Announcement

Collapse
No announcement yet.

Triple Buffering Likely Not Landing Until GNOME 42

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

  • #21
    Originally posted by aufkrawall View Post
    Or make compositing as efficient as Sway's to avoid creating performance issues in the first line?
    The whole issue is that it is too efficient causing the GPU to run into a lower clock energy saving mode.

    Comment


    • #22
      Originally posted by skeevy420 View Post

      Isn't this also one of those quirks that's usually built-in to the Windows driver or Windows software to work around stuff just like this? I swear I've read something like that before.
      Windows's display server also has all kinds of workarounds for various GPUs/Drivers, but you could argue that.

      Comment


      • #23
        FWIW, besides triple buffering, the only other sensible method to avoid performance issues is to explicitly inform the kernel about "low-latency" workloads, to force the driver into a higher GPU P-state. Some Android devices go as far as enforcing a relatively high minimum GPU *and* CPU P-state whenever a finger is on the touchscreen.

        Comment


        • #24
          Originally posted by Alexmitter View Post

          Windows's display server also has all kinds of workarounds for various GPUs/Drivers, but you could argue that.
          It just seems like that's a quirk used everywhere scheduling is inadequate.

          What's kind of funny is that we can argue that the GPUs themselves are flawed if they don't work correctly at their lowest frequencies.

          Comment


          • #25
            Forgive my ignorance, what's inherently wrong with triple buffering?

            Comment


            • #26
              I'm skeptical of this MR making it in at all, it feels like a hack to fix the problem and not a proper solution, even if it does work

              Comment


              • #27
                This is a really dumb stupid hack.

                Mutter should just get better at scheduling it's stuff to meet vblank time.

                Comment


                • #28
                  Originally posted by JackLilhammers View Post
                  Forgive my ignorance, what's inherently wrong with triple buffering?
                  It increases latency by an extra frame - at 60Hz this is a whole 16.667ms.

                  Comment


                  • #29
                    Originally posted by Alexmitter View Post
                    The whole issue is that it is too efficient causing the GPU to run into a lower clock energy saving mode.
                    I really doubt it's particularly efficient. The last time I checked with Gemini Lake, it froze it into an unusable mess.
                    While Intel Gen9 is complete garbage, it worked with Sway though.

                    Comment


                    • #30
                      Originally posted by Alexmitter View Post

                      A Mali 400 MP2 is enough, this is rather a workaround for a bugged Intel GPU energy management.
                      So why this bug does not get fixed instead of adding unmaintainable workarounds to every software it affects?

                      Comment

                      Working...
                      X