https://bugs.kde.org/show_bug.cgi?id=489952
--- Comment #44 from pallaswept <[email protected]> --- (In reply to Zamundaaa from comment #43) > Some jitter is normal - while I also didn't expect it at first, the duration > of each refresh cycle doesn't seem to be perfectly stable. We have code in > KWin to deal with that, and re-schedule frames when a new timestamp comes in. > > For vkcube, do keep in mind that afaik it does not use presentation time to > get the actual presentation timestamps from the compositor. > What you want to use for testing this is weston-presentation-shm. It should > show a p2p that's stable down to a few microseconds. On Nvidia, it might not > be as stable as that, but it should definitely not fluctuate by multiple > milliseconds. Thanks mate very helpful info! I took a bit of a deepdive on this and was starting to feel like it wasn't even possible, and then saw news of VK_EXT_present_timing and realised that it wasn't just me who was trying to solve this problem :D I'm sure glad there are better devs working on it than me! Kwin has been behaving very well in the meantime, so thanks again! -- You are receiving this mail because: You are watching all bug changes.
