Shader Compilation on PC: About to become a bigger bottleneck?

Yeah, and they're struggling for cash and really need budget carefully. Sometimes you can do something because it's Good, even if it doesn't turn a profit.
A commercial company doesn't do anything because it's "good". Neither does it do something because it has money from other projects to lose.
 
Why doesn't Steam have fossilize for Windows? Or does it have it for vulkan titles? Wondering if there's some technical reason why they can offer it for vulkan and not d3d12, or Valve is just keeping it exclusive for their handheld?

Microsoft should be able to do something, at least for their own Xbox store on pc. Fossilize is open source, so I'm not really sure why other companies aren't rolling their own for their own stores. Epic should be able to offer something similar for UE5 games on the epic game store.
 
Why doesn't Steam have fossilize for Windows?
1. D3D12 doesn't have a "plugin" mechanism which allows Fossilize to work on Vulkan ("layers").
2. Even if it would have something similar would Valve want to promote Windows over their own platform?

Or does it have it for vulkan titles?
It doesn't either. Dunno why, probably because there's like 5 games on PC using Vulkan.

Fossilize is open source, so I'm not really sure why other companies aren't rolling their own for their own stores. Epic should be able to offer something similar for UE5 games on the epic game store.
I suspect that the need to support the server backend for this have something to do with other companies not willing to invest into a similar solution.
 
1. D3D12 doesn't have a "plugin" mechanism which allows Fossilize to work on Vulkan ("layers").
2. Even if it would have something similar would Valve want to promote Windows over their own platform?


It doesn't either. Dunno why, probably because there's like 5 games on PC using Vulkan.


I suspect that the need to support the server backend for this have something to do with other companies not willing to invest into a similar solution.

2. I disagree with this one because I'd guess 99% of valve's sales are happening on Windows-based PCs. Unless they have plans to launch a Steam OS for desktop in which case I will install it right away lol.

I would think the epic store and xbox store have quite a lot of cdn capacity. I don't know. But if there is some technical limitation on the D3D-size where a system like fossilize can't work (eg lacking a plugin mechanism) than maybe Microsoft should fix that so the option to offer something like fossilize is viable.
 
A commercial company doesn't do anything because it's "good". Neither does it do something because it has money from other projects to lose.
But should it be that way? Is the world better off because they operate that way? Has it always been that way? Can it only ever be that way?

No to all of those. MS could choose to make it right just to give gamers a better experience and contribute happiness to the world as opposed to a better percentage to the shareholders. They could just be nice for nice's sake.
 
2. I disagree with this one because I'd guess 99% of valve's sales are happening on Windows-based PCs. Unless they have plans to launch a Steam OS for desktop in which case I will install it right away lol.
What's stopping you from installing SteamOS on PC though?

As for sales - people are buying games in Steam right now and the issues are (rightfully so) getting blamed on MS (for DX12 design) and game developers. Would Valve want to interfere here and become a target for complaints instead of MS/developers? Would developers just ignore shader compilation and delegate this to Valve leaving everyone else at a disadvantage?

Windows Steam client has been evolving mostly as a byproduct of their Deck efforts over the last several years which is why I don't think asking Valve to fix Windows is a very prospective idea.

I would think the epic store and xbox store have quite a lot of cdn capacity. I don't know. But if there is some technical limitation on the D3D-size where a system like fossilize can't work (eg lacking a plugin mechanism) than maybe Microsoft should fix that so the option to offer something like fossilize is viable.
So again it's a question of someone's incentive to fix others mess. MS will point you to developers not implementing the PSO processing properly - which would be correct in D3D12 model on PC. Epic will point you to both MS and developers. Inability to find one party who can and have the incentive to solve the issue is the reason why it remains unsolved. Ideally it should be MS of course but since they are a de facto PC gaming monopoly they won't bother unless some other platform will start threatening their dominance (pretty much like it was with D3D12 to begin with).

But should it be that way? Is the world better off because they operate that way? Has it always been that way? Can it only ever be that way?
It is the way it is. You may not like it but this won't change the way every company out there operate, and there are no idea on how this can better aside from making sure that no one company control some market completely. Which in tech has been failing quite spectacularly for half a century now.
 
@DegustatoR I don't think steam os for pc is fully baked yet. There is some kind of image, but I don't think it's really like the Steam Deck experience. You can also build your own linux box and install all of the packages to get your own Proton system going. Basically I'm waiting hoping that Valve will come out with an equivalent of the Steam deck os with more driver flexibility. Something a little more official. Maybe that will never come, but I'm hoping. They also need to get all of the anti-cheat systems working before I can switch.
 
Free upgrades since 2015...


MS gets $0 from games released in Steam (unless it's their own games of course).


Free API download.

As I've said MS doesn't have any incentive to help Valve of all possible players fix shader compilation stuttering.
Maybe if SteamOS would start actively eating up their Windows user base then there would be such incentive. But I don't think it's realistic to expect that.
By this logic MS would have never made DX12 since it doesn’t make them any money supposedly.
 
By this logic MS would have never made DX12 since it doesn’t make them any money supposedly.
Exactly. It's incredible to me that someone can't understand why it's in Microsoft's best interests to "put the work in" for DirectX and ensure it's maintained and has at the very least feature parity with other APIs. If it wasn't important to Microsoft, why create DX12 to begin with? AMD created mantle and Microsoft knew Vulkan was coming from its ashes, and so they rushed out DirectX12 to compete with it.. Why? DirectX is literally MS' way of keeping a tight grip on game development on Windows... that is EXTREMELY important to the entire PC ecosystem.. and all of the gaming companies who work with Microsoft.

Like I've said... nothing is going to get anywhere pointing fingers at others.. but it absolutely becomes Microsoft's issue when Valve and Linux can do DirectX12 better than Microsoft can and people start noticing it. Just like it's embarrassing for Epic and Unreal Engine when developers don't properly utilize all its features. It requires ALL GROUPS to do better. MS and its APIs, game developers actually utilizing features of said APIs, and platform holders.. PERIOD. We need redundancy because not every developer can devote the same amount of resources to the problem.. so for those edge cases, whatever can be done, needs to be.

And this is besides the fact that I literally asked people who work on DirectX about this specific issue about having an equivalent to Fossilize for DX and Windows, and they told me they are very aware of the issue and are working on it but couldn't say anything about what they're planning. Regardless.. what IS needed, is more people calling this stuff out. More people calling out publishers for releasing games which don't do the bare minimum, more calling out of engine developers for lingering issues which are persistent across vast amounts of titles... as well as platform holders for OS and API issues.

With enough bitching.. eventually the right people will hear the message and things will begin to change.
 
By this logic MS would have never made DX12 if not for Mantle and Vulkan both of which has threatened their position of PC gaming monopoly.
By that same reasoning they have plenty of motivation to fix this problem as Valve is coming up with ways to fix this on Linux with Vulkan translation.
 
By that same reasoning they have plenty of motivation to fix this problem as Valve is coming up with ways to fix this on Linux with Vulkan translation.
This isn't a motivation as they are not afraid of gaming demo leaving Windows for Linux anytime soon. If Khronos would fix this in Vulkan on Windows then this would be a motivation.
 
Microsoft doesn't care if PC gamers leave the Windows ecosystem?
They do but I don't think that anyone is leaving Windows for Linux just yet. Steam Deck is a very low volume device and is usually a secondary device in addition to a proper PC anyway.

hw752tji.png
 
The issue will if MS wait and wait until Linux gains momentum, it could be too late. If you get people changing to Linux enough to make MS take notice, by the time they've addressed their issues, there'll be all that existing motivation to switch. I think there are lots of PC gamers who'd be happy to move away from Windows if there was a viable alternative and the transition could rapidly snowball as gamers start switching to something SteamOS like and report it's a real positive. MS would then rush out their improvements and gaming dashboard etc., but if Linux is working, why go back to Windows?

In short, where the competition at the moment isn't any sort of competition and MS doesn't care, if it becomes valid competition, MS will struggle to compete. And they won't be able to just buy Linux to put an end to it. TBH I hope it happens. As iOS found, games drive adoption unlike anything else. A new gaming-focussed OS would open the door to a real Windows alternative and shake up the computing industry from its MS monopoly. The lack of software on Linux has really held it back. If it can not only play all the games Windows get, but actually play them better, that barrier is gone. Then it's a matter of usability. You'd have a real competition again, between MS trying to make Window's simpler and less kludged and good for gaming, versus 'Linux' trying to make the OS more accessible and easier to use for the everyman with more productivity etc.
 
I think there are lots of PC gamers who'd be happy to move away from Windows if there was a viable alternative and the transition could rapidly snowball as gamers start switching to something SteamOS like and report it's a real positive.
Sure but it's not happening now and I don't expect this to happen anytime soon. Windows power is in broad s/w compatibility which Linux doesn't really solve even with things like Proton. Lots of people aren't only gaming on their Windows PCs and for them moving to Linux wouldn't be an option. But generally speaking if some Linux distro (SteamOS? I'm not sure if it's as good for non-gaming stuff as other options) would make inroads into providing similarly broad s/w support then some people would look into switching - and if this percentage will become more than 10 MS may become worried.

I don't think that providing a solution similar to Fossilize would take much effort on their part. Changing the D3D12 model would probably be much harder - albeit less expensive in the long run probably. So as I've said it is unlikely that they'll bother until such threat would actually materialize. They certainly didn't with D3D12.
 
A 1 to 1 replacement won't happen in any foreseeable, but dual-boot certainly could. I'd be happy to have two power buttons on a case, one to boot into a Linux gaming machine and another for Windows for when I need that. Its more a case of 'Linux' getting its foot in the mainstream door. No-one is going to switch when their daily productivity is on Windows, but I think plenty might for a better gaming experience, and that's what would open the door for further developments.
 
Linux has too many sharp edges to totally take over, but it could skim a chunk of users. It’ll always be limited to people more willing to dual boot or learn a more manual operating system. I am one of those people. I just need anti cheat working
 
Linux has too many sharp edges to totally take over, but it could skim a chunk of users. It’ll always be limited to people more willing to dual boot or learn a more manual operating system. I am one of those people. I just need anti cheat working
Linux has become a lot less rough edged, if we saw compatibility improve (anti cheat in particular) I think it could work on a mass market scale
 
Linux has become a lot less rough edged, if we saw compatibility improve (anti cheat in particular) I think it could work on a mass market scale

I'm basically a daily linux user at work. It definitely still has a smoother experience for the most part, but you can get stuck in issues with particular programs that require digging into config files, restarting services etc. And then there are areas like audio that are generally a little more problematic. There are a lot of pc gamers that don't really know much about computers. They just buy a pre-built and play. Linux probably will not be a fit for them unless Valve(or some other game focused company) could really smooth out the experience and make it nearly foolproof.
 
Back
Top