I need to create a group of a bunch of different meshes ranging from 4 vertex planes to 20ish vertex shapes.
Is there any advantage to making these all submeshes of one mesh versus many different GameObjects? Both options provide the functionality I need but with submeshes it seems as though there may be a bit more of a hassle.
Well submeshes will take less memory and processing overhead. GameObjects aren’t free from a performance/memory perspective.
beck
4
So far I’ve found that submeshes actually reserve extra memory in the vertex buffer. Each sub mesh appears to reserve the full vertex count of the mesh, meaning if you have a 300 vertex mesh with 4 submeshes, you would be reserving a buffer for 1200 vertices. This differs from separate game objects, where each mesh would only reserve the vertices it needs. Because of this, I never use sub meshes in my game. The overhead from separate game objects is negligible compared to this memory footprint.
The only real advantage of using submeshes is that you have a single object. If you have a character, for example, where the skin, eyes, hair, and clothes are different submeshes, you only have to move the character around as a single object rather than 4 separate objects. Otherwise performance and so on is pretty much the same compared to having separate objects.
