mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Software

Reply
 
Thread Tools
Old 2009-01-16, 13:04   #1
axn
 
axn's Avatar
 
Jun 2003

3·7·13·19 Posts
Question P-1 & TF overlap

So I was crunching some numbers for P-1 & TF. Currently, server is handing out P-1 & TF exponents around 50M. These are TFed to 67 bits first, followed by P-1 and then 68 & 69 bit TF.
With decent memory allocated, the P-1 limits (B1,B2) are around (600k,20M). Given these, and my "experiments", I estimate a factor space overlap of around 45% between P-1 and 68/69 bit TF. The implications are that 68/69 bit TF should be around 45% less productive than the theoretical 1/68 and 1/69 success rates (should be resp 1/123 and 1/125).
Does the above estimation jell with real world statistics? I know that primenet doesn't show whether a factor was found by P-1 or TF. Perhaps George can give us a break down of the 68 & 69 bit factors (for the current leading edge) based on the method?
Assuming that these numbers do pan out, is it worth doing the final bit of TF, now that it is preceded by P-1?
axn is online now   Reply With Quote
Old 2009-01-16, 13:30   #2
Mr. P-1
 
Mr. P-1's Avatar
 
Jun 2003

7·167 Posts
Default

It should be fairly straight forward to compile a list of factors of the designated length, and determine what proportion are smooth to typical P-1 bounds. You do not need to know how the factors were in fact obtained.
Mr. P-1 is offline   Reply With Quote
Reply

Thread Tools


All times are UTC. The time now is 10:46.


Fri Dec 3 10:46:05 UTC 2021 up 133 days, 5:15, 0 users, load averages: 1.59, 1.24, 1.20

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.