[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [OctDev] linear-algebra: new QR updating functions
From: |
Jaroslav Hajek |
Subject: |
Re: [OctDev] linear-algebra: new QR updating functions |
Date: |
Tue, 19 Feb 2008 17:19:31 +0100 |
On Feb 19, 2008 3:53 PM, David Bateman <address@hidden> wrote:
> Jaroslav Hajek wrote:
> > Hello,
> > to anyone who might be interested:
> > I have just committed an initial revision of functions
> > "qrupdate","qrinsert" and "qrdelete" into the main/linear-algebra
> > package. These functions should provide equivalent behaviour to their
> > Matlab counterparts, i.e. update a QR decomposition after a rank-1
> > update, adding/deleting row/column. qrinsert and qrdelete currently
> > support adding/deleting columns, rows are on the way.
> >
> > enjoy,
> >
> >
> These are core matlab functions (ie not in a toolbox) and there should
> also be considered for inclusion in Octave 3.1.x. Is there any reason
> why you wrote the code itself in F77? Is it just that is encapsulates
> several blas/lapack functions and so to do that in C++ would require
> additional wrapping.
>
Just the reason that I'm quite comfortable with Fortran for numerical work,
especially for BLAS/LAPACK-based linear algebra stuff. Octave's seamless support
of Fortran via mkoctfile was always a big plus for me. Also, I may
need to use the subroutines in a Fortran program.
> I'd suggest reworking this as a patch against octave, with the F77 code
> in a directory libcruft/qrupdate (or something similar) and the C++
> interfaces in src/DLD-FUNCTIONS.. The Makefile.in files will need to be
> modified appropriately.
>
No problem, but I'd put it off at least until row addition/deletion is
supported.
More detailed checking of corner cases would also be advisable.
I'm want to add CHOLUPDATE as well, and possibly SVD updating
(but that needs further studying).
> John, we should include these functions.. Are you happy with the F77
> wrapper to blas/lapack function to have only a single F77 function in
> the Octave C++ wrapper or should that be modified to use a pure C++ wrapper?
>
> D.
>
> --
> 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
>
>
regards,
--
RNDr. Jaroslav Hajek
computing expert
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz