Announcement

Collapse
No announcement yet.

XWayland Lands Another Performance Fix

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

  • XWayland Lands Another Performance Fix

    Phoronix: XWayland Lands Another Performance Fix

    Red Hat engineer Michel Dänzer has uncovered and addressed another performance shortcoming within X.Org's XWayland code...

    https://www.phoronix.com/scan.php?pa...other-Perf-Fix

  • #2
    Great! Is it going to be upstreamed?

    Comment


    • #3
      Originally posted by down1 View Post
      Great! Is it going to be upstreamed?
      That is upstream

      Comment


      • #4
        There are surely a few other small improvements that can still be gained. But overall Xwayland already matches or slightly/moderatly surpasses Xorg on bare metal for me and the future will only be brighter.

        Comment


        • #5
          Does make you wonder how many other tiny little handicaps are lurking about in the X codebase. Are there things which exist that can audit the code to find these little nuggets or is it just dumb luck that it was found?

          Comment


          • #6
            When wayland will be able to support multiple monitors I maybe switch....
            Atm you cant set your main monitor as a primary

            Comment


            • #7
              Perhaps a dumb question, but does this:

              "constant superfluous eglMakeCurrent calls"
              mean, only EGL is going to be benefitted?
              I mean, AFAIK just Nvidia was on the EGL streams train and is going the GBM route now. Or does this have nothing to do with each other?

              Comment


              • #8
                After the arrival of Plasma 5.13 I am steadily on wayland session and it seems to be going really well, I don't use dual monitors which is a complaint that usually affects plasma-wayland so it's ok.
                I also use a lot of applications that use Xwayland such as Chrome for example, however they look ok, no problem so I am happy with these further improvements from Xwayland.

                Comment


                • #9
                  Originally posted by kozman View Post
                  Does make you wonder how many other tiny little handicaps are lurking about in the X codebase.
                  Does not make me wonder at all. Any large code base (especially one from decades ago) has various cruft lurking around. It is simply the nature of such code bases.

                  Are there things which exist that can audit the code to find these little nuggets or is it just dumb luck that it was found?
                  There are various performance profiling tools that can identify "hot spots" to assist one looking for the easy wins, but all too often this comes down to looking at the actual code trace and noticing something does not look optimal and then trying to understand why the code is doing what it is doing, and whether the why still applies (sometimes a "fix" was added to mitigate against an issue that is no longer relevant due to other changes, but the "fix" is still in the code).

                  There are also a few tools that will perform a static analysis of the code and make some recommendations (pass by const reference, rather than copy, for example, although, as always, there may be specific reasons for not following those recommendations).

                  Comment


                  • #10
                    Originally posted by obri View Post
                    Perhaps a dumb question, but does this:



                    mean, only EGL is going to be benefitted?
                    I mean, AFAIK just Nvidia was on the EGL streams train and is going the GBM route now. Or does this have nothing to do with each other?
                    I am not particularly knowledgeable about the graphics stack but i think EGL is an API that gets used for both EGL streams and GBM. So both kinds of buffer APIs will benefit. EGL is an API to connect graphic APIs like OpenGL with the windowing system, EGL streams is just a buffer API made by NVIDIA to use in their binary drivers instead of the more universal GBM API.

                    Comment

                    Working...
                    X