Discussion related to cado-nfs

[paul zimmermann <Paul.Zimmermann@inria.fr>]

       Greg,

I realize your two questions below were not fully answered:

> 1. Is there a way for me to remove whatever is corrupted and restart
> the computation from here?

the best (and recommended way) is to create a new working directory where you
import the polynomial and already computed relations. Cf "Importing a
polynomial file" and "Importing relations" in scripts/cadofactor/README
(git version). Please tell us if this documentation is incomplete.

> 2. When restarting a computation with a new working directory, importing
> polynomials and sieving relations previously computed, the program
> spends significant time on the "Generate Free Relations" step between
> polynomial selection and lattice sieving.  To speed things up, is there a
> way to import the factor base and the free relations previously computed?

this is not possible currently. However with the following you can use (say)
3 threads to generate the free relations, which should speed things up:

$ cadofactor.py ... tasks.sieve.freerel.threads=3

Paul