[Libmesh-users] Re: Eigenvalue Solver

[David X. <dx...@my...>]

Hi Steffen,

I'm looking forward to your example 17 and implementation of
generalized eigen value solver with libmesh.

Thanks,

David

On 2/27/06, Steffen Petersen <ste...@tu...> wrote:
> I have just committed an example 17 including a generalized
> eigenvalue problem. Currently, the largest eigenvalues
> in the spectrum are computed. The code can be easily changed
> in order to compute the smallest values (or use the slepc comman
> line options). However, the Lanczos solver slepc offers with this
> option appears to be somewhat unstable when used to compute
> the smallest values (perhaps this is due to the zero eigenmode in
> this example). An option to work around this problem is to
> configure and install slepc with arpack and parpack (the arpack
> solver works fine for example 17).
>
> Steffen
>
>
>
> David Xu schrieb:
> > Hi Steffen and Ondrej,
> >
> > I read your previous message that you've extended the slepc interface
> > to support generalized eigen problems. I'd like to get the details
> > from you. I'm new to libmesh and I'm wondering if you can post some
> > new tutorial examples on how to solve eigen problems using libmesh
> > (eg. solve 1D or 2D simple harmonic oscillator)? That will a great
> > help.
> >
> > Ondrej, I read about your message about your implementation of libmesh
> > and pysparse's jacobi-davidson solver. That sounds really good. Could
> > you please give me the details too? I really appreciate it.
> >
> > David
>
>