Teleconference 2015-01-07

From VTKM
Jump to navigation Jump to search

Attendees: Ken Moreland (SNL), James Kress (OU), Chris Sewell (LANL), Robert Maynard (Kitware), Jeremy Meredith (ORNL), Dave Pugmire (ORNL)

Copyright Assertion Ken had a chance to talk to Lucy Nowell at SC and asked about the appropriate copyright assertion for VTK-m. She said that the appropriate way to do this was for Sandia to submit a joint copyright assertion with the other institutions. The Sandia lawyers finally started the process. Robert, Chris, Jeremy, and Dave should all have received an authorship memorandum that each should fill out and send either to Ken or the contact on the memorandum. (If you send it directly to the Sandia lawyer, please also send a note to Ken.) Also, let the Sandia lawyer know if there are others in your institution that should be listed as authors.

There is always the question of what happens with code added to the repository. The normal mode of operation is to continue to edit the repository once the copyright assertion is granted. At some point the code will be considered to pass a "major revision." When that occurs we will repeat the process.

Isosurface algorithm Chris recently implemented an isosurface algorithm that works the same as its PISTON counterpart using the features within. This is a good exercise to get familiar with as well as provide an algorithm to start implementing other features within VTK-m.

Data model There is quite a bit of complexity in modifying EAVL data objects to the existing VTK-m code. Jeremy is having a bit of trouble making the first step. The focus of the initial implementation will probably be on the features needed for the isosurface algorithm. The input to that is a uniform 3D grid and the output is zoo triangles. It is probably best to introduce the data models first and then modify the algorithm to use the new data model.

GTC Talk Ken's proposal to talk about VTK-m was granted, but Ken now has a conflict and cannot give it. Robert and Jeremy both expressed interest in giving it, but now Robert has become busy with other GTC activities and would rather not add this to them. Instead, Jeremy is now volunteered for the talk. Ken will contact the GTC organizers and ask to change the presenter. Ken, Robert, and Chris will all help make slides. The initial set of slides is due at the end of the month.

CUDA Robert is working on implementing a CUDA device in VTK-m. He ran into an issue with the current implementation of FunctionInterface being unable to handle the host/device localities. Ken identified the issue and will work to fix it.

No meeting next week Several of us will be in Walnut Creek, CA for the ASCR PI meeting. This will be a good idea for some of us to meet face-to-face and have some technical discussions. Probably Tuesday afternoon/evening will work best.