We are working on porting most of our main shaders to WebGPU in order to have WebGPU as a first class citizen to be future proof.
While doing that we are trying to optimize the size of our package by delay loading shaders (you do not want to load webgpu shaders in webgl and vice and versa).
This should be almost transparent as everything shader related is already asynchronous by nature.
THAT BEING SAID, we will obviously find problems along the way and I would like to ask for everyone patience and understanding.
This is the first step in a long series (more to come later with Render Graph integration :)).
I love this community like a family and I know these changes are painful. But they are necessary and we are working super hard to make them as subtle as possible from the user standpoint.
It makes sense that babylon.js won’t always be able to be 100% backwards-compatible while still making advancements. It would be a pity to prevent improvements for ever, trying to always keep full compatibility. So I’m looking forward to seeing what improvements / optimizations come out of these architecture modernisations!
Yes. I can agree with this. At some point, there’s no other choice. You just have to move on (or you are out of the game ). Just mind to keep with the pillars, vision and DNA…do what you have to do, and it’s gonna be just fine
I found this repo: immersive-web/WebXR-WebGPU-Binding (github.com) which had a commit 5 days ago, so it seems up to date at first glance. I really hope it’s something that can be used to improve WebXR performance.