On Sunday 19 June 2005 21:27, Greg Edwards wrote:
> it wasn't quite clear from my reading of the
> Factoring section of the GIMPS site: does each 2^N stage take twice as long
> as the previous, or is it a non-linear law ?

Basically it's a linear law but there are highly significant "steps" around 
2^64 caused by implementation of multiprecision arithmetic.

> Does a current stage 2^N
> include the previous 2^(N-1), or does the code really look at factors
> "between" 2^(N-1) and 2^(N), ie. doesn't go over the same ground twice.

It would be silly to "go over the same ground twice".... the point is that as 
it's about as likely to find a factor between 2^(N-1) and 2^N as it is 
between 2^N and 2^(N+1), but it takes only half as long to run, it makes 
sense to exhaust the smaller range before starting the larger.

I have a question too. Regarding v24. I have a dual PIII system running LL 
tests on one CPU and factoring on the other. LL efficiency demands sticking 
to v23 but this could result in doing excess factoring according to the new 
cutoffs... advice please.

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

Reply via email to