Announcement

Collapse
No announcement yet.

The First Benchmarks Of Unity On XMir: There's A Performance Hit

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

  • #76
    Originally posted by jayrulez View Post
    I'm going to create a parody game named Angry Nerds. I would like for some of you to send me mugshots at waprave@gmail.com so I can have them catapulted at company/project logos.

    Don't worry, It will be open source.
    Make it BSD just to watch them rage.

    Comment


    • #77
      Originally posted by Sarvatt View Post
      xmir is also using UXA instead of SNA the non xmir results are using, unless it was forced to UXA for those?
      It was using the -intel~xmir ddx for both tests.

      Have you noticed that there is horrible output latency with Xmir? Really unusably horrible, ~1s at times.

      Comment


      • #78
        Originally posted by seb24 View Post
        Not exactly. You link speak about the case of :
        - The CLA don't permit to re-licence the code in proprietary licence. The obligation of publish in the original licence is still here.
        - And in this case the guy explain that it is possible to relicence in BSD licence and after that in Proprietary licence because the BSD licence permit it.
        But in realty I think is wrong. Because if the BSD licence give you this possibility, you still have the obligation to respect the CLA... So you can't.

        To resume they are wrong. The owner (for example Canonical) have the obligation to respect the CLA. And republish the code in the original licence + another licence.
        I'm pretty sure he's exactly right because he's the legal expert here.

        Richard is Intellectual Property Counsel at Red Hat. His work focuses on software licensing (including open source), copyright, patent and trademark issues, and intellectual property issues in standards organizations.

        Prior to joining Red Hat, Richard was Counsel at the Software Freedom Law Center, where he advised free software and open source projects and foundations. He was co-author, with Richard M. Stallman and Eben Moglen, of version 3 of the GNU General Public License (GPLv3), the first update of the most widely used free software/open source license in over 15 years.
        Originally posted by chrisb
        Changing the license of Mir won't eliminate GPLv3 software from Ubuntu Touch.
        It's one of the very few GPLv3 licensed bits in the "core os" (coreutils and other GNU software can be replaced; Tizen for example uses pre-GPLv3 coreutils from GNU). If some one wants to create a competing platform that still supports Ubuntu applications it could be an issue; Unity being GPLv3 not so much.

        But yeah. It speaks mostly against Mir as a standalone project.

        Comment


        • #79
          Originally posted by ickle View Post
          It was using the -intel~xmir ddx for both tests.

          Have you noticed that there is horrible output latency with Xmir? Really unusably horrible, ~1s at times.
          and this is why it's really bad to use Xmir for a DE

          Comment


          • #80
            Originally posted by jayrulez View Post
            I'm going to create a parody game named Angry Nerds. I would like for some of you to send me mugshots at waprave@gmail.com so I can have them catapulted at company/project logos.

            Don't worry, It will be open source.
            I'd humbly suggest you'd dual license it GPLv3 and BSD for maximum effect

            Comment


            • #81
              Originally posted by Teho View Post
              It's one of the very few GPLv3 licensed bits in the "core os" (coreutils and other GNU software can be replaced; Tizen for example uses pre-GPLv3 coreutils from GNU). If some one wants to create a competing platform that still supports Ubuntu applications it could be an issue; Unity being GPLv3 not so much.
              Yeah, but if Canonical allow the user to install a single GPLv3 licensed app, then they have to give the user the ability to replace that app with a modified version. So they would have to remove/filter all GPLv3 software out of the Ubuntu Software Centre if they want to be able to ship a completely locked down device.

              Comment


              • #82
                Originally posted by erendorn View Post
                I'd humbly suggest you'd dual license it GPLv3 and BSD for maximum effect
                Contributors would have to sign a CLA to assign copyrights(+ All their freedoms and their first born) to me so I can release GPLv3 contributions as BSD as well.

                Comment


                • #83
                  Originally posted by spacetoilet View Post
                  and this is why it's really bad to use Xmir for a DE
                  I love how you've made this decision based on the results of one guy with one peice of hardware on one a forum on the internet. That's such a huge sample size. N=1 is total significant. Let's just pretend that Michael's results and that video don't count as evidence.

                  Comment


                  • #84
                    Hmm, does XMir/XWayland start just one X server, or one X server per app? Or can it do it both ways?

                    Comment


                    • #85
                      Originally posted by dh04000 View Post
                      I love how you've made this decision based on the results of one guy with one peice of hardware on one a forum on the internet. That's such a huge sample size. N=1 is total significant. Let's just pretend that Michael's results and that video don't count as evidence.
                      The Video was bad it shows all types of lag bug's etc and Btw The One Guy is from intel....

                      Michael's results are based off from old apps showing a FPS lost on old apps is really bad in Xmir vs Xorg Xmir lost
                      Last edited by spacetoilet; 06-28-2013, 04:28 PM.

                      Comment


                      • #86
                        Originally posted by spacetoilet View Post
                        Michael's results are based off from old apps showing a FPS lost on old apps is really bad in Xmir vs Xorg Xmir lost

                        If an application loses 30% of performace, that comes no were near 1 secon (1000ms) lag.

                        Comment


                        • #87
                          Originally posted by GreatEmerald View Post
                          Hmm, does XMir/XWayland start just one X server, or one X server per app? Or can it do it both ways?
                          The X server inside shouldn't be aware of Mir, so it probably starts only one server.

                          Originally posted by dh04000 View Post
                          I love how you've made this decision based on the results of one guy with one peice of hardware on one a forum on the internet. That's such a huge sample size. N=1 is total significant. Let's just pretend that Michael's results and that video don't count as evidence.
                          It's still useless to do so, and that doesn't depend on samplesize. It's useless because, in the best possible case, starting an X server on other display platform can achieve the same performance as X, if using the whole set of features (like handling composition and such), while it doesn't provide any advantage, since you are running everything from inside this X server. You make no use of any Mir or Wayland provided advantage this way.

                          Originally posted by chrisb View Post
                          Yeah, but if Canonical allow the user to install a single GPLv3 licensed app, then they have to give the user the ability to replace that app with a modified version. So they would have to remove/filter all GPLv3 software out of the Ubuntu Software Centre if they want to be able to ship a completely locked down device.
                          That's as easy as changing the repos. It still uses apt IIRC.

                          Originally posted by Ericg View Post
                          Not necessarily. if you actually look at the architecture of XWayland and XMir you actually avoid a few expensive steps at the end vs real X. In theory, its possible for us to have slightly better performance via XMir and XWayland than Real X. But obviously an alpha-quality release won't have the optimizations and design to actually make that claim true just yet.
                          I don't think that's true when running a full-fledged desktop on top of it.

                          Originally posted by spacetoilet View Post
                          BO$$ you really need to stop sucking off Canonical

                          Redhat Intel Raspberry pi Kubuntu Debian Carsten Munk Tizen Sailfish OS and any non Ubuntu Unity is backing Wayland most of the code in Mir is just a fork of Wayland Android Carsten Munk work Libhybris making it even more buggy shitty you dont know how long it's going to take to remove all the bug's form Mir it may take years do to it trying to move away form the code it's based off from and it's the most Epic Fail of the Linux World Wayland the way Canonical is going is why i jumped ship it's going to be really bad for any one Developing on Linux Canonical just Fucked all Linux Develpers Linux Gamers Linux Users do you know how hard it is going to be to get Windows Developers to look at Linux now?

                          https://www.youtube.com/watch?featur...&v=-V8i8zZPzbU

                          http://www.youtube.com/watch?v=NrjbeE90vVY
                          Can you use periods and commas, please? It's really hard to make any sense of what you wrote this way.

                          Originally posted by seb24 View Post
                          And that relates how?

                          Originally posted by TheOne View Post
                          I have a big query.

                          If Canonical just copy/pasted the XWayland code to make XMir then every person saying that XWayland will be faster that plain X is utterly wrong. We haven't seen any benchmarks regarding XWayland/Wayland, so if XMir is really a copy/paste effort Wayland should be similarly on the same shape.
                          The people who says XWayland will be faster than plain X means using it for apps. Running a desktop implies a lot more work on the X.org server inside that you can ignore otherwise. You are having window management, compositing, etc, inside the X.org server. When you use it for an app, the app inside the server will behave (if I get it right) somewhat as a fullscreen window in a non-compositing environment, and all the real management will happen outside the X.org server, being done by either Wayland or Mir. So you avoid lots of operations. Of course a whole desktop running on XWayland will be slower than at least a native desktop on Wayland, and almost certainly than a desktop on X.
                          And the same way XWayland could be faster than X for particular apps it's likely that, assuming Mir will be faster than X, XMir will be faster than X for particular apps. Whole desktops are out of scope for XMir and XWayland for common sense. There is no reason to not use X if you are going to use an X desktop, since you will be unable to load Mir or Wayland apps.

                          Originally posted by BO$$ View Post
                          I'm curious about performance. Native Wayland and native Mir application. Games especially.
                          If you want a demo of it, I plan to port a Wolfenstein 3D (seriously, I mean it, I know it's not a really advanced game, but it uses OpenGL) to Wayland. I might add Mir, for the sake of an early comparison, at least until someone serious (a professional) decides its engine should run on both Mir and Wayland).

                          Originally posted by seb24 View Post
                          Stop fighting... The both project are free software with different objective. We should be happy to have a dynamic eco-system. that's it.
                          And we will see in the future how the 2 project will evolves.
                          Nope, the problem is they actually seem to share the same objective.

                          Originally posted by seb24 View Post
                          No "version" history here.
                          If your "contribution" is in the Version 2 they have the obligation to publish it in the original licence + other sub-licence.
                          No, they don't. They already published your submission as GPLv3 in version 1, fulfilling their promise. Your contribution has been released with the license the project had when you submitted, so they can sublicense. Version 2 can use the sublicensed version of your code, so they have no obligation to release it with the original one.

                          Originally posted by chrisb View Post
                          So the difference is that, with Mir, only one company could possibly make a proprietary version, whilst with Wayland, every company can make a proprietary version. Neither solution is optimal for open source, but I'm not seeing why the latter would be better.
                          None is better for every scenario. The latter is only more equal, since it gives exactly the same rights to everyone, while the other allows only Canonical or their partners to work on closed derivatives. For a start, this means Mir can use Wayland's code (and it does) but not the other way around.

                          Originally posted by chrisb View Post
                          Changing the license of Mir won't eliminate GPLv3 software from Ubuntu Touch.
                          No, it doesn't. MIT licensed forks doesn't, either. Both cases, though, allow to use the codebase without releasing the changes. One of them, equally for everyone who wants this, and the other only if Canonical lets you.

                          Originally posted by scottishduck View Post
                          Again, wrong. No plans for Mir migration until 14.04 at the earliest.
                          Read the news, Unity will be running on XMir by default for 13.10.

                          Originally posted by jrch2k8 View Post
                          So yes mir got coded faster because wayland + community did all the massive heavy lifting while canonical waited[without 1 freaking commit] until it was good enough for them to start and in some cases take solutions from wayland code[read their bazaar and wayland git and you will see some funny things in there]
                          Actually, they have ONE commit on XWayland, it was in may 2012 IIRC. There was a Phoronix article about it.

                          Originally posted by schmidtbag View Post
                          * MUCH faster development
                          At the cost of being a Unity only solution. That's the sole reason it develops faster.
                          * Supposed to get Android driver support
                          Over Wayland? It achieves this using a library developed for Wayland.
                          * A seemingly more devoted team
                          ?
                          * A fully open source license
                          Both have a fully open source license. In fact, it's easier (i.e., you don't need anyone's permission) to fork Wayland on a closed source project. With Mir you can, if Canonical allows you, because of the CLA, but you can't without their permission, because of the GPL license.
                          * Probably will be more light-weight in the end
                          I don't know what is this based in. Any of the two might be the lighter one. Wayland because of being just the protocol, or Mir because of only implementing what Unity needs, instead of having to be usable for the whole ecosystem.

                          Originally posted by chrisb View Post
                          You misunderstand. Unity will run on Mir. XMir is a compatibility layer that will be used to run old X11 applications that aren't written using a modern toolkit.
                          No, you are ignoring quite a few news. Unity will run on XMir for 13.10, and at the moment it seems flavors will run on XMir, at least eventually. WMs at the very least need to be aware of the display architecture they run on, can't target the toolkit only.

                          http://www.phoronix.com/scan.php?pag...tem&px=MTM5NzM
                          Brief version: 13.10 will ship Mir by default, with Unity 7 running on XMir.

                          I wonder how do people manage to turn that into "Canonical will not ship Mir until 14.10", when it clearly says it will ship it (READ THE TITLE) in 13.10, and will use Unity 8 on 14.04, which runs natively on Mir.

                          Comment


                          • #88
                            Originally posted by mangecoeur View Post
                            Also want to see XWayland for comparison, it should give a good indication of the comparative state of the two technologies. I suspect the fundamental techniques used in both are quite similar so I wouldn't expect there to be a significant performance difference between the two, at least not in the long run. I'd also be interested to see how gnome/kde run under XMir.

                            For X vs Mir vs Wayland it seems that anything QT based would be a good start since there's both QTMir and QTWayland backends (though i don't know what state the wayland one is in these days), and soon enough there should KDE on wayland to do a Unity8 vs KDE comparison (although that said i doubt many people are going to make their DE choice based on raw performance numbers).
                            Do not expect so much. XMir is just the ugly copy of XWayland, but still it's possible that some beginner in Canonical has ruined something in the copy-and-paste process.
                            A whole DE running on top of XMir is just a smoke in the eyes excepts for some advantages like:
                            1) lost in performance
                            2) all user forced to became bug hunters even with an official release and not some alpha/beta/RC
                            3) Another layer for the sake to insert another layer and then start with the propaganda "we are so fast to write code"!

                            Comment


                            • #89
                              Originally posted by mrugiero View Post
                              I wonder how do people manage to turn that into "Canonical will not ship Mir until 14.10", when it clearly says it will ship it (READ THE TITLE) in 13.10, and will use Unity 8 on 14.04, which runs natively on Mir.
                              it says, it is going to ship Unity 8 on 14.10.

                              Comment


                              • #90
                                Originally posted by spacetoilet View Post
                                it says, it is going to ship Unity 8 on 14.10.
                                Right, I skimmed it the last time I reread it to check if I was just misreading when everyone said Mir would be 'at the earliest' in 14.04. It clearly says, tho, Mir (through XMir) will be used by default in 13.10, without X.org server fallback in 14.04. Also, this strengthens my point that Unity will actually run in real life on XMir, not only in 13.10, but in the LTS 14.04.

                                Comment

                                Working...
                                X