2.6 KiB
Minutes from 79th OpenVDB TSC meeting, Feb 9th, 2021, (EDT)
Attendees: Nick A., Jeff L., Ken M., Dan B, Andre P.
Additional Attendees: Johannes Meng (Intel), JT Nelson (Blender), Bruce Chernaik (Intel), Laura Lediaev (ImageWorks), Robin Rowe (Cinepaint)
Agenda:
-
Confirm Quorum
-
Secretary
-
Forum
-
TSC
-
Half
-
Ken Presentation (first quarter objectives)
-
PRs
-
Extrapolate SOP
-
7.2.2 / 8.0.1
-
NanoVDB
-
Next meeting
-
Confirm Quorum
Quorum is present.
- Secretary
Secretary is Nick Avramoussis
- Forum
Ken to look at issue #943. Nick to upload preliminary docs on issue management/ labeling.
- TSC
Andre now officially part of the TSC!
- Half
Jeff has made a feature branch, Nick's PR #961 is up. Once in, a PR can be made from the feature branch (feature/inlinehalf) to master. Still missing CI for build which use Half.
- Ken Presentation (first quarter objectives)
Ken to present to the TAC on objectives this quarter. These include
- AX SOP #952
- Extrapolate SOP #760
- Merge SOP #623
- Morphology improvements #754
- Filter Tiles #843
- Nano VDB
- Making more components of boost optional
- Bumping blosc version
- PRs
- AX SOP #952
- Extrapolate SOP #760
- Morphology improvements #754
- Filter Tiles #843
Noted that, on the morphology PR, the activate SOP has not been updated. Additionally, dilation now correctly turns existing leaf nodes which become dense into active tiles.
- Extrapolate SOP
#760 Changed to VDB Group params for subset of VDBs to extrapolate. Finds first fog volume to use an extension field and uses it's background values. Andre to change some warnings to messages and send Jeff an example scene he is using for tests.
- 7.2.2 and 8.0.1
Releases have been complete. New documentation workflow proposed by Dan - docs are now published to github pages of the openvdb repo (https://academysoftwarefoundation.github.io/openvdb). Website PR #37 required for this change. PR to start development on 8.0.2 #960 to be approved. Dan and Nick to respond to Emily w.r.t collecting images to showcase VDB.
- Update on Nano
Ken is looking at adding in memory compression and updating the node array of NanoVDB to use offsets rather than exact ids (removing the requirement for nodes to be sequential). The latter allows for modifications on the GPU and will require a major version update due to the data layout changing. Also considering bumping the id/offset value to 64bits - removes the hard 128gb memory limit at the cost of ~1-2% performance. Ken to do more profiling.
- Next meeting
Next meeting is Feb 16th, 2021. 12pm-1pm EST (GMT-5).