why are mesh colliders a lot slower to initiate than mesh itself? They are the same shape as the mesh, but I think it takes 5 times longer to rewrite the collider of the mesh with the built-in function, then changing the mesh itself. adding one of them is like adding 5 times the amount of mesh for the frame rate.
Is it inefficient? What parameters take so much time, like normals and tangents etc?
Are there ways to speed up collider calculation? Could it be possible to pass some of it to different cores? divide up into smaller pieces? edit and make it better suited to particular tasks and timings by generating the mesh collider in code? if the mesh collider takes a lot longer than the mesh, and has to be a lot smaller to go as fast in frame. is the terrain mesh collider faster some?
would love to be enlightened! I am surely being naive about something.