mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Msieve (https://www.mersenneforum.org/forumdisplay.php?f=83)
-   -   reduce to 108119486 relation sets and 0 unique ideals (https://www.mersenneforum.org/showthread.php?t=22171)

Alfred 2017-04-01 10:27

reduce to 108119486 relation sets and 0 unique ideals
 
1 Attachment(s)
After collecting 183292403 unique relations for a c231,
msieve reduced to 108119486 relation sets and 0 unique ideals,
leading to a failure in LA.

[CODE]
Sat Apr 01 11:51:22 2017 keeping 113702574 ideals with weight <= 200, target excess is 52698204
Sat Apr 01 11:51:32 2017 commencing in-memory singleton removal
Sat Apr 01 11:51:38 2017 begin with 183292403 relations and 113702574 unique ideals
Sat Apr 01 11:51:50 2017 reduce to 108119487 relations and 1 ideals in 7 passes
Sat Apr 01 11:51:50 2017 max relations containing the same ideal: 2
[/CODE]

Since I should have enough relations, it is possible to use "filter_maxrels" to enlarge successivly the number of relations used, but maybe there is a better way.

If anybody has an idea, what this (IMO) strange behaviour causes, please let me know.
I'm thankful for any ideas how to proceed, for any advice.

Alfred

VBCurtis 2017-04-02 03:49

What large-prime bound did you use? One can guess at a max-rels number to choose, given LP size.

Alfred 2017-04-02 07:01

1 Attachment(s)
In the meantime, the c231 is factored.

I attach the logfile with some comments,
including the params for the polyfile at the top.
comments are included in the middle of the file, too.

The behaviour seems (to me) to be strange at all.
I suppose this has nothing to do with the so called large dataset bug.

msieve finishes with the options "filter_maxrels=182083000 target_density=140",
but fails without.

What I am doing wrong?
I am thankful for any advice.


All times are UTC. The time now is 20:16.

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