general optimizatiin & tuning & profiling on an open source scarcely funded engine, is my first second third & fourth guesses. just a little bit of technical assists can often do wonders in open source. that expert look, is, alas, uncommon.
would be freaking sweet if they'd buy the team some pizza, beer, & coffe but hey we cant expect the world. that'd be basically unprecedentedly nice to open source.
In the console space it is really common to have the platform loan out devs for onboarding to their platform. Maybe they are just providing consultation on common pitfalls when it comes to their architecture.
When the description "just a PC" is applied to Steam Deck, it usually implies not just the x86/IBM-PC-descendant hardware configurations, but also the lack of strict software/hardware lockdown measures to prevent the use and execution of unapproved sw/hw. PS4/PS5/XB1/XSX are indeed similar to PCs hardware wise, but from this definition of "just a PC," they really aren't.
Optimizations on the consoles involve things like adjusting memory access patterns, using gpu-specific intrinsics and hardware specific latency info etc to tune the game engine. It's likely the same with Steam Deck.
There are a lot of generation specific optimizations that are often even exposed on PC but don't make it into any games because games don't often optimize for each specific generation of hardware (unless it's something particularly notable like DLSS/RT). This, I have noticed, is especially common on AMD recently, where some optimizations are very useful on consoles and eventually exposed on PC too, but never actually used there because of their issues until this latest generation.
Valve has probably worked to integrate those into Godot, there's also the consideration for power management, perhaps Valve has exposed additional APIs to make it easier for games to exchange information regarding power availability etc, so while most games just run as if they were running on a PC, there's the possibility for the game to also be Steam Deck aware/optimized.
I don't use Godot, but I don't think there's an official Steamworks integration.
This is speculation, but I think this means they're working on an official module, and it'll come with support that ensures all controller bindings pass through nicely, etc.
That's right you have to merge some code from GitHub and recompile. Which thankfully is a 30' affair unlike UE4 where it's more like a half day wasted.
It’s a Linux pc.. so top notch Linux support I’m assuming is part of it. Plus it has a kind different control system and whatever graphics and sound it has as well.
There were claims that Valve asked some developers to remove their Linux "native" versions of the game since the Proton version is better since the devs didn't support the "native" version very well.
Allowing developers to outsource the work for supporting Linux makes game developer's jobs much easier.
Asking some developers to remove a native version of their game is not the same as asking all developers to target Proton. If anything it's more an indictment of those specific dev's ability to target native Linux.
That doesn't contradict what they said, it reinforces it even:
> Really, it comes down to whatever is the best experience. So if it's easier for the developer to get to a point where the best experience is achieved through Proton we think that's great
Engines can always benefit from optimizations and improvements. They can fine tune it for usage on their APU, improve their Vulkan renderer and what not.
I can imagine at least a "hello world" project with all the settings configured to work out of the box on Steam Deck would be a start. Maybe they have some custom libraries to hook the game in to the OS, and they want to give developers a running start with menu systems etc.
Well, they could be just talking about ensuring Godot-based Windows games running on Proton without any Linux-specific compatibility issues, of course...
Not sure about lots. I got the impression that intersection between PC and mobile games is not that big. And Valve are focused and PC games proper (in this case running on Linux obviously).
If anything, lately less games come out for Linux natively on Steam or GOG may be due to relative decline of marketing push until recently. But now with Steam Deck Valve are a least putting more weight behind marketing again, which is good.
No, the way to drive adoption is to improve wine / dxvk. Allowing developers to target a single set of APIs is key to making it easy for games to work and not rot from developers not testing on Linux.
The effort is on the developer, targeting Proton means the developer writes the code once and can target Windows and Linux. If Valve wants to be the big gaming platform for PC, the best they can do it improve Proton/Win32, and make it the official supported and preferred platform for their store games.
Native Linux is too much of a moving target, and the fact that Win32 is so focused on backwards compatibility makes it a perfect platform to build games upon.
I am a Linux gamer and I'd rather play the Proton version than chase missing libraries and broken audio which often affects bad native Linux ports. And the better Proton is, the longer we'll be able to play old games that are no longer compatible with newer Windows versions.
In a world where tech companies keep disappointing me, Valve continues to be a force for good. Probably because it's still a private company. Publicly traded companies can't see anything further than balance sheets.
this is nice. now amazon is looking at luna for linux. the idea is, people can build games, buy games and microsoft should not be in the loop. they have no justification for forcing people to lock in their customers and the endless loop.
i have been a fan of the whole FOSS movement and never played any games because i don't have a GPU to game on. lol. anyways, this is good for people too, when more people start to not need windows, software developers who have traditionally always relied on microsoft tech stack will have to look around at alternatives. all in all, this looks promising
People can build games, buy games for Android and ChromeOS, both Linux based, without Microsoft on the loop, and it hasn't made GNU/Linux magically get those games.
The problem that GNU/Linux doesn't get games isn't Microsoft.
you know there is a good project called waydroid? anyways, what i mean is when you are talking about amazon level gaming streaming service or steam os, earlier they were forced to keep microsoft in the loop, both in development as well as in end users. now if that monopoly is removed, a game developed for windows works on linux, there is no forcing of the end users to use windows.
like in software. suppose tomorrow adobe releases all its software on linux. a lot of people who were forced to stay on windows BECAUSE of adobe would be free to switch, giving them more choice. right now, there is none.
> software developers who have traditionally always relied on microsoft tech stack will have to look around at alternatives.
Not necessarily. There will be Windows Subsystem for Linux, and Xamarin/MAUI proves that multiple targets with a .net codebase are very possible (desktop win/mac, ios, android). Xamarin/Maui and .net are all OSS.
A MS supported target for linux via Maui is being discussed; and one of the MAUI developers already has a repo to support linux as a target that may possibly become an official branch. https://github.com/jsuarezruiz/maui-linux
There is also the .net Community. Uno Platform (FOSS) & Avalonia (FOSS) are both .net based and target Linux and much more. The code is similar to programming for Xamarin/Maui, you can share most of the core code of your app between these targets.
For Gaming: Unity and Monogame are OSS, Unity and Unity's editor is on Linux. MonoGame (OSS) (stardew valley) targets linux. Stride (3rd party engine) is also .net and targets Linux. Cryengine is also usable from .net.
And last but not least: Godot can officially be programmed using .net.
I am not worried as a "windows software developer" that I would be needing to switch stacks to target Linux. I can already make games or software target Linux from .net.
oh. you are missing my point. cross compatibility is great but you know how microsoft charges datacenter licenses to vps providers? or the small premium in prebuilt new laptops and desktops? that is what i am concerned with.
you are right in saying the stack is cross compatible. that is great for developers but if users are stuck with windows for any reason
A direct shot across the bow of Microsoft. Also against Epic Games and Unity if Valve wants to prevent them from growing into distribution platforms of their own.
This won't deter Amazon, Google, and Nvidia, who are all attempting to run game engines in the cloud. Their strategy is to remove consoles and PCs entirely and have you subscribe to streamed content forever. Latency and partnerships are still the biggest issues for them.
This is shaping up to be the next major battle of the tech wars.
it's likely that godot steals more customers from unity than UE, and it makes strategic sense to fund a competitor to unity, even if it hurts your own EU a tiny bit (if any realistically, since UE targets a different customer base).
let them run their game engines in the cloud but they wont be paying through their nose to microsoft if they use proton/wine. the cost savings might justify their fixing bugs in wine/proton which is good for everyone
49 comments
[ 2.6 ms ] story [ 83.0 ms ] threadwould be freaking sweet if they'd buy the team some pizza, beer, & coffe but hey we cant expect the world. that'd be basically unprecedentedly nice to open source.
There are a lot of generation specific optimizations that are often even exposed on PC but don't make it into any games because games don't often optimize for each specific generation of hardware (unless it's something particularly notable like DLSS/RT). This, I have noticed, is especially common on AMD recently, where some optimizations are very useful on consoles and eventually exposed on PC too, but never actually used there because of their issues until this latest generation.
Valve has probably worked to integrate those into Godot, there's also the consideration for power management, perhaps Valve has exposed additional APIs to make it easier for games to exchange information regarding power availability etc, so while most games just run as if they were running on a PC, there's the possibility for the game to also be Steam Deck aware/optimized.
They are not IBM pc compatible.
But for the most part once you get everything working... The difference is negligible.
What I am getting at is: it is close to a Pc but not. In the ways that matter it is (CPU GPU pcie USB data etc)
Go watch the talk, it's over at media.CCC.de I think at congress 33.
Edit: spelling and grammar
This is speculation, but I think this means they're working on an official module, and it'll come with support that ensures all controller bindings pass through nicely, etc.
Here they state that they have no preference: https://www.gamingonlinux.com/2021/11/valve-answers-the-ques...
Allowing developers to outsource the work for supporting Linux makes game developer's jobs much easier.
> Really, it comes down to whatever is the best experience. So if it's easier for the developer to get to a point where the best experience is achieved through Proton we think that's great
If anything, lately less games come out for Linux natively on Steam or GOG may be due to relative decline of marketing push until recently. But now with Steam Deck Valve are a least putting more weight behind marketing again, which is good.
i applaud this move!
more effort should be put into cross compilation toolchains, including abstracted graphics API
more widely accessible, crossplatform libraries and APIs is the way to go!
targeting linux with godot/unity is a matter of a click
it should be way more encouraged!
valve, next step is a xplatform, ondemand build farm! with a lot of different GPUs and OS, so people can submit and test their game with ease!
if want to keep 30% then invest!
No, the way to drive adoption is to improve wine / dxvk. Allowing developers to target a single set of APIs is key to making it easy for games to work and not rot from developers not testing on Linux.
that makes you non-future proof, and non resilient to failures or niche cases
wine/dxvk is a transition path, a way to supplement your library
look at history, apple did it very well with Rosetta, aid people to transition to the new platform
nothing can beat the ease of use of native product, built for the platform
30 GB SteamOS, something that could have just been few MBs if native linux target was promoted MORE by valve, just think about that ;)
Yes, you can. A "native" port of Among Us would literally not be better than using proton to run it.
with proton you need: proton, wine, dxvk, and bunch of libraries
that already is 10x the size of the game itself
there is also a setup time cost, it's not good for the user having to wait ;)
Native Linux is too much of a moving target, and the fact that Win32 is so focused on backwards compatibility makes it a perfect platform to build games upon.
I am a Linux gamer and I'd rather play the Proton version than chase missing libraries and broken audio which often affects bad native Linux ports. And the better Proton is, the longer we'll be able to play old games that are no longer compatible with newer Windows versions.
i have been a fan of the whole FOSS movement and never played any games because i don't have a GPU to game on. lol. anyways, this is good for people too, when more people start to not need windows, software developers who have traditionally always relied on microsoft tech stack will have to look around at alternatives. all in all, this looks promising
The problem that GNU/Linux doesn't get games isn't Microsoft.
like in software. suppose tomorrow adobe releases all its software on linux. a lot of people who were forced to stay on windows BECAUSE of adobe would be free to switch, giving them more choice. right now, there is none.
Not necessarily. There will be Windows Subsystem for Linux, and Xamarin/MAUI proves that multiple targets with a .net codebase are very possible (desktop win/mac, ios, android). Xamarin/Maui and .net are all OSS.
A MS supported target for linux via Maui is being discussed; and one of the MAUI developers already has a repo to support linux as a target that may possibly become an official branch. https://github.com/jsuarezruiz/maui-linux
There is also the .net Community. Uno Platform (FOSS) & Avalonia (FOSS) are both .net based and target Linux and much more. The code is similar to programming for Xamarin/Maui, you can share most of the core code of your app between these targets.
For Gaming: Unity and Monogame are OSS, Unity and Unity's editor is on Linux. MonoGame (OSS) (stardew valley) targets linux. Stride (3rd party engine) is also .net and targets Linux. Cryengine is also usable from .net.
And last but not least: Godot can officially be programmed using .net.
I am not worried as a "windows software developer" that I would be needing to switch stacks to target Linux. I can already make games or software target Linux from .net.
- edited, clarity
you are right in saying the stack is cross compatible. that is great for developers but if users are stuck with windows for any reason
This won't deter Amazon, Google, and Nvidia, who are all attempting to run game engines in the cloud. Their strategy is to remove consoles and PCs entirely and have you subscribe to streamed content forever. Latency and partnerships are still the biggest issues for them.
This is shaping up to be the next major battle of the tech wars.
https://www.gwern.net/Complement