Announcement

Collapse
No announcement yet.

8-Way Spring 2020 Linux Distribution Performance Comparison With 240+ Benchmarks

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

  • 8-Way Spring 2020 Linux Distribution Performance Comparison With 240+ Benchmarks

    Phoronix: 8-Way Spring 2020 Linux Distribution Performance Comparison With 240+ Benchmarks

    Given the recent releases of Fedora 32, Ubuntu 20.04 LTS, Manjaro 20.0, and others, here is a fresh eight-way Linux distribution benchmark comparison.

    http://www.phoronix.com/vr.php?view=29139

  • #2
    Holy shit, Michael. TYVM.

    These results are making me consider turning off apparmor=1 on Manjaro. AFAICT, that's one of the things the ones at the bottom of the Geometrics have...or the tops of the last place one....

    Comment


    • #3
      Interesting. So the Arch-based distros like Manjaro & EndeavourOS are probably at the bottom because they ship a (mostly) vanilla Linux kernel?

      Comment


      • #4
        I must be in the wrong hemisphere

        Comment


        • #5
          I'm impressed by ClearLinux (in a positive sense) and Manjaro (in a negative sense).

          I was expecting ClearLinux to perform better but to - in some cases - double the performance. Specifically in the video encoding tests. That's insane! Good to know
          But for Manjaro (or rather ArchLinux), they must be able to flip some switches to get more near the performance of ClearLinux. Right?

          Comment


          • #6
            Originally posted by markg85 View Post
            I'm impressed by ClearLinux (in a positive sense) and Manjaro (in a negative sense).

            I was expecting ClearLinux to perform better but to - in some cases - double the performance. Specifically in the video encoding tests. That's insane! Good to know
            But for Manjaro (or rather ArchLinux), they must be able to flip some switches to get more near the performance of ClearLinux. Right?
            Manjaro is a derivative of Arch its not really representative of Arch technically (that's like transposing Ubuntu and Debian performance), if Endeavour is made the same as its predecessor it should be more indicative of Arch performance

            Comment


            • #7
              Originally posted by markg85 View Post
              I'm impressed by ClearLinux (in a positive sense) and Manjaro (in a negative sense).

              I was expecting ClearLinux to perform better but to - in some cases - double the performance. Specifically in the video encoding tests. That's insane! Good to know
              But for Manjaro (or rather ArchLinux), they must be able to flip some switches to get more near the performance of ClearLinux. Right?
              I'm not phased about Manjaro's relative bad performance in gaming or video encoding aspects, but what shocked me was the relatively slow compile times. It appears that Manjaro and Debian-testing were using the same version of GCC and Debain-testing did quite well. Something is very wrong there.

              PS: Thanks for the tests Michael!

              Comment


              • #8
                Thanks for the tests, Michael! I really appreciate the effort.

                Now to my question: why do you think did Manjaro perform worse than the others in these tests (Manjaro KDE user here)? Is there something that can be tweaked to improve the performance relative to the other distros?

                Comment


                • #9
                  Tell you what, the lead Clear Linux is showing in the codec space is making it economically desirable for a video transcode & quality farm.

                  It would be interesting to see how Clear Linux handles the codec-compare test from the IETF.

                  Comment


                  • #10
                    I only found 2 tests in which Debian Testing was last, and again, not by much. It makes no sense for it to be last in the Geometric mean.

                    My guess is that since it could not run many of the tests, it gets a score of 0 (or whatever weight you use) and it then appears to be the slowest, by far.

                    This realization now makes me wonder the interpretation of this graph for many past results (I often look at the last graph for a better overall picture, but I now see that it is misleading). It would be best to normalize the results somehow.

                    Edit: oh, forgot many more tests are included than the ones shown here. I'll look at that later on, perhaps I'm mistaken.

                    Edit2: see Michael's reply in the comments. Disregard this one!
                    Last edited by franglais125; 05-06-2020, 11:03 AM.

                    Comment

                    Working...
                    X