[Dune] [GSoC20015] Not enough projects and mentors
Dominic Kempf
dominic.r.kempf at gmail.com
Mon Feb 16 14:50:16 CET 2015
Dear Christian,
I very much like your proposal because this is on my TODO list anyway. I
need this as part of (a soon to be publicly announced) project on automated
testing in Dune. There, we want to control the variation of system tests
through ini files mainly and solver construction definitely should be done
through a factory class that takes the parameter tree.
That being said, count me in as a mentor/co-mentor on this one. We can talk
about details next week during the dune course.
Best,
Dominic
On Mon, Feb 16, 2015 at 2:40 PM, <christi at dune-project.org> wrote:
> What would be helpful...
>
> currently we lack the possibility to create solvers and
> preconditioners in a general fashion. We are moving towards a state
> where many components can be configured using the parametertree. I
> consider a parametertree based factory for ISTL an important step
> towards a usable solver component.
>
> The following steps are
>
> ...necessary:
> - cleanup the solvers/preconditioners to fully use abstract interfaces
> - update solvers/preconditioners to use parametertrees for the
> configuration
> - introduce a factory class. I have a ParameterizedObjectFactory lying
> around, which can instantiate previously registered classes.
> - add an ISTL library which provides precompiled
> solvers/preconditioners for the most common data types.
>
> ...optional:
> - write a suitable solver backend for PDELab which allows to create
> solvers dynamically
> - add infrastructure to combine 3rd party components for the parallel
> commmunication with dynamic object creation.
>
> Ciao
> Christian
>
> On Fri, Feb 13, 2015 at 10:16:31AM +0100, Christian Engwer wrote:
> > Hi Christoph,
> >
> > I'm away with the kids. I'm willing to mentor and I'll think about
> projects over the weekend.
> >
> > Christian
> >
> > Am 13. Februar 2015 08:44:19 MEZ, schrieb "Christoph Grüninger" <
> christoph.grueninger at iws.uni-stuttgart.de>:
> > >Hi Dune,
> > >we have one week left to prepare our GSoC application. I'll be on
> > >vacation beginning next Thuesday. Thus, I'll have to prepare the
> > >application this weekend.
> > >
> > >But until now we have only three mentors (Oliver, Aleksejs, myself) and
> > >only three projects (IB for CurvilinearGrid, p4est, CMake super build).
> > >If we don't get at least one more mentor and one more project until
> > >tonight, I propose to skip GSoc 2015 and try with ESA again. The latter
> > >is smaller, less crowded and we only get one student.
> > >
> > >Bye
> > >Christoph
> > >
> > >--
> > >GDB does hate your application, expresses its contempt through the
> > >design of its command-line interface. -- Tom Tromey, FOSDEM 2014
> > >
> > >
> > >
> > >------------------------------------------------------------------------
> > >
> > >_______________________________________________
> > >Dune mailing list
> > >Dune at dune-project.org
> > >http://lists.dune-project.org/mailman/listinfo/dune
> >
> >
> > _______________________________________________
> > Dune mailing list
> > Dune at dune-project.org
> > http://lists.dune-project.org/mailman/listinfo/dune
>
> --
> Prof. Dr. Christian Engwer
> Institut für Numerische und Angewandte Mathematik
> Fachbereich Mathematik und Informatik der Universität Münster
> Einsteinstrasse 62
> 48149 Münster
>
> E-Mail christian.engwer at uni-muenster.de
> Telefon +49 251 83-35067
> FAX +49 251 83-32729
>
> _______________________________________________
> Dune mailing list
> Dune at dune-project.org
> http://lists.dune-project.org/mailman/listinfo/dune
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.dune-project.org/pipermail/dune/attachments/20150216/87d86e2a/attachment.htm>
More information about the Dune
mailing list