dolfin team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Mon, Aug 25, 2008 at 03:08:21PM +0200, Evan Lezar wrote:
> (Sorry, I keep missing the list)
> 
> There are also ways to scale the eigensystem so that the desired eigenvalues
> are the largest eigenvalues of the scaled system.  I don't have the references
> available at the moment, but I will track some down later this evening.
> 
> Evan

Yes, I could do that manually, but I was hoping SLEPc could handle
this.

-- 
Anders


> 
> On Mon, Aug 25, 2008 at 2:48 PM, Garth N. Wells <gnw20@xxxxxxxxx> wrote:
> 
>     Anders Logg wrote:
>     > How should one use the eigenvalue solver to compute the N smallest
>     > eigenvalues?
>     >
>     > There seems to be a parameter "Eigenvalues to compute" that is used
>     > internally in SLEPcEigenvalueSolver. Should this be used?
>     >
> 
> 
>     I would say with care. You'll probably need to do some work on
>     SLEPcEigenSolver and look into the SLEPc manual in you want to compute
>     the smallest eigenvalues efficiently. Most algorithms are designed to
>     find the largest and won't work for finding the smallest eigenvalues.
> 
>     If you're interested in small systems, SLEPcEigenSolver can find all
>     eigenvalues using a direct method.
> 
>     Garth
> 
>     >
>     >
>     > ------------------------------------------------------------------------
>     >
>     > _______________________________________________
>     > DOLFIN-dev mailing list
>     > DOLFIN-dev@xxxxxxxxxx
>     > http://www.fenics.org/mailman/listinfo/dolfin-dev
> 
> 
>     _______________________________________________
>     DOLFIN-dev mailing list
>     DOLFIN-dev@xxxxxxxxxx
>     http://www.fenics.org/mailman/listinfo/dolfin-dev
> 
> 
> 

> _______________________________________________
> DOLFIN-dev mailing list
> DOLFIN-dev@xxxxxxxxxx
> http://www.fenics.org/mailman/listinfo/dolfin-dev

Attachment: signature.asc
Description: Digital signature