Fourier transform (DFT) of a size N = N1N2 as a two-dimensional N1×N2 DFT, but only for the case where N1 and N2 are relatively prime. These smaller transforms of size N1 and N2 can then be evaluated by applying PFA recursively or by using some other FFT algorithm.

PFA should not be confused with the mixed-radix generalization of the popular Cooley–Tukey algorithm, which also subdivides a DFT of size N = N1N2 into smaller transforms of size N1 and N2. The latter algorithm can use any factors (not necessarily relatively prime), but it has the disadvantage that it also requires extra multiplications by roots of unity called twiddle factors, in addition to the smaller transforms. On the other hand, PFA has the disadvantages that it only works for relatively prime factors (e.g. it is useless for power-of-two sizes) and that it requires more complicated re-indexing of the data based on the additive group isomorphisms. Note, however, that PFA can be combined with mixed-radix Cooley–Tukey, with the former factorizing N into relatively prime components and the latter handling repeated factors.

PFA is also closely related to the nested Winograd FFT algorithm, where the latter performs the decomposed N1 by N2 transform via more sophisticated two-dimensional convolution techniques. Some older papers therefore also call Winograd's algorithm a PFA FFT.

(Although the PFA is distinct from the Cooley–Tukey algorithm, Good's 1958 work on the PFA was cited as inspiration by Cooley and Tukey in their 1965 paper, and there was initially some confusion about whether the two algorithms were different. In fact, it was the only prior FFT work cited by them, as they were not then aware of the earlier research by Gauss and others.)

Algorithm edit

Let   a polynomial and   a principal  th root of unity. We define the DFT of   as the  -tuple  . In other words,

 
For simplicity, we denote the transformation as  .

The PFA relies on a coprime factorization of   and turns   into   for some choices of  's where   is the tensor product.

Mapping Based on CRT edit

For a coprime factorization  , we have the Chinese remainder map   from   to   with   as its inverse where  's are the central orthogonal idempotent elements with  . Choosing   (therefore,  ), we rewrite   as follows:

 
Finally, define   and  , we have
 
Therefore, we have the multi-dimensional DFT,  .

As Algebra Isomorphisms edit

PFA can be stated in a high-level way in terms of algebra isomorphisms. We first recall that for a commutative ring   and a group isomorphism from   to  , we have the following algebra isomorphism

 
where   refers to the tensor product of algebras.

To see how PFA works, we choose   and   be additive groups. We also identify   as   and   as  . Choosing   as the group isomorphism  , we have the algebra isomorphism  , or alternatively,

 
Now observe that   is actually an algebra isomorphism from   to   and each   is an algebra isomorphism from   to  , we have an algebra isomorphism   from   to  . What PFA tells us is that   where   and   are re-indexing without actual arithmetic in  .

Counting the Number of Multi-Dimensional Transformations edit

Notice that the condition for transforming   into   relies on "an" additive group isomorphism   from   to  . Any additive group isomorphism will work. To count the number of ways transforming   into  , we only need to count the number of additive group isomorphisms from   to  , or alternative, the number of additive group automorphisms on  . Since   is cyclic, any automorphism can be written as   where   is a generator of  . By the definition of  ,  's are exactly those coprime to  . Therefore, there are exactly   many such maps where   is the Euler's totient function. The smallest example is   where  , demonstrating the two maps in the literature: the "CRT mapping" and the "Ruritanian mapping".[1]

See also edit

Notes edit

References edit

  • Good, I. J. (1958). "The interaction algorithm and practical Fourier analysis". Journal of the Royal Statistical Society, Series B. 20 (2): 361–372. JSTOR 2983896. Addendum, ibid. 22 (2), 373-375 (1960) JSTOR 2984108.
  • Thomas, L. H. (1963). "Using a computer to solve problems in physics". Applications of Digital Computers. Boston: Ginn.
  • Duhamel, P.; Vetterli, M. (1990). "Fast Fourier transforms: a tutorial review and a state of the art". Signal Processing. 19 (4): 259–299. doi:10.1016/0165-1684(90)90158-U.
  • Chan, S. C.; Ho, K. L. (1991). "On indexing the prime-factor fast Fourier transform algorithm". IEEE Transactions on Circuits and Systems. 38 (8): 951–953. doi:10.1109/31.85638.
  • Good, I. J. (1971). "The relationship between two fast Fourier transforms". IEEE Transactions on Computers. 100 (3): 310–317. doi:10.1109/T-C.1971.223236. S2CID 585818.