Announcement

Collapse
No announcement yet.

15-Way Linux OS Comparison Shows Mixed High-Performing Linux Distributions

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

  • 15-Way Linux OS Comparison Shows Mixed High-Performing Linux Distributions

    Phoronix: 15-Way Linux OS Comparison Shows Mixed High-Performing Linux Distributions

    Succeeding January's 10-way Linux distribution battle is now a 15-way Linux distribution comparison on an Intel Xeon "Skylake" system with Radeon R7 graphics. Distributions part of this Linux OS performance showdown include Fedora, Ubuntu, Debian, CentOS, OpenSUSE, Antergos, Sabayon, Void Linux, Zenwalk, KaOS, Clear Linux, and Alpine Linux.

    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

  • #2
    Michael the system were updated after the installation before beginning the tests?

    I ask this because you said that "Fedora 23 was tested that with currently available updates", and it is not clear if this "apt-get update & apt-get upgrade" happened to all of the systems in "The out-of-the-box packages, OS settings (CFLAGS/CXXFLAGS, etc), and other defaults were used in making for a comparable and standardized experience for comparing these fifteen Linux distributions due to open-source software being endlessly configurable"

    Comment


    • #3
      Originally posted by andrei_me View Post
      Michael the system were updated after the installation before beginning the tests?

      I ask this because you said that "Fedora 23 was tested that with currently available updates", and it is not clear if this "apt-get update & apt-get upgrade" happened to all of the systems in "The out-of-the-box packages, OS settings (CFLAGS/CXXFLAGS, etc), and other defaults were used in making for a comparable and standardized experience for comparing these fifteen Linux distributions due to open-source software being endlessly configurable"
      Yes, distribution stable release updates were installed at time of testing, sorry for not making it more clear.
      Michael Larabel
      https://www.michaellarabel.com/

      Comment


      • #4
        Very nice. Looks like CentOS is pooping it's pants. I like the average performance of sabayon; considering it's a fat and heavy KDE distro in some cases it's performing surprisingly well. Also surprised me how it was one of the fastest for using the Go language.

        All in all the benchmarks are rather balanced out and pretty much all over the place, don't think there's a clear winner here. But there were some distros that leaned pretty heavily towords the bottom end of charts (CentOS) and some more towards the middle; while yet others were all over the map. Seems like Debian Testing did really well.

        Comment


        • #5
          Originally posted by rabcor View Post
          Seems like Debian Testing did really well.
          Being a fan of Arch, I was surprised how well Debian Testing were on this showdown

          Comment


          • #6
            Michael

            The Arch-based Antergos 16.2-rolling distribution had Linux 4.4.3, Mesa 11.1, and GCC 4.3.0
            GCC should be 5.3.0

            Comment


            • #7
              I realize there are some differences between the distributions but why does the sqllite benchmark perform so different on centos compared to alpine linux? Is it the filesystem? Is there something running in the background messing with the test? If they using the same kernel version and same filesytsem the results shouldn't differ much, right?

              I'm more interested in seeings benchmarks of distributions with applications disabled, fullscreen vs windowed, different desktop environments, etc. I think pure distribution benchmarks are too highlevel. The differences seem to be in gcc versions, kernel versions, etc and not really distributions.

              Comment


              • #8
                Originally posted by CapsAdmin View Post
                I realize there are some differences between the distributions but why does the sqllite benchmark perform so different on centos compared to alpine linux? Is it the filesystem? Is there something running in the background messing with the test? If they using the same kernel version and same filesytsem the results shouldn't differ much, right?

                I'm more interested in seeings benchmarks of distributions with applications disabled, fullscreen vs windowed, different desktop environments, etc. I think pure distribution benchmarks are too highlevel. The differences seem to be in gcc versions, kernel versions, etc and not really distributions.

                Yeah that sqlite result is suspicious. There's so much variation that you really need to understand that in order to trust the rest of the testing.

                Comment


                • #9
                  Originally posted by andrei_me View Post

                  Being a fan of Arch, I was surprised how well Debian Testing were on this showdown
                  Debian Testing is rolling as Arch, nothing surprisingly

                  Comment


                  • #10
                    Originally posted by Sloth View Post


                    Yeah that sqlite result is suspicious. There's so much variation that you really need to understand that in order to trust the rest of the testing.
                    Could it be because of SELinux?

                    Comment

                    Working...
                    X