[Fwd: Re: determinant modulo p of large dense integer matrix]

[David Saunders]

I propose to add this guidance on citing LinBox to the website:

http://www.cis.udel.edu/~saunders/cite-linbox.html

Let me know any corrections, thoughts.

-dave

-------- Original Message --------
Subject: Re: determinant modulo p of large dense integer matrix
Date: Sun, 24 May 2009 16:07:42 -0700 (PDT)
From: John Matrix <johnma...@yahoo.com>
To: David Saunders <saun...@UDel.Edu>
References:
<f625a683-ff47-44cd...@s1g2000prg.googlegroups.com>
<498C218A...@gmail.com>
<feb64afd-1640-4963...@x6g2000pre.googlegroups.com>
<49903AC3...@udel.edu>

I just wanted to thank all you guys for your help. Linbox did a
terrific job on my problem.
When writing a paper and acknowledging use of Linbox, is there a
particular reference that one should cite?

Thanks,
JM

On Feb 9, 7:16 am, David Saunders <saund...@UDel.Edu> wrote:
> John,
>
> You use but one field and it's size is not crucial. But you have many
> field elements. Use sizeof(Modular<int>::Element) and
> sizeof(Modular<double>::Element) when calculating how much memory your
> matrix will consume.
>
> Side remark on matrix size: When blackbox algorithms are to be used,
> the reduction in size of a large sparse matrix may not be worth the
> effort. The dominant parameters are number of nonzero entries and rank
> or degree of minimal polynomial, not matrix order. However, if I recall
> correctly, your matrix is dense and expected rank = order. Still, I
> can't resist throwing out this comment as general background info.
>
> Best, -dave
>
> John Matrix wrote:
>
> > Hi Clement,
>
> > thanks for your further comments. The numbers 12 and 16 are the values
> > of
> > sizeof(Field) in det.C and rank.C, respectively, where Field is
> > Modular<double>
> > in one case and Modular<int> in the other. But either way, I will
> > probably do fine,
> > since some reduction in the matrix size turned out to be trivial.
>
> > Cheers,
>
> > John