When using the Allow CPU Access option on meshes, is there any sort of a performance penalty?
Hey there @Meowcat285! Welcome to the community! This answer to this question is often use case dependent. In theory the access time should be comparable but likely a small performance hit. I’ve never had to stress test it, as generally only meshes that I intend to mess with at runtime are limited.
Ah, I see, out of curiosity, what exactly does Allow CPU Access do on a low level?
If I remember correctly, it just basically retains the data in CPU memory instead of wiping it out once it’s sent to the GPU and continually updates it when changes are made. So the extra performance hit comes from having to pass the updates instead of just loading it into GPU memory and being done. I’d have to take a glance at the source again to verify that, as I hadn’t worked with it since 2016/2017.
If your use case needs a ton of mesh editing at runtime it would probably be best to go ahead and stress test it as hard as you can early.
2 Likes
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.