Discussion related to cado-nfs

[Emmanuel Thomé <emmanuel.thome@gmail.com>]

Hi,

Thank you for your interest in cado-nfs.

In order to diagnose your problem better, we would like to gather more
information regarding your attempts.

 - Could you please provide the full transcripts ? At least the
bwc.log files would certainly be of great help.
 - As you have used mvapich2, I suppose that the environment is some
kind of linux, but can you please indicate which distribution you are
running (e.g. output of lsb_release -a, cat /etc/*_version, cat
/etc/*_release, or so).
 - When you say "It always fails", can you be a bit more specific ?
How many times have you seen bwc fail ? On which matrices ? Does that
correspond to several attempts from scratch, or to resumed attempts on
one given run which had been interrupted ?
 - I have vague suspicion that your bwc run might have been
interrupted in the middle, e.g. because of a disk full condition. This
typically prevents successful recovery, which is admittedly a
nuisance. If you could provide the "ls -l" output of your .bwc
directory, that would help.


Please note that cado-nfs is routinely used for 512-bit
factorizations, and many 600+ have been done, up to 704-bit so far.
This entails *NO* tinkering with the source files ! I'm in fact
extremely surprised that anything happened correctly if you really
changed the typedef for p_r_values_t and index_t from 32 to 64 bits.
This is simply useless, and counter-productive. Now if you encountered
a failure in bwc, that's probably unrelated anyway.

Best regards,

E.


On Mon, Aug 4, 2014 at 4:15 AM, Bin Hu <hubinnibuh1@gmail.com> wrote:
> Dear developers
> Cadonfs is terrific for my learning IF. My Cado works very well for RSA512
> and smaller, but for RSA600+, it always stucks in the gather phase of BWC,
> saying“no solution found, most probably a bug”. The compiler I use is GCC
> 4.8.1 with mvapich2. Is there some additional configuration that I need to
> do before compiling cado for such large integer, other than simply changing
> the first two 32 to 64 in utils/typedef.hand make (thats what I did)? I also
> googled this bug and found this
> thread:http://lists.gforge.inria.fr/pipermail/cado-nfs-ci-reports/2013-May/000524.html,
> is it due to the same kind of problem?
>
> Regards
>
> Bin
>
> _______________________________________________
> Cado-nfs-discuss mailing list
> Cado-nfs-discuss@lists.gforge.inria.fr
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/cado-nfs-discuss
>