Teleconference 2016-03-23

Revision as of 18:00, 23 March 2016 by Kmorel (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Attendees: Erik Bruger (LLNL), Jeremy Meredith (ORNL), Matt Larsen (LLNL), Robert Maynard (Kitware), Chris Sewell (LANL), Dave Pugmire (ORNL)

Chris is working on the example that Matt sent him for the GTC tutorial.

Berk, Chuck, T.J., Chris met last week on the integration of Legion with VTK-m.

Rendering branch is ready to be merged in. Robert will hit the button.

Jeremy removed the field order value. It touched a lot of code, so lots of minor changes everywhere.

Jeremy is taking a look at the dimension parameter on cell sets. We are considering removing it.

Robert has spent an entire week on dashboards. Things are looking much better. Most of the errors have been worked out. One of the big issues left is an issue with the vectorization on some of the code. We need to make sure the vectorization flags are correct for the constraints we give to the code. For a short term solution, we are going to look at disabling vectorization on Intel. We will revisit it afterwords by looking into a hint mechanism that determines when vectorization is safe and effective.

The atomics branch is merged in. Doing performance analysis of various rendering architectures.

Now that atomics are available, Matt would like to merge in his ray casting code. He will wait until after the rendering branch is merged in. The atomics are in vtkm/exec/AtomicArray.h.