On Wednesday 16 March 2005 21:41, Greg Edwards wrote:
> Hi,
> I'm just 0.3% of the way through my first Mersenne check.
> I scanned a good deal of the "Maths" pages but it doesn't explicitly state
> how far the L-L check progresses before success/failure.
> However it does look like it has to go all the way, ie. up to S(P-2), to
> see whether you get that final 0 mod result.
> Can someone confirm ?

Yes. If you find a better way _please_ let us know, we'd all appreciate an 
advance which would make it possible to abandon at least some exponents 
before all p-2 iterations are completed. At present the only known "shortcut" 
is to find a factor using either trial factoring or P-1 before even starting 
the LL test. That's why trial factoring and P-1 are important - the work 
"wasted" on these tasks eliminates more candidates on average than running LL 
tests would do.

It's also valuable to set the CPU memory as high as practical, since this 
makes P-1 more efficient. The full memory allocation will be used only during 
P-1 stage 2; at all other times the actual memory used will be approximately 
that used during LL testing, irrespective of the value set.

Regards
Brian Beesley

_______________________________________________
Prime mailing list
[email protected]
http://hogranch.com/mailman/listinfo/prime

Reply via email to