I didn’t find a lot of trick to improve performance in my system.
Of course, there are.
dtr helps/helped me a lot around pushing distance to cam calculations to GPU with jit.gen. I’m still not totally okay.. but it will be in a close future I guess.
Actually, I’m inactivating OB3D objects when distance is too far.
It is nice, but it also means ALL distances are calculate at each time.
When objects are inactivated because too far, of course, it improves things a lot.
I thought about inactivating visual display of objects "behind" the cam (only visually)
But it would involve more calculations involving camera focal length, frustum etc..
Would you have some tricks for me here …?
or maybe, if I do correctly my distances matrices calculation, it would be enough …?