Shadow of Mordor Performance: Windows 10 vs. Ubuntu Linux

[AndyChow]

Re: Windows 10 keyloggers. This report shows how to disable it: https://www.reddit.com/r/Windows10/c...er_windows_10/ OK, the publicly declared keylogger can be disabled and Cortana removed. You still have an OS from a vendor who has ever installed a keylogger and whose privacy policy explicitly permits it. Thus, for something like this game test I would recommend not logging into any of your normal accounts and not opening any encrypted media, PGP encrypted messages, etc. That way all the information MS gets from open or covert spyware is junk and the test can proceed. If I had to buy a computer that came with Windows 10 I would not power it up until I was either someplace with no wireless access or had temporarily removed all wireless devices from it. That way Windows could not start installing and phone home with device info. I would remove the disk before screwing with the UEFI to get Linux to boot on another disk, so I could take it back unchanged/unactivated if it refused to "disable" secure boot or refused to get to the UEFI screen with Windows never having run.

And don't forget your tin-foil hat! The keylogger is only on the technical preview, and your ideas have been debunked since October of last year at least. Nobody cares about your PGP encrypted love letters to Kim Jong-un.

[johnc]

oof... that's brutal. but thanks for running these benches, Michael.

[iniudan]

And don't forget your tin-foil hat! The keylogger is only on the technical preview, and your ideas have been debunked since October of last year at least. Nobody cares about your PGP encrypted love letters to Kim Jong-un.

Not that the tin-foil crowd isn't going insane, but the keylogger is still there, but it is only used when search stuff with Cortona, I would guess for machine learning purpose.

[quaz0r]

And don't forget your tin-foil hat! The keylogger is only on the technical preview, and your ideas have been debunked since October of last year at least. Nobody cares about your PGP encrypted love letters to Kim Jong-un.

notice how these sorts of reactions are always angry and belligerent ones. whether it is skyscrapers pulverizing themselves to dust at the acceleration of gravity or microsoft encroaching on their users' privacy, you never see a response like this stated in a calm and intelligent manner. if what they were reacting to was really as silly and baseless as believing in the tooth fairy or santa claus, why then are these reactions always so angry and belligerent? the obvious answer: because it is not baseless and the implications are upsetting and possibly even threaten our world view.

[philips]

There is indeed a bottleneck somewhere, the results for Linux are much too low in your benchmark.
With a lesser CPU and GPU I get better result even when I compare my GTX 780 with GTX TITAN:

[justmy2cents]

something is REALLY wrong in this benchmark. i have i7 810 (fucking ancient i7), 8GB, 750Ti and get

- on Lowest 49.48 / 77.17 / 30.75
- on Low 43.18 / 70.04 / 24.50
- on Medium 31.27 / 60.05 / 16.19
- on High 32.98 / 45.39 / 13.94
- on Very high 28.73 / 44.16 / 16.37
- on Ultra 16.60 / 30.81 / 3.46 (2 cores more or less bottleneck here most of the time, my i7 simply cannot handle it)

and that was while running gnome-system-monitor over ssh to check cpu which was 2 cores full time at 100% making this totally cpu bound. something that should be more than expected on my machine

and considering benchmark stresses more than game if i enable fps counter i get 32-40 99% of the time, with only rare dips to 27 (even making shitstorm at Durthang with alarm where there are tons of enemies or doing Elven leap between same ton of enemies) which is better than on my friends ps4

so Michael... your machine is like space rocket compared to mine, yet results are almost 50% worse? 750 and 750Ti is only about 10% or less difference

[philips]

Michael, you really need to find the bottleneck and REDO the benchmark.
The penguins are counting on you to get fair results.

[ruthan]

Thanks for true, that is state of Linux game, some games working, but how.

Why is this game so slow? Linux fanboys would see problem on Nvidia side, but problem it could be whole linux graphics pipeline and architecture to blame and of course Khronos and their absolute OpenGL.

OpenGL is actually good example of how - nonprofit organizations and open standards for big piece of SW sucks.. there isnt any OpenGL company manager, programmer who would be paid based on OpenGL list of features and performance in comparison to Direct3D - that is wrong.

Good example is now PS4 (OpenGL) vs. Xbone (D3D), even if PS4 has +33% GPU raw power there quite few games which are running on Xbone better.

[justmy2cents]

Author of the article compared them, I just comment on a results he got .

And yes if AMD needs to fix their drivers, then nVidia needs to fix their drivers too... but when both have performance issue, it is clear to me something goes more wrong on a game port side and not with the drivers - as you see game does not even have scalable settings...

Of course i also never exclude gremlins as the possible reason, like drivers might missbehave under specific enviroment or on particular DE, etc...

no, you're commenting on two different implementations. they are not same code path you can just compare. one implements directx, while other one tries to do the same thing with opengl. and anyone with a little sense knows that code optimized for one engine would require damn lot of work for other to be equal. port to 100% quality would be costly as hell and probably quite a bit unfair to blame Feral on this. it is a late port where 2 markets are smaller and probably a lot of people already bought windows version where Feral gets no money.

beside that, Michaels benchmark is so broken it is not true. just looking at numbers it says something is wrong.

if any of those results would be valid 70% of people couldn't play it. but, that is what you get when benchmark is done by someone who never plays the game.

[justmy2cents]

There is indeed a bottleneck somewhere, the results for Linux are much too low in your benchmark.
With a lesser CPU and GPU I get better result even when I compare my GTX 780 with GTX TITAN:

bottleneck in game is also obvious when you see better cards having no gain. but, bottleneck only starts to be visible after game is playable. so, there is maybe something Feral can fix there

if Michaels benchmarks would be even close to correct, i would play at 12fps average or less on Very high settings