Announcement

Collapse
No announcement yet.

GNOME 3.38's Mutter Gets More Optimizations - ~10% Lower Render Time In Some Scenarios

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

  • #21
    Originally posted by Azpegath View Post
    But most of the performance enhancements listed is for Gnome Shell, no? I thought most of the work on Intel 4K rendering was Shell-focused.
    I tend to think its both of them.

    Comment


    • #22
      Originally posted by Azpegath View Post

      But most of the performance enhancements listed is for Gnome Shell, no? I thought most of the work on Intel 4K rendering was Shell-focused.
      This post is mutter focused, could be true as well, but really, it GS needs them the most.

      Comment


      • #23
        Originally posted by leipero View Post
        This post is mutter focused, could be true as well, but really, it GS needs them the most.
        If you browse through the older posts regarding performance improvements, I think they are mostly GS-focused.

        Comment


        • #24
          10% faster is 10% less overhead on mutter? 🤔

          Comment


          • #25
            Originally posted by dragon321 View Post

            It would be hard for them because they have less developers than GNOME. If I recall correctly they stated on GitHub they will keep eyes on Mutter improvements. I wonder if it's possible to port Cinnamon Shell to Mutter without massive rewrite.
            I thought Muffin was going to use Mir? In that case, it would also require rewriting Mutter to support Mir. That would be a real hassle.

            Comment


            • #26
              Mutter is already "modular" - Gnome shell is a plugin for it.

              As for Cinnamon, they need to rebase both their mutter and gnome-shell forks. If they cant keep up, they need to return to being a set of extensions on top of gnome shell. otherwise they are letting their users down.

              Comment


              • #27
                Originally posted by Vistaus View Post

                I thought Muffin was going to use Mir? In that case, it would also require rewriting Mutter to support Mir. That would be a real hassle.
                I don't know. Last time I saw on GitHub they stated they are not interested in Wayland now. Mir is different compositor so moving Cinnamon to it would require massive porting. I think they won't do it. Probably they will continue to improve Muffin and someday implement Wayland or maybe they will move Cinnamon Shell to Mutter. Keep in mind these are my predictions not actual Cinnamon developers plans.

                Comment


                • #28
                  Originally posted by Vistaus View Post

                  I thought Muffin was going to use Mir? In that case, it would also require rewriting Mutter to support Mir. That would be a real hassle.
                  Canonical has already buried Mir a long time ago.
                  Last time I read about it a few years ago Canonical wanted to salvage the project by focussing on IoT hardware. But I don't think that turned out well either.
                  When Canonical laid off a big chunk of people a few years ago, the Mir team and the Unity 8 desktop team had to bear almost all of it. So I would consider Mir dead.

                  Check their website out:

                  Their latest stuff is from june 2018, and they were already mentioning Wayland in their articles.
                  That's because they changed Mir from being a competitor protocol to a set of Wayland helper libraries. Lulz.
                  Last edited by Degra; 02 September 2020, 02:43 PM.

                  Comment


                  • #29
                    Originally posted by Degra View Post
                    Canonical has already buried Mir a long time ago. -- Their latest stuff is from june 2018, and they were already mentioning Wayland in their articles.
                    That's because they changed Mir from being a competitor protocol to a set of Wayland helper libraries. Lulz.
                    I wouldn't laugh at it, though. Wayland is a mess until different DEs come together to create a standard implementation which will be shared by everyone. Mir would be an excellent candidate as its origins are in serious desktop environment development. There are some other similar projects but without the same ambition and foundation as Mir.

                    It is quite useless to read every month how Mutter is receiving yet more patches to fix its obvious performance issues. It is frustrating that we don't even really have monthly news about Kwin making actual progress with Wayland. It's sad that smaller community-driven projects like Cinnamon will never be able to make the transition unless they throw away all their own work and revert to upstream Mutter code. (I feel XFCE will have to do something similar.)

                    If there was one solid standard implementation, we would not have this utter mess in our hands. Wayland would be here in 2021, not maybe 2023 on one DE and possibly 2025 on another.

                    Comment


                    • #30
                      144Hz What the hell are "distributors" and what do they have to do with DEs? Only Gnome is using Mutter, and AFAIK nobody else has even considerations about switching to it. I don't care what some random group of irrelevant people think.

                      Besides Mutter does not qualify, really, not in its current state. At least on Arch it has got some serious dependencies to Gnome libraries and even the Gnome desktop environment: gnome-desktop, gnome-settings-daemon, dconf. And of course GTK. My hunch is that Mutter is basically written in GTK and cannot be decoupled without rewriting it altogether.
                      Last edited by curfew; 03 September 2020, 01:03 AM.

                      Comment

                      Working...
                      X