Currently it’s working like this:
form a change list → for each CPU core: (fire up a compute shader then fire up 1 job that makes the mesh) → back to main thread to update the mesh of the chunk
It doesn’t seem that efficient. What’s a better way to do this?
Use the MeshData API. Takes some getting used to but allows you to update multiple meshes in parallel and then apply the changes.
1 Like
Cool thanks.