Discussion related to cado-nfs

[Emmanuel Thomé <Emmanuel.Thome@inria.fr>]

Hi,

Thanks for your report.

There are indeed recent changes in cado-nfs that could be related to the
difficulties that you encounter.

However, I can't find the error stack trace in your log file:

> When they failed, all tests did with the same "corrupted table" error, see
> attached log (c105.log.zip).

I do see that the dup2 program crashed, but no such thing as an
informative stack trace.

In order to best diagnose your problem, the files that are most relevant
are the input and output files of the following command.

PID28397 2020-08-21 14:54:14,619 Error:Filtering - Duplicate Removal, removal 
pass: Command line was: /home/redgolpe/cado-nfs/build/c2018/filter/dup2 -poly 
nfsdata/c105.poly -nrels 2300838 -renumber nfsdata/c105.renumber.gz -t 12 
nfsdata/c105.dup1//0/dup1.0.0000.gz > nfsdata/c105.dup2.slice0.stdout.1 2> 
nfsdata/c105.dup2.slice0.stderr.1

Could you please send us these files ? It would be best if you could post
them in the gitlab issue there:

https://gitlab.inria.fr/cado-nfs/cado-nfs/-/issues/30003

but you might be blocked (not sure) by a counter-intuitive thing that
requires you to request sponsorship for an account (which we're happy to
do, but OTOH I fully inderstand if you don't want to go through this maze
of weird things). Data files by e-mail or from an URL is also fine, if
you don't want to bother.

Best,

E.



On Fri, Aug 21, 2020 at 03:41:07PM +0200, Bernardo Boncompagni wrote:
> Hi all,
> 
> within the scope of the aliquot sequence project at mersenneforum.org I
> have been running several factorizations in the past few weeks with
> CADO-NFS and everything went smoothly until I encountered a bug preventing
> a factorization to complete. I first encountered it on a C114 and then on
> the C107
> 54022122323205311359700529131254845253584832080092810873601245077747279904751944559089001546838958178759103
> = 6892192422790360694669529583587636497846216763819494386433 *
> 7838162228984026472885414974266767581580262988991
> 
> So i ran some tests on two machines with the following specifications:
> - "EC2": an Amazon EC2 instance with 3.3 GHz AMD EPYC processors, 4 cores,
> 8 threads, 16 GiB RAM, Ubuntu 18.04 fully updated
> - "WSL": a Windows 10 PC with Intel Core i7-7800X processors, 6 cores, 12
> threads, 32 GiB RAM, WSL Ubuntu 18.04 fully updated
> 
> All versions of CADO-NFS were installed with git clone
> https://gitlab.inria.fr/cado-nfs/cado-nfs.git.
> All tests were run with the default command line cado-nfs.py <N>
> workdir=<workdir> unless specified.
> When they failed, all tests did with the same "corrupted table" error, see
> attached log (c105.log.zip).
> When failed, the error seems reproducible with the same command line on the
> same machine.
> 
> Test results follow.
> 
> *Failed* on EC2, CADO-NFS with timestamp July 22 (two runs).
> *Failed* on EC2, CADO-NFS with timestamp August 18.
> *Failed* on EC2, CADO-NFS with timestamp August 18, parameters -t 6.
> *Passed* on WSL, CADO-NFS with timestamp March 17.
> *Failed* on WSL, CADO-NFS with timestamp August 18.
> *Failed* on WSL, CADO-NFS with timestamp August 18, parameters -t 8.
> 
> So it looks like whatever it is, it does not depend on the processor type,
> on the number of cores and on the actual cores used, and was introduced
> some time between March 17 and July 22.
> 
> I also attach the requested make cmake output (cmake.log.zip) and gcc -v
> (gcc.log.zip) for the WSL machine. Let me know if you also need those of
> EC2.
> 
> Thank you for your time.




> _______________________________________________
> Cado-nfs-discuss mailing list
> Cado-nfs-discuss@lists.gforge.inria.fr
> https://lists.gforge.inria.fr/mailman/listinfo/cado-nfs-discuss