in development Vulkan API Discussion

Just my opinion, but I can’t see it making any difference. Bearing in mind the complexity of the game, I struggle to see ED devoting what would almost certainly be significant development costs for what is most probably a very small player base.

 

Let me put it another way, would you want them to stop work on Vulkan, the standalone server and a good deal of optimisation to allow ED to redeploy those resources, as that’s the probable outcome, particularly so as it’s not a one off task, there would be a need for ongoing support and development

Well with the question being if Vulkan would extend platform support, there would be no reason to stop work on Vulkan. While no one knows absolutely outside of ED, sticking with DX would mean absolutely no, where as Vulkan makes it plausible even if still unlikely. Vulkan would probably make it easier to run DCS in wine though with less of a performance hit.

Apologies if this has been asked before but would this mean a Mac/Linux build(s) is more likely?

According to the Steam Hardware Survey those usage rates are:

OSX 2.87%

Linux .69%

I would imagine there are many more important priorities for DCS than chasing such a small market.

According to the steam stats there are currently 352 people playing DCS World (Steam edition)... so...

Its not always about statistics but I get your point (which is of course valid).

I would be very happy about Vulkan inclusion because this bring us a little step further to platform independency. There are currently 2 games which prevent my PC to go full Linux. One of these is DCS.

I am a big Linux fan (I use Ubuntu for business and work, Windows only for gaming), and even from that viewpoint I rather see ED focusing 100% on Windows instead.

However, I would love to see a dedicated MP server based on Linux, that's command line driven without the overhead of a Windows-based, GUI driven interface. That's the only viable use-case for Linux and DCS imho.

According to the steam stats there are currently 352 people playing DCS World (Steam edition)... so...

 

 

Its not always about statistics but I get your point (which is of course valid).

I would be very happy about Vulkan inclusion because this bring us a little step further to platform independency. There are currently 2 games which prevent my PC to go full Linux. One of these is DCS.

The question isn’t how many DCS owners are on Steam but what percentage of the gaming demographic runs those particular OS. The Steam Hardware Survey is certainly reflective of the gaming market as a whole considering the majority or a great proportion of games today are sold on Steam.

The question isn’t how many DCS owners are on Steam but what percentage of the gaming demographic runs those particular OS. The Steam Hardware Survey is certainly reflective of the gaming market as a whole considering the majority of games today are sold on Steam.

Its still a broken metric, because of the perspective you're looking at it from. Not every gamer likes the ideology of Steam, and hardcore Linux users are even more likely to be that kind of user. Then there is the fact that so few AAA games have Linux support, that why would even bother to install it on Linux. The steam survey doesn't check to (or have the means to) see if a person dual boots between Windows and Linux, so they're excluded if they haven't installed the client on Linux.

Its still a broken metric, because of the perspective you're looking at it from.

It’s a large enough sample to be valid and relevant to the gamer market. Even looking at the OS market share as a whole, Linux represents only about 2%. It’s almost certainly not worth EDs effort to pursue such a small base at the expense of other things they could be working on.

Steam is apparently going to make all its games Linux compatable anyways

https://arstechnica.com/gaming/2018/08/valves-steam-play-uses-vulkan-to-bring-more-windows-games-to-linux/

Probably this is veering off topic...

It’s a large enough sample to be valid and relevant to the gamer market. Even looking at the OS market share as a whole, Linux represents only about 2%. It’s almost certainly not worth EDs effort to pursue such a small base at the expense of other things they could be working on.

Steam is apparently going to make all its games Linux compatable anyways

https://arstechnica.com/gaming/2018/08/valves-steam-play-uses-vulkan-to-bring-more-windows-games-to-linux/

 

Probably this is veering off topic...

I already know about that initiative, but they aren't quite there yet. DCS using Vulkan in the first place may largely remove the need for ED to build a native Linux client. ED are already changing their DRM scheme to something that might work seamlessly via things like Wine/Steamplay.

If Valve was relying on that metric themselves, then why would they bother investing the effort to make Linux a viable option. The market seeming to not be there is largely influenced by the fact that the games aren't there.

Apologies if this has been asked before but would this mean a Mac/Linux build(s) is more likely?

No Because DCS is still compiled using Windows .NET Framework / Microsoft Visual C++ Runtimes.

They should at least consider the ded. server for Linux. It lowers the cost to rent a rig in a DC.

If Valve was relying on that metric themselves, then why would they bother investing the effort to make Linux a viable option.

Who knows. It doesn’t really seem to make sense. Valve would boosts its sales at the most by 1.5%. Maybe 1.5% of Steam sales is actually a lot of money and worth the cost. But such math wouldn’t make sense to ED spending anything to grab .59% of the market.

The market seeming to not be there is largely influenced by the fact that the games aren't there.

People choose an OS based upon more than gaming.

 

People choose an OS based upon more than gaming.

My wet dream is bootable DCS-OS based on Linux. Streamlined for the sole purpose of serving DCS, in a true client-server scenario.

Steam's Linux is very close to this idea, I would just take it further and exclude anything it doesnt need.

Mecrutio - Vulkan's mgpu will support non-similar gpus. Skatezilla's statement :

 

https://forums.eagle.ru/showpost.php?p=3411910&postcount=79

Skatezilla's statement was made around a month before the article was posted. I should find a recent article and see if Skatezilla's statement still rings true.

Who knows. It doesn’t really seem to make sense. Valve would boosts its sales at the most by 1.5%. Maybe 1.5% of Steam sales is actually a lot of money and worth the cost. But such math wouldn’t make sense to ED spending anything to grab .59% of the market.

 

People choose an OS based upon more than gaming.

Your last statement, while obvious, completely makes my point. Barring the fact that most people don't actually choose an OS in the first place. They realize they need/want a computer, decide on it being a PC (considered synonymous with Windows) or a Mac, and accept whatever their pre-built machine comes with. There are way more linux users than what one can gather from the Steam survey. However being focused on those that would be gamers (the point of the debate), some of them are likely to be gamers with either a complete dislike of Steam, or zero desire to bother with it in Linux. There are many gamers who's only attachment to windows is the absolute reliability of game compatibility/support. These gamers aren't showing up as linux users in spite of their interest because they haven't bothered to even install Steam in Linux when there is only a handful a worthwhile titles guaranteed to work. This is the exact reason (well one of) Valve is bothering. Heck its why in the past Microsoft did what it could to hamper the adoption of OpenGL. The market is there in spite of what the Steam survey would tell you. Canonical alone claims to have at least 20 million Ubuntu users.

Your last statement, while obvious, completely makes my point. Barring the fact that most people don't actually choose an OS in the first place. They realize they need/want a computer, decide on it being a PC (considered synonymous with Windows) or a Mac, and accept whatever their pre-built machine comes with. There are way more linux users than what one can gather from the Steam survey. However being focused on those that would be gamers (the point of the debate), some of them are likely to be gamers with either a complete dislike of Steam, or zero desire to bother with it in Linux. There are many gamers who's only attachment to windows is the absolute reliability of game compatibility/support. These gamers aren't showing up as linux users in spite of their interest because they haven't bothered to even install Steam in Linux when there is only a handful a worthwhile titles guaranteed to work. This is the exact reason (well one of) Valve is bothering. Heck its why in the past Microsoft did what it could to hamper the adoption of OpenGL. The market is there in spite of what the Steam survey would tell you. Canonical alone claims to have at least 20 million Ubuntu users.

And by last count (by Business Insider, for example) there are 1.25 billion Windows OS in the world. So even if triple the 20 million Ubuntu users to 60 million to account for other distros, that's 4.8% marketshare.

NetMarketShare says that Linux accounts for 1.8%.

So again, I can't see anyone going after a TAM of 1.8-4%.

This isn't a Linux is better than Windows is better than Mac. I've used MacOS 6, LISA, the VERY FIRST Linux slackware from sunsite.edu, Coherent (a variant of Unix), BSD, AIX, HP-UX, and of course Solaris. Hell I even had to run SCO when they were busy suing people.

And I've used DOS1 to 6.22, Windows 1 (yeah, it was a runtime to run a game if you can believe that) to Win10, and I've run OS/2 and Warp.

I'm only stating the above to show that I'm not an OS bigot. There was a time when certain programs were only available to a certain OS. Pagemaker to Mac for example.

But no one can credibly expect a game company to go after the Linux market. There's just no money there. The TAM is way way too small.

Oh no, not windows vs linux again…

Oh no, not windows vs linux again…

It's not. The "vs" would imply there is a battle. And math tells us that that's not going to happen. Not when Linux holds single digit marketshare. It's not about Linux being a better or worse OS. It's about marketshare and nothing more.

And by last count (by Business Insider, for example) there are 1.25 billion Windows OS in the world. So even if triple the 20 million Ubuntu users to 60 million to account for other distros, that's 4.8% marketshare.

 

NetMarketShare says that Linux accounts for 1.8%.

 

So again, I can't see anyone going after a TAM of 1.8-4%.

 

This isn't a Linux is better than Windows is better than Mac. I've used MacOS 6, LISA, the VERY FIRST Linux slackware from sunsite.edu, Coherent (a variant of Unix), BSD, AIX, HP-UX, and of course Solaris. Hell I even had to run SCO when they were busy suing people.

 

And I've used DOS1 to 6.22, Windows 1 (yeah, it was a runtime to run a game if you can believe that) to Win10, and I've run OS/2 and Warp.

 

I'm only stating the above to show that I'm not an OS bigot. There was a time when certain programs were only available to a certain OS. Pagemaker to Mac for example.

 

But no one can credibly expect a game company to go after the Linux market. There's just no money there. The TAM is way way too small.

My first computer was TRS-80 Coco, have used every version of Windows since 3.11, and used Apple/Mac OS since at least OS 7 (between school and 15+ years in IT). Didn't really get into Linux until Gentoo came into existence, but have used various distros since. I only currently use linux in brief instances for specific purposes. My only reasons for using Windows is familiarity for work, and gaming (previously used Media center for DVR before it was deprecated). I'm on Windows 10 (besides work) because of DX12. If not for gaming, I would likely only run it in a VM.

I'm playing less and less games lately outside of DCS. My argument is mostly about how flawed using the steam survey to disprove Linux viability because its tainted data. Even with DCS being on Steam, many in the community warn other away from investing in DCS on the platform. Many DCS players aren't really gamers, they're simmers. I'm well aware that Linux is a relatively small market, yet the world's largest computer games distribution service provider (who's statistics are being considered meaningful on the topic) is doing everything they can to support it. Steam only has had a peak concurrency of 18.5 million users. Does that mean there are literally no more than that many PC gamers in the world at any given time? I'm gonna go with no, but Valve is possibly considering their only attachment to Windows is gaming. That 1.25b Windows users includes a very large pool of non-gamers, and EOL versions that current games don't support. Its largely business machines, which is by far largest group actively choosing Windows specifically, still not gaming. A lot of people didn't want to upgrade to Windows 10 when Microsoft was literally giving it away. The handling of 10's various "features" have been very unpopular, and Windows as a service (when it comes) is not likely to be any more popular (for the end user). The PC gaming market is basically made up of companies fighting over scraps anyway. The most successful video game pretty much ever (COD:MW3) sold just over 30mil copies, but a fraction of which were for PC. While the most successful PC game (Sims 3) has sold 7.96mil. It might be more of a disservice to small developer with a niche product to completely ignore Linux. I know a certain mil-sim has chosen to support Linux.

Yeah, again, in the mean time discussions and questions about Vulkan's features gets buried in the :argue: :doh:

I have a solution though,

I created a Windows vs Linux thread... :D

They talking about Vulkan in that thread?

Good question, I am wondering the same thing.

Another question, running along the same lines, and Vulkan API related, will Vulkan's mGPU functionality be supported by SteamVR/DCS World with VR?

:book: :detective:

OMG. The Gods of Simulated Aviation are watching.

Posted today, Vulkan® 1.1.86 - A Specification

Not for the feint of heart. Heavy reading and lots of it. Script Kiddies are required to wear bibs to help prevent the drool from disturbing others.

I posted this the moment I found it. It could give clues. It could be BS. All I know is that I'd rather read ISO standards.

Edit #3; Chapters 7 - Renders, 8 - Shaders and 9 - Pipelines...... where's that barf emoji.... no for real... I admit right here and now I'm not reading this. This could be more for reference.

Variable rate shading is also supported by Vulkan. Massive news for VR users with Turing cards! One developer said it was a dead easy to implement as well.

Variable rate shading is also supported by Vulkan. Massive news for VR users with Turing cards! One developer said it was a dead easy to implement as well.

Yea, this is the sort of stuff I am waiting to see from DCS before I shell out the cash for one of those cards. So far reviews have (to no fault of their own) not been able to take full advantage of the new cards since there aren't really any games that support things like DLSS and RTX.

Variable rate shading is also supported by Vulkan. Massive news for VR users with Turing cards! One developer said it was a dead easy to implement as well.

Dead Easy to Implement in their engine doesnt mean Dead Easy to implement into all engines.

As a VR user desperately needing FPS, I just read through this thread. I'm really dumb when it comes to IT stuff. So as far as I understand, a new APi is coming that will make DCS much, much less GPU needy. I mean the API is out, it's DCS that will start using it. That's great.

But tell me something: everyone keeps saying DCS doesn't use multicore CPUs well, so those always end up being a bottleneck. (I have a Ryzen 1600x). Will Vulkan make DCS easier on the CPU too? What exactly will get better once it's implemented?

PS: Will I have to download/ install this Vulkan thing? (sorry)

The way I understand it is, that Direct X is heavy on draw calls to the cpu. So basically, before the GPU renders anything, it needs an OK from the cpu. So its constantly waiting on the cpu to give it a go ahead to render all that a frame consists of. So when the cpu is stressed, the GPU will end up waiting on the ok to proceed, and thats when we see low GPU usage.

Vulkan remedies this, by GPU not having to ask the cpu for permission all the time, so the GPU is more independent and can just run with the ball. It still needs some synchronization though.

If it'll be all that we hope it will be and free up vast computation resources, well, guess we just have to wait and see.