I’m currently in a testing/optimisation phase for my horror contest game. Between the old and new engines I’ve noticed some significant difference in performance. These screenshots were taken under exactly the same running conditions.
From 2.25:
From 2.34:
Note the physics percentage in 2.34 :o
Hmm, I thought the physics engine for 2.34 was different than that of 2.25’s. That’s maybe why there’s a bigger difference in the physics.
Lol, it wouldn’t matter for you both ways because it seems that you’re getting around 85 fps for both 2.25 and 2.34. As long as the fps stays about the same there’s no complaints.
Although this may mean the physics in 2.34 needs some optimizing.
Jason Lin
85 is my maximum since I have VSync enabled (capped at monitor refresh rate). Problem is my system is quite high end, so it can cope quite well with inefficiencies. I need to make sure it doesn’t crawl on slower systems.
I still find problems with physics in 2.34, including the infamous ‘falling through the floor’ bug. Whilst it happens far less than in 2.33, it does still happen.
Andy, can you please explan the Vsync option and your settings for it? Is it reliable to set and view on other machines?
VSync is basically a function which will lock your maximum fps to the monitors refresh rate. It’s nothing to do with blender. You enable/disable it on your graphics card settings.
Turning it off can give higher fps, but if the monitor can’t refresh that fast there isn’t really any point. On slower systems turning it off can give a performance boost, but it introduces some visual problems.
Here’s an article on it.
I just tested this scene with VSync off and I get ~210 fps for both versions of the game engine. So there really isn’t much difference in performance, even though the profile would have you believe that :-?