Helping Valve to power up Steam devices
igalia.com313 points by TingPing 6 hours ago
313 points by TingPing 6 hours ago
> This is a very difficult combination to achieve, and yet that’s exactly what we’ve done for Valve with Mesa3D Turnip, a FOSS Vulkan driver for Qualcomm Adreno GPUs.
Look at that. Something Qualcomm should have been doing.
Much credit to Valve for pushing that out as FOSS.
> Much credit to Valve for pushing that out as FOSS.
Cynical: Valve doesn't sell hardware or operating systems, they sell games. These devices are merely another storefront.
Optimistic: Valve has also figured out how to turn good will into a commodity. Blowing cash on Steam sales is a bit of a cultural centerpiece of the PC gaming community.
Gabe has proven that you can make stupid amounts of money by [mostly] doing right by the consumer. I'm not sure if there's more to the secret source, her sauce, because we've yet to see another CEO pull their head out of their arse far enough to see how lucrative this approach can be: consumerism is fickle, fanaticism is loyal.
This is what I say a lot. Valve isn't even remotely close to having clean hands here. They invented loot crates. Hats. Etc.
It's just that the bar is so INSANELY low - it's probably somewhere deep in the earth's core at this point - that valve looks like a fucking angel by being only VAGUELY greedy on occasion.
When your competition is EA... it's not hard.
> They invented loot crates. Hats. Etc.
Don't forget the part where they're encouraging kids to gamble with real money on Counter-Strike skins. They rely on an API that Valve freely provides and makes no effort to curtail.
But they like Linux and give refunds so they get a free pass.
> and give refunds so they get a free pass.
They only begrudgingly conceded refunds in 2015 after the no-refunds policy they had maintained for 12 years was found to be illegal in Australia.
It’s incredible how bad driver support is the ARM space. I was looking into some of the various Ambernic handhelds and their Linux firmware. Despite their SoCs being advertised as having Vulkan 1.1 support every firmware for the device ships with it disabled.
So many chipmakers and development board manufacturers see software/driver support as some kind of necessary evil--a chore that they grudgingly do because they have to, and they will do the absolute minimum amount of work, with barely enough quality to sell their hardware.
Come to think of it, for them it is basically customer support.
Most will want to outsource it as cheap as possible and/or push it to the community. They won't care if it takes an eternity for the customer to get their issues solved as long as new customers keep buying.
And a few companies will see an opportunity to bring better customer care as an advantage and/or integrate it in their philosophy.
It bewilders me. Software's gotta be easier than hardware right? Not that either is easy but as a software engineer, the engineering that goes into modern hardware mystifies me.
It's different definitions of "easy."
With hardware, you have about one billion validation tests and QA processes, because when you're done, you're done and it had better work. Fixing an "issue" is very very expensive, and you want to get rid of them. However, this also makes the process more of, to stereotype, an "engineer's engineering" practice. It's very rules based, and if everything follows the rules and passes the tests, it's done. It doesn't matter how "hacky" or "badly architected" or "nasty" the input product is, when it works, it works. And, when it's done, it's done.
On the other hand, software is highly human-oriented and subjective, and it's a continuous process. With Linux working the way it does, with an intentionally hostile kernel interface, driver software is even more so. With Linux drivers you basically chose to either get them upstreamed (a massive undertaking in personality management, but Valve's choice here), deal with maintaining them in perpetuity at enormous cost as every release will break them (not common), or give up and release a point in time snapshot and ride into the sunset (which is what most people do). I don't really think this is easier than hardware, it's just a different thing.
Software is easier than hardware in general but companies generally pay their hardware guys 25-50% less than their software counterparts
Software can always ship a new update for bugs or features.
Hardware not so much
I've done both. There are difficulties with both but overall I would say software is significantly more difficult than hardware.
Most hardware is actually relatively simple (though hardware engineers do their best to turn it into an incomprehensible mess). Software can get pretty much arbitrarily complex.
In a way I suspect it's because hardware engineers are mostly old fogies stuck in the 80s using 80s technologies like Verilog. They haven't evolved the tools that software developers have that enable them to write extremely complicated programs.
I have hope for Veryl though.
What do you think about Atopile? I'm not a hardware person yet, but these seem similar.
But - doesn’t open sourcing it kinda make it someone else’s chore?
Obviously it has to “work” at sale but ongoing maintenance could be shared with the community.
They're stuck in the building model of making semi-custom SoCs for enormous corporations and releasing/developing drivers for them in extreme NDA environments.
It's fine (or arguably not) for locked down corporate devices.
Not so fine for building computers people want to use and own themselves.
I also just love that in open source you can call something “Turnip” because you’re not marketing it to anyone.
Glad too see that while Qualcomm tries to keep things closed shut tightly, Valve and their contractors are trying to do the opposite.
It's incredibly obvious that they're trying to make Steam Deck 2 ARM-based. That's the generational change Valve is waiting for.
This is gonna be fantastic.
There are no ARM chips with enough power. They have said many times that they are not interested in minor performance improvements but rather want a leap. The Snapdragon X2 Elite chip is the leader (I cannot count Apple; they won't share their chips, obviously), but it doesn't even match AMD with their RDNA 3.5, and who knows when they will (or even if).
Apple not sharing their chips extends to Apple keeping their grip on the higher density nodes.
I wonder if it's still the case, but for a while Apple was buying the totality of TSMC's capacity for 3nm nodes, leaving the rest of the world with only 4nm+ chips to grab.
I agree they won’t do a Steam Deck 2 that’s ARM. Maybe in the future?
BUT, what about a “Steam Deck Mini”? Something at/above the current Steam Deck, maybe a little closer to Switch 2, but smaller/thinner/maybe a little cheaper?
Yeah you’re not going to run Cyberpunk 2087: Johnny’s Rent Is Due. But older games, less demanding indie games, and many emulators would still work great. Plus remote play of your big desktop if you have one.
I’m not saying they will, but I could see it as a possibility.
Do you mean there's no ARM chips that they can buy? Surely the ARM chips in Apple's devices are powerful enough aren't they?
I wouldn't go that far but they are clearly poised for that, should it be adventageous.
The Frame is essentially there already, with what should be the top mobile arm setup.
If an x86 chipset dropped that fit their needs better, I don't think Valve would hesitate. I think it's just a matter of Valve trying to enable the best options down the road, whatever they may be.
Huh, I had not connected those (hypothetical) dots, but I could see it..
Or maybe there's 2 next-gen Steam Decks, an ultra-portable ARM-based one that's as small as can be, and a more performant x86 one with AMD's next-gen APU...
Just my thinking, they'll release a "Steam Deck Mini" that's more in line with other current ARM based gaming handhelds like the Ayn Odin.
If they did an AMD CPU using the same TSMC node that Apple uses for Arm CPUs it wouldn't be that much less power efficient and have much great compatibility.
They would realistically gain the most efficiency by getting Nvidia to design a modern super power efficient GPU like what was used in the original switch and Nvidia Shield. AMD GPUs can be great for desktop gaming but in terms of power efficiency to performance ratio Nvidia is way ahead
An AMD CPU and Nvidia GPU might be a hard thing to actually negotiate however given that AMD is big in the GPU space as well. As far as I know most "APU" aren't really that special and just a combo of GPU and CPU
APUs have the GPU and CPU on the same package, or sometimes even the same die (with tiling). If there was to be an Nvidia GPU and AMD CPU type system, they would have to be separate packages.
Yeah, there's a real gap in the market for a relatively compact handheld which can play low-spec PC games. The AMD-based handheld PCs available today are all pretty chunky.
There's plenty of "relatively compact" ARM-based handhelds targeting the retro market already, but many of them are shipping with a pitiful amount of RAM (1GB or so) making them an absolute non-starter, while others (selling for significantly higher prices) run crappy Android-based OS's that will never be updated. There is a gap in the market for a good-quality retro-like handheld shipping with a Linux-native OS (or even just enabling one to be installed trivially after-the-fact, with everything working and no reliance on downstream hacked-together support packages).
Well, apparently there's this project I learned about literally yesterday! https://portmaster.games/games.html
There are handhelds for less than 200$ with very good screens and controls that can play all of these. Not to mention stream (via Steam or other software) from your PC!
Why not just make a performant ARM device? Apple demonstrated to the world that it can be extremely fast and sip power.
Sure, but Apple isn't selling their silicon to anyone else and Valve, successful as they are, don't have Apples money and economy-of-scale to throw at designing their own state-of-the-art CPU/GPU cores and building them on TSMCs state-of-the-art processes. Valve will have to roll with whatever is available on the open market, and if that happens to suck compared to Apples stuff then tough shit.
I'm definitely dreaming but I think it could be a win-win situation if Apple decided to licence its chips to Valve: the resulting handheld and VR headsets would be power/efficiency monsters and PC devs would finally have a good reason to target ARM, which could finally bring native PC gaming to MACs.
The Nintendo Switch already provides >160 million reasons for gamedevs to care about native ARM support, but that hasn't moved the needle for the Mac. Being ARM-based is the least of its problems, the problem is that it's a tiny potential market owned by a company which is actively hostile to the needs of game developers.
Really cool stuff! Especially nice to see the groundwork being laid for what could become very efficient handhelds, considering how much performance Apple's M-series and Qualcomm's Elite series with relatively few watts. Much better than AMD, Intel or Nvidia.
One nit: it's too bad Valve / Igalia choose to completely ignore the lessons from Bazzite.
Bazzite already runs a scheduler like LAVD, called BORE[0]. It would have saved them a lot of work to extend and improve that rather than invent the wheel again. I'm not sure if Valve and Igalia are unaware of Bazzite and BORE or if this is a case of NIH.
Nothing to contribute other than to say that article was an awesome read and now I wish I had the specific skills needed to work at Igalia. :)
Same feeling here! I never really dug much into the low level graphics side of thing.
The Winlator-releated ecosystem already works pretty well, there just isn't a good frontend or integration for it yet. That's what is really exciting here.
Gamehub is a proprietary app by a Chinese controller manufacturer with some suspicious behavior and several LGPL violations that unfortunately works much better then the alternatives. Funnily enough their CDN endpoint is called "bigeyes", which when researching a bit was apparently their (failed) effort to bring x86 VR to ARM almost 10 years ago. Some people have "debloated" the app, but it seems very amateur hour to me and the process isn't very transparent (the GitHub repo is just a readme)
There's also GameNative, which seems promising, but is very buggy.
And Winlator itself, which is a mess of tons of tunables and different forks that I really don't have the patience for when PC handhelds exist today and have a much better ecosystem.
I don't play games almost ever, but I'm going to buy all the products Valve releases soon, just to support their OSS efforts. They seem to be the only vendor that's opening stuff up, rather than locking it down.
I had barely played games for years, and got a steam deck just because it seemed like a cool linux device I could use both for gaming and tinkering. it has definitely gotten me back into gaming in a big way, the experience really is very nice.
Same here! I actually stopped playing when I moved entirely to Linux, and have been running on laptops without a good GPU solution since then.
I bought the SteamDeck because it looked like a cool product and I liked the openness ("it's just running Linux"), and I love it. And it got me back into gaming :-).
Yes! The Deck is the closest I've gotten to getting into gaming. I especially loved the "press the power button and your game is immediately right there" aspect of it.
I ended up selling it to a friend because I enjoy making things much more, but the Deck is such a fantastic device.
It's a great device, I mainly use it for emulation. The fact that it's properly an open platform is amazing.
I just wish my hands wouldn't cramp on hand-held devices. I've never been able to use handhelds for longer than thirty minutes.
Goes for console controllers too.
Apologies if you've already ruled this out, but hand pain is very often caused by strain/injury from further up the arms, the shoulders, or even the neck. You may find that pulling your shoulders back or relaxing them, or adjusting your arm posture, or straightening/relaxingyour neck gives you less pain while playing.
Maybe you need larger controllers ?
Also possible the touchpads are better for fatigue than joysticks
There’s probably a better way to sponsor Valve than to buy physical products you won’t use. That has pretty low monetary efficiency for the purpose.
But maybe I'll use them!
The steam deck, especially the low-spec variant, was sold at very low, likely negative margins. They make huge profit on their games, but if you don't buy the games...
They've implied that they're not going to sell the Steam Machine at a low margin because they're worried about people buying the Steam Machine for general purpose computer use without buying games. I'm not sure that's a rational fear. If you subtract the GPU, you can get an comparable Beelink for ~$350. ~$500 would be the zero-margin price for a Steam Machine. It seems to me that the only people willing to pay an extra $150 for a mid-range GPU that's not good for AI would be gamers.
Not to mention that the Beelink comes with a Windows license, and the Steam Machine doesn't.
> Not to mention that the Beelink comes with a Windows license, and the Steam Machine doesn't.
That's a mark against the Beelink for many :)
> They've implied that they're not going to sell the Steam Machine at a low margin because they're worried about people buying the Steam Machine for general purpose computer use without buying games. I'm not sure that's a rational fear.
I can understand that, OTOH I have a $1500 gaming PC (probably worth far less now--I built it over a year ago) for explicitly that purpose. What I don't have is a modern, low-power living room HTPC with native/first-class Linux support on which to run Kodi (I have a custom one that's quite long in the tooth). If I could dock a steam deck in my living room and use it for Kodi 80% of the time with games for the remaining 20%, why should Valve care? I have already given Valve hundreds, if not thousands of dollars in game sales.
I assume Value is happy if you buy just 1 or 2 games for your Steam Deck or Steam Machine. It's the people that buy exactly 0 games that they claim to be worried about. IOW, not consumers, but companies buying work PC's.
Anybody who has a gaming PC isn't the target market for the Steam Machine. They're going after the console market with the value add of "also it's a real computer that can do real computer stuff".
I do buy quite a few games, which usually end up unplayed. A few times I do binge one, so it's generally worth it for me. I'd like the Steam Machine for playing games in my living room with friends etc, even though it might end up unused, but the OSS support really swings the scale towards "take my money".
That’s what happens when you don’t need to please the shareholders.
Google has contributed more to open source than Valve while being a public company. It's not just Valve who sponsor open source work.
They seem to be skimping out on their contributions to FFmpeg.
Google has contributed many patches to ffmpeg. Valve has contributed 0 as far as I am aware.
But what percentage of what Google has produced has been Free Software vs what percentage of what Valve has produced? Google may have produced more Free Software, but Google also produces way more things.
I don't think this very practical or relevant here, but I expect Google to have a higher percentage. Valve employees are focused on Valve's proprietary software: Steam, SteamVR, their games, etc. Valve more often pays contractorsto work on open source software than work on it themselves.
My comment was more to prove that it possible to do open source while having share holders. My claim that Google does more is auxiliary to it.
I'd like to see that comparison tracking the number of devs and how much open source software each company uses.
If you think publicly-held companies are bad, you should see what private equity gets up to.
I’d wait to see if they open source the Machine, Controller, and Frame before assuming buying their products supports open source that matters for everyone. Right now the Steam Deck is the only product that open source and supports that vision.
Even this article it is not clear how beneficial some of their open source work is for everyone except Valve.
For a few years before I eventually got a Steam Deck, I played a lot of games that I bought outside of Steam, and over the past decade, the experience of doing this on Linux has massively improved. Plenty of their improvements get upstreamed to Wine, and there's nothing stopping you from obtain proton (or even one of the various unofficial tweaks of it) to run games that you don't buy through Steam to get the benefits that aren't upstreamed (or haven't been yet). The article itself mentions that they've implemented a driver for Mesa that has equal or better performance on ARM than the proprietary one from Qualcomm.
It's not clear to me what you're attempting to convey by saying the Steam Deck being the only product they have that supports the open source vision. The Steam Deck is the only new hardware product they've had since 2019, when they released their original first party VR headset that presumably is being replaced by the new one. Other than that, the only other hardware products they've ever worked on were earlier headsets made by other manufacturers or the previous iterations of the other two products announced alongside the new headset. From that standpoint, you could make a credible argument that the only product they even have right now that benefits from the open source work they've done in the past six years they did is the exact one you say supports this vision.
Steam Deck is not free software, is it?
The repo[0] is basically an issue tracker and the hardware is not open either (but they're repair-friendly which is already an improvement over... everything else.)
What do you mean by "open source the Machine"? Valve has stated its a regular open PC. The whole driver stack is open.
The hardware, like they did with the Steam Deck.
I guess you mean the CAD files for the shell? I'm not sure that is the most important part but it would be nice.
Sorry my point is that the Steam Deck is the only product of theirs that really supports beneficial open source in software and hardware. If you don’t think fossing the case is enough then you’re making my point for me that buying the machine, frame, or controller doesn’t do anything for foss.
It’d be like donating to Mozilla and expecting the money to go to Firefox development.
Valve paying for the development of KDE, Wine and adjacent projects is beneficial for FOSS.
I wonder if they(valve) sponsor servo
Igalia is a superhero company doing a lot of great work with surprisingly little fanfare.
Everytime their name pops up it's inevitably "oh some thankless extremely technical low level work leading to impressive/long-awaited features"
Indeed, their work on WebKit, Servo, Mesa drivers, the kernel, and more is seriously impressive!
Their customers, Valve, in this case, deserve credit for being good FLOSS citizens (even if they are building a DRM walled garden on top of it :/), but the actual workers are the real unsung heroes. Them, Codethink, Collabora, and other open-source consultancies I might have missed are doing the community a huge service."
You can ship DRM-free games on it just fine. It's up to the dev/publisher.
Additionally you can get a lot of the benefits of Steam (Proton etc.) even for titles you didn't acquire through Steam - you can add and launch third party executables through the Steam client.
Steam is not exactly a walled garden save for some rather light curation of their own store.
Steam DRM is entirely optional. Blame the publishers for DRM.
Valve doesn't disclose ahead of purchase whether a title has Steam DRM or not. So even if publishers don't take the option, I have no way to know that. Which means the option effectively doesn't exist.
The Steam Frame shows a lot of promise in terms of letting people play games on a massive virtual screen. But with the hardware, even more is possible. I hope they are working on a compatibility layer that allows 2D games to be rendered in 3D, like the 3D TV of the 2010s. In my opinion that would be a killer app.
There are rumours that they are working on this, but I assume they've chosen to keep the exact software experience of the Frame under wraps for now. It would certainly make the experience of gaming on a giant virtual screen even better!
VITURE's Immersive 3D already offers this for several platforms (for VITURE glasses).