mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > NFS@Home

Reply
 
Thread Tools
Old 2021-06-30, 13:12   #111
Xyzzy
 
Xyzzy's Avatar
 
"Mike"
Aug 2002

100000010101102 Posts
Default

We restarted it using the default target density. We didn't keep the log.

We tend to start jobs as soon as there are enough relations to build a matrix, usually before all of the pending relations have been returned. Could this maybe be the problem?

Xyzzy is offline   Reply With Quote
Old 2021-06-30, 13:40   #112
jyb
 
jyb's Avatar
 
Aug 2005
Seattle, WA

24×3×37 Posts
Default

Quote:
Originally Posted by swellman View Post
Anyone willing to attempt this job in order to duplicate this behavior?
I'll give it a shot.
jyb is offline   Reply With Quote
Old 2021-06-30, 21:09   #113
jyb
 
jyb's Avatar
 
Aug 2005
Seattle, WA

24·3·37 Posts
Default

Quote:
Originally Posted by jyb View Post
I'll give it a shot.
I've built a matrix, but it's quite large due to the relations duplicate rate. Can we move 4p3_1341M back to sieving and go for a smaller matrix? We'll see if we can bypass the problem that way.
jyb is offline   Reply With Quote
Old 2021-06-30, 21:24   #114
swellman
 
swellman's Avatar
 
Jun 2012

61338 Posts
Default

Quote:
Originally Posted by jyb View Post
I've built a matrix, but it's quite large due to the relations duplicate rate. Can we move 4p3_1341M back to sieving and go for a smaller matrix? We'll see if we can bypass the problem that way.
We can try it. Moved and Q bumped by 10M.
swellman is offline   Reply With Quote
Old 2021-06-30, 22:32   #115
swellman
 
swellman's Avatar
 
Jun 2012

316310 Posts
Default

Quote:
Originally Posted by Xyzzy View Post
We restarted it using the default target density. We didn't keep the log.

We tend to start jobs as soon as there are enough relations to build a matrix, usually before all of the pending relations have been returned. Could this maybe be the problem?

In general I wait until the Estimated Pending Relations < 1M before downloading the dataset, unless the job is clearly oversieved. But that’s what I do, not necessarily what is optimal. I suspect this job was undersieved even after all the relations had trickled in. Adding more relations may be the path to success.

But I’ve never seen a corrupt state error message. When the requested TD is too large or not enough relations are present msieve will demand more rels, but never a message with the word corrupt.
swellman is offline   Reply With Quote
Old 2021-07-01, 12:11   #116
Xyzzy
 
Xyzzy's Avatar
 
"Mike"
Aug 2002

2·4,139 Posts
Default

We stop and restart the job 2-3 times a day.

Perhaps there is a bug related to that?

Xyzzy is offline   Reply With Quote
Old 2021-07-01, 16:35   #117
jyb
 
jyb's Avatar
 
Aug 2005
Seattle, WA

24×3×37 Posts
Default

Taking 7p5_875L.
jyb is offline   Reply With Quote
Old 2021-07-06, 05:36   #118
jyb
 
jyb's Avatar
 
Aug 2005
Seattle, WA

6F016 Posts
Default

7p5_875L is in LA.
jyb is offline   Reply With Quote
Old 2021-07-06, 21:42   #119
Xyzzy
 
Xyzzy's Avatar
 
"Mike"
Aug 2002

2·4,139 Posts
Default

We were able to finish linear algebra phase on 4p3_1341M without the extra relations.

We think that somehow repeatedly shutting down and restarting the process has a chance of corrupting things.

We are using the BTRFS filesystem on a top-quality SSD with an online UPS so we don't think it is a hardware error.

Anyways, we ran into this problem trying to run the square root phase:
Code:
Tue Jul  6 16:23:47 2021  Msieve v. 1.54 (SVN 1041)
Tue Jul  6 16:23:47 2021  random seeds: 3c46836a 51a73f7c
Tue Jul  6 16:23:47 2021  factoring 798714408583534111218890232290224688792108684535127942573951568105348821635488501822138981923034879063165326737648089547891701098011684158965932673371227082652984500807199785769778899213 (186 digits)
Tue Jul  6 16:23:47 2021  no P-1/P+1/ECM available, skipping
Tue Jul  6 16:23:47 2021  commencing number field sieve (186-digit input)
Tue Jul  6 16:23:47 2021  R0: -671434345011425390364129256815738507
Tue Jul  6 16:23:47 2021  R1: 302918003288600224887967
Tue Jul  6 16:23:47 2021  A0: 855299734549376002681269040716011503195190
Tue Jul  6 16:23:47 2021  A1: -3578725611871361502992472667183513387
Tue Jul  6 16:23:47 2021  A2: -1470548342195016584382531299800
Tue Jul  6 16:23:47 2021  A3: 55439831693748260477406
Tue Jul  6 16:23:47 2021  A4: 5309014851641644
Tue Jul  6 16:23:47 2021  A5: -76090560
Tue Jul  6 16:23:47 2021  skew 12538082.23, size 2.982e-18, alpha -7.736, combined = 3.404e-14 rroots = 5
Tue Jul  6 16:23:47 2021  
Tue Jul  6 16:23:47 2021  commencing square root phase
Tue Jul  6 16:23:47 2021  cannot handle negative leading algebraic polynomial coefficient
Tue Jul  6 16:23:47 2021  sqrtTime: 0
Tue Jul  6 16:23:47 2021  elapsed time 00:00:00
Xyzzy is offline   Reply With Quote
Old 2021-07-06, 23:57   #120
jyb
 
jyb's Avatar
 
Aug 2005
Seattle, WA

6F016 Posts
Default

Quote:
Originally Posted by Xyzzy View Post
We were able to finish linear algebra phase on 4p3_1341M without the extra relations.

We think that somehow repeatedly shutting down and restarting the process has a chance of corrupting things.

We are using the BTRFS filesystem on a top-quality SSD with an online UPS so we don't think it is a hardware error.

Anyways, we ran into this problem trying to run the square root phase:
Code:
Tue Jul  6 16:23:47 2021  Msieve v. 1.54 (SVN 1041)
Tue Jul  6 16:23:47 2021  random seeds: 3c46836a 51a73f7c
Tue Jul  6 16:23:47 2021  factoring 798714408583534111218890232290224688792108684535127942573951568105348821635488501822138981923034879063165326737648089547891701098011684158965932673371227082652984500807199785769778899213 (186 digits)
Tue Jul  6 16:23:47 2021  no P-1/P+1/ECM available, skipping
Tue Jul  6 16:23:47 2021  commencing number field sieve (186-digit input)
Tue Jul  6 16:23:47 2021  R0: -671434345011425390364129256815738507
Tue Jul  6 16:23:47 2021  R1: 302918003288600224887967
Tue Jul  6 16:23:47 2021  A0: 855299734549376002681269040716011503195190
Tue Jul  6 16:23:47 2021  A1: -3578725611871361502992472667183513387
Tue Jul  6 16:23:47 2021  A2: -1470548342195016584382531299800
Tue Jul  6 16:23:47 2021  A3: 55439831693748260477406
Tue Jul  6 16:23:47 2021  A4: 5309014851641644
Tue Jul  6 16:23:47 2021  A5: -76090560
Tue Jul  6 16:23:47 2021  skew 12538082.23, size 2.982e-18, alpha -7.736, combined = 3.404e-14 rroots = 5
Tue Jul  6 16:23:47 2021  
Tue Jul  6 16:23:47 2021  commencing square root phase
Tue Jul  6 16:23:47 2021  cannot handle negative leading algebraic polynomial coefficient
Tue Jul  6 16:23:47 2021  sqrtTime: 0
Tue Jul  6 16:23:47 2021  elapsed time 00:00:00
Try negating all of the algebraic polynomial's coefficients in the .fb file.

Last fiddled with by jyb on 2021-07-06 at 23:58
jyb is offline   Reply With Quote
Old 2021-07-07, 00:45   #121
Xyzzy
 
Xyzzy's Avatar
 
"Mike"
Aug 2002

2×4,139 Posts
Default

Every possible combination fails with this message:
Code:
error: NFS input does not match polynomials
check that input doesn't have small factors
Xyzzy is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
2021 14d post processing reservations and results swellman NFS@Home 226 2021-09-20 14:03
2020 14e post processing reservations and results swellman NFS@Home 544 2020-12-31 15:14
2020 16e post processing reservations and results swellman NFS@Home 20 2020-12-31 05:22
2020 15e post processing reservations and results swellman NFS@Home 112 2020-12-29 22:58
2019 15e post processing reservations and results swellman NFS@Home 220 2019-12-28 16:40

All times are UTC. The time now is 08:25.


Tue Sep 21 08:25:41 UTC 2021 up 60 days, 2:54, 0 users, load averages: 1.38, 1.32, 1.39

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.