Announcement

Collapse
No announcement yet.

X.Org vs. Wayland Linux Gaming Performance For NVIDIA GeForce + AMD Radeon In Early 2023

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

  • #31
    Originally posted by piotrj3 View Post

    Such titles doesn't exist. Currently DXVK and Wine are made assumping Xorg or Xwayland is used. Few native ports have to account that clients only using X server exist so they also don't support Wayland. In close future it is not going to change anytime soon.
    Indeed, these tests are not representative.

    Comment


    • #32
      Originally posted by piotrj3 View Post

      Such titles doesn't exist. Currently DXVK and Wine are made assumping Xorg or Xwayland is used. Few native ports have to account that clients only using X server exist so they also don't support Wayland. In close future it is not going to change anytime soon.
      Umh, hello? Just preload fresh SDL2 (with SDL_VIDEODRIVER=wayland) or GLFW compiled for wayland. Also there are plenty of opensource ports of old games that use SDL2 and therefore support wayland.

      Comment


      • #33
        I want to state that on my notebook (a Renoir + Navi10 notebook) using PRIME, even though framerate counters are roughly equivalent, input latency and smoothness is significantly in Wayland (Plasma KDE)'s favour.
        I had to ensure under X that I never hit any GPU at 100% else it would stutter like mad, but on Wayland I don't need to worry about this at all.

        But to be honest, it's likely due to an implementation detail of how display's are connected on this notebook more than anything else, but Wayland really is a significantly better experience for me, with what appears to be marginally better stability as well. (And I don't have to do bespoke display chaining finagling using XRandr)

        Comment


        • #34
          Originally posted by grigi View Post
          I want to state that on my notebook (a Renoir + Navi10 notebook) using PRIME, even though framerate counters are roughly equivalent, input latency and smoothness is significantly in Wayland (Plasma KDE)'s favour.
          I had to ensure under X that I never hit any GPU at 100% else it would stutter like mad, but on Wayland I don't need to worry about this at all.

          But to be honest, it's likely due to an implementation detail of how display's are connected on this notebook more than anything else, but Wayland really is a significantly better experience for me, with what appears to be marginally better stability as well. (And I don't have to do bespoke display chaining finagling using XRandr)
          not your notebooks problem and no need to cap 100% usage to notice. anything that popups up on the screen (tooltips, notifications, windows, animations, video) has a good chance of halting everything else's rendering briefly on nvidia linux desktop, making EVERYTHING else stutter for some significant time. literally anything or any DE, unless you're using noveau. It's worse on x11 as wayland clients seem to behave better but still shitty. On noveau at min clocks everything is smooth rofl...

          waiting for people to blame anything else but the nvidia stupid driver...

          Comment


          • #35
            Speaking of X.org "insecurity" - not a single documented incident for over 30 years of its existence.
            Speaking of Wayland "security" - Wayland compositors are executed by the user account (not root and its own memory space) and can be trivially hijacked.

            Speaking of

            Originally posted by qarium View Post

            the real reason is security IBM/Redhat always did want to sell redhat/fedora linux to Governments and big Coproations and Banks and secret intelligence services and so one and so one and linux was always rejected by the government and big corporations because X11/x.org is a security hole and the security holes are essencial parts in the protokoll means no one can fix it.
            This is blatant lies. The US government and three-letter agencies do use RHEL. Here's some useless agreement which I've just made up.

            More info: Red Hat Enterprise Linux 7.x: EAL4+
            Last edited by avis; 31 January 2023, 08:28 AM.

            Comment


            • #36
              Originally posted by Velocity View Post
              If Wayland performance is so bad, what is the reason for its existence?
              Wayland was faster than X in all tests except GTA and Dirt Rally (probably a shitty game implementation), didn't you see? That's when using XWayland. So, the conclusion is, Wayland + XWayland is a bit faster than X in general, while native Wayland is clearly ahead.

              Comment


              • #37
                Thats why my first question is "do you use nvidia?" if a shill is complaining about the "much worse" situation of Wayland.

                Comment


                • #38
                  Originally posted by RejectModernity View Post

                  Umh, hello? Just preload fresh SDL2 (with SDL_VIDEODRIVER=wayland) or GLFW compiled for wayland. Also there are plenty of opensource ports of old games that use SDL2 and therefore support wayland.
                  Because so many games that can actually be gpu bound on modern graphics card uses those....

                  Does Unity support Wayland? Unreal Engine? Wine? DXVK? Currently only unofficial forks of Wine support Wayland sort of. Everything else is "no".

                  Comment


                  • #39
                    Originally posted by Myownfriend View Post

                    XWayland is still an additional layer that's not present in the X11 tests that can have it's own bugs and performance issues so it's relevant to mention and can only, in theory, be slower than the game using Wayland directly.

                    Also since SDL2 supports Wayland, some games can use it natively so being more specific would communicate to people when they're seeing something using Wayland vs using X11 under a Wayland session.
                    Well yeah, but DRI … shouldn't it talk directly to the kernel for rendering?

                    I mean, obviously there are some bugs and regressions, but in theory, the additional layer shouldn't matter for rendering via DRI, right?
                    In the end, that's the whole point about DRI, so the display server doesn't get in the way.

                    Comment


                    • #40
                      Originally posted by Velocity View Post
                      If Wayland performance is so bad, what is the reason for its existence?

                      Wayland has given us desktop users so many problems, so many bugs, so many disasters. What is it all good for? What does Wayland offer that is worth the enormous cost for end users who see their systems not boot, have error messages, lower performance, and other terrible effects. We are now so many years later that most distro's have switched or are in the process of switching, but i see no clear benefits for ordinary users. If they would know that some games / applications drop more than half in performance, what would be the reason to use it? What was wrong with X.org in the first place? Its complicated design may lower performance in theory i guess, but with Wayland having so many years of development it still struggles against X.org in some cases big time. The performance is unacceptable for some titles.
                      As Daniel Stone said in The Real Story Behind Wayland and X, they've hacked around so much with things like GLX, DRI, MIT-SHM, compositing window managers, etc. and things like GTK and Qt doing their own internal widget compositing that X.org today does nothing but "terrible, terrible IPC". (IIRC, 100+ round-trips (with no latency guarantees) to initial window show for GEdit. 1000+ round-trips to initial window show for Chrome.)

                      Comment

                      Working...
                      X