Re: eigs arpack arpack++ and Debian

[David Bateman]

Christophe Prud'homme wrote:

>All, David
>
>it seems that eigs is not supported in octave 2.9.4 (at least in debian 
>version)
>I am the maintainer of arpack and arpack++ in Debian. arpack++ is a C++ 
>interface to arpack and simplifies tremendously the use of arpack.
>
>
>unfortunately arpack++ is not maintained anymore by the original authors. I 
>have taken over the maintenance (non officially) and ported arpack++ to the 
>newest version of g++ and use now 
>autoconf/automake/libtool. There is also a complete testsuite that can be run 
>automatically with a large set of sparse matrices and problems (standard and 
>general) to test the library
>
>the implementation of eigs should perhaps be simplified with arpack++
>and the testsuite could perhaps be ported to octave. 
>
>here is the link (at the bottom the the page)
>http://packages.qa.debian.org/a/arpack++.html
>
>Could that be helpful ?
>
>Best regards
>C.
>  
>
Christophe,

I spent a large time about a year ago trying to get eigs running in
octave. I have a full implementation but it is unstable. There is a
thread about it on the maintainers list in April and May last year.

I believe there is something I'm doing incorrectly with arpack that
makes the results unstable, and I'm not sure what. For real PD matrices,
this means that the algorithm seems to be slower to converge than the
matlab equivalent, whereas as for other types of matrices, the incorrect
eigenvalues are returned. Someone who knows more about the Arnoldi
algorithm or even better arpack itself than me really needs to look at
this code to fix.

I attach my latest version of eigs.cc to this message. To build it run

% mkoctfile eigs.cc -larpack

There are integrated tests in the code that can test its full
functionality. that can be run from the octave prompt as

% test eigs.cc

though you'll either need the octave CVS (or 2.9.5 that is coming out
very shortly), or octave-forge to have the test function available.

If you have experience with arpack, then perhaps you might like to take
a look at this code and see if you can see what I did wrong...

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

eigs.cc.gz
Description: application/gzip