[Dune] parallel solve with distributed data

Ruth Schöbel r.schoebel at fz-juelich.de
Tue Dec 5 16:01:19 CET 2017


Dear Dune community,

I like to solve a system like (mass + dt * stiffness) * u = mass * u_0 in parallel.

I start constructing my basis functions out of a parallel overlapping grid and with them I call the assemblers from dune-fufem. So now every process has a part of the global mass and stiffnessmatrix? Is this so far a proper procedure?

In the next step I am at least confused about dealing with the index sets (remote, local, global) for the parallel dune-istel solver:

The OwnerOverlapCopyCommunication can take an IndexInfoFromGrid object, which I probably need? How did I get all the index sets for that from the grid or do I need to set something up by hand? Are there any example codes about how to deal with remote matrices?

Regards Ruth



------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
Forschungszentrum Juelich GmbH
52425 Juelich
Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzender des Aufsichtsrats: MinDir Dr. Karl Eugen Huthmacher
Geschaeftsfuehrung: Prof. Dr.-Ing. Wolfgang Marquardt (Vorsitzender),
Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt,
Prof. Dr. Sebastian M. Schmidt
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.dune-project.org/pipermail/dune/attachments/20171205/50736525/attachment.htm>


More information about the Dune mailing list