[Dune] AMG and performance

[Markus Blatt]

On Thu, Apr 30, 2009 at 10:26:49AM +0200, Arne Rekdal wrote:
> The timings is performed on an AMD Opteron 2216 with 32 GB memory.
>
> S is a matrix coming from a 3D Mixed-Hybrid Element formulation of an  
> anisotropic Poisson problem. S is the result of performing two  
> Schur-complement reductions of the original system. The unknowns in the 
> transformed system (Sr=b), is the pressure on the element's faces.
>

You are using the default values of the criterion (which are
reasonable for an isotropic 2d problem).

As I seed it, your problem would require 3d ansitropic coarsening. 
I agree, that up to now setting the parameters was rather magic.
Therefore I just introduced utility functions in the class
AggregationCriterion.

So add a 

criterion.setDefaultValuesAnisotropic(3,2);

for 3d and aggregate side length 2 (will lead to roughly building
aggregates 2x2x3).


I also changed the way to get debugging output. Per default you will
get some statistics about the aggregation process. It would be nice to
have a look at that.

Be sure to use the latest istl from trunk before trying again.
Cheers,

Markus