Re: elltors() bug

["Prof. J. E. Cremona" <John.Cremona@maths.nottingham.ac.uk>]

Karim,

Darrin Doud implemented a much faster torsion algorithm in GP which
uses (1) reduction moddulo small primes, (2) Mazur's classification.
It is described in a paper which appeared recently in Manuscripta
Math; he also sent me his code, which is in ~cremona/gp/doud.gp in
Bordeaux where you can find it.

Doud's program was so much faster than mine that as soon as I received
it I implemented the same tricks in my own code (before that I had
just been too lazy to do so).  It would be great to have that in GP as
the default -- and I am sure  you can convert it from GP to C+pari
faster than I can!

John