Re: [Lapackpp-devel] Complex Matrix Support

[Andrew O. <end...@gm...>]

Hey

Attached is the C++ code and the M file I used. The C++ code has some
more detailed results in the commenting.

The machine I ran this on is a SuSe 10.0 from Novell. This is what I
get when I type "uname -a"

Linux mnp 2.6.13-15.8-smp #1 SMP Tue Feb 7 11:07:24 UTC 2006 i686 i686
i386 GNU/Linux

But the RPMs I installed for SuSe all had the i586 suffix on them.

These are the versions of the relevant compilers/rpms:

gcc-4.0.2_20050901-3
gcc-fortran-4.0.2_20050901-3
f2c-0.11-1052

When I compile test.cpp I type:

g++ test.cpp -I/usr/include -I/usr/include/lapackpp -L/usr/lib -llapackpp

and execute by typing "./a.out".

The m-file is just two simple lines and you can see the results in
MATLAB and compare to the values printed out by test.cpp. Thanks for
your help.

Andy


On 5/7/06, Christian Stimming <sti...@tu...> wrote:
> Am Samstag, 6. Mai 2006 22:25 schrieb Andrew Oh:
> > I was wondering how I could access and make use of the complex matrix
> > libraries. On your website you state:
> >
> > Note: To switch on the support for complex-valued matrices, you need
> > to define the macro LA_COMPLEX_SUPPORT in your application.
> >
> > How should I go about doing this?
>
> #define LA_COMPLEX_SUPPORT
>
> Which OS, which compiler is this?
>
> > On a separate note. I tried finding the eigenvalues of some matrix A
> > using LaEigSolveVecIP(), they did not match with the results I
> > obtained from MATLAB.
> >
> > Also, when I looped LaEigSolveVecIP() several times for the same
> > matrix A, the resulting sets of eigenvalues did not match with each
> > other.
> >
> > What could be causing these discrepencies? Any help would be greatly
> > appreciated. Thanks.
>
> Please give detailed instructions for reproducing this, i.e. please copy =
the
> Matlab and C++ code here, so that we can reproduce. Thanks a lot.
>
> Christian
>