...if the server runs in any given frame then the renderer doesn't run in that frame.
Exceptions are: if things are running slow ("slow" is defined as less than 72 FPS), or if you're at the console or in the menus.
The effect is to smooth out performance and gameplay a LOT. In any given frame there are two things that take up most of the time: the server and the renderer. Everything else is small fish by comparison. So rather than having two big hitches in every frame in which both run ("big" being relative, of course) now we only have one.
Obviously only relevant in SP games, of course.
What's doubious is: should frames in which the renderer doesn't run be added to the FPS counter? Part of me says "yes" because there's still a lot of significant things going on in these frames which should be reflected in the counter. Part of me says "no" because DirectQ's FPS counter will then be at variance with tools such as FRAPS.
Another part of me is wondering if this should even be done at all...
Thursday, October 6, 2011
Possibly Dubious Performance Boost
Subscribe to:
Post Comments (Atom)
3 comments:
I'dd say no, unless you'dd make 2 seperate counters.
I'm also voting for 2 counters, one counter counting two different things doesn't really convey what is going on.
Also, I'm sure 99% of gamers consider 'frames' the stuff that gets displayed on their monitor.
When being a developer you might have a different view on the matter, but adding them up would just serve confusion to the average end user.
Post a Comment