DFT with gp

Dear all,

I should compute the DFT and the inverse DFT of a large vector of real numbers computed

using gp; I wrote a trivial implementation of it which is fine when the

main parameter q (an odd prime) is tiny but it becomes very slow as soon as q is a bit larger.

I read here

about the existence of some FFT-related functions in libpari

but they are not publicly available.

So I am wondering if these functions can be installed in gp ?

(I mean using some install() directive).

If so, some of you can send me an example of using them?

Thanks in advance. Best regards,

Alessandro