- Brecht Van Lommel (Blender)
- Kévin Dietrich (Blender)
- Jeroen Bakker (Blender)
- Clément Foucault (Blender)
- William Leeson (Blender)
- Patrick Mours (NVIDIA)
- Brian Savery (AMD)
- Christophe Hery (Facebook)
- Feng Xie (Facebook)
- Most of the meeting was Jeroen & Clément presenting a proposed 3.x series roadmap and discussion related to that. The details of that will be communicated elsewhere.
- OpenSubdiv acceleration likely will not make it into 3.0 given the close deadline. Brecht proposes to take time to finish it well for 3.1 then, and make time to get a clean implementation based on feedback from other developers.
- Brecht proposes Kévin works on Eevee generic attribute rendering for geometry nodes, since this is closely related to his work on Alembic generic attribute support and familiarity with this area due to the OpenSubdiv work. This might make it into 3.0 still.
- Cycles X landed in master. During the bcon2 stage we still plan to solve a few important issues, and of course fix reported bugs in general.
- Optimizing transparent hair rendering to avoid performance regressions, by baking transparency per curve vertex
- Improve ambient occlusion to work with transparent shadows and hopefully improving performance
- Improving random walk SSS for some mesh shapes, or restoring Christensen-Burley if not possible.
- Fixing some rendering artifacts introduced by new shadow terminator
- HIP device support for AMD GPU rendering landed in master. However this not yet ready for user testing or enabled in daily builds. Next steps are enabling building on the buildbot and waiting for the Windows driver to be released. (task)
- William found that PMJ was not using the better PMJ02 pattern, this seems to get convergence closer to Sobol, so if further testing confirms this we can make this change.
- William keeps working on improving self-intersection handling. Some issues still with dynamic/static BVH. Brecht suggest making intersection object/prim the same in
ShaderData, since this is mostly a legacy optimization from BVH2.
- Feng investigated multi-GPU scaling more, and found that the scaling bottleneck is mostly in code outside path tracing. Instead of 10 GPUs giving a 10x speedup as before, it’s now around 5x. This is being investigated further.
This is a weekly video chat meeting for planning and discussion of Blender rendering development. Any contributor (developer, UI/UX designer, writer, …) working on rendering in Blender is welcome to join and add proposed items to the agenda.
For users and other interested parties, we ask to read the meeting notes instead so that the meeting can remain focused.