Teleconference 2018-02-21

From VTKM
Jump to: navigation, search

Attendees: Ken Moreland (SNL), Matt Letter (SNL), Dave Pugmire (ORNL), Rob Maynard (Kitware), Berk Geveci (Kitware), Sujin Philip (Kitware), Allie Vacanti (Kitware), James Kress (ORNL), Matt Larsen (LLNL), Tom Otahal (SNL), Mark Kim (ORNL), Ollie Lo (LANL)

Dave P wrapping up the cross product and dot product filters. This is a bit of an issue because it takes two fields and the filter interface does not handle that. For the short term Dave will make a field in the filter class to store the string name of the secondary field.

James has been using VTK-m and VTK-h within ADIOS staging. Been working with Rob and Matt Larsen to resolve some issues with CUDA 7.5. Big issue is that CUDA 7.5 does not support static const array variables. But they do support constexpr arrays.

Mark is working on a random array generator.

Rob has been working on discretized color transfer. To get that to work, you have to implement several different color spaces. Adding tests with ParaView.

Allie is working on an OpenMP background for VTK-m.

Sujin recently added a probe filter for VTK-m.

Rob submitting GTC VTK-m draft next week.

Matt Larsen was held up from a bug in the bounds where including an empty bounds caused infinite bounds.

Rob brought up that one of the changes we need to make to VTK-m is to remove exception throwing from templated methods. This happens to add a lot of object code because a different exception has to be recorded for each instantiation.

Ollie looking into extending the cell set connectivity filter from only cell set single type and it now works for any explicit cell set.