Teleconference 2017-06-07

Attendees: Ken Moreland (SNL), Abhishek Yenpure (UO), Rob Maynard (Kitware), Berk Geveci (Kitware), Dave Lonie (Kitware), Sujin Philip (Kitware), Matt Larsen (LLNL), Alberto Villarreal (Intel), Manish Mathai (UO), Tom Otahal (SNL), Ollie Lo (LANL), Mark Kim (ORNL)

Abhishek has been doing performance monitoring of the particle streaming. Has found no inconsistencies between architectures.

Manish is working on space splitting based off of LANL's earlier work.

Sujin working on surface normals. There is a merge request that implements the filter.

Dave is working on getting cell data through Marching Cubes filter.

Rob is working on the gradients filter. He is trying to eliminate the need of writing around the intermediate gradients when doing something like voracity.

Matt is taking another look at annotation in the rendering package and are going to implement it for ray tracing.

Matt continuing to troubleshoot MPI + OpenMP + TBB. A problem is that as soon as you set the affinity, OpenMP grabs all threads and turns of threading for everyone else. Looking into possible solutions.

Looking into supporting multi-image rendering for ALPINE. You can be more efficient if you do all the multiple renderings as one rather than treat them as independent renders.

Alberto has offered to set up a meeting between VTK-m folks and TBB developers if we would like it.

LANL getting kd-tree to work. Can have a bunch of points as input and create a kd-tree in parallel. Can find points and nearest neighbor in parallel. Need to come up with an interface for the kd-tree and other search structures.