Valve have released the Steam Hardware Survey results for October, so it is also time to release our own survey results. 857 people completed the survey, making our sample size 187 people larger than last time, so thanks everyone!
The changes in the questions seemed to have worked well and the data should hopefully be far more accurate this time. There are still a few minor things to iron out, but the questions will mostly remain the same so we can see trends developing a few months down the line.
I will not compare results with those from the first “experimental” survey since many of the questions have changed now that the survey has taken a more permanent form - mostly the fact that the first survey was not time-specific. For now, you can either check out the previous survey results and try and draw your own conclusions, or wait until next month when comparisons can be done.
There were also two comments sections which yielded mixed results (mostly not being used for what they were intended) however, many comments give interesting insights so I’ll also make a few observations about those, but will get rid of them for the new survey.
Do you currently use Linux as your primary PC gaming platform?
Did you exclusively buy Linux-supported games last month?
What proportion of games did you play through Wine last month?
What proportion of PC games did you play on your Windows partition last month?
If you stopped using Windows last month completely for gaming, please state your reasons why (ie. a new game release, realised you never use it, etc.)
Unfortunately, this section was filled with "windows is crap" type comments which - although true - makes it very difficult to distinguish whether or not the people commenting stopped using Windows or not last month. This will therefore be replaced with a "did you stop using Windows last month?" and "if so, why?" style question where you have to respond to the first part to be able to respond to the latter.
However, some of the most common reasons cited for stopping Windows usage altogether were the release of Civilization V, XCOM: Enemy Unknown and Dota 2. Getting the format of this question right may mean that we see spikes in people leaving Windows altogether that coincide with big releases - like CS:GO and Borderlands 2 last month.
What distribution do your primary Linux gaming PC?
What Desktop Environment do you use on your primary Linux gaming PC?
This was a new question for this survey, and it has yielded some very interesting results. Although distributions seem to be mostly Debian-based, Desktop Environments are a completely mixed bag. The most popular was Unity, suggesting many out there are using vanilla Ubuntu.
These results sort of sum up the DE situation at the moment, with many major DEs having changed their design radically over the last few years and many more having appeared as a result - usage is completely fragmented over all of them. The "other" responses are mainly people who don't use a DE and use windowing managers like i3, FluxBox and OpenBox - so I will add none as an option.
Did you change your primary Linux gaming distribution last month?
What graphics card do you use on your primary Linux gaming PC?
Which drivers do you use for that graphics card?
This was another new question. Open Source drivers (AMD, Intel and Nvidia combined) made up just 18% of the total. I have heard that the Open Source AMD drivers are quickly catching up to the proprietary ones in terms of performance, so this may quickly change.
There were also comments suggesting that "updated" and "official" drivers can be quite subjective depending on the distribution, so this will now be limited to two choices (Open Source and Proprietary).
Personally the reason I don't use Open Source drivers is that I paid a good amount of money for my graphics card so want to get the most performance out of it possible, however if the nouveau drivers get within a 10% performance margin of the proprietary ones some day, then I will most likely switch. Why do you use the drivers you use? I'm quite curious.
What CPU do you use on your primary Linux gaming distribution?
Which of these retailers did you use to buy your Linux games last month?
How many Linux games did you buy last month?
Unique question - Have you bought Tropico 5?
Not surprising that it was quite low considering strategy isn't the biggest of markets. If you are still on the fence about this one, I recommend getting it! The campaign was very enjoyable and the game really did a lot to add replay value which previous iterations didn't have. However, if you have either Tropico 3 or 4, I would recommend getting this one on sale since although the changes are plenty, they are quite subtle and gameplay is essentially an improved version of the same thing.
Conclusions
The results were certainly interesting and more refined than last time. I think one issue which this survey has shown is some pretty major differences with Valve's survey. For instance, the AMD CPU market share is 8% higher on our survey and they don't seem to show GPU manufacturers on there. Like Valve's survey, ours isn't perfect either and obviously gaming websites as a whole tend to have a certain demographic as a readership (ie. less "casual" gamers) so such differences could be attributed to people playing a few 2D games on their integrated card simply not reading GOL. Nevertheless, this gives us the best insights into Linux gamers we have at the moment.
The new Survey for November is available here - so please fill that in if you have the time. Article taken from GamingOnLinux.com.
The changes in the questions seemed to have worked well and the data should hopefully be far more accurate this time. There are still a few minor things to iron out, but the questions will mostly remain the same so we can see trends developing a few months down the line.
I will not compare results with those from the first “experimental” survey since many of the questions have changed now that the survey has taken a more permanent form - mostly the fact that the first survey was not time-specific. For now, you can either check out the previous survey results and try and draw your own conclusions, or wait until next month when comparisons can be done.
There were also two comments sections which yielded mixed results (mostly not being used for what they were intended) however, many comments give interesting insights so I’ll also make a few observations about those, but will get rid of them for the new survey.
Do you currently use Linux as your primary PC gaming platform?
Did you exclusively buy Linux-supported games last month?
What proportion of games did you play through Wine last month?
What proportion of PC games did you play on your Windows partition last month?
If you stopped using Windows last month completely for gaming, please state your reasons why (ie. a new game release, realised you never use it, etc.)
Unfortunately, this section was filled with "windows is crap" type comments which - although true - makes it very difficult to distinguish whether or not the people commenting stopped using Windows or not last month. This will therefore be replaced with a "did you stop using Windows last month?" and "if so, why?" style question where you have to respond to the first part to be able to respond to the latter.
However, some of the most common reasons cited for stopping Windows usage altogether were the release of Civilization V, XCOM: Enemy Unknown and Dota 2. Getting the format of this question right may mean that we see spikes in people leaving Windows altogether that coincide with big releases - like CS:GO and Borderlands 2 last month.
What distribution do your primary Linux gaming PC?
What Desktop Environment do you use on your primary Linux gaming PC?
This was a new question for this survey, and it has yielded some very interesting results. Although distributions seem to be mostly Debian-based, Desktop Environments are a completely mixed bag. The most popular was Unity, suggesting many out there are using vanilla Ubuntu.
These results sort of sum up the DE situation at the moment, with many major DEs having changed their design radically over the last few years and many more having appeared as a result - usage is completely fragmented over all of them. The "other" responses are mainly people who don't use a DE and use windowing managers like i3, FluxBox and OpenBox - so I will add none as an option.
Did you change your primary Linux gaming distribution last month?
What graphics card do you use on your primary Linux gaming PC?
Which drivers do you use for that graphics card?
This was another new question. Open Source drivers (AMD, Intel and Nvidia combined) made up just 18% of the total. I have heard that the Open Source AMD drivers are quickly catching up to the proprietary ones in terms of performance, so this may quickly change.
There were also comments suggesting that "updated" and "official" drivers can be quite subjective depending on the distribution, so this will now be limited to two choices (Open Source and Proprietary).
Personally the reason I don't use Open Source drivers is that I paid a good amount of money for my graphics card so want to get the most performance out of it possible, however if the nouveau drivers get within a 10% performance margin of the proprietary ones some day, then I will most likely switch. Why do you use the drivers you use? I'm quite curious.
What CPU do you use on your primary Linux gaming distribution?
Which of these retailers did you use to buy your Linux games last month?
How many Linux games did you buy last month?
Unique question - Have you bought Tropico 5?
Not surprising that it was quite low considering strategy isn't the biggest of markets. If you are still on the fence about this one, I recommend getting it! The campaign was very enjoyable and the game really did a lot to add replay value which previous iterations didn't have. However, if you have either Tropico 3 or 4, I would recommend getting this one on sale since although the changes are plenty, they are quite subtle and gameplay is essentially an improved version of the same thing.
Conclusions
The results were certainly interesting and more refined than last time. I think one issue which this survey has shown is some pretty major differences with Valve's survey. For instance, the AMD CPU market share is 8% higher on our survey and they don't seem to show GPU manufacturers on there. Like Valve's survey, ours isn't perfect either and obviously gaming websites as a whole tend to have a certain demographic as a readership (ie. less "casual" gamers) so such differences could be attributed to people playing a few 2D games on their integrated card simply not reading GOL. Nevertheless, this gives us the best insights into Linux gamers we have at the moment.
The new Survey for November is available here - so please fill that in if you have the time. Article taken from GamingOnLinux.com.
0 Likes
Some you may have missed, popular articles from the last month:
The comments on this article are closed.
Patreon
PayPal
and maybe a question about monitor count? ..but it might make it too bloat :)
Not sure about multi-monitor because (so I've heard) it's very hard to run well under Linux and not many games support it.
But Laptop vs Desktop is a good one! I'll do that next time as a unique question :)
its not really hard when it comes to running game on 1 monitor and having other monitors for different things.. but having game over 2-3 of those that can be trouble with some games indeed
It is indeed variable - I use Zaphod Mode for my two screens specifically because of the fact it keeps games from interfering with them, but there are an increasing amount of people actually playing games on Linux with more than one head, although still probably not enough to truly justify adding it to the survey as of yet.
It's usually no hassle to set up, however it can be pain with some games which doesn't support multi-monitor properly (like running on wrong monitor, force-disabling other monitors, and such). Still an interesting question though, especially in case of people stretching games onto 3 monitors.
It certainly doesn't "just work" 9 times out of 10, but as with many things in Linux, you can usually work around it.
I now know that the biggest hurdle for me at least is the Nvidia drivers and SLI support in Linux. SLI support is technically there but hasn't been touched since they first introduced it in 2008 and it didn't work very well even then. Pretty much any gamer on any platform running 3+ monitors for gaming has/needs multiple cards to support it. So with my second card not really working these newer titles are extremely rough on a single GTX580. I'm hoping for a couple gtx980's for Christmas so hopefully one 980 will help me playing this way in Linux a bit better at least until Nvidia gets off their butts and fixes multi gpu setups for us.
That would allow to see data change over time in a consistent way.
Yeah, all the main questions are staying as they are now. Really, there is no difference between these ones and the last ones, all I did in a couple of questions was group things together, so they can still be compared with the previous results. The unique questions will come and go - but they're just things which are interesting as a one off, so don't really need to get trends out of them.
There's a lot of talk of multi-monitor usage, so I'll do that as a one-off. With these kinds of things it might be good to do a unique question now and then come back to it in a few months time and see if it's changed rather than month-by-month (will do that with the original Steam Machines question).
From my experience multi-monitor support is still a pain.
Think of combining displays of different geometry. Think of combinations of rotated displays / pivot. Think of different approaches how displays are addressed: Extended Desktop, Xinerama, Merged Framebuffer, Zaphod, etc... While all that is typically supported by the drivers and X, it can (and will) give you headaches in all thinkable ways (and beyond).
Programs that take the wrong display number. Programs that get the right display but taking the geometry of the other display (in case of different geometries). Programs SEGFAULTing in fullscreen because the other(!) screen is rotated. Inconsistent treatment of the parameter "primary display". I guess, I could go on for another 5 minutes.
I worked with a dual screen setup for nearly 5 years (Radeon, Catalyst, intel). One display rotated (that's really great for writing and reading/viewing stuff) while the displays had different geometries. I'm now better off with single a 27" monitor of higher resolution.
I actually exploit that in order to game on all 3 with 'Option "nvidiaXineramaInfo" "False"' so more modern DE's and WM's such as XFCE, KDE or Gnome will still receive info about the individual monitor boundaries and then not so modern ones like Openbox (recently figured out cinnamon as well) will see only a single 5760x1080 screen and games (normally) happily span all three. So nowadays I use XFCE for normal computing and then switch to openbox to play games and use cinnamon for the few (usually UE3 engine games) that tend to fail on openbox when ran at such wide resolutions.
Taken from the Nvidia docs.
Option "nvidiaXineramaInfo" "boolean"
The NVIDIA X driver normally provides a Xinerama extension that X clients (such as window managers) can use to discover the current layout of display devices within an X screen. Some window mangers get confused by this information, so this option is provided to disable this behavior. Default: true (NVIDIA Xinerama information is provided).
On X servers with RandR 1.2 support, the X server's RandR implementation may provide its own Xinerama implementation if NVIDIA Xinerama information is not provided. So, on X servers with RandR 1.2, disabling "nvidiaXineramaInfo" causes the NVIDIA X driver to still register its Xinerama implementation but report a single screen-sized region. On X servers without RandR 1.2 support, disabling "nvidiaXineramaInfo" causes the NVIDIA X driver to not register its Xinerama implementation.
Due to bugs in some older software, NVIDIA Xinerama information is not provided by default on X.Org 7.1 and older when the X server is started with only one display device enabled.
For backwards compatibility, "NoTwinViewXineramaInfo" is a synonym for disabling "nvidiaXineramaInfo".
EDIT* does anyone know if wayland is likely to make such setups easier to deal with?