If the rendering is this controlled (and sparse), blocking the thread with it isn't really a concern. And with this data size, serialization/deserialization bet

If the rendering is this controlled (and sparse), blocking the thread with it isn't really a concern. And with this data size, serialization/deserialization between the main and worker thread would probably become nontrivial.