13-05-2020, 02:01 PM
Hi Emma,
Make sure you're not limited by vsync, target frame rate, or other artificial frame-capping in your production machine. The fact that you get similar performance on linux/windows in the slightly faster machine is indicative that differences in the linux/windows binaries are not the issue. Must be something specific to your production machine.
Could be, but I agree that it is weird. Check how many times per frame is FixedUpdate() being called, that's the best way to tell for sure if death spiraling is the culprit. 1 call, or maybe 2 some frames, is ok. More than 2 calls most frames are a red flag.
My best guess is some sort of artificial framerate-capping is taking place, though.
Quote:With real-world tests, on our production machine we get 40-50 fps (same particle count and parameters as before, debian).
But on windows with a slightly faster machine we get 350+ fps (on the same machine but ubuntu around 300).
Make sure you're not limited by vsync, target frame rate, or other artificial frame-capping in your production machine. The fact that you get similar performance on linux/windows in the slightly faster machine is indicative that differences in the linux/windows binaries are not the issue. Must be something specific to your production machine.
Quote:So it might just be that the production machine is just slow enough to reach the death spiraling threshold... but seems weird that it "stops" at 50,
Could be, but I agree that it is weird. Check how many times per frame is FixedUpdate() being called, that's the best way to tell for sure if death spiraling is the culprit. 1 call, or maybe 2 some frames, is ok. More than 2 calls most frames are a red flag.
My best guess is some sort of artificial framerate-capping is taking place, though.