Each frame I construct a global scene data buffer. It holds all data needed to render my scene. Its done in main thread.
Now, if I would like to do multithreaded rendering, is it possible to set this buffer to be shared between all deferred device contexts? Like some kind of global state?
I know that I can just set it in each deferred device context, but Im afraid it will be copied to GPU many times then. Am I right about that?
I thought I can just set it once in immediate device context in main thread, but I am not sure.
Aucun commentaire:
Enregistrer un commentaire