Announcement

Collapse
No announcement yet.

KDE Is Down To Just One Wayland Showstopper Bug Remaining

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

  • Originally posted by oiaohm View Post

    If you go though the JAN results X11 did not always win.
    which of those results do you think show X11 losing?

    remember, you are trying to show
    "What does Wayland (currently!) offer that X.Org can't?
    + Better performance​"

    Isn't a lie, so it needs to be absolutely unambiguously better performance you could actually notice,

    not

    "wayland might offer better performance if you run it on an AMD card on a tuesday at low settings on an application no one has ever heard of or used".

    just by the nature of running opengles wayland will sometimes get better fps - simply because opengles doesn't support all the functionality of a full opengl graphics pipeline. that isnt better performance from wayland, its what you get from lower quality graphics, and would be the same if you crippled X11 with opengles.
    Last edited by mSparks; 29 November 2023, 03:09 PM.

    Comment


    • Originally posted by mSparks View Post
      just by the nature of running opengles wayland will sometimes get better fps - simply because opengles doesn't support all the functionality of a full opengl graphics pipeline. that isnt better performance from wayland, its what you get from lower quality graphics, and would be the same if you crippled X11 with opengles.
      Opengl es in fact does not forbid having a full opengl implementation.

      EGL and GLX from Nvidia and AMD in fact expose the same opengl extensions. Something you did not consider that you can be opengl 4.6 and opengles 3.2 at the same time.

      Please note applications like firefox and chrome have moved to using openglES under X11 because this performs better than using GLX.

      You may (or may not) have heard that Firefox is moving from GLX to EGL for the Linux graphics stack. It’s an indicator of which way the tides are moving in the software world. Let’s loo…


      The idea that opengl es is low performance is a big mistake.

      The old GLX interface does not support zero copy buffers. EGL/opengles does support zero copy buffers. This is a major difference in performance. Of course all applications that switch to EGL under X11 don't work with X11 TCP/IP or ssh X11 forwarding.


      Isn't a lie, so it needs to be absolutely unambiguously better performance you could actually notice,
      No that the problem. Person has omitted data how they came up with their result. Wayland does offer better performance in over 50% of the tests in January. The maths merging performance results hid this fact.

      So how do you come up with unambiguously better performance. You do 20 tests and like 15 are faster and 5 is slower does this not that equal umambiguously better perform for the one that won 15 times had better performance. You average the performance of those 20 tests you can now end up with the one that only won 5 times showing better performance. Just count wins and losses and the result in JAN read very differently to the average and this is the problem. So one way of measuring the Jan results Wayland wins the other way X11 wins. So what measurement method did the person who made that statement use.

      I have been telling you that you need to take a close look at that complete JAN report not cherry pick it. Not cherry picking it tells a very important story that a person may be claiming Wayland has better performance and absolute believe it based on how they calculated the result.

      This is a case of truth lies and statistics. Its like the 3 different ways to perform average they give different results.

      The fact you can have a stack of Wayland use cases win over X11 means a person workflow might only use applications that gain from using wayland.

      At the moment I would not say that in performance that X11 or Wayland is universally better or worse it will depend on your workflow/use case. In fact the current numbers don't support claim of x11 being unambiguously better than Wayland either.

      Comment


      • Originally posted by oiaohm View Post

        Wayland wins the other way X11 wins.
        X11 didn't lose to wayland once in those jan tests.
        Or any of my tests

        therefore it is simply a lie to say wayland has better performance.

        Comment


        • Originally posted by mSparks View Post
          X11 didn't lose to wayland once in those jan tests.
          Or any of my tests

          therefore it is simply a lie to say wayland has better performance.
          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


          Someone need to stop lieing. On the very page you quoted from.

          DDraceNetwork, Quake 2 RTX, Tesseract, Xonotic,, on AMD hardware with Wayland won over X11. Unvanquished was a draw on AMD hardware. So 4 wins and a draw.


          Yes Page 4 AMD Wayland does not lose once. So another 3 wins.

          Page 3 1 game that does not run with AMD at all. Then another win and draw.

          Page 2 another 2 wins with AMD and the only loss.

          4+3+1+2, So 10 wins.
          2 draws.
          1 loss
          1 do no finish.

          12/14 anyone. So 85% chance of being equal or better off using Wayland on AMD hardware is what that Jan number says.

          Your tests where with Nvidia hardware.

          Jan 30 tests say if you are on AMD hardware most likely you will get better performance using Wayland than bare metal X11..

          See I did not cherry pick mSparks. AMD user could for sure be claiming better performance and be telling the truth.

          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


          Now we come forward to June numbers. With Jun numbers after many Nvidia driver fixes we start seeing Nvidia Wayland performance winning over X11 bare metal performance.

          Yes something with Dirt rally with AMD on Wayland does not get along.

          Of course you want not look at the June tests that show that your claims were based on faulty Nvidia drivers as I told you at the time. By the way Nvidia still tells us they have not 100 percent fixed their drivers yet so Wayland with Nvidia is still running slightly slower than it should be.

          Wayland not being poor performance even when using Xwayland that kills particular argument dead and the June test showed that. Poor performance with Wayland and Xwayland has been shown over and over again to be driver issue that driver vendor at some point fixes.

          So anyone using newer numbers than you is going to claim Wayland gives them better performance in general be they AMD/Nvidia user.

          Comment


          • Really... it's useless to come here or to other forums to discuss bugs and then maybe not report them. Reporting them not only allows the devs to find out about them, but also to have details via debugging that are not possible here. Some bugs may be hardware specific...​

            Comment


            • Originally posted by oiaohm View Post
              on AMD hardware with Wayland won over X11.
              Saying

              "wayland might offer better performance if you run it on an AMD card on a tuesday at low settings on an application no one has ever heard of or used".

              =

              better performance

              is still not being truthful

              Originally posted by oiaohm View Post
              Yes Page 4 AMD Wayland does not lose once.
              Ahh, so you don't see it as a blatant lie because you understand, "wayland better performance" to actually mean



              You should probably familiarise yourself with the English language definition of "better", it doesn't mean what you seem to think it means.

              The "truth" here is
              "What does Wayland (currently!) offer that X.Org can't?
              + equal or worse performance​"​
              Last edited by mSparks; 30 November 2023, 10:47 AM.

              Comment


              • Originally posted by mSparks View Post
                The "truth" here is
                "What does Wayland (currently!) offer that X.Org can't?
                + equal or worse performance​"​
                No that still a lie.
                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

                Because there are benchmarks like GravityMark where Wayland has the advantage it does not matter if you are running Nvidia or AMD.

                You are still falling back to the Jan numbers instead of using the june numbers that have updated Nvidia drivers. Why the june numbers don't support you argument.

                When you compare the test the same between the Jan and June it clearly shows that the Jan Nvidia drivers were broken and that between Jan to June most of Nvidia performance problem was fixed..
                Last edited by oiaohm; 30 November 2023, 02:13 PM.

                Comment


                • Originally posted by oiaohm View Post

                  No that still a lie.
                  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

                  Because there are benchmarks like GravityMark where Wayland has the advantage it does not matter if you are running Nvidia or AMD.

                  You are still falling back to the Jan numbers instead of using the june numbers that have updated Nvidia drivers. Why the june numbers don't support you argument.

                  When you compare the test the same between the Jan and June it clearly shows that the Jan Nvidia drivers were broken and that between Jan to June most of Nvidia performance problem was fixed..
                  Where's the update? I don't see it anywhere.

                  Comment


                  • Originally posted by Panix View Post

                    Where's the update? I don't see it anywhere.
                    June 2023 numbers https://www.phoronix.com/review/ubun...nd-xorg-gaming
                    January 2023 numbers https://www.phoronix.com/review/wayland-nv-amd-2023

                    I have been including the links.

                    Comment



                    • Originally posted by oiaohm View Post

                      June 2023 numbers https://www.phoronix.com/review/ubun...nd-xorg-gaming

                      I have been including the links.
                      still "equal or worse performance" which is not "better performance".
                      And even then exclusively full screen without any networked display involved which is the "best case" for wayland.

                      Persist the lie all you want I guess, it seems no one is going to convince you this isn't wayland winning

                      or that this is not "better"


                      Which is still "equal or worse"

                      The only difference anyone would ever notice in that entire article is if you are on a 120hz display with a RX7900XTX playing Dirt Rally 2, when they will abandon wayland and switch back to xorg, distro hopping if needed.

                      Comment

                      Working...
                      X