Announcement

Collapse
No announcement yet.

Power Use, RAM + Boot Times With Unity, Xfce, GNOME, LXDE, Budgie & KDE Plasma

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

  • Originally posted by Noee View Post
    TW has the same issue Martin describes above. For every TW install I've done recently, when first installing the GNome DE and then installing the Plasma5 DE, I see a many of the GNome services running while I'm using Plasma. It works the other way too, meaning in GNome, you will see Plasma components running. I don't know about other distros and I don't know how Michael installed the DEs, but on TW, you have to manually remove the DE using zypper or Yast/Software Mgt and delete the components manually.

    I literally just did this on the last TW update on a very, very old machine (2x250Gb WD blacks and a BE-5000+ with 3Gb RAM). I had to remove all the GNome stuff manually (control center, GOA stuff, etc.) and add locks to taboo the GNome stuff, and then when I get to the Plasma desktop, I'm sitting right there under 500MB.
    I have related issues on Arch but there it seems Gnome's fault.
    If you happen to have KDE and Gnome installed in parallel and run a KDE session, Gnome glorious services (especially my friend tracker) run in background.
    I didn't had these issues when running the Gnome session. So KDE's services was fine for me.
    Seems like the desktop-files are wrong configured.
    You can fix them for every service by hand e.g. define in which DE not to start or report the issues upstream.
    If I switch the DE for testing, I just uninstall the other DE's completely. It's pretty easy with pacman.

    Comment


    • I prefer Plasma to gnome desktop because of layout besides it seems that some utility has problem on gnome (Fedora) as VLC and the possibility to install DVDs codec to reproduce them.

      Comment


      • Originally posted by GhostOfFunkS View Post

        Except that a session internalize the needed extensions and QA to the same level as normal distribution packaging and dependency resolving. So you basically have no clue about how this works. You should consult the Ubuntu developers and users, at least they understand this.
        Perhaps next time when you reply to something you should actually address the points. Better still perhaps you could learn some basic programming skills, so you can be a bit more informed when you are trolling instead of coming across as an obsessional lunatic.

        Comment


        • I don't see importance of RAM usage and why people should focus on that instead of general "feel" of the DE. It also depends how you measure RAM usage, I would assume that any other method (system monitoring tools etc.) than "free -m" in terminal is invalid method. Using this method, I can confirm that in VIrtualBox with 4GiB of RAM alocated to the machine, KDE Plasma uses less RAM, but not by much. That was tested with Ubuntu 17.04 (to be precise with Ubuntu GNOME 17.04 and Kubuntu 17.04).

          So, Kubuntu 17.04 results for "free -m" after fresh boot and wait time of 4 minutes are:
          total=3949, used=402, free=2819, shared=7, buff/cache=727, available=3321

          Results for Ubuntu GNOME 17.04 under same conditions as above (same wait period):
          total=3949, used=697, free=2727, shared=8, buff/cache=524, available=3029

          So according to free memory, diffeence is about 100MB in favor of KDE, and according to available memory, difference is about 300MB in favor of KDE Plasma. I would argue that real difference is in free, but, i can accept availalbe memory as difference also...

          Other difference is that KDE Plasma runs very slugish under VIrtualBox compared to Ubuntu GNOME, and just to confirm that Ubuntu maintainers are not biased towards GNOME, and that they equaly mess up things for both GNOME and KDE is the fact that Ubuntu GNOME "booted when it wanted to boot" in VirtualBox, it more often didn't boot... (with OK dmsg - Starting Firmware Update Dameon, Reached Target Network Online etc. I simply did not wait more than minute to see if it will boot or not...), that could be VirtualBox bug, but since this problem can be traced back to Ubuntu GNOME 15.xx versions, and since it only happens on Ubuntu GNOME (other distributions I've tested in past did not had that problem) It's safe toa ssume it is Ubuntu problem.

          So, as you can see the difference isn't that big to begin with considering RAM usage.

          Comment


          • Originally posted by GhostOfFunkS View Post

            ... You should understand the basic concept of a session. Ubuntu developer Didier posted a plethora of blogs how they did it in Ubuntu.
            Wow I'm totally blown away that their default session loads JavaScript extensions. Not. Try harder next time.

            Comment


            • Originally posted by ArchFanatic View Post
              My KDE Plasma eats only 382 MiB with Dolphin, Kate and System Monitor running and it is much smoother than gnome on both wayland and x11.
              https://image.ibb.co/jyCjHQ/Screensh...831_153509.png
              Well, every time I try a new KDE iteration it's pretty and everything but always sluggish, even after tweaking cpu & effect options... then i go back to Gnome and everything is smooth and fast, though not as pretty, but I can live with that.

              Comment


              • Originally posted by GhostOfFunkS View Post

                Nah it is you who need to try understand WHY this is important and how it impacts branding, maintenance and QA.
                Oh dear, now you are plucking things from thin air. You couldn't make it up.

                Comment


                • Kde5/Plasma all the way for me. I just don't get along with the Gnome (aka oss OsX) philosophy so that's out and the rest lack some important features (wobbly windows just a few clicks away and so on..... yeah I want my bling). The ram use difference is tiny since I have 32 gb and Chromium eats much more with a 100+ tabs (no I don't need them but that's how I browse) anyway and most of the boot time is luks checking the pw (I set the number of iterations so that it takes 90s), everything else is nothing compared to that, I like using some of the lighter ones in vms although even there I'll mostly just use plasma/kde5.

                  I tried the various tiling windows managers like i3 but again just didn't get the appeal. Maybe if I had 3 8k screens.

                  Comment


                  • Originally posted by boxie View Post

                    really? that sounds like you are spouting shit that you know nothing of

                    Been using KDE's plasma in anger now for more than a year on the neon distro, can't say I find anything you said here to be true.

                    RAM is there to be used, it makes things fast - why not use it?
                    How much RAM is enough? There are plenty of laptops being sold which have just 2 gigs of RAM and no upgrade option. For their users every meg counts. TDE/KDE 3.5.10 consume less than 120MB of RAM. Your shiny modern KDE has this process cold Plasma which takes up to 300MB of RAM with zero plasmoids running. Why are people so full of shit?

                    Also go fuck yourself, dick. Unlike you I've been using Linux for almost 20 years now. What about you, a recent Windows 10 convert?

                    Also, using more RAM has never equated to being faster - actually the opposite is true. Old applications written in assembler/C run circles around their modern shiny JavaScript-everywhere counterparts which consume gigs of RAM.

                    You should drink a lot less Kool Aid, dude. More RAM usage means faster applications, my ass. Never read such a heresy in my entire life.
                    Last edited by birdie; 02 September 2017, 06:37 PM.

                    Comment


                    • There's plenty of resources in modern computers, but it does not explain why one application needs significantly more resources than another. From my point of view, resource hogs most probably do something wrong, their quality as pieces of software might be lower overall. I recently did some benchmarking of Libreoffice Calc vs gnumeric. Surprisingly, LO Calc was several times and several dozens of seconds slower than gnumeric when importing some quite normal looking .csv files. (For those who don't know, gnumeric is a light weight spreadsheet program while LO Calc is the de facto standard in the Linux world - also notorious for it's fucked up development style with german comments, dependencies on several scripting languages, several years of bug fixing just to make it usable in day to day jobs). I'm going to analyze it further and file some bugs, but IMO this just shows that when something kind of works on modern machines, but isn't very close to being optimal, it can be a show stopper on older hardware. Imagine things like building kiosk PCs with Raspberries or some clones. It's just not possible if the machine has 1 to 2 gigs of RAM in total, with few hundred megabytes dedicated to graphics/GPU. I'm pretty tired of tuning some low end machines to use openbox when one needs to save resources. I used KDE in 2005 and my relatively okayish machine had 1 GB, my next laptop had 1 GB of RAM. I bought it as new. Now, they're telling me the desktop uses all of my RAM. It's getting ridiculous.

                      Comment

                      Working...
                      X