XDC 2019 conference live stream starting now.

1
Hi all,

it's beginning of October and the XDC 2019 X-Developers conference
2019 ist starting at 9 am today in Montreal, Canada. This is the main
conference about all things open-source graphics and display, ie.
about the components that make up the Linux graphics and display
stacks, but also touch on topics of potential relevance for the less
fortunate Windows users and the unfortunate macOS users. As usual,
many interesting talks and topics that are important for those
fortunate users of Psychtoolbox that can use Linux for superior
functionality. This year i can't be there, because money and time, so
i'll follow all important bits via the live stream.

For those technically inclined and interested, here's the schedule -
also with links to presentation material:

https://xdc2019.x.org/event/5/timetable/#20191002.detailed

The Twitter feed:

https://twitter.com/xdc2019

Most importantly the live stream which will hopefully start any moment now:

https://www.youtube.com/watch?v=rV0P1ChE_5o&feature=youtu.be

Some especially interesting bits:

"Now" - Zink - OpenGL on Vulkan.
The Vulkan WSI layer

Enjoy
-mario
2
Hi Mario, I saw the titles of a couple of the talks were focussed on frame display timing, and one of the speakers linked to this general-interest article explaining how lack of timing feedback in APIs can make even high-performance GPUs stutter, which may be interesting for PTB users:


What do you think is the most useful new developments for PTB users, e.g. could PTB benefit from Zink or other new frameworks?

Cheers, Ian
3
Thanks for that link! Interesting read.

I wonder if it applies to PTB users though. We schedule our flips and do not accept jitter? Or could it be that if flip timestamping is off, then our next scheduled flip misses a presentation deadline, thus causing stutter?

I tried animating a simple display at 240Hz with a pc and gfx card that should be by far powerful enough. Still got stutter sometimes. Just a student project, didn't care to get out of my "at a loss to explain this" state. Never considered looking at both requested and actual frame timestamps, i wonder if small errors start to matter when you have that little margin (just over 4ms). Also never tried to run on Linux with its superior timestamping (nvidia gfx card notwithstanding), perhaps that may shed some light on things.

Cheers,
Dee

On Fri, 4 Oct 2019, 16:05 iandol@... [PSYCHTOOLBOX] <PSYCHTOOLBOX@yahoogroups.com> wrote:


Hi Mario, I saw the titles of a couple of the talks were focussed on frame display timing, and one of the speakers linked to this general-interest article explaining how lack of timing feedback in APIs can make even high-performance GPUs stutter, which may be interesting for PTB users:


What do you think is the most useful new developments for PTB users, e.g. could PTB benefit from Zink or other new frameworks?

Cheers, Ian