Re: toeplitz solver

[David Bateman]

Gorazd Brumen wrote:
> Hi,
>
> Another question,
>
> Does the A\b operation recognize that the matrix A is toeplitz and
> does it solve it the "appropriate, fast" way? I have done some work on
> it and am willing to post separate code on it (I will switch to
> developers list as soon as I get an answer to this). I do not have
> the clue how to incorporate it into the \ solver.
>
> Is anybody interested in it?
>
> Gorazd
>
>   
I'm setting the followup flag to maintainers... No the solvers don't
recognize toeplitz matrices yet. If you have C++ code for this then it
can be added by modifying the liboctave/MatrixType.{h,cc} code to probe
for toeplitz matrices (both full and sparse?), and then modify the
{dMatrix,CMatrix,dSparse,CSparse}::solve methods to has a case for the
newly probed toeplitz type and call your code there. Finally, you'll
need to modify the src/DLD-FUNCTIONS/matrix_type.cc function to include
both the reporting and setting of the toeplitz type..

In any case can we see the code you are proposing as I might be able to
help in the above process..

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