Announcement

Collapse
No announcement yet.

How KDE On Wayland Will Lead To Better Power Management

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

  • How KDE On Wayland Will Lead To Better Power Management

    Phoronix: How KDE On Wayland Will Lead To Better Power Management

    Aside from "every frame being perfect" with Wayland among its many advantages over X11, another benefit will be for greater power-savings in properly implemented software when a display is turned-off/sleeping...

    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
    Another reason to abandon x11 besides its inefficiency
    Last edited by Azrael5; 01 September 2015, 10:52 AM.

    Comment


    • #3
      Originally posted by Azrael5 View Post
      Another reason to abandon x11 besides efficiency
      Not every shortcoming of a platform is a reason to jump ship. Most of them can be reworked (I don't think adding two-way communication to X is impossible, more likely a matter of ROI).

      Comment


      • #4
        Originally posted by bug77 View Post

        Not every shortcoming of a platform is a reason to jump ship. Most of them can be reworked (I don't think adding two-way communication to X is impossible, more likely a matter of ROI).
        x11 is deprecated it is good like the shit.

        Comment


        • #5
          The protocol will live on for compatibility and until better toolkits are out there that do "X" like things. But you are right in saying that the X Server (as we know it) will probably go away. Perhaps more interesting would be if X11 continues to evolve... that would be interesting if not a surprise (I hear things).

          Comment


          • #6
            Originally posted by cjcox View Post
            1.) The protocol will live on for compatibility and until better toolkits are out there that do "X" like things. But you are right in saying that the X Server (as we know it) will probably go away. Perhaps more interesting would be if X11 continues to evolve... that would be interesting if not a surprise (I hear things).

            2.) Not every shortcoming of a platform is a reason to jump ship. Most of them can be reworked (I don't think adding two-way communication to X is impossible, more likely a matter of ROI).

            1.) Sadly, X11 cannot evolve without breaking the X11 protocol definition to the point of being incompatible with X11, that is precisely why the Xorg Devs and others went to design Wayland(it wasn't just for the lulz) and set X11 to maintenance mode outside of Xwayland ofc.

            2.) Check the mailing list, over the years everyone tried and not one of them could fix X11 shortcoming without breaking X11 compatibility entirely(or plainly just breaking X).

            The point here is, X11 was designed to work with 80's graphics card dumb framebuffer with fixed functions for 2D acceleration(not GPU's or even less 3D), with a single monitor(more was not even possible at that time), with a single threaded model(this is years before the first dual core reached the market) and the render process was designed for blit pixmaps into a root pixmap to emulate an framebuffer(quite a novel in the 80's), input was designed with absolute 0% of security(Internet wasn't a thing back then) in mind and only for mouse and keyboard mechanical events, etc. etc. etc.

            Xorg developers didn't wanted wayland because was the new shiny thing, it was because X11 reached it's technical limit and they needed an entire new model optimized for this century and able to keep supporting new hardware generations for years like X did.

            If you want some interesting areas to check for yourself why, i leave some here:

            1.) Check input ddx code like xf86-input-libinput and check how much time the CPU have to waste pack event to Xevents
            2.) Try to understand Xinput code (if you do so, you'll get a nice job i guarantee it)
            3.) Try to understand why all new extensions(Xdamage, Xcomposite, etc.) require more than 10 callbacks between the app and the Xserver to put a fragging pixmap on a screen everytime, in every pixmap, in every vblank event, in every Xevent, in every input event.
            4.) Try to add advanced feature to X like dpms, true fullscreen, etc., Hint is not posible because X can't understand more than 1 screen, today works due a hack that generate a root pixmap with N.. screen sizes added and a fragging lot of hacks to compute how to show them to the outputs, this is basically the same reason why keep vsync working with more than 1 monitor is not possible and the same applies to Xinput since it simply don't know what screens are and this make impossible multiGPU solutions since X cannot share the root pixmap between devices and this make impossible to work with overlays directly(you need to capture images and pass them to Xserver) and so on and on and on.

            Comment


            • #7
              I think KDE+wayland+EGL be the real modernity on linux systems able to develop the useful hardware acceleration on the whole system optimizing hardware usage as well. Hardware optimization is the key.
              Last edited by Azrael5; 01 September 2015, 12:59 PM.

              Comment


              • #8
                I hope it's not going to work like switching from tty to xorg...

                Comment


                • #9
                  Originally posted by m132 View Post
                  I hope it's not going to work like switching from tty to xorg...
                  it will at first, but is not wayland fault since in that particular scenario is actually VT_CONFIG faults but once kmscon lands properly or find a way to bypass it. go to a tty from wayland should be completely mess free.

                  actually im not sure is kmscon will be or a new solution is coming since last commit was in 2014(it could be done tho)

                  Comment


                  • #10
                    Originally posted by bug77 View Post
                    Not every shortcoming of a platform is a reason to jump ship. Most of them can be reworked (I don't think adding two-way communication to X is impossible, more likely a matter of ROI).
                    X11 is already hacks and workarounds on top of hacks and workarounds on top of hacks and workarounds on top of hacks and workarounds. Another layer of hacks and workarounds won't make much of a difference.

                    Comment

                    Working...
                    X