OpenGL is badly in need of replacement inside Windows
There was an old saying about OpenGL; it was the best choice when you had no other choice :wink:
--but let's face it--when it comes to AAA computer games that require decent horsepower to run--the Windows market is the place to be. Valve's splinter of Linux distribution is so small it is nearly non-existent--indeed
Indeed, the stats on the Steam Hardware Survey point to this; Win7 controlling just over 30% of the market and Win10 just under with total Windows dominance at 95% of the share.
(and that uptake of Win10 is why the "gamers wont have to update past Win7" argument people have made holds very little water imo).
Linux based OSes don't even make up 1% of the share.
Nothing in the mobile markets has the horsepower, and indeed, probably never will have the horsepower when compared to state-of-the-art desktop computer-gaming tech.
While true mobile is the one area where, potentially, Vulkan will be a massive win.
OpenGL|ES had quickly become a mess, in terms of the API, driver support and indeed the overhead on the runtime - this is part of why Apple made Metal and, assuming everyone falls in to line, should be something Vulkan solves.
(I spent just over a year supporting Android at my last place; at the start I was a card carrying Android fan, now you couldn't pay me to use an Android device.. hateful hateful things...)
As for the consoles... [...] So where does Vulkan come in?
This remains my question; I know some people have asked for Vulkan support on the PS4, I even heard potential of maybe happening, but I have to ask why?
The API layer is, without a doubt, the thinnest bit and the PS4's Gnm will look, from a high level point of view, like Vulkan - so you don't save complexity and indeed you'd still have to have extensions to allow for the hardware of the PS4 where it differs significantly from a PC. (Pretty sure some NDA stuff still exists on this, thus vagueness...)
And as I mentioned before, most places have a PS4 back-end up and running and have for some time now - over two years since console release and it was maybe 18 months before release I got to see the specs (I think, memory is a little vague..) - so the PS4 ship has largely sailed and for everyone else coming in now, well, Unity and UE4 are the solutions.
[source]
As I mentioned earlier, the critical balance for whether Vulkan makes it--and I hope it does--will be developer support and that is 100% dependent on the level of support the Khronos Group gives developers through its development tools. The market is not necessarily cross-platform at all, really--it's the Windows market primarily in which Vulkan will sink or swim (as Vulkan is introduced as a Windows API, specifically.)
[/quote]
You will have games with Vulkan support out there, UE4 has already demonstrated the API running (although there was nothing particularly magically 'Vulkan' about it) and Unity have in house work going on although I get the impression it isn't a priority and they have bigger fish to fry stability and performance wise.
(Heck, knowing what I know about the UE4 code base I'd be surprised if the performance delta was really that great as much of that code wasn't great when I last saw it...)
My own opinion is that Vulkan's adoption rate is probably going to parallel Mantle's--but that's a maybe.
I'm not so sure.
Mantle filled a need, a cross platform need amusingly, which with the release of D3D12 largely got swallowed up on Windows - if you are a big AAA developer then I'm not sure where your motivation is for using it.
The only, only, way I can think it makes sense is if you are starting out basically now and are writing your low overhead renderer from the ground up and, importantly, weren't planning on hitting the Xbox - then it might make sense. Otherwise, I'd still say D3D12 first - hits two platforms right off the bat.
Vulkan is being launched squarely against d3d12, and it will be the tools for developers primarily that will determine the ultimate success or failure of the Vulkan API.
The tools and plans for the tools look sane; the problem is it launched when it did; months behind Win10 and D3D12's public release.
Everything is still beta where as D3D12 has had some settling time now.
The good news for Vulkan is that D3D12 has probably found many of the underlying issues so things might go smoother.. but it was still late to the party and while not bad it hasn't brought enough with it to make everyone stop hanging around the host and hang with it instead.
I have nothing against Vulkan, but whenever I hear someone preaching a new API as if it's going to save the world I get a bit recalcitrant...