Valve have released the Steam Hardware Survey results for September and thus, GOL’s first survey comes to an end. 670 people completed the survey, which is quite a nice sample - so thanks all of you who took the time to do it.
As soon as I published the survey, there was a lot of feedback on changes which should be made, and that has been taken on board and incorporated into the new survey. The main change is that all questions are related to “last month” (ie. how many games did you buy last month) rather than being more general.
The new Survey for October is available here - so please fill that in if you have the time.
Here is a question-by-question analysis on the results. This will likely be shorter in the future,but given that this is the first time we are seeing these results, this on will be more detailed:
Q1: Do you use Gnu/Linux as your primary gaming platform?
Yes - 63.43% (425)
No - 12.54% (84)
Yes, and I only buy games with Linux support - 24.03% (161)
This is kind of expected from readers of a website devoted to Linux gaming - still 87.46% is quite high for Linux overall. The question has been clarified to state “primary PC platform” (rather than just platform) in the new survey.
Q2: What proportion of games do you play through Wine?
Nearly all games through Wine - 0.60% (4)
More games on Wine than native - 3.74% (25)
About half and half - 7.17% (48)
More native games than through Wine - 15.10% (101)
Nearly all games native - 40.36% (270)
I don't use Wine / don't use Wine for gaming - 33.03% (221)
The total users using Native Linux more than Wine is 88.49%, with 4.34% using Wine more often than Linux to play games. One point of clarification in the next survey is that games using a Wine wrapper should still be counted as native since on Steam it would count as a Linux sale/download.
Q3: What proportion of games do you play on your Windows partition?
Nearly all games on Windows - 3.89% (26)
More games on Windows than Linux - 8.52% (57)
About half and half - 7.32% (49)
More games on Linux than Windows - 8.37% (56)
Nearly all games on Linux - 17.49% (117)
No Windows partition / don't use Windows for gaming - 54.41% (364)
It was quite surprising to see that over 50% of GOL readers don’t have a Windows partition, and given all the new releases coming out from when this survey was published until December (Tropico 5 and Borderlands 2 already), the number of people dual booting will most likely decrease - which we may have already witnessed this month in the Linux market share increase from Steam Hardware Survey.
Q4: What distribution do you use as your main gaming distribution?
Debian based - 58.98% (394)
Slackware based - 1.20% (8)
Red Hat based - 3.59% (24)
Suse based - 3.89% (26)
Arch Based - 25.45% (170)
Other (please specify) - 6.89% (46)
It is perhaps no surprise given most developers test on and officially support only Ubuntu that most use a Debian-based distro, however, Arch still has a pretty strong showing. Those who put “other” mostly specified Ubuntu or Mint (which I have now clarified in the new survey that they are Debian-based) and others put Gentoo, which I missed out and have now included. I have also added a question about Desktop Environments.
Q5: Are you a distro hopper? (change distribution more than once every two years)
Yes - 13.75% (92)
No - 80.12% (536)
Yes, but not for gaming - 6.13% (41)
Quite surprisingly, an overwhelming majority said no and not for gaming (which would leave Steam’s results unaffected) - so either a lot of people lied and said no so as not to annoy Liam, or this may well be less of an issue than we thought.
Q6: What graphics card do you use on your main gaming PC?
Nvidia - 69.91% (467)
AMD - 23.80% (159)
Intel - 6.29% (42)
This is to be expected considering the bad reputation AMD have on Linux - though with the open source drivers improving quickly and AMD promising to improve their drivers, this may gradually change over time. Just as a comparison, Steam’s results show 51.56% for Nvidia, 29.47% for AMD and 18.63% for Intel - where the increase Intel users can most likely be attributed to the fact that Steam’s survey applies to everyone on Steam while this site appeals more to “enthusiast” gamers who are less likely to use integrated cards. The next survey also asks what drivers you use.
Q7: What CPU do you use on your main gaming PC?
Intel - 69.21% (463)
AMD - 30.79% (206)
Other - 0.00% (0)
The AMD share is 5% higher than Steam’s results across all platforms, which suggests Linux gamers are more into AMD CPUs than Windows gamers. I don’t know what I was expecting with the “other” response - maybe I thought some of you might duct tape a graphics card to a ARM Raspberry Pi and call it a gaming PC?
Q8: Which of these retailers do you use to buy your Linux games?
Steam - 96.69% (642)
Desura - 21.08% (140)
GOG.com - 45.93% (305)
Humble Store - 72.14% (479)
Other (please specify) - 7.98% (53)
What we can take from that is that a lot of people use Steam. I have added Indie Gala, KickStarter, GamersGate and developer website to the list of options in the new survey since they appeared often under Other.
Q9: How many Linux games do you buy per month?
Less than 1 - 17.09% (114)
1-2 - 44.68% (298)
2-5 - 28.19% (188)
5-10 - 7.95% (53)
10+ - 2.10% (14)
This is one result which will most likely vary heavily from month to month, depending on bundle offerings, Steam summer/winter sales, new releases, etc. Given the usual drought in this period and how Humble Bundles have had less and less Linux games lately, it is understandable that most people answered 2 or less. Again - the new survey refers specifically to the previous month - ie. October’s survey will ask how many you bought in September.
Q10: Will you be getting a Steam Machine any time after release?
Yes - 11.49% (77)
No - 29.25% (196)
Still undecided - 30.60% (205)
I already built my own (or planning to build my own) - 28.66% (192)
These were very interesting results, and I would have separated the will/have built if I had known that result would be so high. Rather than seeing how this one changes on a monthly basis, I think this one will remain out of the next survey and appear again on release of the Steam Machine. Instead, there will be a monthly “one off” question relevant to that month. For example, September’s big release was Tropico so the question for October will be “Did you or are you planning on buying Tropico” or this could be CS:S or Borderlands 2 next month. Suggestions welcome!
Conclusions
In the monthly Steam Hardware Survey articles, Liam has stated Linux users not using Steam or distro-hoppers as possible causes to the changes in Linux users we see month to month on Steam Hardware Survey, and up until now I also thought the same, but this survey has thrown much of that out the window. This shows one of the main reasons why having such a survey is important to Linux Gaming - because there are things we are still in the dark about given that Linux gaming is still a fairly new thing.
Hopefully, this will provide some interesting reading for us as we watch the stats change on a monthly basis, and also provide some useful information for developers looking to make games for Linux since currently there isn't much out there. If you can keep filling out this survey on a monthly basis, it will most likely be a big help to all of us so remember to fill in this month's survey.
UPDATE:
There seems to be a bit of confusion with one of the new questions on this month's survey. "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.)" is specifically aimed at people who stopped using Windows altogether LAST MONTH since it would be interesting to get an idea over time of why people stop using it and how the release of certain games affects that. Currently people are saying why they don't use Windows generally for games, which is interesting but not the point of the question. Article taken from GamingOnLinux.com.
As soon as I published the survey, there was a lot of feedback on changes which should be made, and that has been taken on board and incorporated into the new survey. The main change is that all questions are related to “last month” (ie. how many games did you buy last month) rather than being more general.
The new Survey for October is available here - so please fill that in if you have the time.
Here is a question-by-question analysis on the results. This will likely be shorter in the future,but given that this is the first time we are seeing these results, this on will be more detailed:
Q1: Do you use Gnu/Linux as your primary gaming platform?
Yes - 63.43% (425)
No - 12.54% (84)
Yes, and I only buy games with Linux support - 24.03% (161)
This is kind of expected from readers of a website devoted to Linux gaming - still 87.46% is quite high for Linux overall. The question has been clarified to state “primary PC platform” (rather than just platform) in the new survey.
Q2: What proportion of games do you play through Wine?
Nearly all games through Wine - 0.60% (4)
More games on Wine than native - 3.74% (25)
About half and half - 7.17% (48)
More native games than through Wine - 15.10% (101)
Nearly all games native - 40.36% (270)
I don't use Wine / don't use Wine for gaming - 33.03% (221)
The total users using Native Linux more than Wine is 88.49%, with 4.34% using Wine more often than Linux to play games. One point of clarification in the next survey is that games using a Wine wrapper should still be counted as native since on Steam it would count as a Linux sale/download.
Q3: What proportion of games do you play on your Windows partition?
Nearly all games on Windows - 3.89% (26)
More games on Windows than Linux - 8.52% (57)
About half and half - 7.32% (49)
More games on Linux than Windows - 8.37% (56)
Nearly all games on Linux - 17.49% (117)
No Windows partition / don't use Windows for gaming - 54.41% (364)
It was quite surprising to see that over 50% of GOL readers don’t have a Windows partition, and given all the new releases coming out from when this survey was published until December (Tropico 5 and Borderlands 2 already), the number of people dual booting will most likely decrease - which we may have already witnessed this month in the Linux market share increase from Steam Hardware Survey.
Q4: What distribution do you use as your main gaming distribution?
Debian based - 58.98% (394)
Slackware based - 1.20% (8)
Red Hat based - 3.59% (24)
Suse based - 3.89% (26)
Arch Based - 25.45% (170)
Other (please specify) - 6.89% (46)
It is perhaps no surprise given most developers test on and officially support only Ubuntu that most use a Debian-based distro, however, Arch still has a pretty strong showing. Those who put “other” mostly specified Ubuntu or Mint (which I have now clarified in the new survey that they are Debian-based) and others put Gentoo, which I missed out and have now included. I have also added a question about Desktop Environments.
Q5: Are you a distro hopper? (change distribution more than once every two years)
Yes - 13.75% (92)
No - 80.12% (536)
Yes, but not for gaming - 6.13% (41)
Quite surprisingly, an overwhelming majority said no and not for gaming (which would leave Steam’s results unaffected) - so either a lot of people lied and said no so as not to annoy Liam, or this may well be less of an issue than we thought.
Q6: What graphics card do you use on your main gaming PC?
Nvidia - 69.91% (467)
AMD - 23.80% (159)
Intel - 6.29% (42)
This is to be expected considering the bad reputation AMD have on Linux - though with the open source drivers improving quickly and AMD promising to improve their drivers, this may gradually change over time. Just as a comparison, Steam’s results show 51.56% for Nvidia, 29.47% for AMD and 18.63% for Intel - where the increase Intel users can most likely be attributed to the fact that Steam’s survey applies to everyone on Steam while this site appeals more to “enthusiast” gamers who are less likely to use integrated cards. The next survey also asks what drivers you use.
Q7: What CPU do you use on your main gaming PC?
Intel - 69.21% (463)
AMD - 30.79% (206)
Other - 0.00% (0)
The AMD share is 5% higher than Steam’s results across all platforms, which suggests Linux gamers are more into AMD CPUs than Windows gamers. I don’t know what I was expecting with the “other” response - maybe I thought some of you might duct tape a graphics card to a ARM Raspberry Pi and call it a gaming PC?
Q8: Which of these retailers do you use to buy your Linux games?
Steam - 96.69% (642)
Desura - 21.08% (140)
GOG.com - 45.93% (305)
Humble Store - 72.14% (479)
Other (please specify) - 7.98% (53)
What we can take from that is that a lot of people use Steam. I have added Indie Gala, KickStarter, GamersGate and developer website to the list of options in the new survey since they appeared often under Other.
Q9: How many Linux games do you buy per month?
Less than 1 - 17.09% (114)
1-2 - 44.68% (298)
2-5 - 28.19% (188)
5-10 - 7.95% (53)
10+ - 2.10% (14)
This is one result which will most likely vary heavily from month to month, depending on bundle offerings, Steam summer/winter sales, new releases, etc. Given the usual drought in this period and how Humble Bundles have had less and less Linux games lately, it is understandable that most people answered 2 or less. Again - the new survey refers specifically to the previous month - ie. October’s survey will ask how many you bought in September.
Q10: Will you be getting a Steam Machine any time after release?
Yes - 11.49% (77)
No - 29.25% (196)
Still undecided - 30.60% (205)
I already built my own (or planning to build my own) - 28.66% (192)
These were very interesting results, and I would have separated the will/have built if I had known that result would be so high. Rather than seeing how this one changes on a monthly basis, I think this one will remain out of the next survey and appear again on release of the Steam Machine. Instead, there will be a monthly “one off” question relevant to that month. For example, September’s big release was Tropico so the question for October will be “Did you or are you planning on buying Tropico” or this could be CS:S or Borderlands 2 next month. Suggestions welcome!
Conclusions
In the monthly Steam Hardware Survey articles, Liam has stated Linux users not using Steam or distro-hoppers as possible causes to the changes in Linux users we see month to month on Steam Hardware Survey, and up until now I also thought the same, but this survey has thrown much of that out the window. This shows one of the main reasons why having such a survey is important to Linux Gaming - because there are things we are still in the dark about given that Linux gaming is still a fairly new thing.
Hopefully, this will provide some interesting reading for us as we watch the stats change on a monthly basis, and also provide some useful information for developers looking to make games for Linux since currently there isn't much out there. If you can keep filling out this survey on a monthly basis, it will most likely be a big help to all of us so remember to fill in this month's survey.
UPDATE:
There seems to be a bit of confusion with one of the new questions on this month's survey. "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.)" is specifically aimed at people who stopped using Windows altogether LAST MONTH since it would be interesting to get an idea over time of why people stop using it and how the release of certain games affects that. Currently people are saying why they don't use Windows generally for games, which is interesting but not the point of the question. 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.
All posts need to follow our rules. For users logged in: please hit the Report Flag icon on any post that breaks the rules or contains illegal / harmful content. Guest readers can email us for any issues.
Patreon
PayPal
How to set, change and reset your SteamOS / Steam Deck desktop sudo password
How to set up Decky Loader on Steam Deck / SteamOS for easy plugins
I think you're over thinking it. If it's a PC that runs the SteamOS distro, then it's a "Steam Machine" IMO. Doesn't matter if you use a steam/PS/Xbox controller or a KB and mouse. Doesn't matter if you buy it or build it. Not 100% positive but I'm Pretty sure ALL retail steam machines will come with SteamOS. But obviously it's just a PC so you can always install win on it if you want, and I can imagine some of the companies will offer it as an option. The second there is not a SteamOS distro installed on it though it's no longer a Steam Machine, it's just a "pretty" console looking PC. But that's just my opinion of course.
If it's a PC that runs any distro other than steamOS and is used for gaming then it's just a Tux Powered Gaming PC™ ;)
Which is why I ticked "no" to buying a steam machine when I did the survey. I build all of my PC's and I wont be running SteamOS on any of them. I may have a "SteamOS session" installed on my Arch powered HTPC at some point but I wouldn't call it a steam machine.
Yes, please.
2 reasons, but only one pertains to most. The first reason is resources. When I am gaming I don't need a whole desktop environment running wasting system resources. I want to give my games all the horse power my rig has. With my fairly customized (5 conkys, guake and cairo-dock)XFCE desktop there are a bunch of things running eating cpu cycles and RAM. RAM usage with nothing but the desktop sits around 800-900mb and there are about 190 processes running.
Openbox is not a DE, it's just a very light weight window manager. When you start it for the first time you may think somethings broken. It's not. All you will get is a black screen and a small no frills app menu when you right click anywhere on the desktop. On my system with just openbox running, the whole system is using about 380mb of ram and about 40 less processes(150) than XFCE, XFCE is a damn light weight DE so I can only imagine the usage of something like KDE or Unity.
You can then add your own dock apps, taskbar etc to it if you wanted to basically build your own customized environment, but I leave it basically a blank slate. Most of the commonly needed apps and even games I have mapped to key combinations so no dock/launcher needed. It's kinda funny though I feel like I spend more time in openbox than XFCE these days, and not just because of games, I'm just use to my key combos and edit most configs at the config file level these days anyways. So other than my XFCE desktop being "pretty" I don't really have a need for all of its "advanced" features anymore. But that's a discussion for another day.
TL;DR Openbox is extremely light and fast.
The second reason is due to my multiple monitors(3) and that many games will refuse to go fullscreen across all of them at once in most DEs/WMs. They will only see a single 1920x1080 monitor and just wont go higher(bigger?). The way openbox handles per monitor boundaries I can manipulate it in a way to force it to see nothing but one big 5760x1080 monitor forcing those games to work basically. Like I said this one doesn't really pertain to many. Though I highly suggest 3 monitor gaming to anyone who can manage it and I would be glad to assist anyone who goes down that path, it certainly doesn't "just work" in linux.
DRM is also an important issue. I don't like Steam very much compared to GOG, but I also don't want to blindly trust Valve or rely solely on them. Some of their games are really good, but you have to always measure what you're willing to give up *after* you pay money, if there's DRM.
That's another reason why I proposed splitting up the Debian based distros.
You might read it as mostly Ubuntu, but perhaps, there's lot of Debian and Debian-but-not-Ubuntu-based stuff hiding in the big piece of pie...
I used to use xfce+Arch for gaming and found games very snappy. General ram usage was sitting at around 500mb idle. My current setup is Mint 17 (cinnamon) and I now sit at nearly 2gb idle. If you have limited ram then the benefits of a lightweight desktop for gaming are actually really noticeable. I've found it most noticeable for games in Wine as wine's limitations are exaggerated by low system resources.
OpenSUSE KDE FTW!
Also, give two votes because most people do not play just one genre.
The options could be:
Action
Adventure
Arcade
MMORG
RPG
Simulation
Strategy
Racing
Platformers
Puzzles
Shootemup
etc...
It seems that I get less random "stutters" while playing. Which is really mainly in more modern demanding games. I account it to less things in the background to randomly grab the cpu's attention. It's not like I'm running out of RAM with 8gb on any linux game or any win game that works in wine. My cards are pretty dated to be pushing 3 monitors in games the way I do, and even with an fx-8350 running the show it's still not the best processor for multi card setups so I just try to give it everything I've got ;).
Technically I can swap out XFCE's xfwm for openbox and get the same effect with my monitors for gaming without ever leaving the XFCE session by doing "openbox --replace" and just swap back to xfwm by doing "xfwm4 --replace" when I was done.
This is my Arch laptop, and systemd-journal is taking up 60MB for who knows what.
Right now, in OB, after firefox (390mb), X is the next largest ram user at 63mb and openbox is using 42mb. Plus the terminal I'm running HTOP in is eating 50mb. the rest is normal low ram system processes.
spiffy simple OB session you have there :)
Can you make a possible reply:
No, but i only buy games with Linux support... (because at this moment i'm forced to use Windows but I plan to ditch it in a future)
If I put every possibility that people have suggested on here, the survey would be 3 pages long and the data would not be any better - it would just satisfy a few people who are exceptions. With surveys there will always be people that don't exactly fit the mold, but those people just have to put whatever fits them best.
The people who use two window managers/DEs should just put the one they use for gaming. Those with optimus cards should put Nvidia (I have a laptop with optimus and if that was my main gaming device, I would certainly put Nvidia since that part of the card is what I would use most for games). Those who have the latest drivers should put "updated", however you installed them - the xorg-edgers PPA was just one example of how one might update them.
Those on Ubuntu should simply put Debian-based (because that's what it is) and based on distrowatch numbers, we can be almost sure that the 60% will be made up of mostly Ubuntu derivatives and Mint. It's great that people want to show off their favourite distro, but that's really not the point of the survey - we just need to show the main family since most things are compatible within them and from a gaming point of view, that's all that interests us.
In your case, you simply answer NO for the first question and YES for the second, there's no adjustments needed for the survey for you to be accurately represented. If in a couple of months you ditch Windows, then you take that survey and answer YES on both questions. If you feel you need to clarify something, that's what the comments are for.
That said, we have had some really great suggestions - personally I think a lot of them would be great for a "poll of the week" or something like that which a lot of other sites have - but I doubt Liam s the time for that, so if there's someone who would like to set that up then it would be of great benefit for the site as a whole.
People using Debian 7 (Wheezy) should know that they need a newer libc and get it some way or the others (there are scripts for it, or they fetch it from Jessie, or just download it, ...). You can rely on a glibc > 2.13 if you need to.
But I've a question (as a developer): What's the big difference between these versions?
Also not clear on why people object to Steam. These are non-open source games, so who cares if I buy them on Steam or directly from the dev site? I don't get it, what am I missing? And as far as I can tell, Steam/Valve is almost single-handedly responsible for us GETTING these games in the first place. I will happily support Steam, even if they take a cut of the profits, because these publishers wouldn't have even bothered with Linux if not for Valve saying that Windows is on its way out.
Otherwise, the results are more or less unsurprising, and very informative. I feel like I know our gaming community much better now.
Off to take the new survey!