The routine you want to look at is DGEEV. But this is really complicated. I do
not advise you to do that.
You might want to develop your own 6-by-6 solver based on the QR algorithm
(Check any numerical linear algebra book.) For 6-by-6, this is not that hard
and most of the
recent development have been targeting larger sizes. The plain QR algorithm
from the 70s
is just fine for 6-by-6. So you reduce your matrix to Hessenberg form, find a
shift, apply a QR step
on the Hessenberg form, deflate when needed, etc. This is a standard project
for graduate students
when they take a Numerical Linear Algebra class. It is probably better to go
that road than to try to
translate the LAPACK code. J.
On Jul 20, 2012, at 1:50 AM, Roly Whear wrote:
I was wondering if you could help with some advice on which lapack routine I
should use for the following problem (the maths is getting a bit beyond me).
I'm looking to get the rigid body modes of a mass on a set of springs. Sorting
out the mass and stiffness matricies to get the characteristic equation has
been relatively straight forward which leaves me with a 6x6 non symmetric
matrix. All I'm after are the real Eigenvalues and vectors.
The plan is to rework the C or fortran code into VB in Excel if it will work.
Any advice gratefully received
Lapack mailing list
-------------- next part --------------
An HTML attachment was scrubbed...