John Mousel <[email protected]> writes:

> I'm trying to mark two distinct regions of a domain by solving the Laplace
> equation with constant Dirichlet BC's. Because I'm just rounding to the
> nearest integer, I just need cheap AMG like behavior to move the constant
> across the domain quickly. 

What is the physical propagation mechanism?  Is there a concept of
conservation that should be satisfied?  What's wrong with a normal AMG
configuration for the Laplacian?

Attachment: signature.asc
Description: PGP signature

Reply via email to