Announcement

Collapse
No announcement yet.

2D Performance Also Impacted By Unity On XMir

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

  • #16
    Originally posted by kokoko3k View Post
    How can kwin in 4.11 work with wayland then?
    It doesn't really. It's just an experimental hack, where KDE still uses X to render onto a Wayland backend - kind of similar to how Unity on XMir works, actually. The difference is, it's not enabled by default, it's just there as an experimental feature so that anyone who wants to play with it - not meant for production use, like in Ubuntu 13.10.

    The purpose of the KDE hack is, according to Martin, more as practice - to get familiar with Wayland. After the actual transition to Wayland, KDE will use it's own Wayland compositor to render to a Wayland backend, and XWayland will be used for legacy apps.

    Comment


    • #17
      Originally posted by dee. View Post
      It doesn't really. It's just an experimental hack, where KDE still uses X to render onto a Wayland backend - kind of similar to how Unity on XMir works, actually. The difference is, it's not enabled by default, it's just there as an experimental feature so that anyone who wants to play with it - not meant for production use, like in Ubuntu 13.10.

      The purpose of the KDE hack is, according to Martin, more as practice - to get familiar with Wayland. After the actual transition to Wayland, KDE will use it's own Wayland compositor to render to a Wayland backend, and XWayland will be used for legacy apps.
      Thank you.

      Comment


      • #18
        Originally posted by dee. View Post
        It doesn't really. It's just an experimental hack, where KDE still uses X to render onto a Wayland backend - kind of similar to how Unity on XMir works, actually. The difference is, it's not enabled by default, it's just there as an experimental feature so that anyone who wants to play with it - not meant for production use, like in Ubuntu 13.10.

        The purpose of the KDE hack is, according to Martin, more as practice - to get familiar with Wayland. After the actual transition to Wayland, KDE will use it's own Wayland compositor to render to a Wayland backend, and XWayland will be used for legacy apps.
        There is also a new native software implementation to replace Xrender:
        http://blog.martin-graesslin.com/blo...e-compositing/

        This one is only compatible with wayland for the moment.

        Comment


        • #19
          Originally posted by BO$$ View Post
          OK so probably in 13.10 we'll take a performance hit. Hopefully the impact will be smaller, much smaller than it is today.... Get to work Canonical, only a few months left. Get it right a lot of people will think better of you. Get it wrong and haters will complain about this performance loss for another decade, even if it will have been fixed in 14.04.
          http://www.olli-ries.com/first-mir-benchmarks/

          One of the reasons for this result set is missing composite bypassing support, which we are aware of since January. Composite bypass helps when apps/benchmarks run fullscreen because… well, because they don’t need to be composited. Gamers out there… there is hope and a plan in place to get you your precious FPS back. This feature/bug is currently scheduled once other key functionality landed. Also, in order to make FPS based benchmarks really count, we need eglSwapInterval(0) implemented, which is currently in progress

          Comment


          • #20
            Originally posted by TheOne View Post
            http://www.olli-ries.com/first-mir-benchmarks/
            One of the reasons for this result set is missing composite bypassing support, which we are aware of since January. Composite bypass helps when apps/benchmarks run fullscreen because… well, because they don’t need to be composited. Gamers out there… there is hope and a plan in place to get you your precious FPS back. This feature/bug is currently scheduled once other key functionality landed. Also, in order to make FPS based benchmarks really count, we need eglSwapInterval(0) implemented, which is currently in progress
            All of this would have been a non-issue if they'd have stuck with Wayland.

            Comment


            • #21
              Originally posted by dee. View Post
              All of this would have been a non-issue if they'd have stuck with Wayland.
              I second this!
              Also, it explicitelly talk to "gamers".
              But even if i were not a gamer and just wanted a smoother scroll out of my browser or watch an high fps video (think to deinterlaced TV content with doubled framerate) do i need to made it fullscreen on mir?

              Comment


              • #22
                Originally posted by kokoko3k View Post
                I second this!
                Also, it explicitelly talk to "gamers".
                But even if i were not a gamer and just wanted a smoother scroll out of my browser or watch an high fps video (think to deinterlaced TV content with doubled framerate) do i need to made it fullscreen on mir?
                Gamer's is how Windows win!
                btw is TheOne dh04000 sound like the same guy to me
                Last edited by spacetoilet; 06-29-2013, 10:04 AM.

                Comment


                • #23
                  Originally posted by kokoko3k View Post
                  I second this!
                  Also, it explicitelly talk to "gamers".
                  But even if i were not a gamer and just wanted a smoother scroll out of my browser or watch an high fps video (think to deinterlaced TV content with doubled framerate) do i need to made it fullscreen on mir?
                  Actually XMir acts as a fullscreen window to render the desktop environment, thats why the canonical guys is mentioning that.

                  Originally posted by spacetoilet View Post
                  Gamer's is how Windows win!
                  btw is TheOne dh04000 sound like the same guy to me
                  Errm, thats kind of funny, why would you think I have another account? Maybe you do that yourself and so you think everyone else play those kid games

                  Comment


                  • #24
                    Which... has zero bearing on 2D performance, which this article showed taking a performance hit as well.

                    Comment


                    • #25
                      Originally posted by GreatEmerald View Post
                      Which... has zero bearing on 2D performance, which this article showed taking a performance hit as well.
                      This^

                      Comment


                      • #26
                        Originally posted by GreatEmerald View Post
                        Which... has zero bearing on 2D performance, which this article showed taking a performance hit as well.
                        Since I haven't looked at the source I lack knowledge of the technical details surrounding its functionality but as other sources said XMir is rendering the complete desktop as a fullscreen window which output is controlled by X. What are the technical implications of that or how it correlates with the cited article? I'm not sure.

                        Comment


                        • #27
                          Originally posted by d2kx View Post
                          On the other side, Compiz 0.9.10 in Ubuntu 13.10 will probably introduce buffer_age support, which will speed some things up.
                          Doesn't buffer age affect only for image quality and frame skipping?

                          Originally posted by BO$$ View Post
                          OK so probably in 13.10 we'll take a performance hit. Hopefully the impact will be smaller, much smaller than it is today.... Get to work Canonical, only a few months left. Get it right a lot of people will think better of you. Get it wrong and haters will complain about this performance loss for another decade, even if it will have been fixed in 14.04.
                          I don't think so. A lot of people will probably think better of Canonical if they don't run their desktop on XMir on a release. That's not what Mir or XMir are designed for, and that provides none of the advertised advantages, so the tech people will not think better, and the not so tech people doesn't care that much about what is running on the background, as long as it works. Even with no performance hit, a desktop on XMir is a bad idea. Personally, I will think a lot better of them if they abandon that idea. Or even if they don't use that on the LTS (I agree with the need of whatever they use in the LTS to be tested, but could just use X.org on the LTS, and test this combination on 13.10 and Mir+Unity8 on 14.10).

                          Originally posted by ickle View Post
                          Even worse. If we use cairo as a yard stick for measuring acceleration across the different solutions, then cairo-xlib is many times faster (when supported by a good driver i.e. sna or nvidia, and even when supported by bad drivers such as uxa and fglrx) than cairo-gl (with their respective OpenGL drivers). We are back to looking at using the CPU and pushing images around.

                          Pure client side rendering also has higher memory overheads as what the display server does cache between multiple clients, is now allocated separately in every client.
                          I think the idea is to use them in similar conditions, i.e., if cairo-xlib is only available on X, cairo-gl should be used in all three benchmarks.

                          Comment


                          • #28
                            Originally posted by TheOne View Post
                            Since I haven't looked at the source I lack knowledge of the technical details surrounding its functionality but as other sources said XMir is rendering the complete desktop as a fullscreen window which output is controlled by X. What are the technical implications of that or how it correlates with the cited article? I'm not sure.
                            The composite bypass they mean is in the Mir side. It means Mir is trying to realize all of the operations needed for compositing, even when working with a fullscreen app, so this affects everything.
                            The rendering of the desktop as a fullscreen window (independently of who controls the contents, as long as it's not Mir) implies that Mir doesn't really need to handle any compositing, so you can avoid those operations (that's why it's called a bypass). X still needs to do its inner compositing, as long as your real app is not in fullscreen mode, in which case a well thought window manager bypasses it (GNOME, Unity, KDE, XFCE, and maybe some others include this feature), too.

                            EDIT: On a sidenote, does anyone know if Xubuntu will be running on XMir for 13.10? I'm thinking of giving Mir a try on one of boxes, but I don't really like Unity (and I would have no ground to make a comparison, since it was never installed on that machine, I would have no way to say if there was a loss of quality or speed).
                            Last edited by mrugiero; 06-29-2013, 02:12 PM.

                            Comment


                            • #29
                              Originally posted by mrugiero View Post
                              The composite bypass they mean is in the Mir side. It means Mir is trying to realize all of the operations needed for compositing, even when working with a fullscreen app, so this affects everything.
                              The rendering of the desktop as a fullscreen window (independently of who controls the contents, as long as it's not Mir) implies that Mir doesn't really need to handle any compositing, so you can avoid those operations (that's why it's called a bypass). X still needs to do its inner compositing, as long as your real app is not in fullscreen mode, in which case a well thought window manager bypasses it (GNOME, Unity, KDE, XFCE, and maybe some others include this feature), too.

                              EDIT: On a sidenote, does anyone know if Xubuntu will be running on XMir for 13.10? I'm thinking of giving Mir a try on one of boxes, but I don't really like Unity (and I would have no ground to make a comparison, since it was never installed on that machine, I would have no way to say if there was a loss of quality or speed).
                              far as i know Xfce's Developer said it will work on Wayland once it get's ported to GTK+ and from that i can see Xbunutu will run on Xorg

                              Comment


                              • #30
                                Originally posted by spacetoilet View Post
                                far as i know Xfce's Developer said it will work on Wayland once it get's ported to GTK+ and from that i can see Xbunutu will run on Xorg
                                Two things. The first, I thought they said they'd stick to X, but I'm not sure where, though (this is good news, because I wanted to switch to Wayland eventually and thought I'd have to switch desktop, which I'm not willing to since I'm very comfortable with XFCE).
                                The second, what XFCE's devs say has nothing to do with Ubuntu's and flavor's decisions, that's why I asked. I'm aware they don't plan to switch to Mir, so it's either X.org or XMir.
                                There must be a way to set it up otherwise, though, for the sake of testing.

                                Comment

                                Working...
                                X