With the new Volumetric Clouds right around the corner for 4.26 (I seriously can’t wait to use it right away - my current solution have severe shader complexity issues and have no artistic control tbh) and looking more promising than I initially thought, how scalable it is to lower-end less powerful hardwares? Would be 30 FPS with the clouds using hardware equivalent to Nintendo Switch viable, regardless of resolution compromises?
Have you tried enabling the built-in Volumetrics plugin in 4.25? https://youtube.com/watch?v=Rhr2B81zS4M
Should be functionally close enough to what’s shipping with 4.26 to know. I’m impressed by the performance of it overall, but I doubt it will be viable for Switch.
I already tried that one from the Volumetrics plugin, but it’s so sluggish on GTX 750 Ti (which is closer to a base PS4). I only get around 10 FPS on 720p High, compared to my (crude) solution which stays on 60 FPS top. I would assume the sluggish performance from the built in one have yet to see meaningful optimisation.
Unless it’s only meant for RTX 2080 Ti (PS5 equivalent) or above, but then PS4 games already achieving volumetric clouds with stable performance since 2017.
Dont make performance assumptions about the 4.26 clouds based on the earlier volumetric plugin, they are not the same system and I believe that making it performant enough to be practical in more situations is big part of the work for 4.26. There are also quite a lot of cvar settings for the 4.26 clouds that could have an impact on performance. Also there are a lot of different possibilities in terms of visual results and performance that depend on the cloud material your use.
I tried the work in progress via github a while ago and although it was too early to be reaching conclusions about performance, it didnt feel sluggish and heavy in the way the experimental plugin stuff does.