[Dune-devel] Semantics of overlapSize and ghostSize

Peter Bastian peter.bastian at iwr.uni-heidelberg.de
Wed Aug 3 12:18:27 CEST 2016


Dear Oliver,

what about the definition in the dune grid paper (part 1)?

My understanding is this:

Codim 0:
- interior forms a partitioning of the set of all elements
- overlap are elements to form a set of overlapping subdomains in the
sense of domain decomposition methods (I think there is no strict 
definition)
- ghost is similar but typically ghost elements only serve as the minimum
additional elements needed in addition to interior elements
- A grid might have only interior elements(e.g. Yasp), interior+overlap 
(e.g. Yasp),
interior+ghost (UG, ALU), interior+overlap+ghost(none)

Codim>0:
- For any codim 0 entity all subentities of codim>0 are also stored on 
that process
- In general, a codim>0 entity has the same partition type as the codim 
0 entity that it is part of
- since this may be non-unique, two additional partition types exist: 
border and front.
- border are entites on the boundary of interior to X where X is either 
overlap or ghost
- front are entites on the boundary from overlap to ghost

Probably this is still ambiguous at some places, but it should give the 
general idea.

-- Peter



Am 03.08.16 um 11:14 schrieb Bernd Flemisch:
> Hi Oliver,
>
> there has been some discussion on the dune-devel list, see for example
> http://lists.dune-project.org/pipermail/dune-devel/2013-November/000648.html
>
> People agreed that this should be discussed in a meeting or a small 
> task force, but I am not aware of that this actually happened.
>
> Kind regards
> Bernd
>
> On 08/03/2016 11:00 AM, Oliver Sander wrote:
>> Dear Dune experts,
>>
>> I am trying to explain the meaning of the overlapSize and ghostSize methods to someone,
>> and I've come to realize that I don't know it myself.  I know it means something like
>> "the thickness of the overlap/ghost layers, measured in terms of elements", but is there
>> anything more precise?  In particular, boths methods take integer arguments 'codim',
>> but I have no idea of how that argument is supposed to influence the result.  What is
>> the most precise definition of the semantics of overlapSize and ghostSize that you know of?
>>
>> Thanks,
>> Oliver
>>
>>
>>
>> _______________________________________________
>> Dune-devel mailing list
>> Dune-devel at dune-project.org
>> http://lists.dune-project.org/mailman/listinfo/dune-devel
>
>
> -- 
> _______________________________________________________________
>
> Bernd Flemisch                         phone: +49 711 685 69162
> IWS, Universität Stuttgart             fax:   +49 711 685 60430
> Pfaffenwaldring 61            email:bernd at iws.uni-stuttgart.de
> D-70569 Stuttgart            url:www.hydrosys.uni-stuttgart.de
> _______________________________________________________________
>
>
> _______________________________________________
> Dune-devel mailing list
> Dune-devel at dune-project.org
> http://lists.dune-project.org/mailman/listinfo/dune-devel


-- 
------------------------------------------------------------
Peter Bastian
Interdisziplinäres Zentrum für Wissenschaftliches Rechnen
Universität Heidelberg
Im Neuenheimer Feld 205 (Mathematikon) Raum 1/401
D-69120 Heidelberg
Tel: 0049 (0) 6221 5414517
Fax: 0049 (0) 6221 548884
email: peter.bastian at iwr.uni-heidelberg.de
web: http://conan.iwr.uni-heidelberg.de/people/peter/

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.dune-project.org/pipermail/dune-devel/attachments/20160803/05a2307f/attachment.htm>


More information about the Dune-devel mailing list