Quoting: Shmerl

Quoting: HoriAlso, you have to understand that companies as big as MS plan things many years in advance. DX12 was probably planned and approved as a long term strategy for a long time before that happened.

It was a reactionary effort driven by Mantle proposal from AMD. MS literally used Mantle itself for DX12 (I linked the evidence above). So they could simply join Vulkan collaboration instead, but their lock-in mentality still prevailed there.

It worked out better for OpenXR - MS for a change did participate in it.

MS probably knew about Mantle before it has been made public and by then, they could have work on DX12.

View this comment - View article - View full comments

Quoting: Alm888

Quoting: ArtenI don't think that windows game runing under proton can access linux filesystem.

It can (by default root directory is mounted as Z: drive). So a Linux-aware Windows virus can access ones "host" filesystem and, let's say, encrypt all the files it can grab.

WINE developers warn everybody: WINE is NOT a sandbox!

Thanks for info, i missed it.

View this comment - View article - View full comments

Quoting: Lofty

Quoting: CatKillerI expect those videos to end with "Windows on this device kinda sucks."

why ? i mean watching this bit of the steamdeck hands on they state that the new UI is coming to windows shortly after. All things being equal the performance and compatibility should be higher on windows. I don't expect a mountain of people to do this honestly. But maybe when the novelty of using steamOS wears out perhaps.

https://youtu.be/jb6OWxORfY0?t=502 [External Link]

Quoting: CatKillerBut if people install Windows on it and genuinely prefer it, good for them. That's exactly what I'd want for installing Linux on all my computers. If it turns out that Valve haven't made Linux gaming compelling enough for the masses, so their customers are going to go through the hassle of installing fresh Windows on a device that comes with a different OS, then they aren't safe from Microsoft and they need to Linux harder.

Sure. At that point im not sure what more valve could do anyway for those kinds of people. Some people are just never going to compromise on absolute performance and compatibility by running Linux.
I remember a thread on a windows centered forum a year or so back when proton was really taking off, some news articles were dropping and the conversation got onto how far Linux had come, people were impressed for sure but even then the consensus was that if there was even a less than 5% drop in performance no matter how much freer, secure and private Linux was they would not switch. Not to mention the lack of other game stores (although im sure that's something that can be improved if this thing takes off)

I do like the idea that the deck(linux) gets the UI first at least but there has to be some reason to use the deck other than technical inability or laziness over installing windows+steam.

For performance side, can help that Valve know specific hardware? So Everything (kernel, wine,...) can be built for ZEN2? It's not great benefit to performance, but in some games is wine equal and rerly better then Windows even now.

View this comment - View article - View full comments

Quoting: F.Ultra

Quoting: Alm888New API? I wonder, what will it be?
In my impression developers just need to target WIN32 API releasing for Windows and Valve will do the "magic". Well, that and Vulkan is preferable. At least Valve has said as much…

Well to be honest in order to fulfil "that will tell games if they're being run on the Steam Deck" that API could be just /etc/os-release :-)

I don't think that windows game runing under proton can access linux filesystem.

View this comment - View article - View full comments

Quoting: Purple Library Guy

Quoting: Guest

Quoting: Purple Library Guy

Quoting: slembckeOn the other hand, Proton is really running a fine line as Microsoft could really screw them over if they wanted to. They will also be playing catch up indefinitely.

It's not as simple as that. The problem for Microsoft is, they change Windows to make new Windows software (temporarily) incompatible with Proton, I believe they've just abandoned backwards compatibility with Windows software too. They could find themselves in a situation where all the old Windows software runs better on Proton than on latest Windows. And I'm not just talking about games. That would be bad for them.
And then Proton would catch up, and it probably wouldn't even take that long because Windows is a big old beast that isn't simple to make workable changes to so whatever they did couldn't be that big a deal. Then Proton would be working with the new stuff and working with the old stuff, making it definitively a better Windows than Windows.

Windows breaks compatibility between versions sometimes. To their credit, Microsoft put an effort in to minimise such things, but it does happen.
Microsoft can make it very difficult for future games to run through wine. That's enough to bury Valve's efforts, before even considering legal challenges they could make. Don't even have to be particularly valid legal challenges - Microsoft could just drag it on, and that alone would also be enough to bury Valve's efforts.

Windows do try to avoid breaking compatibility, though. They know it's a big problem for them when they do, because one of the biggest things holding people to Windows is all the back catalogue of software that runs on Windows but not elsewhere (well, as far as they know, or easily). As I understand it, the last time they did was because they were desperately trying to get Windows 7 out of the way because nobody wanted to upgrade.
Breaking compatibility exactly because something exists that can run all that old stuff might make people think about taking a look at that something. Could backfire. Imagine Wine ran existing versions of Microsoft Office, but the latest Windows didn't, and everyone knew that was because Microsoft was afraid of Wine.
No doubt they can make it very difficult for future games to run through Wine--but at what cost? What else stops working? How difficult does it make it to write those future games for Windows? I just don't think it's as simple as some people suggest. You can't actually move the target that much, because doing so doesn't just mess with the emulators, it messes with your whole ecosystem.

As to the legal thing . . . they can drag things out in the courts, but that only matters if they can make an injunction stick. Otherwise Valve can just go along doing their thing while fighting it. I don't know, but I would imagine that you have to have at least a sort of workable case to get an injunction to stick if the defendant has good lawyers and is not indigenous or something. Otherwise every company would just sue on some spurious basis every time a competitor started to eat their market share, and shut down for years whatever it was that the competitor was beating them with. Maybe MS could muster something good enough to actually get an injunction that would stop Steam Deck production, but I have doubts. If all they could manage was to be annoying in court for a few years and cost a few million in legal fees, that would be irrelevant.
Now if they could win, that would be pretty disastrous--presumably that would make Wine itself illegal, and arguably the whole concept of reverse engineering, every emulator of everything. But I really doubt we've gone all this time without that kind of thing ever being tested in court, and there are still emulators.

For legal thing, we have favorable precedent:
https://arstechnica.com/tech-policy/2021/04/supreme-court-sides-with-google-in-api-copyright-battle-with-oracle/ [External Link]

View this comment - View article - View full comments

Quoting: TeodosioI would like to see a new game from Valve, released on GNU/Linux only.

Better solution is Linux nativ exlusive, but with WSL as linux compatibility layer on windows :-)

View this comment - View article - View full comments

Quoting: dubigrasu

Quoting: Arten

Quoting: DorritWhy do you guys think they went with Arch? Is it for not needing to eventually reinstall a new OS version?

They don't trust canonical after their initial plans to drop 32bit i guess. Arch begun emerging in Valve projects after that (futex), if i remember correctly.

Canonical planning to drop 32bit support with Ubuntu 19.10 onwards

Valve looking to drop support for Ubuntu 19.10 and up due to Canonical's 32bit decision (updated)

They weren't using Ubuntu for SteamOS though, but Debian (stable).

Ok, my mistake.

View this comment - View article - View full comments

Quoting: DorritWhy do you guys think they went with Arch? Is it for not needing to eventually reinstall a new OS version?

They don't trust canonical after their initial plans to drop 32bit i guess. Arch begun emerging in Valve projects after that (futex), if i remember correctly.

Canonical planning to drop 32bit support with Ubuntu 19.10 onwards

Valve looking to drop support for Ubuntu 19.10 and up due to Canonical's 32bit decision (updated)

View this comment - View article - View full comments

Quoting: Samsai

Quoting: ArtenEvery new chip design for new proces node is more expensive with every generation of process node and there is no silver bullet like machine learning - because machine learning slowly moving to specialized hardware and so they have they own chips. We need make more GPUs and ethereum do this job for us. Righ now, there is shortage, because eth price fluctuation, but that is only short term problem. But without is limiting development of GPUs for PC, so we can in the end be tied to professional segment and game consoles chips as high-end and integrated GPU as low-end and nothink in-between.

What kind of corporate welfare are we planning here? GPU manufacturers ought to make decent products that meet the demand. There is, and always has been, demand for mid-tier GPUs. They aren't even prohibitively expensive to make, it's an industry standard practice to make mid-tier and low-end products by scrounging together half-broken chips from higher tiers and disabling the broken bits. I don't see the logical connection from Ethereum miners to mid-tier GPU availability, and even if such a connection did exist, I don't think subsidizing chip manufacturers by having their products be bought in massive volumes for the purpose of crunching useless numbers and gathering dust in some warehouse next to a newly reopened coal power plant that will power said warehouse is the way to do it.

What corporate welfare? Crypto is free market. There is no welfare here.

RTX 3060 die 276mm^2, RTX 3070 die 392.5mm^2, RTX 3080 die 628.4mm^2. So its clearly different chip.
Link with ethereum is mining cards. NVIDIA CMP 30HX uses TU116 (GTX 1660), NVIDIA CMP 40HX uses TU106 (RTX 2060). And they use damaged chips, which are not usable for GPU.

About what imaginary coal power plant do you talk? I know only about one powerplant which has been rebuilded from coal to gas, so there is no reopened coal powerplant. If you read leftist propaganda in arstechnica, pleas read whole article, not only misleading title.

View this comment - View article - View full comments

Quoting: Samsai

Quoting: ArtenSo, actual rising cost of development for every process nodes will be on the shoulders of the players. End of desktop gaming and hardware diversity is coming.

Wat?

The cryptocurrency market hasn't been the driving factor behind process nodes ever, mining gains have been a byproduct of the increased compute that can be put to work for actually useful purposes. You may have heard of this little thing called machine learning, which seems to be selling a fair amount of GPUs and lately specialized hardware (so I guess hardware diversity will continue, who knew!). Apparently they can also accomplish stuff other than just burn watt-hours, so that's kinda cool.

Also, smaller nodes are running into a problem of diminishing returns anyway, crypto or no crypto. Laws of physics are tough to break, so future performance gains will come from architecture, both in terms of hardware and software. And both still have a fair amount to give for the purposes of desktop gaming. So I just don't find the doomer vision of desktop gaming dying because Ethereum won't be providing nonsense crypto-puzzles for people to solve very convincing. I'd say it's more likely that desktop gaming will thrive if some gamers can actually get their hands on some damn GPUs.

When i say it's driving force behind process nodes development? [My point is chip development](https://www.extremetech.com/computing/272096-3nm-process-node ) for new proces node. Every new chip design for new proces node is more expensive with every generation of process node and there is no silver bullet like machine learning - because machine learning slowly moving to specialized hardware and so they have they own chips. We need make more GPUs and ethereum do this job for us. Righ now, there is shortage, because eth price fluctuation, but that is only short term problem.
But without is limiting development of GPUs for PC, so we can in the end be tied to professional segment and game consoles chips as high-end and integrated GPU as low-end and nothink in-between.

View this comment - View article - View full comments