Boost logo

Ublas :

Subject: Re: [ublas] Eigenvalues & -vectors
From: Kraus Philipp (philipp.kraus_at_[hidden])
Date: 2010-07-04 17:20:15

Am 04.07.2010 um 20:11 schrieb Thomas Klimpel:

> Marco Guazzone wrote:
>> not sure but (seeing the number of errors) I think Philipp has my
>> same problem.
>> [snip]
>> Still waiting for an answer :(
> Sorry, I no longer monitor the boost user list. If any of you still
> monitors the boost user list, please redirect any requests regarding
> ublas or the bindings to this ublas mailing list (that's what I also
> did when I was still monitoring it).
> Inclusion of "bindings/lapack.hpp" doesn't work yet. I tried to
> clean up "bindings/lapack/driver.hpp" recently, but "bindings/lapack/
> computational.hpp" and "bindings/lapack/auxiliary.hpp" are
> definitively not clean yet. Your best bet right now is to directly
> include the header of the lapack routine you want to use. This is
> "bindings/lapack/driver/syev.hpp" in case of Philipp, as he wants to
> compute the eigenvalues of a symmetric matrix.

That was a perfect hint with inclunding the sysev.hpp directly. I can
compile my source without any problems.

I had compiled the LAPACK into / with ATLAS, so I had 4 static lib
(libatlas.a, libcblas.a, liblapack.a, libptcblas.a) and I use the g++
with -D BOOST_NUMERIC_BINDINGS_BLAS_CBLAS.Is that all right so far?

I have looked at the example for the dsyev call under
Can I call boost syev in the same way?

I want to calculate for an arbitrary square matrix the eigenvectors
and -values (like the Matlab command
  ). Do I understand the steps in the correct order?
1. I create a lower bounded matrix of my sqaured matrix
2. I call the boost command syev with the bounded matrix

Can I use LAPACK commands for creating the lower bounded matrix? How?
What do I use the workspace in the syev function call?