KDE Plasma 5.25 Beta Released With Many Improvements, Wayland Support Maturing

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

  • MrCooper
    replied
    Originally posted by schmidtbag View Post
    Xwayland adds overhead [...]
    [citation needed]

    Xwayland incurs no overhead to speak of compared to Xorg for fullscreen apps (assuming the Wayland compositor can directly scan out client buffers, which all the main ones can by now) or windowed apps with compositing. Xorg could be lighter only for windowed apps without compositing, which hasn't been really relevant for many years.

    Leave a comment:


  • extremesquared
    replied
    That task switcher order-randomizing bug from 2019 still exists. Sort of embarrassing considering how fundamental a feature alt-tab switching is to an OS. For a lot of people, window switching is primary workflow.

    Leave a comment:


  • cmakeshift
    replied
    The same company that made that awesome performance menu could surely add a checkbox somewhere that says "Use Wayland on desktop mode". And surely it'd be disabled by default, because when a casual user has issues using tool windows in libreoffice, or when drag and drop between the two sides of the same dolphin window is broken, that's no good.

    Leave a comment:


  • SSJBurger
    replied
    Originally posted by theriddick View Post
    I've been testing out wayland this past week (not this version) and its pretty easy to cause conflicts/crashes/UI alignment issues or DPI problems.

    The Fractional scaling system they have for System Displays STILL causes blurring if you select anything fractional.
    Probably a GTK/QT problem but still; you don't experience this on X11 version of the desktop which I often run at %125 scaling. (4k)

    Additionally proton/game performance on Wayland was not better; in fact most games had frame timing stutter issues even if the framerate was above 60fps. Who knows when it will improve.
    This is strange I have multi monitor 1080p 240hz and a 60hz. I only run arch KDE wayland due to it having less stutter and better frame timing on my 6800xt, 5800x3d computer. My girls computer is one monitor 1440p 165hz Vega 64 with a 5600x arch KDE wayland and also seems smoother on wayland. we both use Proton GE for steam with DXVK_ASYNC=1 set for every game though. Never have any crashes or issues and have been on wayland strictly since plasma 5.24.

    Leave a comment:


  • vegabook
    replied
    Originally posted by schmidtbag View Post
    .... too long to read or quote given low signal:error ratio
    Basically the problem is you don't have hands on time with this complex custom software/hardware stack, but you still think you know everything even though you're making poor assumptions and errors. Some humility would make you look smarter.
    Last edited by vegabook; 20 May 2022, 08:12 AM.

    Leave a comment:


  • theriddick
    replied
    I've been testing out wayland this past week (not this version) and its pretty easy to cause conflicts/crashes/UI alignment issues or DPI problems.

    The Fractional scaling system they have for System Displays STILL causes blurring if you select anything fractional.
    Probably a GTK/QT problem but still; you don't experience this on X11 version of the desktop which I often run at %125 scaling. (4k)

    Additionally proton/game performance on Wayland was not better; in fact most games had frame timing stutter issues even if the framerate was above 60fps. Who knows when it will improve.

    Leave a comment:


  • AmericanLocomotive
    replied
    You're acting like the Steam Deck is some choked up Intel Atom powered netbook from 2010. It's a 4c/8t Zen2 APU with 16GB of LPDDR5. The CPU can turbo to 3.5 GHz, and GPU up to 1.6 GHz. It runs AAA games like Doom: Eternal at 60 FPS and high settings. The deck is fast. It's far more performant the vast majority of mobile systems out there.

    Secondly, Steamdeck's native "Gaming Mode" uses Wayland. That means any non-native Wayland game you play (which is most of them), are using XWayland. Squeezing every single drop of performance out of this thing is paramount to Valve, and if XWayland was a problem, they wouldn't be using it. You keep talking about SOME regressions in performance in those benchmarks, while ignoring that in most cases, Wayland was faster. Hence likely why Gamescope uses a Wayland session instead of X.

    Thirdly, Desktop mode on the Steamdeck isn't particularly intended for gaming. So any extra overhead XWayland MIGHT have is not even relevant.

    I will reiterate: I own a Deck, and it's fast. It's not a 2c/2t Celeron in a Chromebook screaming for dear life just checking your email.


    Leave a comment:


  • schmidtbag
    replied
    Originally posted by vegabook View Post
    Then, please explain why you started out your responses explaining that Wayland had no place on a system with limited resources? Is it lighter weight, or not?
    I did explain it, several times. Not my fault you deliberately ignore inconvenient truths.
    Please simply admit, that you did not know that the Steam Deck has a wayland gaming mode, that Xwayland overhead is therefore irrelevant since it will only every be used in desktop mode, where even a Pi does fine.
    I have no problem admitting I didn't know gaming mode used Wayland. I assumed that since desktop mode didn't have it, and since the xwayland overhead does matter in some cases, it would have been sensible that they didn't yet use it. So no, the overhead is not irrelevant.
    I don't know how many times I have to tell you that using Wayland on a Pi makes sense. You're so caught up in disagreeing with me and being overly offended that you don't even realize when I agree with you. Maybe if you would calm down and stop thinking the world is out to get you, this could have been a more civil discussion. My original reply to you wasn't aggressive. You decided to escalate this.
    You've been running around in circles trying to avoid this obvious mistake that you have made, thereby confusing _actual_ noobs who might come to these forums for information.
    Which mistake? I made 3 primary points, each of which you haven't disproven:
    1. Xwayland can have detrimental effects on performance
    2. Wayland is not stable enough for everyone to use yet
    3. If you don't like how the system behaves, then either fix it yourself or use it the way it was intended (for gaming)
    Just because you don't like my answer, doesn't mean it's wrong.
    You were better off spending all this time figuring out how to enable Wayland than whining. I wouldn't be surprised if you could run this and trick whatever the DM is to use Wayland:
    ln -s /usr/share/wayland-sessions/plasmawayland.desktop /usr/share/xsessions/plasma.desktop
    Don't be an idiot and copy plasma.desktop first in case it fails. Don't say I didn't warn you.
    Moreover by asserting, without even owning the device, that you know better than the makers of the device, you've lost credibility completely and utterly.
    Did you actually read this?:
    https://www.phoronix.com/forums/foru...32#post1324532
    That's a best-case scenario with a CPU possessing more resources than a game would ever know what to do with and a GPU that is leaps and bounds better than the Deck's. In most of those tests, the GPU is the bottleneck. And yet, there are cases where Xwayland causes regressions. Tell me, all-wise and all-knowing one: how do you think a 15W APU is supposed to yield better results when the CPU could be the bottleneck? The Deck is expected to play modern titles, even AAA titles. In case you're not aware, 4c/8t isn't always enough for a lot of modern games. So when you have the added overhead of wine and DXVK, stop lying to yourself thinking that Xwayland isn't going to have an impact. Even DXVK is known to hurt performance if you don't have enough threads.
    That article also exemplifies why KDE wasn't the best choice.

    So - what do you have to prove with your credibility? "Because Valve did it!" is not a good reason. Pick any tech company at random and you'll find a sub-optimal decision for a product.

    Wayland has come a long way and considering how well it performs in a lot of cases, it very likely could some day be the best option for gaming. Right now, it isn't.

    Leave a comment:


  • pWe00Iri3e7Z9lHOX2Qx
    replied
    I'd say some other stuff like refactoring how screens are handled in kwin is the most important bit as it could have big improvements for multi-monitor users.

    We’re past the soft feature freeze of the next Plasma release, so it’s a good time to step back and a have look at the work that has been done in KWin during 5.25 development cycle. Ges…

    Leave a comment:


  • vegabook
    replied
    Originally posted by schmidtbag View Post
    Wayland is lighter-weight, Xwayland is inherently heavier. I shouldn't have to spell this out for you.
    .
    Then, please explain why you started out your responses explaining that Wayland had no place on a system with limited resources? Is it lighter weight, or not?

    Please simply admit, that you did not know that the Steam Deck has a wayland gaming mode, that Xwayland overhead is therefore irrelevant since it will only every be used in desktop mode, where even a Pi does fine.

    You've been running around in circles trying to avoid this obvious mistake that you have made, thereby confusing _actual_ noobs who might come to these forums for information.

    Moreover by asserting, without even owning the device, that you know better than the makers of the device, you've lost credibility completely and utterly.
    Last edited by vegabook; 19 May 2022, 04:54 PM.

    Leave a comment:

Working...
X