2025-04-29 Render & Cycles Meeting

Attendees

  • Brecht Van Lommel (Blender)
  • Christophe Hery (Meta)
  • Leonhard Rannabauer (Intel)
  • Michael Jones (Apple)
  • Sahar A. Kashi (AMD)
  • Sergey Sharybin (Blender)
  • Thomas Dinges (Blender)
  • Weizhen Huang (Blender)

Notes

  • Blender 4.5
  • Brecht is working on texture cache
    • Bistro goes from 5 to 2 GB :slight_smile:
    • Update related tasks. Once ready to collaborate make a PR or move to the main repository.
  • Weizhen is working on volume sampling
    • Should be ready for review soon.
  • HIP SDK is public now. Need to update HIP-RT and OIDN.
    • After that we can look into the balanced BVH option to lower the memory usage.
  • HIP-RT supports binary compression. Might take a bit to become available in the public HIP-RT repository.
  • Look into moving the meeting to allow Christophe Hery join more meetings.
    • Will send mail with suggestions.
  • Discussion went into camera node modified bits. Should be possible to reduce the number of bits to allow having more parameters in the camera node.
    • Christophe will also look into adopting OSL camera.

Practical Info

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.

13 Likes

As a professionnal lighting artist, I just wanted to mention that I am very excited by the list of Light improvements and by the addition of a texture cache and mipmapping system. The latest will finally resolve some of the scalability issues cycles had with gpu rendering and big scenes.
I guess the goal is to have the cache system totally hardware agnostic ? Meaning it would work out of the box for Apple silicon, Nvidia’s, Amd’s and Intel’s gpus ?

13 Likes

Indeed that is the goal.
Thing to keep in mind is that we might get there in steps, its a big project so some thing might come to CPU first. But it will be more like CPU vs GPU. The code on GPU is quite unified, so it should work on all GPUs once we implement it.

16 Likes