When developing a game that supports user-built assets, the standard approach seems to use dynamic mesh components instead of static mesh components (which cannot be created during runtime).
It’s not clear what the tradeoffs in pure rendering are. If we do not want to update the geometry during gameplay, only in some build mode, and we do not want nanite, lumen, or LOD support, what are the rendering performance differences between static and dynamic meshes?
It seems there is some packing and updating differences but it’s not clear if a dynamic mesh that is not updated (the geometry data is static but the object may itself be transformed) is a premium feature that should be used sparingly or if there is little rendering overhead and whole levels can be made of them?