Loading ...
Sorry, an error occurred while loading the content.

Hello!

Expand Messages
  • ggnfsmonico
    Greetings all! So I finally realized that it was time to do something - a mailing list or something similar. At some point, GGNFS may wind up on sourceforge,
    Message 1 of 1 , Sep 17, 2004
    • 0 Attachment
      Greetings all!

      So I finally realized that it was time to do something - a mailing
      list or something similar. At some point, GGNFS may wind up on
      sourceforge, in which case we'll get a little more sophisticated. But
      for now, this might be handy.
      In the meantime, I rather hope that we can put this group to some
      interesting use. Perhaps we can share statistics and parameters from
      various factorizations, and tips-and-tricks and such. We can also use
      it for bug reports and such. Or whatever strikes our fancy.
      The one downside of these groups is that they often get heavily
      spammed. To try and prevent that sort of garbage, I have used the
      setting that requires a user's first post to be moderator-approved. In
      my experience, though, spammers can find ways around that. So, I have
      also made several of you moderators. If you happen to see any spam or
      non-GGNFS stuff coming though here, feel free to immediately ban that
      member.
      And, since I'm posting anyway, I should mention that the latest
      version of GGNFS, 0.54.4, seems to be free of all known major bugs
      (with the exception that it still doesn't work under MingW32; but I'm
      starting to wonder if that might be an issue with MingW and not
      GGNFS). I just got some updates for the VC project files which I will
      incorporate and post sometime today or this weekend.
      Also, just FYI, there are two things on my short-term plan:
      (1) Speed up Lanczos, and make the MPI-version.
      (2) Find out why I can't get Franke's siever to spit out relations
      with more than 3 large prime factors. GGNFS should get a noticable
      speed boost on larger numbers if I can figure this out. I have noticed
      that it will crash if MPBR or MPBA are set too large. This suggests
      that as soon as these are large enough to produce numbers with 4 large
      prime factors, gnfs-lasieve goes and tries to factor the extra stuff
      and promptly crashes. But it seems to crash somewhere in the assembly
      code, badly corrupting the stack in the process, and so I haven't had
      much luck tracking it down (admittedly, I haven't tried very hard yet
      - I was merely curious.

      Just thought I'd drop you an update.

      Cheers,
      Chris
    Your message has been successfully submitted and would be delivered to recipients shortly.