From: jdd@aiki.demon.co.uk (Jim Dixon)
Date: Fri, 26 Aug 94 06:30:48 PDT
To: karn@unix.ka9q.ampr.org
Subject: Re: DSPs
Message-ID: <8051@aiki.demon.co.uk>
MIME-Version: 1.0
Content-Type: text/plain


In message <199408260422.VAA16628@unix.ka9q.ampr.org> Phil Karn writes:
> I'm not really an expert on DSPs, but it does seem like they should be
> good for modular exponentiation given the number of multiplies required.
> But I don't think I'm the first to point that out.

DSPs are actually optimized for the operation required in the FFT: integer
inner products.  A DSP will have hardware for extremely fast summations of
expressions like
	    c[i]*x[i]
where the c[i] are constants, and new x vectors are introduced at a high
rate.

This is somewhat different than the kind of fast multiplication you are
looking for.
--
Jim Dixon