Announcement

Collapse
No announcement yet.

GNOME On Wayland Lands Improved Handling For Direct Scanout Support

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #41
    Btw. funny how those measurements by Xaver Hugl busted fake news (lies) by Alexmitter and others (including oiaohm?) that Xorg would always have higher latency, except that numbers for immediate uncomposited Xorg are probably as low as even possible with 120Hz and such a testing procedure. So much make believe nonsense that people sacrifice their own integrity for, it's sad...

    Comment


    • #42
      Originally posted by aufkrawall View Post
      Btw. funny how those measurements by Xaver Hugl busted fake news (lies) by Alexmitter and others (including oiaohm?) that Xorg would always have higher latency, except that numbers for immediate uncomposited Xorg are probably as low as even possible with 120Hz and such a testing procedure. So much make believe nonsense that people sacrifice their own integrity for, it's sad...
      Read Xaver Hugl numbers again and take note of the ** bits this time.
      A considerable amount of people assume Wayland isn’t particularly suitable for gaming, usually because you can’t turn off the compositor. This post will challenge that assumption and see how the current state of gaming on Wayland is, with a focus on KWin, KDEs compositor.


      fifo on X11 with compositing is basically shot to hell. The fifo on wayland still had a extra buffer in it. When you subtract a buffer worth of milseconds there is a serous chance that fifo on X11 server without compositor will be slower than wayland and xwayland.

      Xaver Hugl numbers made those claims about wayland false it self is a false claim. The difference between mailbox and freesync is too close to measure between Wayland, Xwayland and X11 without compositing.

      Fifo there is a issue that will be fixed that will will result in either X11 without composting losing or being too close to measure.

      Yes read Xaver Hugi stuff closely you will notice that the Wayland immediate is modified mailbox that suggest that Wayland mailbox mode can go faster than what is.

      Xaver Hugl is testing the existing wayland implementation. The ways the wayland protocol works in theory it should either match of exceed x.org in performance in everything bare immediate mode. Xaver Hugl numbers suggest that the case other than area where there are issues.

      Proper work to support for immediate could get a lot closer than Xaver Hugi hacked up patch.

      Really go back and read those numbers way more closer and the notes on those numbers of Xaver Hugl. Remember X11 server on bare metal is fairly much tapped out for optimisations.

      Remember tear free is complete on in all the wayland things test. Xaver Hugl tested without tearfree turned on with the drivers in X11. Reality here you turn tearfree on yes fifo gains extra frame of latency under X11 bare metal server without compositor for the same mesa reason. Yes tearfree on with X11 compositor makes matters worse.

      X without compositing with tearfree on ties in fifo and loses in immediate because tearfree by open source drivers has some serous overhead being a complete frame.

      That the catch that Xaver Hugl did not make clear enough if you want a tearfree solution KDE wayland compositor is wining.

      Do also notice X with compositing did not have any freesync measurements. When you have X11 with compositing this screws up your interfaces to use freesync from application. So no freesync no immediate mode when using X11 desktop with a compositor.

      Comment

      Working...
      X