mersenneforum.org  

Go Back   mersenneforum.org > Prime Search Projects > Sierpinski/Riesel Base 5

Reply
 
Thread Tools
Old 2009-07-17, 02:52   #1
geoff
 
geoff's Avatar
 
Mar 2003
New Zealand

13·89 Posts
Default Prime95 25.11

Times for PRP test of 59912*5^325037+1 on my 2.66GHz Core 2 Duo (1 core):

mprime 25.9: ~51 minutes
mprime 25.11: ~18 minutes

Almost 3 times faster!
geoff is offline   Reply With Quote
Old 2009-07-17, 04:06   #2
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

3·2,083 Posts
Default

These very radical speed increases are the result of v25.11 of the gwnum math library (used by Prime95, PRP, LLR, and PFGW) being optimized for non-power-of-2 numbers. (Previously it was only optimized for base 2.) At this time, both Prime95/mprime and PFGW have been updated to take advantage of this, though at the moment PFGW still suffers from a minor issue where the first character of its residuals is incorrect (thus it wouldn't work for SR5, since it would be flagged as incorrect during doublechecking).

Thus, for now, Prime95 would probably be the best way to go for SR5 members wishing to utilize one of the new, faster programs for manual testing. PFGW should be fixed soon to remedy the residue bug, and I expect LLR will be updated as well to utilize the new math library.
mdettweiler is offline   Reply With Quote
Reply

Thread Tools


All times are UTC. The time now is 09:20.


Sat Jul 17 09:20:13 UTC 2021 up 50 days, 7:07, 1 user, load averages: 1.41, 1.76, 1.68

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.