SteamOS updated, Valve drops AMDGPU-PRO for Mesa

Quoting: Whitewolfe80

Quoting: ghemAwesome news <3
I sense an AMD APU based Steam Machine in a not so distant future

Aww thats cute you think Steam machines are still a thing

People fail to realize that Valve is in for the long term (I know, not a popular notion these days with shareholders impatient to get their 300% ROI in 6 months).
People also fail to realize Steam machines are not only there to protect Valve's market against microsoft, it's equally a tool to grab market share from consoles.

And no Steam Machines are not dead.
Valve did launch them because it was be the only way to actually work out in the open and to generate an initial interest so as to get things going.
Currently, they're doing all the required, laborious work on the ecosystem: removing roadblocks (AMD drivers), improving the experience (Steam controller, BPM), pushing Vulkan.
Though they still are lagging severely behind in the video department (Netflix, Youtube, etc.)

Only when the ecosystem is ready will it be the time for a big push.
Tough I don't really have a clue as to how to do that.
Release a standard, officially endorsed Steam Machine?
Release a new Valve game with a temporary SteamOS / Linux exclusivity? (I know they don't like exclusive games, but who knows)

They pushed Steam by making it a requirement for Half-Life but they can't do the same with SteamOS because it would be too intrusive for the user. So I wonder how they would entice people to switch, how they would differentiate themselves in the eyes of the average consumer.
What could you do on a Steam machine that you wouldn't be able to on a windows PC or a console? Or that you would be able to do in a better or different way?
So far, I only see the Steam controller, but I wonder if that's enough

Quoting: ghem

Quoting: Whitewolfe80

Quoting: ghemAwesome news <3
I sense an AMD APU based Steam Machine in a not so distant future

Aww thats cute you think Steam machines are still a thing

People fail to realize that Valve is in for the long term (I know, not a popular notion these days with shareholders impatient to get their 300% ROI in 6 months).
People also fail to realize Steam machines are not only there to protect Valve's market against microsoft, it's equally a tool to grab market share from consoles.

And no Steam Machines are not dead.
Valve did launch them because it was be the only way to actually work out in the open and to generate an initial interest so as to get things going.
Currently, they're doing all the required, laborious work on the ecosystem: removing roadblocks (AMD drivers), improving the experience (Steam controller, BPM), pushing Vulkan.
Though they still are lagging severely behind in the video department (Netflix, Youtube, etc.)

Only when the ecosystem is ready will it be the time for a big push.
Tough I don't really have a clue as to how to do that.
Release a standard, officially endorsed Steam Machine?
Release a new Valve game with a temporary SteamOS / Linux exclusivity? (I know they don't like exclusive games, but who knows)

They pushed Steam by making it a requirement for Half-Life but they can't do the same with SteamOS because it would be too intrusive for the user. So I wonder how they would entice people to switch, how they would differentiate themselves in the eyes of the average consumer.
What could you do on a Steam machine that you wouldn't be able to on a windows PC or a console? Or that you would be able to do in a better or different way?
So far, I only see the Steam controller, but I wonder if that's enough

Last I heard Valve was not a publicly traded company, so no stockholders to worry about. Another of the reasons I really like Valve.

Quoting: ghem

Quoting: Whitewolfe80

Quoting: ghemAwesome news <3
I sense an AMD APU based Steam Machine in a not so distant future

Aww thats cute you think Steam machines are still a thing

People fail to realize that Valve is in for the long term (I know, not a popular notion these days with shareholders impatient to get their 300% ROI in 6 months).
People also fail to realize Steam machines are not only there to protect Valve's market against microsoft, it's equally a tool to grab market share from consoles.

And no Steam Machines are not dead.
Valve did launch them because it was be the only way to actually work out in the open and to generate an initial interest so as to get things going.
Currently, they're doing all the required, laborious work on the ecosystem: removing roadblocks (AMD drivers), improving the experience (Steam controller, BPM), pushing Vulkan.
Though they still are lagging severely behind in the video department (Netflix, Youtube, etc.)

Only when the ecosystem is ready will it be the time for a big push.
Tough I don't really have a clue as to how to do that.
Release a standard, officially endorsed Steam Machine?
Release a new Valve game with a temporary SteamOS / Linux exclusivity? (I know they don't like exclusive games, but who knows)

They pushed Steam by making it a requirement for Half-Life but they can't do the same with SteamOS because it would be too intrusive for the user. So I wonder how they would entice people to switch, how they would differentiate themselves in the eyes of the average consumer.
What could you do on a Steam machine that you wouldn't be able to on a windows PC or a console? Or that you would be able to do in a better or different way?
So far, I only see the Steam controller, but I wonder if that's enough

Valve doesn't want to use MS's and Sony's anti-consumer strategies to force people into using Linux because they want people to use whatever the heck they want. They have no reason to push people into linux, if Windows becomes a jail SteamOS is their Plan B which is already pretty great, a lot of games, one big push with BPM upgrade to have more features and third party "apps" and some contracts with game devs and they're automatically in the console market EASILY + bringing all the goodies of Open platform with that.

Quoting: LeopardActually we're all depending on Gaben's attitude. I spoke based on " he will never cut a deal with MS" guess.

If MS approaches him like a solution like that ; Windows 10S is our future and this will not let anything outside of our store and Steam is a piece of software that needs to be installed outside of store. So what're you saying that Steam in Windows Store? If you're saying no ; then you can't sell games anymore on Windows.

If Gaben says yes to that , you can consider that Linux gaming will be dead. But this also means that MS will get their share from Gaben , because they exist on store and it has a price.

We're all depending on Gaben's attitude.

You REALLY think that MS would even bother approaching Gaben and asking him to put DIRECT COMPETITION to their UWP on their own Store... I doubt that, they will just ignore Steam and push their stuff and eventually even try making Steam uninstallable on Windows, it would be good for their store and terrible for consumers

Quoting: MaCroX95

Quoting: LeopardActually we're all depending on Gaben's attitude. I spoke based on " he will never cut a deal with MS" guess.

If MS approaches him like a solution like that ; Windows 10S is our future and this will not let anything outside of our store and Steam is a piece of software that needs to be installed outside of store. So what're you saying that Steam in Windows Store? If you're saying no ; then you can't sell games anymore on Windows.

If Gaben says yes to that , you can consider that Linux gaming will be dead. But this also means that MS will get their share from Gaben , because they exist on store and it has a price.

We're all depending on Gaben's attitude.

You REALLY think that MS would even bother approaching Gaben and asking him to put DIRECT COMPETITION to their UWP on their own Store... I doubt that, they will just ignore Steam and push their stuff and eventually even try making Steam uninstallable on Windows, it would be good for their store and terrible for consumers

Possibility of that is very low , i agree. But MS will be getting a share anyway , so this is not impossible but also very low possibilty.

Quoting: Leopard

Quoting: MohandevirI'm really curious to see what the desktop landscape will look like in 5 to 10 years from now.

For sure Win10S and Surface devices is the way of the future in Microsoft's mind. They are really stubborn about the Windows Store and the lock down that comes with it, but what's in it for us PC Gamers? Streaming devices for Xbox Scorpio? Is this what they are thinking? Is this Microsoft's vision of the post traditional desktop era?

If that's the case, Valve needs SteamOS more than ever.

Let me explain MS moves from my point of view.

1-) MS Store is an important figure to get things right for Xbox and Windows Mobile.

MS Store games are growing day-by-day and usually they have Play Anywhere feature. This means you can play your purchased games on Windows and Xbox. So that is not killing Xbox ; instead strenghtening that.

Let's think. You bought bunch of games when you're on Windows and after years you decided to become a console ( couch) player. If you choose Sony's Playstation you will start with zero games. But if you choose Xbox , you can immediately start enjoying your pre purchased games on Xbox.

That is for killing PS.

2-) Mobile side. Last year ; MS announced that Windows Phones will be able run x86 apps in the future. You can see that ; Whatsapp , Telegram , Vlc and other kind of stuff is moving into MS Store.

Plan is ; " we couldn't get attention for WP's from developers ; so if our phones can benefit from x86 apps there will be no need for another WP port effort. "

MS Store is a key element for rising of WP. In the mobile area ; rising = app variety and number. If they come head to head with Android and Ios at app basis ; rise will begin.

3-) Developer side. You know that all stuff about Linux subsystem in Windows.

Linux is better for developing languages and tools side than Windows. So they're moving to Linux. MS saw that and they offer developers to Linux based tools. Actually , there is no need for further explaining.

Just look at that Stackoverflow Developer Survey 2017.

https://insights.stackoverflow.com/survey/2017

Look at the Technology sections ; Most Loved, Dreaded, and Wanted Platforms chart. That will tell so much thing.

4-) But you are probably asking ; how is this a threat for Steam? People invested for Steam so much ; they can't move to other store.

Answer is ; GOG Connect.

People are mostly deciding by their previous investments. And that causes a little lockdown - an example : Apple Store. People are so much invested to Apple Store , the thought of buying everything again on Windows is simply freaking them and they continue to use Apple ecosystem.Even if you are developing something for Apple Store , you must sent it to store via a Mac.- on Steam . But Steam is platform agnostic ; so you can move to another platform without buying your games again ; thanks to Steam Play. Ms Store is providing that cross play for only MS based solutions.

So why is the answer is GOG Connect. GOG saw that ; if we want people attraction , we must lighten their burden of previous purchases. So they invented it. If you are binding your Steam account to GOG , you can get free GOG versions of selected titles. GOG 's library is not that much big so they can offer limited games.

MS has the power for doing that GOG Connect solution on a much bigger basis. When MS Store was big enough to get provide much desired games to players with zero pay ; gamers will move to their store automaticially. That is at least 2-3 years job but it will eventually come to this.

That effort will kill the Steam.

5-) If all of that i wrote at previous happen ; that will mean that market was succesful and MS can enjoy with market incomes.

So what is the fifth step? Free Windows for everyone. Because that market incomes can afford that OS development.

That's an interesting set of moves/tactics/strategies. All seems fairly plausible. Of course, MS wanting things to go that way doesn't necessarily mean things WILL go that way. Linux and Free Software aside, in each of those areas they are up against big opposition. In phones they are up against Apple and Google and (various manufacturers such as Samsung who have bet on Android). Those are big guys who know them very, very well from battles past, and who on phones have up until now squashed them with extreme prejudice. Right now nearly all Microsoft's phone revenue comes from patent trolling everyone shipping Android. Microsoft can make moves, but we can expect there will be counter-strategies, and strong ones. Convergence can go in different directions--Google will probably leverage Android into a software ecosystem for ChromeOS, for instance.

In consoles they are up against Sony and I suppose to some extent Nintendo. Sony is not, as near as I can make out, terribly agile or imaginative in its strategies. But they are a gorilla with as much poundage as MS itself; their portfolio is wide. They are damned good at simply rolling out the hardware, making sure there's games, and marketing the **** out of everything at rollout. That is surprisingly effective; clever doesn't always win over hammering the basics real hard. Plus, one thing Sony are good at is monkeywrenching the corporate politics just like MS does; there's reasons DVD players went with Blu-ray. If they got annoyed they might use their multimedia muscle to cut the Xbox out of some stuff; imagine the Xbox suddenly not having a contract with Netflix or something. Plus, if pushed really hard they might just turn the next PS generation into something like de facto Steam Machines.

Your number 3 is a fallback move for MS; it creates vulnerability. Ol' chairthrowing Ballmer would be livid--handing the "developers, developers, developers" over to a rival platform. They may think they can manage this so that everyone develops on Linux but for Windows and remains dependent on Windows, but I don't see it. If the developers are all working with platform A, they are not going to retain loyalty to platform B, and if platform A has a software ecosystem, those developers will become accustomed to it. There will be a certain gravitational pull for developers to make products for the platform they use. As those developers become executives at tech companies . . . I can see why MS is finding they have to do this, but it's bad for them.


Last edited by Purple Library Guy on 24 May 2017 at 5:38 am UTC

Quoting: sr_ls_boyWhile they're at it, they could update the copy ofthe std c++ lib
in their runtime.

libstd++ should not be bundled in the runtime, it's an abysmally bad idea (unless you make sure all linking is limited to exactly the same version, like for example using static linking).

One of the libraries that's affected by this is Mesa-DRI. Example:

ldd /usr/lib/x86_64-linux-gnu/dri/radeonsi_dri.so
...
        libLLVM-3.9.so.1 => /usr/lib/x86_64-linux-gnu/libLLVM-3.9.so.1 (0x00007f4a0c254000)
        libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f4a0bed0000)
...

Note how it depends on libstdc++. And it's an important part of the graphics stack.

GNU C++ ABI guarantees (mostly) forward compatibility. I.e. if you are using application built against older libstdc++, dynamic linking with new one will work. But, there is no backwards compatibility. If you are using something built against newer libstdc++, and it dynamically links against older one, stuff can go haywire.

And that what happens, if in Steam runtime (or any other such bundle) you have a stale libstdc++, the game links against that, and then system library like Mesa-DRM links against newer one, and in result both envs are mixed, and something ends up calling stuff from new and from old in a sequence that causes a segfault (consider for example allocating memory with new one, and then deallocating with old one or any other such scenario).

https://gcc.gnu.org/onlinedocs/libstdc++/manual/abi.html

QuoteVersioning gives subsequent releases of library binaries the ability to add new symbols and add functionality, all the while retaining compatibility with the previous releases in the series. Thus, program binaries linked with the initial release of a library binary will still run correctly if the library binary is replaced by carefully-managed subsequent library binaries. This is called forward compatibility.

The reverse (backwards compatibility) is not true. It is not possible to take program binaries linked with the latest version of a library binary in a release series (with additional symbols added), substitute in the initial release of the library binary, and remain link compatible.

Last edited by Shmerl on 24 May 2017 at 2:14 am UTC

Quoting: ghemSo I wonder how they would entice people to switch, how they would differentiate themselves in the eyes of the average consumer.
What could you do on a Steam machine that you wouldn't be able to on a windows PC or a console? Or that you would be able to do in a better or different way?
So far, I only see the Steam controller, but I wonder if that's enough

Compared with PCs:

• Maintenance-free, no need for a keyboard & mouse for administrative tasks.


• No desktop OS raising its ugly head: no OneDrive notifications, GeForce Experience pop-ups, other processes stealing window focus, less components, thus less risk of something breaking during update etc.


• (Potentially) less overhead. This is not the case today, as the BPM itself is a performance hog, but there is significantly less stuff happening in the background, so when Valve fixes the BPM, more CPU/memory should be available to games on SteamOS than on Windows.

Compared with consoles:

• Mods. No, limited modding being slowly introduced on consoles doesn't count, though this may change in the future.


• Steam itself: community, marketplace, workshop, sales, free multiplayer, cheaper games etc., etc.


• Lower barrier for entry: more indies (and more crap :) ).


• Open system underneath, if you wish to take a deep dive (important for a very narrow audience, as many people who would appreciate it scream "Steam DRM" and refuse to use Steam at all).

Is it enough to compel people to switch? So far not, but that's mostly because of performance and lack of games. If UWP is made mandatory though, we may see that some issues that consoles have will come to Windows, all while Valve, game devs and Nvidia/AMD work to improve gaming experience on Linux/SteamOS.

Heck, Valve could create momentum with some collectible cards and hats alone. I don't understand it, but people seem to care about collectibles in steam a lot.
Regularly annoncing to drop some in this or that game on linux on a regular basis might just generate enough interest once the ecosystem is ready.

Quoting: constHeck, Valve could create momentum with some collectible cards and hats alone. I don't understand it, but people seem to care about collectibles in steam a lot.
Regularly annoncing to drop some in this or that game on linux on a regular basis might just generate enough interest once the ecosystem is ready.

That's silly--nay, ridiculous--and yet probably true.
Heck, setting up a Linux dual boot and installing some games is probably way less effort than a lot of stuff people do just to get an achievement in a single game.