On 12 Feb 2002, at 12:41, Aaron Blosser wrote: > > Have the predictions on the work eliminated by P-1 factoring been pretty > much confirmed by the # of large factors found? In other words, is the > extra processing time paying off? > > I'd hazard a guess that the time saving is indeed appreciable, but I > wonder if anyone has done some cold hard stats on it.
Small sample, just my current account report: 15 factors found - 11 trial factoring, 4 P-1 on LL assignments, 0 P-1 on DC assignments 49 LL assignments & 42 DC assignments completed - almost all of the LL assignments and about half of the DC assignments most of these have included the P-1 factoring phase. So it looks as though running P-1 on DC assignments has been wasteful, but on the other hand one shouldn't expect to get many "successes" with a predicted factoring rate of around 2% and a sample size of only about 20. Conversely, with a predicted factoring rate of around 5%, I've found almost twice as many factors as expected when running P-1 prior to LL assignments. On balance, I'm beating the odds - the factors I've found by running P-1 prior to LL & DC assignments have saved about 3.5 times the amount of time I've put into running P-1. Regards Brian Beesley _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
