[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: command \
From: |
David Bateman |
Subject: |
Re: command \ |
Date: |
Wed, 08 Mar 2006 15:57:34 +0100 |
User-agent: |
Mozilla Thunderbird 1.0.6-7.5.20060mdk (X11/20050322) |
Christophe Prud'homme wrote:
> Hi Dave
>
> thank you so much for the detailed info !!
>
> basically we have the same functionalities except that :
>
> o octave (2.1.72, 2.9.4) does not support symbolic calculation
> o octave-2.1.72 is not interfaced with umfpack
>
>
octave-forge has a symbolic package that links to ginac/cln/gmp for
symbolic calculation. However, it doesn't handle matrices and so the
question of how the \ operator works doesn't arise.
The octave-forge sparse solver that can be used with 2.1.72 only uses
SuperLU for the solver.
> [ Wednesday 8 March 2006 15:17 ]
> | Hi Christophe,
> |
> | Your description below is pretty much the goal of the octave, though I
> | think Matlab also special cases at least three other matrices you don't
> | include, these being
> |
> | 1) If A is symbolic, matlab uses maple for symbolic factorization
> | 2) If A is sparse and diagonal (or permuted diagonal), the result is
> | just a scaled version of the input
> | 3) If A is dense Hessenberg matrix, reduce to upper triangular and
> solve.
> |
> | I haven't gotten to implementing the dense upper/lower triangular and
> | cholesky solvers yet (this is on my todo list for octave 3.0), but once
> | that is done the only missing things from matlab relative to octave
> will
> | be the symbolic and Hessenberg factorization. That is at the moment
> | dense matrix are only treated by LU or QR factorization. The current
> | situation for sparse matrices in the octave CVS tree is the following
>
> does that apply to some extent also for octave 2.9.4 or only to CVS ?
>
octave 2.9.4 is missing the sparse QR solvers and so is slightly more
limited than the CVS. So 2.9.4 only handles square non-singular
matrices. Note that the octave CVS also updated the diagonal, and
triangular solvers to handle non-square cases which is why there is no
step 0)
"0) If the matrix is not square go to step 8)".
Regards
David
--
David Bateman address@hidden
Motorola Labs - Paris +33 1 69 35 48 04 (Ph)
Parc Les Algorithmes, Commune de St Aubin +33 6 72 01 06 33 (Mob)
91193 Gif-Sur-Yvette FRANCE +33 1 69 35 77 01 (Fax)
The information contained in this communication has been classified as:
[x] General Business Information
[ ] Motorola Internal Use Only
[ ] Motorola Confidential Proprietary
-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.
Octave's home on the web: http://www.octave.org
How to fund new projects: http://www.octave.org/funding.html
Subscription information: http://www.octave.org/archive.html
-------------------------------------------------------------
- command \, Christophe Prud'homme, 2006/03/08
- Re: command \, David Bateman, 2006/03/08
- Re: command \, Christophe Prud'homme, 2006/03/08
- Message not available
- Re: command \,
David Bateman <=