mod p limit

9 views

Skip to first unread message

Martin Albrecht

unread,

Sep 26, 2011, 5:12:36 PM9/26/11

to libsingular-devel

Hi,

at http://trac.sagemath.org/sage_trac/ticket/11829 Paul Zimmermann and I are
wondering why the limit for GF(p) is chosen at 2147483629 instead of

sage: previous_prime(2^31)
2147483647

? Just wondering.

Cheers,
Martin

--
name: Martin Albrecht
_pgp: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x8EF0DC99
_otr: 47F43D1A 5D68C36F 468BAEBA 640E8856 D7951CCF
_www: http://martinralbrecht.wordpress.com/
_jab: martinr...@jabber.ccc.de

Hans Schoenemann

unread,

Sep 27, 2011, 8:01:59 AM9/27/11

to libsingu...@googlegroups.com

On Mon, Sep 26, 2011 at 10:12:36PM +0100, Martin Albrecht wrote:
> Hi,
>
> at http://trac.sagemath.org/sage_trac/ticket/11829 Paul Zimmermann and I are
> wondering why the limit for GF(p) is chosen at 2147483629 instead of
>
> sage: previous_prime(2^31)
> 2147483647
>
> ? Just wondering.
>
> Cheers,
> Martin
>
>

Thre are 3 limits for Z/p:
- for the singular kernel: <2^31 (the wrong prime is a copy-and-past typo)
- for factory routines (factorize, det, resultant, gcd, ..): <2^29
- for the NTL-routines < NTL_SP_BOUND (2^30 on 32bit machines)