Never had access to easy-mode GUI before, so now I"m abusing the hell out of it
Never had access to easy-mode GUI before, so now I"m abusing the hell out of it
LImGui also drops framerates by a fair amount :^)
Land BreakneckSleep doesnt actually account for extra time spent in the loop from what im reading
Lso if anything takes any time in your loop, it will run too slow
KAlright now that I have a sensibly slow updating FPS counter lets see...
With all the fixes to breaknecksleep:
500 requested - Getting 335 - 358 FPS/UPS
Without the fixes:
500 requested - Getting overall 335 with changes differing from fractions to +- 2fps/ups
With all the fixes to breaknecksleep:
500 requested - Getting 335 - 358 FPS/UPS
Without the fixes:
500 requested - Getting overall 335 with changes differing from fractions to +- 2fps/ups
KSo it seems using the original stuff is less "bouncy" in terms of stably staying at a certain FPS/UPS rate, but still lower than the requested number
Lits likely because of this
KAh yeah... I am doing lots of fucky things with string.format and using a queue for averaging of FPS
Lits not a queue problem, its a "you are doing anything at all" problem, and perksey's BreakneckSleep doesnt account for extra time already spent
Kyeah it needs a bit of tweaking there
Lhttps://github.com/ppy/osu-framework/blob/master/osu.Framework/Timing/ThrottledFrameClock.cs
osu!framework has code to handle this better
osu!framework has code to handle this better
Misn't osu xna
Lthat hasnt been the case since like
L2015?
Moh
Lalso osu!framework isnt based on the stable codebase
Mmonogame? fna?
Lits a full rewrite
Lno
Msharpdx?
Lno
Mslimdx?
Lno
Mopentk?
KAlso I feel like I'm forgetting to clean something up, cause i"m getting protected memory exceptions whenever I close the window
La fork of it, yes
Alwjgl? 
Lcalled osuTK
Mwhy peppy need to fork opentk
L¯\_(ツ)_/¯
Mthis is why lazer never gonna be done
M/j
Losu!stable also uses a fork of OpenTK
Linfact thats where the fork comes from
Lthe fork originates from stable
L
Moh
KHuh, interestingly the lower you set the loop cap, the less jittery it becomes
Mfair cop
Lyes that makes sense, given "FPS" is an inverse relationship with actual time
Mhigher framerate increases the precision requirement
K600 = 300 - 400 more or less
60 = Stays pretty much glued to 55 most of the time
60 = Stays pretty much glued to 55 most of the time
Lits actually the same amount of error, its just that fps is a really bad measure in this case
Lyou should be measuring frametimes
Mis your frame counter accurate
KI am measuring frametimes, and getting the FPS from that (1f/Frametime)
Muhh
Mhm
Lwhat i mean is in this case, dont convert to FPS
Cyou should use frametime for your measurements rather than fps
L^