Discussion related to cado-nfs

[Alexander Kruppa <alexander.kruppa@inria.fr>]

----- Original Message -----
> From: "Paul Leyland" <paul.leyland@gmail.com>
> To: Cado-nfs-discuss@lists.gforge.inria.fr
> Sent: Monday, 25 November, 2013 4:27:39 PM
> Subject: [Cado-nfs-discuss] Curious non-monotonicity
> 
> Running "grep tasks.sieve.rels_wanted pa*" in params_py yields
> curious
> output. Spotted this because I've started a c130 immediately after a
> successful c131.  The relevant lines from that grep are:
> 
> params.c130:tasks.sieve.rels_wanted = 29000000
> params.c131:tasks.sieve.rels_wanted = 20000000
> 
> 
> FWIW, the c131 needed a little over 24M relations.
> 
> I'm guessing that, for instance,
> 
> params.c132:tasks.sieve.rels_wanted = 1
> 
> indicates that no experimental data is yet available.  As time goes
> by I
> may be able to provide such data.
> 
> 
> Paul

Yes, a few of the parameters are still, let's say, at an educated-guess level 
of accuracy. Now that the Python script is mostly working, my next task is to 
automate parameter selection. Hopefully the parameters will then become more 
systematic and monotonous across number sizes.

Alex