Windows 10 S might alarm Valve into boosting SteamOS again

Quoting: qptain Nemo

Quoting: appetrosyanI understand what you mean by overpreparing, but in out particular case this isn't really a risk. Right now, Linux in the minds of the 75% of the Desktop market share is this obscure, hacker only OS, with little to accommodate anything beyond software development. Not advertising just leads to status quo, nothing lost nothing gained.

On the other hand, advertising it right now would be difficult: the average consumer would want to be able to do what they can do on pretty much any other device and right now all of those things are getting a facelift (getting, not got). Advertising now is like inviting people to a recital when you've just started playing an instrument.

We are just transitioning to Wayland: it's not feature complete, only one mainstream DE has workable support, almost no applications work under it and to top it all off, nVidia refuses to support it. And wayland would be huge for gaming. Making mainstream audiences switch now would create a bunch of myths about Linux that we can't easily shake off. It makes sense for Valve to wait until some of the dust has settled, and lay the groundwork in other areas.

I see your point and I appreciate the fact that there are times when you have to patiently prepare but at the same time I don't agree with the current state being quite so hopelessly unappealing and rough. You actually can do everything you can on other platforms and have a pretty good time doing it. But oh well. I do wonder at which point would you start advertising it heavily? Like what marks that moment when the dust settles in your opinion? Flawless adoption of Wayland in mainstream distros like Ubuntu?

When Wayland works. In my own experience, while it's way less versatile than x.org, where it works, it gives you a massive performance advantage. You will agree that the frame-rate is important for gaming, so it makes sense to advertise this gaming-specific OS when you actually can produce at your best.

I don't mind the dropped frames, and the extra work to make things run on Linux, but that doesn't mean that there aren't objective disadvantages to gaming on it. Right now the biggest one I face, is the fact that in a complex scene, the rendering speed drops. this all due to the way in which frames are buffered in X.org, and is one of the main reasons why the switch is being made.


Last edited by appetrosyan on 10 February 2018 at 9:42 pm UTC

Quoting: webcreature

Quoting: appetrosyan

Quoting: julespetrikov

Quoting: webcreatureI disagree! When the propriator of the one OS that is pre-installed on almost every sold gaming ready PC, decides to pre configure this OS to a default that makes Steam incompatible, and when the user has to change that configuration against warnings that tell him this step might be dangerous, then a big percentage of mostly new users will not do that. And that will shrink the user-base potential for Steam, not actual numbers of course.
OR.. people decide to use an alternate OS instead, which I'd like very much of course.. Do you believe in it?

"pre-installed on almost every sold gaming ready PC"

That's the case, they can't do that. That's bad business. That's exactly why Windows 10 S is not the choice for gaming or production, because it's not intended to run all the games and it's not commercialised as one. Same goes for ChromeOS, Android etc. It's much like how people were asking "How the hell SteamOS will replace Windows" in the early days and got the most obvious answer: It's not intended to do that. S Mode itself is not a profitable way of selling gaming ready PCs, since consumer is not always the idiot, but it's a good way of ripping off people.

There's absolutely no possible future in which Microsoft won the gaming industry by such a move. It's just plain dumb. You can't just force an incompatible OS in a market that demands a compatible OS. That won't force people to buy things from their Stores, it will force consumer away from Microsoft and Windows. To Mac and Linux, mainly.

I'm not trying to justify their point here. Just telling people: SteamOS and Windows 10 S or Windows 10 S Mode or whatever stupid thing they could put forward has no relevance at all.

Spoken like someone who knows business. Seriously though, they won't go that route, as they'd be shooting themselves in the foot. And Valve are still not pushing Steam OS, as they're likely waiting for the Wayland Conundrum to finally wear down and get to work.

MS definitely made Windows S for a reason. I think it comes close to what MS would find ideal for their future, and what they might try to achieve with Windows CoreOS. MS can't control Win32, but they definitely want to control things on Windows more.
On the other side no gamer in his right mind would buy S for gaming, and no gamer in his right mind would switch to S-Mode for gaming. Non gamers, or "not yet gamers" on the other hand might never even try out Steam or GOG or certain open source software because they're in S Mode and they don't want to switch for "security reasons" or because they even had to pay for it.
Even if MS had not gaming specifically in mind when they developed S, Win32 gaming conflicts with their interests. For them it would be best if gaming would come to UWP and let the rest be lagacy. They made UWP for a reason...
Now, of course I'm not saying the will succeed with it, I just say if they could choose, it would be that way, or something similar. Fortunately MS has a history of "wanting to much at once and p!ssing people of with it", so all of this could make some more people switch to Linux. If they do not succeed with it, as you say, I'd say they do have to change plans... again..
If it is in fact untrue, that they try to influence OEMs to pre-configure S-Mode on certain game capable devices, then all of the said things are irrelevant of course. We'll see..

But isn't Microsoft alreadly making S-Mode or UWP games - pretty well everything they publish on the Xbox is cross compatible with UWP.

Quoting: appetrosyanRight now the biggest one I face, is the fact that in a complex scene, the rendering speed drops. this all due to the way in which frames are buffered in X.org, and is one of the main reasons why the switch is being made.

OpenGL and Vulkan drivers on Linux make use of direct rendering, communicating directly with the hardware. Frames are not presented through an X.org buffer. What you said is mostly true, but it doesn't affect gaming that much.

Quoting: Samsai

Quoting: nitroflow

Quoting: Samsai

QuoteApparently, Windows 10 S can be upgraded to a "normal" version of Windows 10 Home for free, but the problem is that Microsoft has said around 60% don't even bother to do the upgrade keeping them locked into the Windows Store.

This stuff right here is just Microsoft preying on the ignorance of its users. Just like the free upgrades to Windows 10, the free upgrade to Windows 10 Home will be gone eventually and all of the people that didn't know to upgrade their OS by then have accidentally locked themselves into using a technically inferior product for no benefit.

Actually I took the opportunity to upgrade my parents POS from windows 7 pro to windows 10 pro for free because touch screen support is vastly superior but to each is own :P

You misunderstood my point. My point is not that Windows 10 is an inferior product but that the upgrade opportunity expired. And I predict the upgrade opportunity from S to Home will also expire and S is an objectively inferior product to Home.

If you wanted to the free option is still there all you need is a vaild win 7 key still they closed the web portal but downloading the iso and using a win7 key still works for free i did an install of it on a rig a friend wanted so he could play pubg.

Quoting: tuubi

Quoting: appetrosyanRight now the biggest one I face, is the fact that in a complex scene, the rendering speed drops. this all due to the way in which frames are buffered in X.org, and is one of the main reasons why the switch is being made.

OpenGL and Vulkan drivers on Linux make use of direct rendering, communicating directly with the hardware. Frames are not presented through an X.org buffer. What you said is mostly true, but it doesn't affect gaming that much.

Last I checked Vulkan was supported by Croteam (who are doing a great job an away) and Bethesda, (who make great games, and don't support Linux). When we get native support that circumvents the issues of X.org and the common ways of porting games (e.g. SDL) start making use of Vulkan, we're in business.

Quoting: appetrosyan

Quoting: tuubi

Quoting: appetrosyanRight now the biggest one I face, is the fact that in a complex scene, the rendering speed drops. this all due to the way in which frames are buffered in X.org, and is one of the main reasons why the switch is being made.

OpenGL and Vulkan drivers on Linux make use of direct rendering, communicating directly with the hardware. Frames are not presented through an X.org buffer. What you said is mostly true, but it doesn't affect gaming that much.

Last I checked Vulkan was supported by Croteam (who are doing a great job an away) and Bethesda, (who make great games, and don't support Linux). When we get native support that circumvents the issues of X.org and the common ways of porting games (e.g. SDL) start making use of Vulkan, we're in business.

SDL doesn't really "make use of Vulkan" (unless you use just SDL's built-in graphics routines and even then I think it might just use OpenGL). What you do with SDL is make a window, load up Vulkan and create a Vulkan context/surface/whatever onto that window and start pushing data and commands to the GPU to do things with that surface. The support for doing this was released in last September.

Vulkan and Wayland are quite probably not going to be some kind of silver bullets that will solve the performance issues. Vulkan will allow devs to possibly tweak their games more and parallelize their rendering more than OpenGL allows but that'll largely depend on the devs' ability to make efficient code. As for Wayland, it's not really much of a performance boost. I've used it for a couple of months now and the only real difference between X.org and Wayland is that window movement is a bit smoother, 3D performance is more or less equal between the two.

Quoting: GrazenBut isn't Microsoft alreadly making S-Mode or UWP games - pretty well everything they publish on the Xbox is cross compatible with UWP.

Yes, they made UWP, and they made special Windows 10 S versions, which for gaming are UWP exclusive.
What seems to be coming next are pre-installed Windows versions which could default to a new S mode, if/when OEMs want to save money and do as MS pleases. This S mode is UWP exclusive if not switched off by the user. Switching off S mode (I believe) will be accompanied by security warnings, and (as they say) will cost 50$ on everything above Home Edition. On Home Edition it is said to be free.

Quoting: Samsai

Quoting: appetrosyan

Quoting: tuubi

Quoting: appetrosyanRight now the biggest one I face, is the fact that in a complex scene, the rendering speed drops. this all due to the way in which frames are buffered in X.org, and is one of the main reasons why the switch is being made.

OpenGL and Vulkan drivers on Linux make use of direct rendering, communicating directly with the hardware. Frames are not presented through an X.org buffer. What you said is mostly true, but it doesn't affect gaming that much.

Last I checked Vulkan was supported by Croteam (who are doing a great job an away) and Bethesda, (who make great games, and don't support Linux). When we get native support that circumvents the issues of X.org and the common ways of porting games (e.g. SDL) start making use of Vulkan, we're in business.

SDL doesn't really "make use of Vulkan" (unless you use just SDL's built-in graphics routines and even then I think it might just use OpenGL). What you do with SDL is make a window, load up Vulkan and create a Vulkan context/surface/whatever onto that window and start pushing data and commands to the GPU to do things with that surface. The support for doing this was released in last September.

Vulkan and Wayland are quite probably not going to be some kind of silver bullets that will solve the performance issues. Vulkan will allow devs to possibly tweak their games more and parallelize their rendering more than OpenGL allows but that'll largely depend on the devs' ability to make efficient code. As for Wayland, it's not really much of a performance boost. I've used it for a couple of months now and the only real difference between X.org and Wayland is that window movement is a bit smoother, 3D performance is more or less equal between the two.

I would suggest you read my comment a bit more carefully. You are completely right, but I feel that I need to point out a few things.

First of all, when game development shifts towards Vulkan in lieu of DirectX we will not have a distinct disadvantage, as porting a game will no longer involve porting the most difficult part of the rendering pipeline. It will bring Linux and Windows onto the same footing in terms of graphical optimisations. It's not a silver bullet, but it's sure going to help us make a good case for linux gaming. It will definitely help out with the emulation: wine gives you Windows level performance on Doom and Wolfenstein using Vulkan, meaning that we can shrug off - "But you don't have as many games" by saying "Yeah we do. You can run everything you could on windows".

Secondly, Wayland's improved performance stems from a few factors. Most of the applications and games still use X.org, which means that you couldn't detect a performance increase, mostly because you were running it through the same graphical pipeline. If you don't have Wayland running in the background, you will see the difference. Also, Wayland decouples some of the hardware from the software, which (if the developers listen to reason) will allow us to do Windows level customisation to Gaming peripherals. Right now, the only way of creating keyboard macros is bound to x.org. Once we have enough people running wayland full time, we will be able to convince them to let us do some of those things: pass keypresses to specific applications (in gamer world terms 'tis a must).

Quoting: appetrosyan

Quoting: Samsai

Quoting: appetrosyan

Quoting: tuubi

Quoting: appetrosyanRight now the biggest one I face, is the fact that in a complex scene, the rendering speed drops. this all due to the way in which frames are buffered in X.org, and is one of the main reasons why the switch is being made.

OpenGL and Vulkan drivers on Linux make use of direct rendering, communicating directly with the hardware. Frames are not presented through an X.org buffer. What you said is mostly true, but it doesn't affect gaming that much.

Last I checked Vulkan was supported by Croteam (who are doing a great job an away) and Bethesda, (who make great games, and don't support Linux). When we get native support that circumvents the issues of X.org and the common ways of porting games (e.g. SDL) start making use of Vulkan, we're in business.

SDL doesn't really "make use of Vulkan" (unless you use just SDL's built-in graphics routines and even then I think it might just use OpenGL). What you do with SDL is make a window, load up Vulkan and create a Vulkan context/surface/whatever onto that window and start pushing data and commands to the GPU to do things with that surface. The support for doing this was released in last September.

Vulkan and Wayland are quite probably not going to be some kind of silver bullets that will solve the performance issues. Vulkan will allow devs to possibly tweak their games more and parallelize their rendering more than OpenGL allows but that'll largely depend on the devs' ability to make efficient code. As for Wayland, it's not really much of a performance boost. I've used it for a couple of months now and the only real difference between X.org and Wayland is that window movement is a bit smoother, 3D performance is more or less equal between the two.

Secondly, Wayland's improved performance stems from a few factors. Most of the applications and games still use X.org, which means that you couldn't detect a performance increase, mostly because you were running it through the same graphical pipeline. If you don't have Wayland running in the background, you will see the difference. Also, Wayland decouples some of the hardware from the software, which (if the developers listen to reason) will allow us to do Windows level customisation to Gaming peripherals. Right now, the only way of creating keyboard macros is bound to x.org. Once we have enough people running wayland full time, we will be able to convince them to let us do some of those things: pass keypresses to specific applications (in gamer world terms 'tis a must).

3D applications that use system SDL already have access to native Wayland and I've also built some games from source so that they utilize system SDL and thus shouldn't be running through XWayland. In both cases I have seen no noticeable performance benefits. X.org just isn't the bottleneck with game performance, sub-optimal OpenGL is.

Did SDL solve relative mouse cursor positioning issue on Wayland?