Vulkan Present Timing Extension With Aim To Avoid Stuttering

[phoronix]

Phoronix: Vulkan Present Timing Extension With Aim To Avoid Stuttering

The Vulkan System Integration working group has decided to publish their work-in-progress extension on VK_EXT_present_timing as their effort to expose the presentation engine's display details and better allow scheduling a present to happen at a specific time...

http://www.phoronix.com/scan.php?page=news_item&px=VK_EXT_present_timing

[Gusar]

mpv devs should find this useful.

[artivision]

Most monitors flushing during big change on the framerate anyway. Such a useless technology.

[ms178]

Is this effort related to Keith Packard's work sponsored by Valve?

[Zan Lynx]

Most monitors flushing during big change on the framerate anyway. Such a useless technology.

You might need to rewrite that because I don't think anyone knows what you meant to say. Flushing? And what exactly is the "useless technology?"

[shmerl]

Wayland compositors should become more Vulkan/WSI focused and less tied to OpenGL/EGL.

[mercuriete]

Is this effort related to Keith Packard's work sponsored by Valve?

Just to remember that amazing speech....

https://www.youtube.com/watch?v=ihWLqQciDEg

and...

https://keithp.com/blogs/present-period/

and...

Alen Ladavac, Croteam speech

https://www.youtube.com/watch?v=n0zT8YSSFzw

So I think (correct me if I am wrong) is an agnostic specification of VK_MESA_present_period.

I love how vulkan is getting better and better

[ssokolow]

Most monitors flushing during big change on the framerate anyway. Such a useless technology.

Taking a quick glance through, it looks like it's meant to provide information about VRR/FreeSync/GSync timing to the game engine so it can fine-tune what it renders to avoid juddering.

Have you never written even a little toy game loop in something like SDL? The purpose of this API is equivalent to when you realize you need to replace your sleep calls with something like clock and some calculation to adapt for how much time has passed since the previous frame.

[ms178]

Just to remember that amazing speech....

https://www.youtube.com/watch?v=ihWLqQciDEg

and...

https://keithp.com/blogs/present-period/

and...

Alen Ladavac, Croteam speech

https://www.youtube.com/watch?v=n0zT8YSSFzw

So I think (correct me if I am wrong) is an agnostic specification of VK_MESA_present_period.

I love how vulkan is getting better and better

Thanks for listing all of these in one place, I was thinking on these items while reading this news. But it is still not clear to me if this is Vulkan extension is based on or inspired by Keith's work or if it is a different approach from a different group of people solving the same problem. I haven't found his name yet in the contributor section so it seems to be a different approach.

[rawr]

That goal sounds really useful. Hope they can make it happen.

It seems also that the right people might come together to define something proper.

Getting good and consistent latency can get tricky quickly. I think Android, albeit much better now, still has issues with audio latency for many reasons but mainly too much buffering.

I think it's a good thing that the Vulkan group moved this process into the public.
Sounds like a promising way in general as long as it doesn't inhibit their ability to make decisions.