Worse performance on latest Unity LTS, how to fix?

Hi!

I was using Unity 2020.1.16f for my project, then I updated my version to the currently latest 2020.3.24f. I'm using HDRP.

My problem is I'm getting worse performance than before.
In the editor -40 FPS, (100 FPS instead of 140)
in builds: -100 FPS (300 FPS instead of 400 FPS)

I mean 300 FPS is still fine, but it is just a basic scene, what would happen in later stages of development, with more populated and dense scenes?

What could cause this issue?
Any idea how can I fix this?

Thank you.

 

My proposition is: stop staring at the statistics window and work on your game. When you will have performance problems you can measure what happens and you can fix it. Until then you should focus on your game.

 

If what you have is an empty scene, then it's too early to worry for performance, wait until your game enters alpha / beta stages and then double down on performance.

Because in the meantime, say there's something wrong with the specific version you are using and you spend days / weeks figuring it out... Is that the version you will be releasing your game with? Or will you keep updating Unity? Because if you plan to update even to minor (same year LTS) versions, you will have spend a lot of time finding an issue in a version of Unity you will not be releasing your game with.

 

Make sense to me, thanks.

 

Not an empty scene, but a playable prototype scene, just cubes for environment instead of the real models.
I needed some fixes and new features, that's why I updated, but not sure if I'm releasing my game with this version.
Probably not, so you guys are right, I can worry about performance later.

 

If I remember correctly, you are also getting more accurate measurements in the newer builds, so your project might not have lost as many FPS as you thought, you're just getting a "more correct" number now.

 

And that's why bloated unoptimized software continues to flood the market.

 

Eh, at the prototype phase is definitely too soon to be optimizing. And doing things in kind of an optimized manner from the start requires experience and to get it you must have an actual game in your hands that needs optimizing (and not just a prototype).

I think not optimizing early on is good, just be prepared to re-write big chunks of stuff in late development.

(and the reason unoptimized software is flooding the market is that people use crappy engines)

 

You should also be able to justify changing engine versions after the project has begun. If there isn't a compelling reason to do so, it's better not to. It causes many problems.

I don't see how that follows.

People avoid premature optimization because they understand many things in the project will change so its wasted effort to try and organize things before you even know what you'll end up with. How does this effect unitys incentive for quality control?

 

ROFL. No, not that's why. Because people don't care about optimization that much, that's why, but it doesn't mean you need to concentrate on it day and night. Besides it's extra funny throwing it on my post. I think I am one of the biggest proponent of performance optimizations even at the begin with on these forums. But if you don't have a game, you don't have anything to performance optimize.

 

I have a prototype project that became so slow in the editor that it was no longer fun to work on, so I was forced to look at optimisation before I would even want to continue.

 

I believe that’s just called “running Unity with Collaborate”.

 

Doing disruptive optimizations early on and not being mindful of performance during development are different things. First of all in this case cleary it's an engine issue and if you wait to report it until late in development you're gonna have a bad time, specially with how long Unity takes to solve bugs.

Secondly, waiting too late to think about optimization is how you end with either a mess (a la Cyberpunk) or having to downgrade (a la Ubisoft).