shmerl

Hi!

How do you see Vulkan affecting Linux graphics stack in the light of OpenGL → Vulkan transition? For example will it fit into Mesa project or produce something new? Also, what can happen with APIs like EGL which assume OpenGL stateful architecture, when Vulkan presents new stateless approach? Will EGL have to be replaeced with Vulkan based alternative? And how might that affect projects like Wayland?

Thanks!

The Vulkan window system interface is still being discussed.

That's interesting.

OpenGL will live on...it serves a key purpose. EGL will live on with OpenGL.

What purpose do you see for OpenGL in the future besides supporting legacy applications? I got an impression that Vulkan proposes new application design that is geared towards increasing hardware parallelism, while OpenGL is restricted by design in that regard. So I'd assume that for any new projects it's advisable to use Vulkan (and build whatever is needed on their own on top of that). I.e. no matter whether it's games or graphics servers and windowing systems. Or you consider OpenGL useful for future projects too in some cases?

On mobile every bit of performance is significant because it increases battery life. So Wayland will care I hope.

My point is that OpenGL is still not ideal if something can be higher level than Vulkan while not compromising potential parallelisms capabilities. Of course simply nothing like that exists yet.

I guess most core projects like Wayland do care about performance enough to go with Vulkan route.

That makes sense, on the other hand engines and projects which care about performance will probably want to avoid limitations of OpenGL. May be some middle layer can emerge which can provide higher level abstraction than Vulkan and address some common needs, but on the other hand won't be restricted by OpenGL design?

Using OpenGL just for the sake of simplicity sounds like poor man solution in this case because of the lack of alternaives (since major tradeoff is performance).