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

Issue in lasieve?

Expand Messages
  • Mark Rodenkirch
    I have been trying to factor a 133 digit composite for a couple of weeks and ran into a problem in ggnfs-lasieve. It appears to be stuck in a loop in redu2.c
    Message 1 of 5 , Jul 1 3:32 PM
    • 0 Attachment
      I have been trying to factor a 133 digit composite for a couple of
      weeks and ran into a problem in ggnfs-lasieve. It appears to be stuck
      in a loop in redu2.c I say this because it is stuck on q=13871119 for
      many minutes without finding new relations. It isn't a tight loop and
      I don't know the code that well, so it is a mystery to me. Should I
      let it continue at this q value for a few hours hoping that it will
      get past it or does anyone else here think there might be a problem?
      This happens on both PowerPC and a Core 2 Duo.

      --Mark
    • gchil0
      I ve seen this a couple of times. Very annoying! I saved the files somewhere to track it down later, but never did. I not sure I still have those files
      Message 2 of 5 , Jul 1 5:38 PM
      • 0 Attachment
        I've seen this a couple of times. Very annoying! I saved the files
        somewhere to track it down later, but never did. I not sure I still
        have those files now...

        Greg

        --- In ggnfs@yahoogroups.com, Mark Rodenkirch <mgrogue@...> wrote:
        >
        > I have been trying to factor a 133 digit composite for a couple of
        > weeks and ran into a problem in ggnfs-lasieve. It appears to be stuck
        > in a loop in redu2.c I say this because it is stuck on q=13871119 for
        > many minutes without finding new relations. It isn't a tight loop and
        > I don't know the code that well, so it is a mystery to me. Should I
        > let it continue at this q value for a few hours hoping that it will
        > get past it or does anyone else here think there might be a problem?
        > This happens on both PowerPC and a Core 2 Duo.
        >
        > --Mark
        >
      • Mark Rodenkirch
        I have zipped up the directory. It is about 600 MB if anyone has the skill to track this down.
        Message 3 of 5 , Jul 1 5:59 PM
        • 0 Attachment
          I have zipped up the directory.  It is about 600 MB if anyone has the skill to track this down.

          On Jul 1, 2008, at 7:38 PM, gchil0 wrote:

          I've seen this a couple of times. Very annoying! I saved the files
          somewhere to track it down later, but never did. I not sure I still
          have those files now...

          Greg

          --- In ggnfs@yahoogroups. com, Mark Rodenkirch <mgrogue@... > wrote:
          >
          > I have been trying to factor a 133 digit composite for a couple of 
          > weeks and ran into a problem in ggnfs-lasieve. It appears to be stuck 
          > in a loop in redu2.c I say this because it is stuck on q=13871119 for 
          > many minutes without finding new relations. It isn't a tight loop and 
          > I don't know the code that well, so it is a mystery to me. Should I 
          > let it continue at this q value for a few hours hoping that it will 
          > get past it or does anyone else here think there might be a problem? 
          > This happens on both PowerPC and a Core 2 Duo.
          > 
          > --Mark
          >


        • Sander Hoogendoorn
          ... I regularly have the siever / perl script get stuck on some special q s. Just increase the special q in the .job file with a couple of thousand and
          Message 4 of 5 , Jul 2 11:23 AM
          • 0 Attachment
            > I have been trying to factor a 133 digit composite for a couple of
            > weeks and ran into a problem in ggnfs-lasieve. It appears to be stuck
            > in a loop in redu2.c I say this because it is stuck on q=13871119 for
            > many minutes without finding new relations. It isn't a tight loop and
            > I don't know the code that well, so it is a mystery to me. Should I
            > let it continue at this q value for a few hours hoping that it will
            > get past it or does anyone else here think there might be a problem?

            I regularly have the siever / perl script get stuck on some special q's.

            Just increase the special q in the .job file with a couple of thousand
            and continue from there. Don't forget to save the relations of the last
            run.
          • Mark Rodenkirch
            That got me past this roadblock. Hopefully someone will find and solve this problem.
            Message 5 of 5 , Jul 3 4:16 AM
            • 0 Attachment
              That got me past this roadblock.  Hopefully someone will find and solve this problem.

              On Jul 2, 2008, at 1:23 PM, Sander Hoogendoorn wrote:

              > I have been trying to factor a 133 digit composite for a couple of
              > weeks and ran into a problem in ggnfs-lasieve. It appears to be stuck
              > in a loop in redu2.c I say this because it is stuck on q=13871119 for
              > many minutes without finding new relations. It isn't a tight loop and
              > I don't know the code that well, so it is a mystery to me. Should I
              > let it continue at this q value for a few hours hoping that it will
              > get past it or does anyone else here think there might be a problem?

              I regularly have the siever / perl script get stuck on some special q's.

              Just increase the special q in the .job file with a couple of thousand 
              and continue from there. Don't forget to save the relations of the last 
              run.


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