You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Possible (modest) perf improvement to Cannon wrapper by not syncing positions of static objects to the physics engine (cos they are static, so should never move).
Requires introducing a distinction between kinematic & static objects (as per the Ammo driver), which doesn't currently exist. Kinematic objects will require syncing to physics every tick. That will be the only difference.
The relevant part of #15 quoted here for convenience:
Cannon doesn't distinguish between kinematic & static objects. So you could say that all static objects in Cannon are kinematic - hence we should update them every frame.
We could introduce a "kinematic" concept at the wrapper level in Cannon, where the sole difference between static & kinematic would be that kinematic objects would have their position sync'd to the physics engine every frame, whereas static objects wouldn't.
I'd suggest a new issue to track that potential improvement. That said, I'm not sure there's that much value as the before/after phases are already much cheaper relative to the engine with Cannon than they are with Ammo.
The text was updated successfully, but these errors were encountered:
Offshoot of #15.
Possible (modest) perf improvement to Cannon wrapper by not syncing positions of static objects to the physics engine (cos they are static, so should never move).
Requires introducing a distinction between kinematic & static objects (as per the Ammo driver), which doesn't currently exist. Kinematic objects will require syncing to physics every tick. That will be the only difference.
The relevant part of #15 quoted here for convenience:
The text was updated successfully, but these errors were encountered: