Karim Belabas on Wed, 19 Jun 2013 00:01:11 +0200


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

Re: FlxqM_gauss, FlxqM_inv and FqM_inv


* Bill Allombert [2013-06-18 23:50]:
> On Tue, Jun 18, 2013 at 02:54:33PM +0200, Peter Bruin wrote:
> > Hello,
> > 
> > For personal use I needed the functions FlxqM_gauss and FlxqM_inv.
> > Probably it would be nice to have these in PARI, so I am attaching a
> > patch that adds these functions (as straightforward adaptations of the
> > existing equivalent functions for other base fields).  I also added
> > FqM_inv since I noticed it was missing and it was trivial to implement.
> > The patch is based on version 2.6.1 (development git-1e829d2).
> 
> Hello Peter,
> 
> Mostly the reason I did not add these functions when I added the generic
> interface was that there were no way to test them in GP. So I would very
> much welcome some use for such functions inside the library.

They seem useful to me.

We can add a call to FlxqM_gauss in FqM_gauss when p is small, as
suggested in Peter's first email. I expect this to be an improvement.

And the functions would be tested at the same time as FqM_gauss.

Hum.

Scratch that: FqM_gauss is currently never used in libpary when p is small...
But we could at least test it through the nfsolvemodpr() GP function.

Cheers,

    K.B.

P.S. Your point about untested unused functions remains quite valid. I
just wasted 10 minutes trying to understand why my adaptation of the new
"inverseimage" code to Flm inputs was failing horribly. Turned out the
Flm_Fl_mul(x, y, p) function actually returned x, indepently of y ...

(The function that's being used and tested is Flm_Fl_mul_inplace)

--
Karim Belabas, IMB (UMR 5251)  Tel: (+33) (0)5 40 00 26 17
Universite Bordeaux 1          Fax: (+33) (0)5 40 00 69 50
351, cours de la Liberation    http://www.math.u-bordeaux1.fr/~kbelabas/
F-33405 Talence (France)       http://pari.math.u-bordeaux1.fr/  [PARI/GP]
`