Announcement

Collapse
No announcement yet.

KWin Can Cause A Performance Hit Too

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

  • KWin Can Cause A Performance Hit Too

    Phoronix: KWin Can Cause A Performance Hit Too

    Last week we published Arch Linux vs. Ubuntu benchmarks to finally lay to rest that for the overall system performance the speed of the rolling Arch Linux distribution is not too different from that of Ubuntu when running with similar package versions. One of the areas, however, where the performance was different with the "out of the box" experience was the OpenGL gaming where Ubuntu was using Compiz by default where as Arch had Metacity. This surprised many so we published another article entitled The Cost Of Running Compiz where we showed the performance penalties of a compositing window manager with different hardware and drivers. This led some to ask whether the performance of KWin also causes the OpenGL frame-rate to drop, so here are those KDE benchmarks.

    http://www.phoronix.com/vr.php?view=14958

  • #2
    Judging by the results, perhaps the title should be changed to:
    The Cost of Running Compositing on NVIDIA drivers.

    Looks like ATI and Intel actually BENEFIT from compiz or kwin in most cases. From this, we can deduce it is not compiz or kwin reducing performance but compositing on Nvidia.

    Comment


    • #3
      Could you rerun the nVidia test with their newly released 256 series beta drivers?

      I seem to remember kwin having an option to disable compositing for fullscreen windows too, which perhaps could make life a bit easier for users with nvidia+kde+gaming.

      Comment


      • #4
        for me kwin effects are really slow and choppy with fglrx.
        compiz on the other hand runs smooth
        (im using the backclear patch)

        Comment


        • #5
          I really would be interested in seeing if the speed ups with compositing is because the cards are being dynamically clocked.

          By this i mean if the catalyst driver is maintaining a higher clock speed when compositing in order to cope and as a side effect is making the game run faster.

          An interesting read anyhow.

          Comment


          • #6
            Is it possible to add a base line ie without compositing to see how much of a negative impact it makes

            Comment


            • #7
              Originally posted by andrnils View Post
              I seem to remember kwin having an option to disable compositing for fullscreen windows too
              yes, I remembered something like that, too. See here:

              http://www.kdedevelopers.org/node/3639
              Second improvement is unredirecting fullscreen windows - that is, if a fullscreen window is not covered by something else, such as a game or a movie, that window is excluded from compositing and allowed to draw normally. This lets it draw at the full speed and should also help with avoiding tearing.
              has been implemented almost 2 years ago.

              So either this feature was disabled during the tests, or it didn't work in some cases. Interesting.

              Comment


              • #8
                As I posted under the compiz thread, unredirect of fullscreen windows has to be manually enabled in ~/.kde/share/config/kwinrc in kde4. For me (nvidia proprietary driver) it solves choppiness issues when running Unigine.

                Comment


                • #9
                  Does the performance hit occur if the desktop effects are disabled too? Or is it linked to the effects?

                  Comment


                  • #10
                    Originally posted by andrnils View Post
                    I seem to remember kwin having an option to disable compositing for fullscreen windows too, which perhaps could make life a bit easier for users with nvidia+kde+gaming.
                    This is an option or this should work out of the box?

                    There's usually no PulseAudio running in KDE, so games should be even a bit faster compared to Gnome + PulseAudio.

                    Comment


                    • #11
                      I am not sure but I think that all theses tests are mood.

                      When you do compositing the window manager decides what to paint when, so the fps outputed by a game do not have to match the fps the user sees then. [1] indicates something similiar.

                      [1] http://www.kdedevelopers.org/node/3365

                      Comment


                      • #12
                        Originally posted by Pfanne View Post
                        for me kwin effects are really slow and choppy with fglrx.
                        compiz on the other hand runs smooth
                        (im using the backclear patch)
                        Same here, kwin is also really bad for me with the open source drivers too :/

                        Comment


                        • #13
                          Does the performance hit occur if the desktop effects are disabled too? Or is it linked to the effects?
                          It has to do with compositing, not any effects you switch on.

                          KWin can easily suspend compositing (there is a shortcut), but like others have said, it should turn off compositing automatically for fullscreen GL apps. I have certainly never edited any rc files to achieve that.

                          I seem to remember a GUI option too, but I can't find it now.

                          Comment


                          • #14
                            Originally posted by mat69 View Post
                            I am not sure but I think that all theses tests are mood.

                            When you do compositing the window manager decides what to paint when, so the fps outputed by a game do not have to match the fps the user sees then. [1] indicates something similiar.

                            [1] http://www.kdedevelopers.org/node/3365
                            You probably did notice the guying say a multitude of times that the video he linked was a fake, right?

                            Comment


                            • #15
                              Originally posted by nanonyme View Post
                              You probably did notice the guying say a multitude of times that the video he linked was a fake, right?
                              You probably did notice the updates right?

                              [...] most people probably don't have the knowledge to know that, since compositing is an additional pass, it cannot make things faster in general.

                              That conflicts with some of the benchmarks Michael did.

                              Comment

                              Working...
                              X