LARGE numbers

[Paul Rubin]

aleax at mail.comcast.net (Alex Martelli) writes:
> As the author of gmpy, I'd like to point out that the speed difference
> isn't all that large, if all you're doing is ordinary arithmetic -- a
> few times at most (it can be better if you need some of GMP's
> functionality which gmpy exposes, such as primality testing).

For numbers of this size, won't gmpy use FFT-based multiplication?
That's potentially orders of magnitude faster than ordinary n**2
multiplication.