Re: use LAPACK routine for triangular systems?

[Evan Monroig]

On 11/28/05, Paul Kienzle <address@hidden> wrote:
>
> On Nov 27, 2005, at 8:22 PM, Evan Monroig wrote:
>
> > On 11/26/05, Paul Kienzle <address@hidden> wrote:
> >> Octave-forge has trisolve.m in main/splines which uses the LAPACK
> >> routines.  Octave 2.9.x has this built into the sparse matrix
> >> solver routines.
> >>
> >> - Paul
> >
> > This is for tridiagonal systems, and I am looking into using a LAPACK
> > routine for upper-triangular systems ;)
>
> In octave-forge/extra/linear-algebra there is a triangular
> matrix type and implements a solver.

Yes, I just looked at it and saw that
octave-forge/extra/linear-algebra/ov-re-tri.cc implements a triangular
matrix type, and also a method for solving a linear system with such a
matrix by recursive back-substitution (I'm not sure if this is the
right name).

After reading the corresponding functions in LAPACK and BLAS
(lapack3-3.0.20000531a/blas/src/dtrsm.f), the same method seems to be
used. I don't quite fluently read fortran but it looked like that.

I thought that there would be special methods to avoid troubles with
badly-conditionned matrices when dividing by the diagonal numbers of
the matrix ^^.

Evan



-------------------------------------------------------------
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
-------------------------------------------------------------