[Dune] Setting and optimising AMG parameters in ISTL

[Bath Uni]

Hi Markus,

>> thanks a lot, I'm still having trouble installing the latest version
>> of pdelab (you might have seen my email on the pdelab list). 
> 
> Sorry for not being precise. I should have said revision r1996 of
> dune-pdelab and not latest.
ah, ok, thanks, I can install r1996 without problems. What I should probably do first is extract the parameters and print them all out, that will at least give me a better idea of what is used at the moment, and we can think about how this can be tweaked for our specific problem.

> 
>> As far as I can see the coarsen criterion is only used in the
>> apply() method of the class ISTLBackend_AMG in backend.hh, where the
>> AMG object is instantiated, so I guess I could always derive a class
>> from ISTLBackend_AMG and overload the apply() method with the
>> UnSymmetricCriterion (and then also derive ISTLBackend_AMG_SSOR
>> etc. from this). Or even hack the ISTL_Backend_AMG class in my Dune
>> installation... 
> 
> Yes, but you should probably use new names for ISTLBackend_AMG_SSOR,
> etc. 
Of course, what would be a good name?
> 
> How about instead of (or after) hacking you create a clean
> backend for the UnsymetricCriterion and send us patch. We would highly
> appreciate this and apply it to PDELab
Sure, if I manage to get it to work can create a patch relative to r1996.

Eike
> .
> 
> Cheers,
> 
> Markus
> 
> 
> -- 
> Do you need more support with DUNE or HPC in general? 
> 
> Dr. Markus Blatt - HPC-Simulation-Software & Services http://www.dr-blatt.de
> Hans-Bunte-Str. 8-10, 69123 Heidelberg, Germany
> Tel.: +49 (0) 160 97590858  Fax: +49 (0)322 1108991658 
> 
> _______________________________________________
> Dune mailing list
> Dune at dune-project.org
> http://lists.dune-project.org/mailman/listinfo/dune