Hi, am I the only one experiencing performance variability in games from install to install?
I periodically install other distros to see if the green is grassier on the other side and I decided to do a cleanup, migrate to lvm for my arch install while I’m at it.
I have gone through 5-6 installs in the last 3 weeks, several iterations of arch as well. I always play the same 3-4 games and I know how they should perform already and I am experiencing some weird performance variability, what do I mean by this?
- Some installs have lower fps but stable 1% lows
- Sometimes if feels stuttery, even though the frametimes are stable
- Other times its just perfect
I am not doing anything out of the ordinary, archinstall with my setup, install steam, start my games.
- All AMD system
- single NVME
- Plasma with wayland
I’m stumped, I feel like an idiot looking at this issue because it makes no sense whatsoever. I finally nailed the setup after figuring out how to get LVM running, got everything going and now, it’s a stuttery mess.
Edit: Things eliminated or fixed:
- stutters removed by lowering mouse polling rate to 1000hz
- wayland as a possible overhead by launching my games from the tty with gamescope
Random uneducated guess: Could it be some cache/shaders stuff? Depending on the game, it could take a while until the cache is warm and everything runs smoothly.
Well, one thing was the high polling rate, I just figured that one out, reduced to 1000hz, so the stutters are gone.
There still a performance delta of about 30% vs windows for some of my games.
Its annoying since I did reach performance parity during one of my previous installs, but I never figured out what it was that I did.
There still a performance delta of about 30% vs windows for some of my games.
I know Proton is often touted as the magic beans that evens out performance between windows and linux systems. But most games will likely still perform poorly on Proton compared to WIndows native.
So the question is. Does this performance gap matter in the games you play? If I can still play and have fun playing the game, then it would not matter to me. If you play competetive, where that performance does matter, then stick to the system that has better performance. Even if it might be Windows for some games.
To that end, I decided to test out this exact premise, its been years since I had a windows installed natively.
I was compqring the numbers with those from memory. Well… My memory has mislead me. The performance gap is 2x for the lows in my favorite game.
So as you said, the windows partition stays for some of the games. I dont care what it is, Ill use the right system for the right usecase.
I’m still digging through the issues though, I remember clearly I reached parity once, I just dont remember how or what was it that I did.