On Wed, Jun 11, 2008 at 02:43:23PM -0500, Mike McCarty wrote: > It can't be fast if it can't be run at all, which, I think, > is the point. The subject matter here is GIMPS. Since a > sieve just isn't feasible in the range GIMPS is exploring, > a sieve is actually pretty much off topic. THAT is the > reason for lack of interest. It isn't that what you have > might not be interesting elsewhere, it just isn't a matter > of interest to GIMPS afficionados.[*]
I tend to disagree. There's been several posts here over the years that are interesting but does not help GIMPS much. The reason he's not getting much interest is, at least if you asked me, that nobody believes he really has found something noteworthy. Part of this reason is that he's done extremely flawed "comparisons" of his own algorithm with the main competitors (mainly, by implementing both in a programming language with very atypical performance characteristics); part of it is effectively claiming "it will be a lot faster if you only read my paper and implement the things I didn't bother to". There are other reasons as well. /* Steinar */ -- Homepage: http://www.sesse.net/ _______________________________________________ Prime mailing list [email protected] http://hogranch.com/mailman/listinfo/prime
