dolfin team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Fri, Oct 07, 2005 at 01:32:17PM +0200, Johan Jansson wrote:
> On Thu, Oct 06, 2005 at 03:31:12PM +0200, joakim.odqvist@xxxxxxxxxxxx wrote:
> > Hi,
> > I'm interested in a good numerical scheme to solve the Cahn-Hilliard equation. As you may know it's a non-linear fourth-order in space, 
> > first-order in time parabolic PDE, usually with periodic or zero Neumann boundary conditions.
> > Is this something that DOLFIN/FFC can handle?
> > 
> > Best regards,
> > 
> > Joakim
> > 
> 
> Hi Joakim,
> 
> Garth Wells has (successfully it seems) tested this specific equation
> with DOLFIN/FFC, check out the Fenics gallery:
> 
> http://www.fenics.org/gallery/
> 
> Perhaps he has some more information.
> 
>   Johan

There are two ways to solve a PDE with DOLFIN/FFC: either you use one
of the existing solvers or you implement your own. In the case of
Cahn-Hilliard, you would have to implement your own solver, since
there is no solver/module for that equation in DOLFIN.

Garth, maybe you could implement your solver for Cahn-Hilliad as a
DOLFIN module if there is interest?

/Anders