koobas.hobune.stream
Question on Static Primative vs Mesh Collider Performance
Hello everyone, I was looking at this Unity Ask post and that made me think about another related question: http://answers.unity3d.com/questions/576565/is-it-efficient-to-split-up-very-large-static-mesh.html
Say that I had a primitive sphere collider rolling across a static primitive box collider. Then say the sphere was rolling across a static mesh collider in the shape of a box with say 1,000 tris on each face. I understand that there would more bounds checking against the higher poly static mesh collider, but would that difference in time even be noticable?
Essentially, how much work should be done in optimizing static mesh colliders? Are there any noticeable performance increases that can be gained by reducing a static mesh collider by a certain % or amount? Any insight is appreciated.
I can't really give solid information about how much more performant primitives vs complex mesh geometry are. It's the kind of thing you'd be better encountering during fleshing out the design of your game. So if you are using complex mesh colliders then notice they're causing a performance hit, you'd swap them out for primitives where possible.
The more salient thing with mesh colliders vs primitive colliders is accuracy of physics simulation and collision detection. If your mesh is concave, there's a chance mesh collisions will not work as desired. If this occurs then you want to use primitive colliders to approximate your geometry.
