Skip to content

Commit 65a39c8

Browse files
authored
Merge pull request #1518 from sideeffects/send_upstream_tscnotes20221101
Meeting notes for Nov 1st, 2022
2 parents 6432d93 + ec7a67f commit 65a39c8

File tree

1 file changed

+113
-0
lines changed

1 file changed

+113
-0
lines changed

tsc/meetings/2022-22-01.md

Lines changed: 113 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,113 @@
1+
Minutes from 153rd OpenVDB TSC meeting, November 1st, 2022
2+
3+
Attendees: *Jeff* L., *Andre* P, *Dan* B., *Ken* M.
4+
5+
Additional Attendees: Greg Hurst (United Therapeutics), Sebastian Gaida
6+
7+
Regrets: *Nick* A., *Rich* J.
8+
9+
Agenda:
10+
11+
1) Confirm quorum
12+
2) Secretary
13+
3) OpenVDB release
14+
4) PyBindings 11
15+
5) Audit of DCCs
16+
6) File IO
17+
7) VDB Max
18+
8) CI
19+
9) Next meeting
20+
21+
------------
22+
23+
1) Confirm quorum
24+
25+
Quorum is present.
26+
27+
2) Secretary
28+
29+
Secretary is Jeff Lait.
30+
31+
3) OpenVDB release
32+
33+
Success, thanks to Andre! Mostly worked smootly, but we have to
34+
remember to spell check changes. It was very useful that we had
35+
pre-run the weeklies.
36+
37+
A PR is present for the website update for Greg's help docs.
38+
39+
CHANGES.txt needs the 10.0.1 change and spell checking.
40+
41+
Greg will post a intro tutorial on the Wolfram community when website
42+
is updated.
43+
44+
We'll wait for Wolfram to okay with being listed as an adopter before
45+
we add them to the adopter list.
46+
47+
We will not update example version numbers in the release check list.
48+
49+
4) PyBindings 11
50+
51+
Issue 1124. matthewdcong is looking at building python bindings for
52+
NVidia and plans on adding this to the OpenVDB.
53+
54+
ILM has boost::python but is a different approach. But doesn't help
55+
us move away from boost::python.
56+
57+
Hope is the new one would be good enough for ILM to eventually switch;
58+
but their version probably is too deeply integrated to do.
59+
60+
Matthew should join and present his work so far and what we'd like it
61+
to look like. Eg, vdb_tool-like?
62+
63+
ILM has "VDB Group" - associated vdb primitives, like a velocity field
64+
and distnace field. Needs to be some mechanism to pair up/collate
65+
VDBs. Hard to integrate into Houdini where VDB are atomic. Within a
66+
group topology does not have to be shared, but by convention they
67+
often are. Groups are disjoint. Could be a directory name per vdb.
68+
69+
5) Audit of DCCs
70+
71+
Larry Gritz has started a spreadsheet for how long things are
72+
supported by various DCCs and what studios are still using. We tend
73+
to be on the leading edge in deprecating quickly.
74+
75+
Most are following VFX Platform and Houdini.
76+
77+
6) File IO
78+
79+
Can we remove all the IO code from tree/Hierarchy into a separate
80+
.h/.cc? Then do all the IO through that mechanism. This will make a
81+
new file io easier. Hopefully user still uses Archive, it just is
82+
implemented elsewhere.
83+
84+
This may require always initialization openvdb, even for writing. A
85+
lot of root node methods require an absolute coordinate, so makes
86+
working with offsets very hard in the file io.
87+
88+
Can friend be added without changing ABI?
89+
90+
Plan is to remove support for the pre-open-source file formats.
91+
92+
Support for 1.0 on will remain. Saving into the new format would be
93+
opt-in, so people can use once their pipeline has migrated.
94+
95+
7) VDB Max
96+
97+
This is pending, has been tested internally and is ready to push up
98+
soon. Designed around fog volumes.
99+
100+
8) CI
101+
102+
Are some of the machines disappearing soon? At some point we have to
103+
pay for these. Goal is to use November stats to figure out how much
104+
we have to pay for. ASWF is tracking costs, but currently it is for
105+
free. Weekly CI outside of a release can use lower powered machines
106+
as it is not critical. There are GPU resources now available; we
107+
might want to add GPU unit tests.
108+
109+
The latest weekly build cost 12 hours of billable compute.
110+
111+
9) Next meeting
112+
113+
Next meeting is on November 9th, 2022. 2pm-3pm EDT (GMT-4).

0 commit comments

Comments
 (0)