mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   15k Search (https://www.mersenneforum.org/forumdisplay.php?f=16)
-   -   LLR performance on k and n (https://www.mersenneforum.org/showthread.php?t=5465)

robert44444uk 2006-02-08 23:56

LLR performance on k and n
 
Sorry guys and lassies for dropping into you forum but what is the value of k that LLR can handle that is faster than pfgw at large value of n?

The reason for asking this is that I am dying to get some hot prp software that can take the many Payam numbers k (Riesel and Proth) which are hugely prime in the n>260K<400K range.

Axn1 has software to generate such Payam k, there are billions to choose from, but they are quite large in absolute k terms maybe too large for LLR to be efficient.

Pfgw seems to suffer a huge slow down at n appox 260K for such large k and the effort beyond this point seems not worthwhile in terms of building a solid portfolio of prps which will last beyond next year in the top 5000.

So who knows this stuff?

Regards

Robert Smith

Prime95 2006-02-09 01:43

LLR should be as fast or faster than pfgw for all k/n that it can handle. The underlying assembly code routines can handle k values up to 50 bits or so.


All times are UTC. The time now is 17:36.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.