I am using 2019.2 and i am debugging what causes fps drops and i realized that sometimes i disable an object and get near 90fps and then play again and disabling it gives me around 50. I tried deleting every object but 2 and i got 50 fps, then disable an object and get 90fps then enable it again and get 90fps anyways, dont understand anything
Check the profiler in the different FPS rates. What does it show?
I get Semaphore.waitforsignal using most cpu. Now i dont get 90 fps again, i get 70 with an empty screen and i use a modern gaming computer
Can you place a screenshot of the profiler in both cases?
Those screen is just having a camera in scene
Looks like it’s waiting for the GPU. never worked with XR devices before though.
you should check frame rates and profiler data when building and running from the actual device. The editor is unreliable for performance measurements.
1 Like
In 2018.3 Unity editor showed me near 90fps, with 2019.2 is horrible fps
Again - ignore the editor. Check on the device.
The FPS measurement in the editor is half noise.
1 Like
But i got with 2018.3 good fps in editor. It shows the editor loop and the player loop, so player loop should be what is going to happen in build. In build you always get higher fps cause dont have the editor loop which is using about 30% of cpu, maybe get 90fps in build but with 2018.3 probably would get more
Well, you can keep guessing or you can actually check.
Unless you plan your users to play the game from the editor.
1 Like
Even if Unity 2019 didnt give a worse performance i couldnt optimize properly the game if i need to make a build which takes an hour every time i want to check the performance
The code running in the editor in play mode is intentionally not performance optimized and runs in the same thread as the rest of the editor. Editor performance should only be used for spotting really obvious problems, or to give you a rough idea how a change might actually affect performance in a build. The editor profiler should not be used for investigating performance issues any further than that.
Comparing performance of your game between different editor versions is pretty meaningless, unless your goal is to report an editor performance regression to Unity.
Aren’t you regularly making builds just as part of the typical development process already?
Profiler shows a editor loop and a player loop so you can check how much CPU uses the process not related to editor, so yes, dont need to make builds to check game performance as already shows it in player loop. In build will always have more performance cause have not the editor loop but it doesnt mean that you need to make regular builds, that is a bad practice that waste lot of time unless you have a small project
Sounds like you’ve got this all figured out then.
Read the starting post, that is what i want to figure out