Re: Hilbert class polynomial roots mod q

[Andreas Enge <andreas.enge@inria.fr>]

Am Sat, Jul 09, 2022 at 08:30:09PM +0200 schrieb Bill Allombert:
> > You may want to try my software CM:
> >    https://www.multiprecision.org/cm/
> > I just checked with one of my favourite discriminants,
> > -23512271 or 4 times this, both with a class number of 10000,
> > with a Weber invariant. CM is a little bit faster on my laptop than GP:
> > ? f=polclass (-6961631, 1);
> > cpu time = 1min, 16,564 ms, real time = 1min, 16,620 ms.
> For the record: this is not the case, the time on my laptop is
>   cpu time 3,665 ms, real time 3,736 ms.

Yes, sorry, I meant f=polclass (-23512271, 1), of course.
A silly copy-and-paste error. Otherwise we would compare apples and pears.

> Indepently, in commit 15c47e34a2e1a1865d0 I have changed polclass to be parallel.
> So for example, on a 20 cores system:
> ? polclass (-23512271, 1)
>   cpu time 1min, 55,255 ms, real time 8,886 ms.

Excellent!

Andreas