[aspect-devel] meeting notes

Timo Heister heister at clemson.edu
Mon Mar 30 10:05:53 PDT 2015


Hey all,

here is a quick summary of today's meeting:

- For these meetings in general:
  - no pre-planned topics, but sending questions to Timo/Wolfgang/Mailing list
    before hand is helpful
  - reporting what you are currently working on / struggling with

- free surface bugs? (Cedric):
  - please report problems (and send simple test cases)
  - Ian needs to fix some bugs
  - not tested in 3d, not tested with new manifold description
  - free surface computations can not use higher order mappings/etc. and need
    very fine mesh on surface to begin with

- Bug with grouping vtk output? (Scott):
  - there was a bug with intel MPI and MPI I/O that is used to group output
    files (should be fixed, though)
  - if it is still broken with newest development version, please report error
    messages
  - for now: disable grouping (only necessary if you use a large number of
    CPUs)

- Solid body rotation:
  - there is a feature to do that: see "Remove nullspace" in the manual and
    cookbooks/future/

- Please explain manifolds:
  - Timo explained manifold descriptions in detail (all faces in interior and
    on surface are represented using higher order polynomials in the
    computation)
  - vtk output is only a linear approximation, nothing we can do about it
  - this will create triangle shaped holes near hanging nodes in the output
    when using adaptive refinement
  - We can implement refining into many more patches for the output (very
    costly for high resolution)
  - Another option: move hanging vertices back onto surfaces for output
    (hackathon topic)

- tracer status:
  - Rene is working on attaching data (->hackathon)
  - Gerry's student has some work done. Instead of preparing a finished pull
    request (difficult right now), we should try to share the current state,
    so we can take a look before the hackathon.

- How to output stress (Sarah):
  - not done yet, but easy enough to implement (hackathon topic)
  - full tensor vs. only norm? maybe both?
  - => Sarah creates issue on github

- thanks for your participation and all your questions!

- next meeting: April 20th, 9am

-- 
Timo Heister
http://www.math.clemson.edu/~heister/


More information about the Aspect-devel mailing list