mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Marin's Mersenne-aries (https://www.mersenneforum.org/forumdisplay.php?f=30)
-   -   Unreserving exponents(these exponents haven't been done) (https://www.mersenneforum.org/showthread.php?t=6799)

jasong 2006-12-18 21:44

Unreserving exponents(these exponents haven't been done)
 
I keep getting errors with Prime95, but not much else. I'm unreserving the following exponents, and will stop running Prime95 until I figure out what's going on.(I may consider double checks)

The following exponents still need to be run.
[code]Pfactor=40018607,68.000000,4137260
Pfactor=40018609,68.000000,4137260
Pfactor=40018639,65,0
Pfactor=40018669,65,0
Pfactor=40018907,65,0[/code]

harlee 2006-12-18 22:31

I'm getting the same type of errors with Prime95 v24.14. Could it a bug???

harlee 2006-12-19 03:34

These are my observations with this problem.

By default, Prime95 v24.14 wants to trial factor these exponents to 68 bits before starting P-1 testing. The trial factoring runs fine and when finished, updates the Pfactore statement as shown below and then crashes.
[QUOTE]Pfactor=40018607,68.000000,4137260[/QUOTE]
I tried rewriting the Pfactor statement
[CODE]Pfactor=40018639,65,0[/CODE]
to be
[CODE]factor=40018639,65
Pfactor=40018639,68,0
[/CODE]
but the same thing happens. Did notice that the crash happens if there is a Pfactor statement int the worktodo.ini file for the exponent that was trial factored. I've mixed factor statements and Pfactor statements without any crashes as long as the exponents are different
[CODE]factor=40019467,65
factor=40019489,65
Pfactor=40019227,68,0[/CODE]

jasong 2006-12-20 01:50

I don't know whether it's a bug or not, but I do know that I am now running a Prime95 torture test on both cpus for about 3 days to see if it's my cpus.

Please post if you're positive it's a bug.

Edit: Reread your post and decided to bring my computer back online, although on a different project. Thanks for the heads-up, harlee.

(You wouldn't happen to be black, would you? I have a black cat named Harle, and think it would be really cool if I met someone online with the same name and the same "race"[technically it's fur, rather than skin color] as my cat)

harlee 2006-12-20 03:16

I posted the problem in the Software forum and "Prime95" issued a response: [URL="http://mersenneforum.org/showthread.php?t=6801"]http://mersenneforum.org/showthread.php?t=6801[/URL]

Harlee was the name of the kids pet rabbit. She passed away almost 11 years ago, but I still remember that she had two different eye colors (one brown and one blue). I'm using the different pets names as the account and computer name.

jasong 2006-12-20 20:50

[QUOTE=harlee;94383]Harlee was the name of the kids pet rabbit. She passed away almost 11 years ago, but I still remember that she had two different eye colors (one brown and one blue). I'm using the different pets names as the account and computer name.[/QUOTE]
Even better. :smile:

jasong 2006-12-22 02:31

If I may ask a possibly stupid question:

Wouldn´t it be easier to make it so there´s one line that JUST tells it to trial-factor, and another line that JUST tells it to P-1? Even if you only have a Gig of hard drive space left, there´s still plenty of room for a worktodo file that´s doubled in size.

harlee 2006-12-22 21:59

Tried that and Prime95 still crashes (see the 19 Dec 06, 03:34 AM posting).

If you want to do the two lines then don't add the Pfactor line for a given exponent until after that Factor line is completed and deleted from the worktodo.ini file.


All times are UTC. The time now is 17:08.

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