--- Richard Traynham <therichardt@...
> re: Why did Phil C. cross the road, etc.
> Primatosis -- that most dreaded of diseases, that
> is caused by the in-ability of the mind to give
> up thinking about primes.
Believe me, if while sitting in a 90C sauna, as relaxed and limp as a
human can be, you start weighing up various prime-hunting algorithms,
then you've really got a bad case of it.
> Yes, there is no cure for primatosis,
> but there IS hope.
I know this is on-topic, and thus maybe in the wrong thread, but Paul
& Jack - the improvements from a big-steps/baby-steps are probably
less than those that are available from doing a vast number of k
simultaniously (which actaully makes it a fixed-n sieve CC-esque with
vast number of powers (the n's [*])). If you want to do the sieve for
many k, then it's better to do them all in parallel rather than using
the potentially faster for a single k (who knows) discrete logarithm.
However, you need loads of RAM, and need to look at the
gather-read/scatter-write cache issues to actually be practical.
However, half-gig machines aren't that rare nowadays, and with it
appears that 'potential candidate removal list' technique is one
entering the sieving field since my recent GFNSieve communications,
and maybe that can be of use.
Do You Yahoo!?
Yahoo! Health - your guide to health and wellness