Plan Mercury Intensive Workshop 2016
Here's a list of discussion points for tomorrow; please add any topics you think I forgot.
I preliminarily assigned the issues to certain people; this is not definite and will be discussed tomorrow.
- M>1: Quaternions (Thomas), superquadrics (Deepak),
stay convex, otherwise we require multiple forces per particle,
Deepak, add paper reference - M>1: Making handlers into factories: remove addObject, pointers need to be gotten from factory, for saver restart (Ant)
- M>1: Parallelisation: see mesh decomposition library (Marnix, Deepak)
Domain decomposition - MPI, not OpenMP - M1: Check todo's, bug's (Ant)
- M1: Check documentation (Kit)
- M1: Write down the procedure for committing (run fullTest), selftesting (10 sec rule), document the different cmake options (how to run a single test) (Thomas)
logging (how?), (Marnix) - M>1 SPH particles (Thomas)
- M>1 upgrade CG (flexible set of variables) (Thomas)
- Binary output
- Merging ED support (Ant)
- Write down and apply the coding style: f.e.
- we made member classes for handlers and files public,
- where to comment member variables (in the set function or the declaration)
- how to set defaults (set to nan or sensible values)
- check values before solve (we need to introduce that check)
- changes should be made in the trunk, except for bug fixes.
- Prepare the Take-off phase 2 grant application: http://www.stw.nl/en/node/6726 (MercuryLab)
- M1: Ant: move alpha commits to Trunk
- M2: DPMBase templated on ParticleType, InteractionType
- Update paper list of the MErcuryDPM website
Low priority
- Implicit time integration: For large time steps; look into sundials library
Thomas: 6 Write documentation for cmake, 1 Quaternions
Marnix: Logger, Parallelisation
Ant: Checking the documentation, 13 Merging, Review Email
Kit: Documentation
Deepak: