Building a gaming PC

[doktor_fleck]

Which should change. There is no reason today to continue using Windows for gaming. If you are stuck with something that's Windows only and simply can't be used otherwise - a different story. But for gaming, Windows is not a necessity anymore.

Oh, it's the Linux year again?

But seriously: it will never happen. Yes, you can play on Linux. Yes, you can get similar performance on Linux. No, it will never be as easy as on Windows. I've read multiple articles, written by people who tried to switch from Windows as a gaming platform to Linux and it was always connected to some troubleshooting, compromises, complete failures and almost universally some Linux knowledge was required, which is a no-go for like 99% of Windows users.

Hell, even Steam project to push Linux as a gaming platform failed miserably. in other words: without some heavy financial support, all the fan-based open-source tries to make Linux for gaming are doomed.

 

[Gilrond-i-Virdan]

I've been hearing Linux taking over as long as I've gamed on PC, so roughly 15 years and I've yet to see it happen, or even gain traction during these years.

Hearing is really not telling much, until you try it. Does it work for you for gaming today, or you are still having issues? That's the way to measure how feasible it is. Also, not all use cases are exactly the same. Some edge cases exist, such as a few multiplayer games using some weird anti-cheat, that borks in Wine and so on. Those don't apply to me though, so I was never affected by such. Also, Valve are working with those anti-cheat solutions to address this, it's quite a lengthy project from what I've heard.

Post automatically merged: Jan 9, 2020

Hell, even Steam project to push Linux as a gaming platform failed miserably.

Not sure what's your information source. Steam's Linux effort is quite successful - their Linux developer team grew over the years. In fact, their investment into Linux gaming stack only ramped up. From Wine, dxvk and etc. which made the games like TW3 playable on Linux (without CDPR lifting a finger), to VR and AMD OpenGL (radeonsi) and Vulkan (radv) work, and to literally a shader compiler for AMD (ACO), which is an amazing project. Thanks in part to Valve, situation for AMD users is a lot better on Linux than on Windows. So I see nothing failed, if anything, they achieved a huge progress in just a few years.

Another elephant in the room is Google of course, who unlike Valve also poured a ton of money into marketing, to the point that developers like CDPR jumped on their effort and announced CP2077 for Stadia. That simply means that Linux expertise among gaming developers is going to become a norm. So Linux gaming progress has been very good in the recent years.

And not exactly accidentally, Google are focused on AMD as well.

 

[SigilFey]

I've been hearing Linux taking over as long as I've gamed on PC, so roughly 15 years and I've yet to see it happen, or even gain traction during these years.

But seriously: it will never happen.

Not sure what's your information source. Steam's Linux effort is quite successful - their Linux developer team grew over the years.

That's the crux of it, I think.

"Taking over" is probably not going to happen. End result is that Linux just doesn't have the gaming library or support of either Windows, MacOS, the consoles, or any of the mobile platforms. It does require more on the user end to get things working, and the way it works can be a little weird compared to what most people are used to. The fact that they left it open-source and never really "competed" in the industry means that things happen more slowly, generate little to no income, and we never see the advertising and general, public presence of the Microsoft / Apple world. There are still people that don't even know it exists. "What is Linux?" is a question I've answered a fair few times. Visibility is critical in order to either be competitive or take over. All of that combines to mean that Linux is still plugging along in the background, quietly being pretty awesome.

But!

"Taking over" is not necessary for success. A venture doesn't need to be "#1" in order to be successful. A business can come in dead last in competition and still be generating stable revenue. Something can be totally free, unknown, unassuming, and still absolutely brilliant. I think the main force behind Linux is that it just does its own thing. And that has definitely kept it rolling along all these years.

The one thing that may bring it more into the forefront is Stadia. We'll have to see.

 

[Gilrond-i-Virdan]

To demonstrate that mainstream gaming developers are now starting to take Linux seriously, take a look at this:

Measuring Mutexes, Spinlocks and how Bad the Linux Scheduler Really is

This blog post is one of those things that just blew up. From a tiny observation at work about odd behaviors of spinlocks I spent months trying to find good benchmarks, (still not entirely successf…

probablydance.com

And further back and forth with Linus Torvalds himself:

Real World Technologies - Forums - Thread: Nuances related to Spinlock implementation and the Linux Scheduler

www.realworldtech.com

Real World Technologies - Forums - Thread: Nuances related to Spinlock implementation and the Linux Scheduler

www.realworldtech.com

Real World Technologies - Forums - Thread: Nuances related to Spinlock implementation and the Linux Scheduler

www.realworldtech.com

Very interesting read, though quite technical, going deep into OS scheduler details.

Post automatically merged: Jan 9, 2020

Visibility is critical in order to either be competitive or take over. All of that combines to mean that Linux is still plugging along in the background, quietly being pretty awesome.

Yeah, surprisingly Valve despite putting a lot of development backing and effort, didn't really do any major marketing (except in the early days of Steam Machines project, which they basically put on the back burner today). I'm not sure what exactly their plan is, may be first taking everything to even higher quality than it is now, and then do the marketing push?

Clearly, during Steam Machines times, the technology wasn't quite ready in various aspects (Vulkan wasn't even out yet). Today it's way better, but a few things are still moving (like ACO isn't really fully released yet, it's still work in progress, also OpenXR and etc.). My guess is that once it all falls into place, Valve could make some product (like their own console or anything of the sort), that will get a lot more marketing from them, providing the presence you mentioned above. They clearly are heavily invested, so we should expect them to use their own efforts.

"Taking over" is not necessary for success. A venture doesn't need to be "#1" in order to be successful. A business can come in dead last in competition and still be generating stable revenue. Something can be totally free, unknown, unassuming, and still absolutely brilliant. I think the main force behind Linux is that it just does its own thing. And that has definitely kept it rolling along all these years.

Also, the major force is that anyone (like Google, Valve and etc.) can take it, and use it for their product. With Windows they can't, without paying MS.

 

[SigilFey]

Also, the major force is that anyone (like Google, Valve and etc.) can take it, and use it for their product. With Windows they can't, without paying MS.

Sort of makes me wonder what it would be like if a game were specifically written for Linux. Would it be easier or more difficult to port over to Windows that it is to go from Windows to Linux?

 

[doktor_fleck]

Yes, every year there is a new, amazing technology, which will make Linux an amazing gaming platform. Guess what? It will never happen:

- Linux share on home desktops is abysmal. Developing something for 0,01% of players is not worth it.
- People familiar with Linux and it's quirks are using it for work, and not for fun, mostly.
- Graphic drivers are still issue, after all those years. NVIDIA is dominating the graphic card market and they don't give a crap about Linux.
- MS is moving strongly into open source. Linux is now a sub-system of Windows 10. Which means it's easier for people who need Linux to do serious stuff to just install Windows 10, use Linux subsystem for work and if they want, play on Windows.
- Dual boot is still easier to do, than making a pure Linux-based gaming rig.

 

[SigilFey]

Yes, every year there is a new, amazing technology, which will make Linux an amazing gaming platform. Guess what? It will never happen:

- Linux share on home desktops is abysmal. Developing something for 0,01% of players is not worth it.
- People familiar with Linux and it's quirks are using it for work, and not for fun, mostly.
- Graphic drivers are still issue, after all those years. NVIDIA is dominating the graphic card market and they don't give a crap about Linux.
- MS is moving strongly into open source. Linux is now a sub-system of Windows 10. Which means it's easier for people who need Linux to do serious stuff to just install Windows 10, use Linux subsystem for work and if they want, play on Windows.
- Dual boot is still easier to do, than making a pure Linux-based gaming rig.

It was and still is an amazing gaming platform. Granted, I've not personally used it since the '90s, but even back then, when you got the installation sorted and games configured -- Holy Choco! Stuff that would take Windows 60+ seconds to launch would pop up instantly on Linux. And game performance was very, very good. Like, impossibly good, as it felt to me sometimes. (Nothing compared to today's tech, but ridiculously good for the Daggerfall days.)

So (pre-empting this): Why was it never adopted? I mean, if it was sooo good, surely -- surely -- someone would have picked it up and...capitalized...on it.

^ Bingo.

Linux always was, and shall likely remain, open-source. That means I can do what I want with it...but I can't:
1.) charge because it's on Linux. No kick-backs for exclusivity.
2.) ignore cutthroat business tactics. Microsoft and Apple went to some lengths to discourage people from using it. (See #3.)
3.) expect to maintain trade secrets. It's open...source. That means anyone with Linux and know-how is very capable of reverse-engineering my source code. (That might sound like...well...der...to the modern world, but back in the '90s, this was a HUGE, @#$%!ng concern.)
4.) expect it to successfully compete with Microsoft or Apple. Microsoft was so strong in these days, their only, real competition was Apple...and Microsoft was crushing them. (Gods, I hated Macs during this period. Horrid machines. Great for art, though.)
5.) expect people to deal with it. This was a time when plugging a printer cable into the correct serial port was a challenge beyond pretty much anyone without a step-by-step manual in front of them. Computers were weird and scary. Software was weird and scary. Gaming was the domain of nerds and geeks. And...yeah...here's an OS that requires you to write files in order to get a game to launch without errors or graphical corruption...[...*entire conversation fades into broken static*...]

So no one ever bit.

But, to Linux's credit, they just kept on pluggin'. I say, "Good!" Wish more ventures went this way. Just goes to show you what you can accomplish with quiet determination and focus.

I don't think we'll ever see them "dominate" anything though. I do think we may eventually see them sell out, and Linux become another, commercial OS. Or...Linux will be picked up and utilized to corner a whole, new market of gaming. The thing is, Stadia is not that impressive an option for Western markets (people that can afford gaming PCs or consoles). The Eastern market will flood to it if it works. People making less than $5,000 USD / year on average. It puts an entire world of entertainment in their hands with the technology they already own. I've lived in these regions, and I know how different the Western world is. If it works...it's going to explode in these areas. That might be enough to bring Linux into the arena as a major presence.

 

[Flyddon]

That simply means that Linux expertise among gaming developers is going to become a norm.

I have heard that Half Life III is going to be on Linux.

 

[Gilrond-i-Virdan]

Sort of makes me wonder what it would be like if a game were specifically written for Linux. Would it be easier or more difficult to port over to Windows that it is to go from Windows to Linux?

I think, it would actually make it easier. Those who start developing for Linux, already likely are going to use something that's portable and cross platform enough. Linux discourages lock-in. On the other hand, those who start making something just for Windows are likely stuck with some Windows only junk they need to replace to make their game portable.

Post automatically merged: Jan 9, 2020

I have heard that Half Life III is going to be on Linux.

Better listen to actual developers. See links I posted above.

Post automatically merged: Jan 9, 2020

Yes, every year there is a new, amazing technology, which will make Linux an amazing gaming platform. Guess what? It will never happen:

You don't seem to be up to date on things. So let's correct it your misconceptions:

- Linux share on home desktops is abysmal. Developing something for 0,01% of players is not worth it.

Should we care about share? It's irrelevant, what's important is the size, i.e. absolute number of Linux gamers. It's big enough for market to be attractive to developers, and it's constantly growing for your reference. It's the reason developers are making games for Linux today to begin with.

However, it's a smaller market, therefore some developers (primarily controlled by legacy publishers) don't care to release for Linux even though it's profitable already. They are used to dealing with big markets and known "vendors". Venturing into new markets isn't their thing, unless they are promised some crazy number of sales. That's why Google convinced them with thin client idea (Stadia), while they don't care to make release for normal Linux users.

Compare it to developers who release in order to increase the reach of their art. As long as it's profitable, they'll do it, no matter if the market is small. So we see this difference quite clearly. Innovative studios and independent developers release for Linux quite often, even though for them it's of course harder than for huge publishers to find more development resources. Legacy publishers on the other hand commonly don't, due to common greed logic (i.e. releasing for Windows - more users, releasing for Linux - less users, even though both make profit, they make more profit by releasing more stuff to the first group.

Good thing we have developers who don't follow this "I want moar $$$$" logic all the time.

- People familiar with Linux and it's quirks are using it for work, and not for fun, mostly.

You probably don't know how many Linux gamers there are today. I use it both for work and gaming.

- Graphic drivers are still issue, after all those years. NVIDIA is dominating the graphic card market and they don't give a crap about Linux.

Nvidia market share on Linux is gradually falling for several years already. Due to the fact that AMD drivers are open source, and also backing of other developers including Valve. So drivers situation is sorted out for the most part already. Nvidia usage will fall to the point of them opening up their drivers, or falling into complete irrelevance on Linux. I simply don't care about them already. AMD works just fine.

- MS is moving strongly into open source. Linux is now a sub-system of Windows 10. Which means it's easier for people who need Linux to do serious stuff to just install Windows 10, use Linux subsystem for work and if they want, play on Windows.

WSL is handicapped in comparison with Linux proper. Why should I ever bother with it, if I can use normal Linux? And as for gaming, MS is still stuck in the old '80s - '90s mentaility of lock-in, thanks to Phil Spencer I suppose. Once they'll start supporting Vulkan on Xbox, I'd agree with you, that they started moving in the right direction when gaming is concerned.

- Dual boot is still easier to do, than making a pure Linux-based gaming rig.

According to GOL stats, dualbooting is also gradually falling, thanks to rapid advacement of Wine/Proton, dxvk and etc. which allow running Windows games on Linux without the need to dualboot. That's also something Valve invested a lot of effort in.

 

[Restlessdingo32]

Should we care about share? It's irrelevant, what's important is the size, i.e. absolute number of Linux gamers. It's big enough for market to be attractive to developers, and it's constantly growing for your reference. It's the reason developers are making games for Linux today to begin with.

Well, what's important is the number of additional products a developer would sell by supporting Linux. Take me for example. If every game I wanted to play past, present or future worked natively on Linux I wouldn't be using Windows for any games. I wouldn't be using Windows period, in fact. By natively I mean specifically designed for Linux from the ground up (Wine, for all it's ingenuity, has it's faults... or did when I had used it). But if this were not the case I would still be running those same games on Windows. So whether the developer provides a native Linux version or not, they're still selling the product. They wouldn't sell an extra one for providing Linux support.

I think when people say the Linux gamers are a small market this is accounted for in the statement. The "Linux gamers", labels be damned, are the Linux purist crowd. Open source, free software, real choices, limited bloat, zero spying on users (not sure if this applies anymore... given how ubiquitous it has become), etc. It's probably not a stretch to say these users are a minority. So the question isn't how many game copies would sell on Linux. It's how many additional copies would sell from providing support for the OS.

The other hurdle is getting people to use the OS. Linux has this stigma of being unfriendly to the user. It takes work to make things go on it. While this certainly used to be true it's arguably not anymore. Various Linux distros are nearly as brain-dead to "use" as Windows. The Ubuntu's of the Linux world. I wouldn't personally recommend them but they do exist. Even so, this stigma still persists. In other words, PC users are intimidated by the OS.

In any case, this is all getting kind of off topic. I'm still curious if anyone can attest to AMD driver quality for Windows at this moment in time. For whatever reason I have an urge to go with the underdog across the board. Fuck the man . Not that AMD is necessarily the underdog on one front anymore. I'm apprehensive on the GPU front based on past information.

 

[Gilrond-i-Virdan]

By natively I mean specifically designed for Linux from the ground up (Wine, for all it's ingenuity, has it's faults... or did when I had used it). But if this were not the case I would still be running those same games on Windows.

Wine today is very high quality, considering dxvk now supports both DX11/10 and DX9 already. So while I surely prefer native games, Wine allows playing Windows ones without touching Windows. Some edge cases still remain when it doesn't work, but they are shrinking pretty fast. It just shows that given enough effort, even such lock-in can be broken, despite MS trying to make it hard, and developers ignoring Linux users for a long time.

I.e. give it a try and see how it works today. Making assumptions based on the past tests isn't useful, because things are progressing really fast these days.

It's probably not a stretch to say these users are a minority. So the question isn't how many game copies would sell on Linux. It's how many additional copies would sell from providing support for the OS.

I think we can agree that the bottom line is simply whether it's profitable for the studio or not. Today it is profitable as those game developers who release for Linux can attest, the main blocker is not the absence of profit, but simply the fact that legacy publishers view big markets as bigger profit generators, and therefore don't care to address smaller ones, even if support costs are reduced to almost nothing (like Stadia cases demonstrate, when all those legacy publishers who ignored Linux users until now, run to make versions for Stadia, but still don't care to release them for desktop Linux).

 

[doktor_fleck]

Since when GLOBAL share of Linux gamers and GLOBAL share of NVidia cards, and you know, the "big picture" perspective is irrelevant for the devs, who want to make money? I just state politely that you are closed in your Linux bubble and just don't want to see that the GLOBAL market dictates the rules. Can some indie dev make amazing games only for Linux? Yes. Will he be able to live from that? Absolutely not. You can write down walls of text about how amazing Linux is, but as long as MS sells Windows with almost all PC sold globally, every year will be a "Linux year" on home PC.

I will tell this once more, and for the last time: there is no technological "magic bullet" which will make Linux a gaming platform. Only economy matters and the truth is, that making AAA titles to run on Windows and Linux is not economically feasible, because Linux players are minority globally.

Oh, and Stadia is basically dead already.

 

[Gilrond-i-Virdan]

Since when GLOBAL share of Linux gamers and GLOBAL share of NVidia cards, and you know, the "big picture" perspective is irrelevant for the devs, who want to make money?

Since they cared about number of sales that comes from the actual number of users, not about some abstract percentages, that don't tell them anything about the former.

Share is the ratio, number of users on the other hand are your potential customers. I.e. again what difference does the share make? I.e. let's say you have a hundred million. Is 1% a lot or a little? It doesn't matter, because percentage is a relative, not an absolute measure. A million is a lot, it's an absolute number. Or to give you the opposite example. Let's say you have 10. Is 90% a lot? Doesn't matter, because 9 is quite a small number.

So now you hopefully get, that shares is not what estimates profitability. Number of potential users does.

I will tell this once more, and for the last time: there is no technological "magic bullet" which will make Linux a gaming platform.

Because it has been one for a while already. You just missed it, by not paying attention. And economy of it is good, only greedy legacy publishers ignore it. If the market wouldn't have been there, developers wouldn't have been making Linux releases, which you see coming out quite regularly. As I said before, not every developer is using "only big market" logic of the legacy publishers who care about their greed more than about their art.

 

[doktor_fleck]

Now you are just grasping at semantics. The core issue is stil there: it's unprofitable to make games for Linux. You can bend this however you want, it will still be the truth.

 

[Gilrond-i-Virdan]

The core issue is stil there: it's unprofitable to make games for Linux.

False. You are trying to substitute "not as profitable as for Windows" with unprofitable. That's demagoguery. Unprofitable means your expenses are higher than your income. That's not the case here. What you are probably talking about is called profit margins. I.e. you are trying to say, that Linux profit margins are smaller. And that's true - it's a smaller market.

It's not the reason not to release for it for gaming developers though. Because their natural intent as art creators is to reach more users.

And this idea was even expressed by CDPR CEO himself in the past.

 

[doktor_fleck]

Talking about wanting to do something is cheap. Hiring Linux devs, porting assets, developing stuff from scratch for a minuscule (in comparison to Windows and consoles) gamer population is expensive.

Still, we can discuss this all along. I stated the facts, you decided to not accept them. It's fine. 2% of Linux PC gamers can still hope for a miracle, rest of us will be playing on Windows or consoles. I'm sure as hell that 2021 will be a "Linux Year" too.

 

[Gilrond-i-Virdan]

Hiring Linux devs, porting assets, developing stuff from scratch for a minuscule (in comparison to Windows and consoles) gamer population is expensive.

Yet, they are already doing it, because Stadia. We are passed that step already, you missed it. Today it's not about yes profit or no profit. It's about profit vs bigger profit.

Still, we can discuss this all along. I stated the facts,

You actually seem to be unaware of the facts. I recommend first to analyze how some developers are already releasing for Linux and making profit doing it. Then we can indeed return to this, otherwise I see no point.

 

[doktor_fleck]

Let me say that one more time: Stadia is dead. DEAD. Check the numbers.

Please bump me again when the number of PC Linux players go from 2% to 3%.

 

[Gilrond-i-Virdan]

Let me say that one more time: Stadia is dead. DEAD. Check the numbers.

I don't really care about Stadia itself, I'm not using it and don't plan to. But the effect on studios developing for Linux because of it is obvious. You can call it "dead" as much as you want, you seem to have some problem with Linux or something? Things are moving regardless of that, and if you don't like it for some reason, then well, deal with it. Trying to pretend, won't change the facts.

 

[doktor_fleck]

I don't really care about Stadia itself, I'm not using it and don't plan to. But the effect on studios developing for Linux because of it is obvious. You can call it "dead" as much as you want, you seem to have some problem with Linux or something? Things are moving regardless of that, and if you don't like it for some reason, then well, deal with it. Trying to pretend, won't change the facts.

Linux is amazing system, and I'm not saying this ironically. In business.

As a gaming platform, viable for players and developers? Not really. It never was and it never will be.