mersenneforum.org GNFS
 Register FAQ Search Today's Posts Mark Forums Read

 2022-08-14, 18:31 #34 charybdis     Apr 2020 23·107 Posts I was able to reproduce the crash on RSA-100 with the latest version of CADO. It doesn't appear on every number so is likely matrix-dependent. Now to find out when it was introduced and report it.
 2022-08-14, 18:44 #35 charybdis     Apr 2020 85610 Posts @storm5510, while we wait for the bug to be fixed, I suggest downloading an older revision of CADO and using that instead: Code: git clone https://gitlab.inria.fr/cado-nfs/cado-nfs.git cd cado-nfs git checkout 7eb15697 make make check (note this will still have the annoying bug where lingen is incredibly slow when anything else is running on your system. but hey, at least lingen works!)
2022-08-14, 23:09   #36
storm5510
Random Account

Aug 2009
Not U. + S.A.

2×5×233 Posts

Quote:
 Originally Posted by charybdis @storm5510, while we wait for the bug to be fixed, I suggest downloading an older revision of CADO and using that instead: Code: git clone https://gitlab.inria.fr/cado-nfs/cado-nfs.git cd cado-nfs git checkout 7eb15697 make make check (note this will still have the annoying bug where lingen is incredibly slow when anything else is running on your system. but hey, at least lingen works!)
I only run one process at a time. I will give it a go and see what happens. Where are the resultant factors stored, if any?

2022-08-14, 23:35   #37
charybdis

Apr 2020

23×107 Posts

Quote:
 Originally Posted by storm5510 I only run one process at a time. I will give it a go and see what happens. Where are the resultant factors stored, if any?
The factors are in the c130.log (or similar) file in the directory you set as workdir. (Hence why it's important to set workdir - if you let CADO put the factors in /tmp then they're liable to be lost!) They're about 45 lines from the end of the logfile, on a line like
Code:
PID2642974 2022-08-15 00:22:17,180 Info:Square Root: Factors: 37975227936943673922808872755445627854565536638199 40094690950920881030683735292761468389214899724061
Searching for the string "Factors" will find the line.

 2022-08-14, 23:50 #38 charybdis     Apr 2020 23·107 Posts In the course of investigating the bug, I found that resuming CADO - i.e. just running it again with the same command line - actually completes the factorization successfully after the crash. But of course it's more convenient not to have to do this, so I've reported the bug to the developers.
2022-08-14, 23:57   #39
chalsall
If I May

"Chris Halsall"
Sep 2002

2·5,333 Posts

Quote:
 Originally Posted by charybdis In the course of investigating the bug, I found that resuming CADO - i.e. just running it again with the same command line - actually completes the factorization successfully after the crash. But of course it's more convenient not to have to do this, so I've reported the bug to the developers.
Just to say... There's something to be said for having new people work with stuff...

"Why did they do it that way?

That's actually a two-part question...

"Why?" involves a whole lot of very complicated engineering decisions that might, or might not, have made sense at the time.

"Did they do it that way?

Yes.

2022-08-14, 23:58   #40
EdH

"Ed Hall"
Dec 2009

131A16 Posts

Quote:
 Originally Posted by charybdis I was able to reproduce the crash on RSA-100 with the latest version of CADO. It doesn't appear on every number so is likely matrix-dependent. Now to find out when it was introduced and report it.
If it helps, I tried:
Code:
$./cado-nfs.py 1522605027922533360535618378132637429718068114961380688657908494580122963258952897654000350692006139 with the following revision: Code: commit 35c7392c14263e7cda088938ce8e81a3ff081915 (HEAD -> master, origin/master, origin/HEAD) Merge: 7014dcfc1 92a499fa0 Author: Emmanuel Thomé <emmanuel.thome@inria.fr> Date: Wed Apr 13 05:40:04 2022 +0000 and, it ran fine: Code: Info:Complete Factorization / Discrete logarithm: Total cpu/elapsed time for entire Complete Factorization 9350.01/587.253 Info:root: Cleaning up computation data in /tmp/cado.gox0ibz8 40094690950920881030683735292761468389214899724061 37975227936943673922808872755445627854565536638199 I can probably try a June revision tomorrow, if interested. 2022-08-15, 00:05 #41 charybdis Apr 2020 23×107 Posts Quote:  Originally Posted by EdH If it helps, I tried... Thanks. I also tried with an April revision - from later in the same day, in fact - and it ran fine. However, it doesn't produce the same matrix, as there have been some revisions to the merge code since then, so it's not a direct comparison. I gave up trying to find out which revision introduced the bug after that. 2022-08-15, 00:21 #42 storm5510 Random Account Aug 2009 Not U. + S.A. 2×5×233 Posts It looks like I because really behind in the time it took to write this. Quote:  Originally Posted by charybdis The factors are in the c130.log (or similar) file in the directory you set as workdir. (Hence why it's important to set workdir - if you let CADO put the factors in /tmp then they're liable to be lost!) They're about 45 lines from the end of the logfile, on a line like Code: PID2642974 2022-08-15 00:22:17,180 Info:Square Root: Factors: 37975227936943673922808872755445627854565536638199 40094690950920881030683735292761468389214899724061 Searching for the string "Factors" will find the line. I had a theory about /tmp so I tested it. Anything not pertinent to the OS gets deleted on a restart, it seems. I created a folder in /cado-nfs titled "temp." I used it as workdir=temp. It worked. Any use of folder references higher up created a flurry of errors. Example: /home/cado-nfs/temp. I just looked in the folder I made and there are a lot of items in it. I am running RSA-100 as a test. At least, I know where to look for results now, and how.  2022-08-15, 02:30 #43 EdH "Ed Hall" Dec 2009 Adirondack Mtns 2·3·5·163 Posts Glad you got that ironed out. There is a little potential confusion with Ubuntu. The folder referenced as Home in the file manager is not /home in the directory structure. It is actually /home/userid in a normal installation. The user does not have permissions to write into the /home directory. Try echo$HOME to see what the true path is for "Home" in the file manager. If you open the file manager and go to Other Locations > Computer > home you will see a folder with your userid as its name. That is the folder that is called Home in your default file manager window. BTW, in my recent test, the /tmp/cado.folder actually was deleted without a reboot.
2022-08-15, 05:57   #44
VBCurtis

"Curtis"
Feb 2005
Riverside, CA

22·372 Posts

Quote:
 Originally Posted by EdH BTW, in my recent test, the /tmp/cado.folder actually was deleted without a reboot.
The last step CADO takes in a factorization job with workdir in /tmp is to delete the work folder. If CADO does not finish for any reason, the folder isn't deleted.

 Similar Threads Thread Thread Starter Forum Replies Last Post pinhodecarlos NFS@Home 0 2014-12-24 19:13 pinhodecarlos NFS@Home 34 2014-04-01 21:27 fivemack Factoring 84 2011-04-26 10:22 10metreh Factoring 48 2009-04-08 01:54 Andi47 Msieve 5 2009-01-26 18:19

All times are UTC. The time now is 06:38.

Fri Sep 30 06:38:02 UTC 2022 up 43 days, 4:06, 0 users, load averages: 1.30, 1.09, 1.00