Re: symbol catenation and montgomery

[Torbjorn Granlund]

address@hidden writes:

    [Snipped discussion of modexp speeds: GMP is now as fast or faster
     than the OpenSSL modexp library].
  
  I see that the GMP manual "references" section cites Montgomery's
  paper on modmult without trial division, so I guess GMP is definitely
  using Montgomery reduction for modexp.
  
It sure does.  It was written and integrated just before the GMP 3.0
release.

  By the way, I'm a little depressed now that GPG is so slow.  I was
  hoping it could be sped up by simply changing it to use a faster
  modexp library instead of GMP.  But it looks like GMP is basically as
  fast as anything else out there.  So not much can be done about GPG's
  speed without changing the algorithms in it.
  
It is frustrating to me if GPG's speed gives GMP poor reputation.  I
think GPG uses a modified version of GMP 2, which could explain part
of its performance problems.

-- 
Torbjörn