[Cado-nfs-discuss] bug in polyselect2l.c parameters
Paul.Zimmermann at loria.fr
Mon Oct 1 20:39:43 CEST 2012
I've decreased the value of P in the params.c* files of the development
version to avoid the issue you reported.
I've chosen a value of P such that the maximal value of l = p_1*p_2*q does
not exceed m0/P^2.
In some cases I had to decrease P by a factor of 10 or more. For example in
params.c155 we now have P=52000 instead of P=1e6 before.
We could also decrease lq (the number of special-q primes) to avoid decreasing
P too much.
Thank you again for your detailed report,
More information about the Cado-nfs-discuss