Announcement

Collapse
No announcement yet.

GNOME 3.34's Mutter Lowers Output Lag On X11 To Match Wayland Performance

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

  • GNOME 3.34's Mutter Lowers Output Lag On X11 To Match Wayland Performance

    Phoronix: GNOME 3.34's Mutter Lowers Output Lag On X11 To Match Wayland Performance

    Adding to the list of positive changes with GNOME 3.34 due out this September is lowering possible output lag when running GNOME's Mutter on X11/X.Org...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Would be nice to see more his patches cleaned up and merged, GNOME's performance whilst better is still kinda sucky.

    As far as I know, some gjs performance upgrades (removal of big hammer) might come in 3.34 too.

    Comment


    • #3
      Back in 2017 before I devised my solution I used Mutter on KDE (since KWin had too much lag) and it was already sort of lag-free...

      Also, "output lag"? Guess my terms were wrong all the time...
      Last edited by tildearrow; 21 May 2019, 01:17 PM.

      Comment


      • #4
        Good. It's hard to be enthusiastic about gnome 3 while it is not particularly performant, featureful or lightweight.

        Comment


        • #5
          Originally posted by tildearrow View Post
          Back in 2017 before I devised my solution I used Mutter on KDE (since KWin had too much lag) and it was already sort of lag-free...

          Also, "output lag"? Guess my terms were wrong all the time...
          Same thing. Input lag, output lag, it’s the time between the input occurring and a reaction happening on the output, so both terms are fine.

          This is is one of two of vanvugt’s long-standing patches to be committed in the last few days. Pretty soon we won’t need to patch Mutter to get the lag to an acceptable level.

          Comment


          • #6
            Vanvugt has done some really great work to make GS better. Unfortunately he doesn't seem to coordinate very closely with the rest of the gnome community, unlike some other Canonical fellows. Therefore his work sometimes ends up not matching the direction of the rest of the project, taking longer to get reviewed or doesn't get merged at all.
            Still I'm always happy when one of his MRs finally crosses the finish line

            Comment


            • #7
              Glad to see performance optimizations for X11 as it's not going anywhere.

              Comment


              • #8
                Originally posted by bearoso View Post

                Same thing. Input lag, output lag, it’s the time between the input occurring and a reaction happening on the output, so both terms are fine.

                This is is one of two of vanvugt’s long-standing patches to be committed in the last few days. Pretty soon we won’t need to patch Mutter to get the lag to an acceptable level.
                NO input lag and output lag are two different things.

                Input lag is the time the input takes to get to the application.
                output lag is the time the generated output takes to get from the application to screen.
                Total latency is input lag+ output lag + application processing lag.

                Rough eyeballing is total latency and a guess where the problem is.

                Depending on what form of lag alter what area of code needs fixing. If the problem is output lag and you are attempting to fix input lag developer will never find the problem. Same with the reverse.
                Last edited by oiaohm; 21 May 2019, 06:16 PM.

                Comment


                • #9
                  Will it solve that weird half-second freeze when you press a "normal" key after a multimedia key?

                  Comment


                  • #10
                    Originally posted by angrypie View Post
                    Will it solve that weird half-second freeze when you press a "normal" key after a multimedia key?
                    The answer would be most likely no. This would most likely be input lag side. To be correct the input side stalling.

                    Output lag and input lag are different problems. Yes it would be good to have solid measurements on both.

                    Comment

                    Working...
                    X