Edit: The video of that’s the original source of the test results had flawed testing for some of the games and seemingly didn’t set the same tdp limits when switching between the OSs. I’m guessing we’ll see other people test this out later, but for the time being it’s best to assume these results are exaggerated or wrong.
You must log in or register to comment.
This is purely from a broken test. Watch the video and you’ll see that its not testing with the same power limits.
17w tests are actually 16w vs 20w+ and 35w test is 25w vs 35w
This leads to drastically higher clock speeds on both cpu and gpu as seen in the video and thus higher fps (and power draw, so lower battery life)
Please help stop the spread of misinformation
Good catch, results didn’t seem suspicious because Linux has had major performance improvements over windows on previous handhelds.
https://youtu.be/OwWRCrGoXV0?t=284 Digital Foundry tested few games and found no significative differences. But it’s still a big win for linux gaming since those game aren’t native.
I’m also suspicious of those tests since they don’t show details like power usage. Same problem Tom’s did here, took the data at face value missing the clear signs in the original video showing how it was tested improperly.
There are many tweaks and bugs like 17w being 20w+ with this version of the os.
Here’s another video on it that goes more in depth: https://youtu.be/CJXp3UYj50Q
Basically the SteamOS version gets better fps in most games while also have way better battery life. Lower power games would get 2x the battery life when fps when both versions were targeting the same settings/fps.
Basically windows naturally has more overhead than Linux, which really impacts performance/battery life in low power games. But vulkan also typically outperforms directX, so many directX see a noticable performance increase on SteamOS as well (although some amount of this performance gain can be had on windows as well by using dxvk).
I’ll take a look later but quick check he is using two different devices which means its not identical hardware/ firmware.
Also there are many many many more cases where performance has been the same
And countless others who see no difference between windows and Linux.
Most times if there is is due to dxvk / Proton fixing a game bug/bottleneck, or invalid testing with not identical settings/hardware.
There is no way windows is using more power and giving lower clock rates without some other bug effecting it.
Edit: finished the video. He didn’t show any actual testing its basically as in depth as the Tom’s article.
When testing Kingdom Come: Deliverance 2, Cyber noticed a shockingly significant jump in FPS, with Linux generating ~32% more FPS compared to Windows. This trend follows at lower wattages, albeit with less noticeable differences, and the delta actually plateaus in Hogwarts Legacy to the point that both Bazzite and the Xbox FSE offer the same FPS at 13W. That being said, those frame rates are much more consistent on Linux, according to Cyber, who shows that the FPS graph on Windows fluctuates regularly, while staying mostly flat on Bazzite.
The irony of it considering Microsoft’s tweaks to make it a “gaming mode” windows (which results in miniscule FPS gains).
I guess in that case it’s not irony but rather a shoddy job on Microsoft’s part as per.
I mean, its ironic that one of the developers of the graphics pipeline that most consoles use (DirectX) and one of three major consoles cant make their own OS perform better than Valve and a bunch of developers working in a cave on Linux, lol
I get that. Maybe the irony is lost on me because I have yet to see Microsoft make an effort towards something that an end user would care about. Windows playing Windows games worse than another operating system is entirely expectable.
Won’t help with titles like Destiny 2 which require Windows.
I do not know if there is a game I would want to play enough to go back to windows.
