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

11989Re: [PrimeNumbers] has SB been tested?

Expand Messages
  • Michael Bell
    Apr 1, 2003
      Hi,

      I think we are getting confused here between roundoff errors (which are
      implementation independent assuming the exact same FFT algorithm, limb size,
      etc. being run on a processor with the same precision) and coding errors.
      The latter are nothing much to do with size.

      I believe SoB is using George's code which is fairly well tested, but I
      think they may have been tweaking it a bit. Certainly it is beginning to
      seem odd that there've been no primes, but I'm not sure its that far out of
      line with the expectations.

      I agree with David that running a couple of test cases isn't going to harm
      throughput much and will give a bit more confidence to the results. But
      here probably isn't the best place to suggest an idea for SoB, they have a
      forum at http://www.free-dc.org/forum/forumdisplay.php?s=&forumid=38

      Mike.
    • Show all 8 messages in this topic