Subject: Discussion related to cado-nfs
List archive
- From: Paul Leyland <paul.leyland@gmail.com>
- To: cado-nfs-discuss@lists.gforge.inria.fr
- Subject: [Cado-nfs-discuss] SNFS suggestion.
- Date: Tue, 09 Apr 2013 11:51:46 +0100
- List-archive: <http://lists.gforge.inria.fr/pipermail/cado-nfs-discuss>
- List-id: A discussion list for Cado-NFS <cado-nfs-discuss.lists.gforge.inria.fr>
Today is Tom Lehrer's 85th birthday. He states that Lobachevsky taught
him the secret of success in mathematics: Plagiarize.
In honour of both mathematicians, may I suggest that CADO-NFS stea...
adapts an idea from msieve and provide semi-automatic selection of
sieving parameters for SNFS factorizations? Only be sure always to call
it please 'research'.
The GNFS parameters are chosen pretty well automatically but choosing a
params/params.* for, say, a SNFS-240 is not obvious. Rhetorical
question: do I use the two-thirds rule and use params.c180 or something
else? If params/ contained params.s200, params.s210, ... a simple
computation of SNFS difficulty and, where necessary, interpolation
between data stored in those files would make life rather easier.
We could get an initial set of parameter files by "researching" how
factMsieve.pl or factmsieve.py does it. (Yes, I know I could do that on
a case by case basis but it should not be necessary to do it more than
once.)
Paul
- [Cado-nfs-discuss] SNFS suggestion., Paul Leyland, 04/09/2013
- Re: [Cado-nfs-discuss] SNFS suggestion., Zimmermann Paul, 04/09/2013
Archive powered by MHonArc 2.6.19+.