I was making a quick vid to ask for help regarding sound events since they seem kinda expensive on performance…
But then at the same time i manage to get on video the most random and inconsistant performance in my project… any ideas about any of this at all?
Do you have performance issues in preview mode while the GDevelop scene editor is open? That’s been an issue for me in the past, particularly if it was an external layout.
I’m not sure how much impact the checking if a number of objects are on screen has much of an impact, but you’re doing the on screen check twice - first time it’s inverted, the second time it’s normal. I’d consider the following:
It dosent matter, none of this matters.
I just finished doing all the performance adjustments on my events, its all nice and tide up, nothing running when it shouldnt, all the checks in place, i even went as far as removing all the “isOnScreen” events and removed the extension.
The events themself are much better in performance, but it seems the more performance i get out of them, the more shows up elsewhere for no reason what so ever…
i really dont get this, iv done so much testing and changes, iv fixed everything up and nothing works…
When it feels like it, it will run perfectly smooth, then out of no where it goes bad, then back to good, and so on…
Every video iv made showcases this issue, some times it great with as low as 7.70ms, then with the same everything, quite leterally goes to 15.00ms just because it felt like it.
My computer performance is pretty much the same, so it really is an engine thing, more so since theres more people that complain about it…
I give up, iv done all i could to make it work, i squeezed all i could but the issue is elsewhere…
I even went as far as deleting all my events just for science and then ran the profiler, performance was still crap… so it really aint anything to do with that.
I did learn a lot about performance tho and that most people dont have a clue about what really taxes performance, the engine itself can hangle a whole bunch, it just as long as its working right.
It doesn’t? But in your other post you conclude that the behaviour affects performance, and here it’s used twice, in succession, every game frame.
The other thing is that to properly test it, you need to remove all other irrelevant events. Just a have a bare bones scene with the lava objects and an object that is moved which the camera focuses on, and then run the performance testing. As it stands, you have a lot of other things going on in the scene that could somehow be impacting on the bit you are testing.
Yeah, and that’s frustrating. What I do when that happens is use the performance profiler to identify where the surge in processor usage occurs.
For those reading this and unsure what I mean by that, that’s where those event groups come in - the profiler collates performance on the event groups. From this you get the general area where performance takes a hit, and then add event groups in that to narrow it down. to the specific event.
