mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Prime Sierpinski Project (https://www.mersenneforum.org/forumdisplay.php?f=48)
-   -   Quality of results (https://www.mersenneforum.org/showthread.php?t=2873)

ltd 2004-08-09 15:05

Quality of results
 
Hi,

i made some checks in the database to get some information about the quality of the received results.

We have made 89227 PRP tests so far
Double checks done 17213.

Form these double checks 13 have been different to the first time checks.
For 12 of them the reason is known. These are tests which where calculated with an llr.exe version which contained a bug.

So at the end we have 1 out of 17213 test with a real wrong residue.

Conclusion: At the moment there is no need to do more double checks.
But if somebody still wants to do some due to the fact that he has a very slow PC no problem there are enough double checks left.

Lars

Mystwalker 2004-08-09 15:11

Well, one advantage of doing doublechecks is to raise the lower bounds of sieving, slightly improving performance.
So, it's definitely not bad to do double checks. :)

Citrix 2004-08-10 22:09

Mystwalker, I agree with Lars. Changing the bounds from 100K to 200K will make very little difference and so on. It might take a second less for a billion, such small time changes don't make a significant difference.

But, if we find a prime, we might gain a few minutes for each billion, hence it is best to look for new primes than to double check for right now. Later in the project when we have more computers and can expend some computers, we can use some for the double check effort.

But still if some one has slow computers they can definately participate in the double check effort.

Citrix
:cool: :cool: :cool:


All times are UTC. The time now is 10:45.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.