[Dune] Dune in socis?

Christoph GrĂ¼ninger foss at grueninger.de
Mon Mar 4 00:37:31 CET 2019


Hi Ria,
great to hear people get excited when they discover Dune.

2) There is no particular math level required for contribution.
Commenting other people's merge requests, investigating reported issues,
 fixing bugs, getting rid of warnings, or eliminating deprecations might
be first steps.
Dune is not only about difficult numerics but about programming or
computer science problems, too. It also depends on the area of your
interest.

1) There are no actual plans to apply for SOCIS 2019. It has not been
announced anyway. More important, there were not enough volunteers to
serve as potential mentors for GSoC, so I believe we have a similar
problem for SOCIS.
Nevertheless, if you find a topic interesting enough to work for and
someone from the Dune community is willing to mentor, I am happy to
prepare and manage Dune's application.
Get some confidence with Dune and search for some area you could help.
It might pan out, if some gets interested in mentoring you. But if you
are really looking for a SOCIS project, I advice you to check also with
other organizations.

Some ideas that comes to my mind:
* Improve the Python binding. (more precise project idea needed)
* Add support for latest Scotch PT and keep binding to ParMETIS working
(not enough for a project, but might serve as a start)
* Create a parallel binding for the direct solver of MUMPS (requires
some experience with MPI parallel code)

Please feel free to ask further questions.

Bye
Christoph

-- 
Unfortunately, plots are notoriously hard to get right. Partly, the
default settings of programs like gnuplot or Excel are to blame for
this since these programs make it very convenient to create bad plots.
                        -- Till Tantau, "The TikZ and PGF Packages"




More information about the Dune mailing list