Announcement

Collapse
No announcement yet.

Experimental Wayland Support For Wine Now Sees More Functionality Working

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

  • #31
    Originally posted by Luke_Wolf View Post

    Like I said... not digging through years of forum posts, but the Wayland/Xorg guys that were here... way back when Wayland was originally announced were making claims that Wayland should be faster than X11 for a variety of reasons. It's worth pointing out that what I'm going to call the "Wayland stack" to refer to Wayland, xdg-shell, etc and the concepts of how they integrate with the compositor that replaces the X11 Stack has changed pretty dramatically from when this was just a proof of concept with Weston not even implemented in any of the toolkits. However I'm also going to note that said performance claims only applied to end to end Wayland, not applications running on XWayland in a Wayland Environment.
    Way back at the beginning of the project X was still using DRI2. As I recall, DRI3 was essentially a port of many of the performance enhancements people had come up with for Wayland back to the X architecture.

    I think today you will still see performance improvements in niche scenarios - mostly when you are on an extremely limited device, like a Pi. On any typical desktop nobody is going to notice if the CPU % drops from 2% to 1%, and application performance is limited by other factors anyway. The other case is when using more specialized capabilities like hardware planes, which aren't hooked up through the desktop compositors yet, but probably will be eventually. That could probably be ported back to X as well, but nobody understands X well enough to do it these days without some major corporate sponsor backing the project.
    Last edited by smitty3268; 19 February 2021, 08:47 PM.

    Comment


    • #32
      Originally posted by duby229 View Post
      Improved performance hasn't been demonstrated at all, not even once in all these years.
      moron, how improved performance in wayland could be demonstrated without wayland support(subj is about adding wayland support in future)?

      Comment


      • #33
        Originally posted by aufkrawall View Post
        This is true, though in return this doesn't make games run worse on Xorg (Or why would you want to play GPU demanding titles in windowed mode?)
        you are basically saying that x11 is fast, but only if you run fullscreen. well, it means that in general x11 is slow. and why would you be forbidden to play gpu demanding titles in windowed mode?

        Comment


        • #34
          Originally posted by muncrief View Post
          One day Wayland will "just work" on everything like X does
          x doesn't "just work" on everything, i.e. that day is in the past already

          Comment


          • #35
            Originally posted by duby229 View Post
            If you claim haven't heard nonsense performance claims about Wayland before today
            i'm pretty sure he also heard claims of walking on water, should i give you google search link for that?

            Comment


            • #36
              Originally posted by duby229 View Post

              Bro, really? There are Hundreds of references... If you claim haven't heard nonsense performance claims about Wayland before today, then I have to assume your claim to have never heard them is a straight up lie...

              A damn search would suffice...
              https://duckduckgo.com/?q=Wayland+pe...=v248-1&ia=web
              None of those point to any claims made by Wayland developers. So this question remains unanswered by you. Feel free to try again.

              Comment


              • #37
                Originally posted by smitty3268 View Post

                Way back at the beginning of the project X was still using DRI2. As I recall, DRI3 was essentially a port of many of the performance enhancements people had come up with for Wayland back to the X architecture.

                I think today you will still see performance improvements in niche scenarios - mostly when you are on an extremely limited device, like a Pi. On any typical desktop nobody is going to notice if the CPU % drops from 2% to 1%, and application performance is limited by other factors anyway. The other case is when using more specialized capabilities like hardware planes, which aren't hooked up through the desktop compositors yet, but probably will be eventually. That could probably be ported back to X as well, but nobody understands X well enough to do it these days without some major corporate sponsor backing the project.
                Yeah, I'm sure. Basically the entire ecosystem of Linux has gotten rewritten in preparation for Wayland so it would be no surprise if the performance gains had found their way back into X11 and otherwise. However the question also has to be asked as you're kinda pointing out: If it was faster would it be measurable? And the answer is: Probably actually no in the real world when you're comparing Apples to Apples. Everything else layered on top of it requires so much more CPU and GPU Time.

                That said sitting here claiming "oh they never said that" would be like claiming that the devs never claimed "perfect frames" or a "secure input model"... we were here, we saw them say it. I'm pro wayland but I'm not going to sit here and let that one fly.

                Comment


                • #38
                  Originally posted by Luke_Wolf View Post

                  Like I said... not digging through years of forum posts, but the Wayland/Xorg guys that were here... way back when Wayland was originally announced were making claims that Wayland should be faster than X11 for a variety of reasons. It's worth pointing out that what I'm going to call the "Wayland stack" to refer to Wayland, xdg-shell, etc and the concepts of how they integrate with the compositor that replaces the X11 Stack has changed pretty dramatically from when this was just a proof of concept with Weston not even implemented in any of the toolkits. However I'm also going to note that said performance claims only applied to end to end Wayland, not applications running on XWayland in a Wayland Environment.
                  So while I don't have a source for this one, this is a much more tepid assessment that someone saying Wayland has performance as a goal.

                  A lot of the Wayland infrastructure improvements are at this point shared by Xorg and that's not limited to performance improvements but also other core components like say libinput. It is going to be hard to benchmark core Wayland components against Xorg because mostly people care about something like a specific desktop environment running on Xorg vs Wayland, out of which the latter is newer and there is usually a whole host of performance improvements that is feasible but all of them in place yet. Desktop performance often comes down thing to things like interactive input latency which typical benchmarks don't even capture. Linux kernel has also struggled with accurately representing this for ages.

                  Comment


                  • #39
                    Originally posted by Luke_Wolf View Post

                    That said sitting here claiming "oh they never said that" would be like claiming that the devs never claimed "perfect frames" or a "secure input model"... we were here, we saw them say it. I'm pro wayland but I'm not going to sit here and let that one fly.
                    No, that isn't the case at all and I don't think you can put it all in the same category. I have seen the latter two claims but I haven't seen any Wayland developer claim performance by itself is a goal. Happy to see any sources for it. Otherwise, merely saying they did say that doesn't count as a reference.

                    Comment


                    • #40
                      Originally posted by RahulSundaram View Post

                      So while I don't have a source for this one, this is a much more tepid assessment that someone saying Wayland has performance as a goal.
                      That's both splitting hairs and not what anyone is claiming. What the Devs claimed is that the X11 architecture was garbage in quite the laundry list of ways and that one of the advantages (not goals... advantages) of Wayland was that because it fixed those things it would be faster. To use a bad car analogy what you're doing is the equivalent of taking the fact that a Cross Over SUV with an I4 doing 180HP is slower than a Truck with a V6 rated at 300HP, with an auto engineer involved in both projects saying "The truck is faster than the CUV" and claiming "well he never said speed is a goal".

                      What people are saying is that the stated advantage has not materialized.

                      Comment

                      Working...
                      X