Karim BELABAS on Wed, 11 Sep 2002 02:30:13 +0200 (MEST)

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: bnfisnorm() broken again

On Tue, 10 Sep 2002, Igor Schein wrote:
> ? rnfisnorm(rnfisnorminit(y^2 + 29*y + 21,x^3 - 10*x^2 - 3*x - 5),2)
>   ***   precision too low in isunit.
> Looks like internal precision is used to compute units

Not really. I've slightly modified the logic to jump quickly to the "right"
precision instead of doubling a few times before giving up. Is is better ?

It's now a bug that such large units enter the picture. The extension field
simply has large fundamental units (see T[2].fu, with T the rnfisnorminit)

> and user cannot influence it.

Yes you could. But it should have been about \p2000 or higher to stand a

> Also, I can't figure out which flags to use to speed up rnfisnorminit() in
> the example above.

It's not slow, it takes a few seconds! The algorithm requires computing the
bnfinit() of the absolute extension, which may be very expensive.

Karim Belabas                    Tel: (+33) (0)1 69 15 57 48
Dép. de Mathematiques, Bat. 425  Fax: (+33) (0)1 69 15 60 19
Université Paris-Sud             Email: Karim.Belabas@math.u-psud.fr
F-91405 Orsay (France)           http://www.math.u-psud.fr/~belabas/
PARI/GP Home Page: http://www.parigp-home.de/