In windows I got 1800 points
You must log in or register to comment.
For a CPU benchmark like this, something is definitely weird because wine shouldn’t be translating anything. I wonder if the benchmark might be doing weird things with the Windows API.
That sounds like you’re hitting an edge case and it might not be representative of the actual performance you can expect out of Wine.
Have you tried other benchmarks?
No
I’m thinking of testing Cinebench
Are there any other tests I could check?
Find something which has native linux version as well, e.g. passmark performance test.
Run on win, linux and win version in wine. Comparing the 3 results you can figure out if wine is the problem, or some settings in linux if wine and linux results are similar
I check Cinebench and Passmark
Cinebench
Wine - 4112 points
Windows - 3874 points
Passmark
Wine - 7053 points
Windows - 6864 points
Linux - 7145 points
It looks like the problem is in cpu-z and not in wine or linux
Why is windows slacking so much there? Or is lower better on these ones?
It’s not unusual to see better performance on Wine compared to native Windows. Wine is a compatibility layer, not an emulator. So there’s not a lot of overhead. Additionally, vanilla Windows has a lot of background bloat consuming resources.
Internet searches show many instances of people reporting higher FPS in games on Wine vs vanilla Windows (on the same machine).
That’s actually excellent. I knew that ‘Wine Is Not an Emulator’ from their web page but I didn’t know I could expect better performance in anything running with it.
That’s a test I’d like to see the results of
Yeah this is weird. CPU instructions aren’t being translated by Wine/Proton, as I understand it. It’s not emulation.
I suspect there’s something in the reporting which is actually giving you false figures.
Do you get better numbers with Windows on the same PC?
Yes, 500 points more
I’m curious if this is caused by the fact that wine/proton is an API emulation layer. Whatever API calls this app uses for benchmarking may be less efficient, or maybe emulated, rather than talking directly to hardware. It should also be noted that these benchmark numbers are probably not indicative of actual game performance, as games likely use different API calls that are well or better optimized.
