Announcement

Collapse
No announcement yet.

FreeBSD 13.0 vs. DragonFlyBSD 6.0 vs. Linux On AMD EPYC 7003 Series "Milan"

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

  • FreeBSD 13.0 vs. DragonFlyBSD 6.0 vs. Linux On AMD EPYC 7003 Series "Milan"

    Phoronix: FreeBSD 13.0 vs. DragonFlyBSD 6.0 vs. Linux On AMD EPYC 7003 Series "Milan"

    For those wondering how well the likes of FreeBSD 13.0 and DragonFlyBSD 6.0 performance on AMD's EPYC 7003 "Milan" processors launched earlier this year, here are some initial benchmarks of those BSDs alongside a few Linux distributions. With recently having a Tyan 1U server in the lab with EPYC 7543 32-core processor, I've been running a number of BSD benchmarks on it given these recent BSD releases have been running well on this 1P server.

    https://www.phoronix.com/vr.php?view=30284

  • #2
    The usual & expected outcome:

    1.
    CentOS in the lead because it uses the performance CPU governor & always enables transparent hugepages, which can increase computational performance but also lead to higher (i.e. worse) latency, which is usually not a problem on a server system setup.

    2.
    As always, the various *BSDs are not even worthy of consideration when serious work needs to be accomplished!
    (When again was the last time any BSD showed up in the Top500 list?)

    Comment


    • #3
      Linuxxx
      The usual & expected outcome:

      always: Ubuntu the worst possible defaults, CentOS good defaults, but older than everything else.

      P.S. It seems Shuttleworth does everything to make Linux more m$ like in terms of performance.

      Btw. geometric mean of all tests is so useless it's just depressing to see it included.
      Last edited by Volta; 22 June 2021, 10:03 AM.

      Comment


      • #4
        Competition is good

        Comment


        • #5
          It would be interesting comparing a tuned FreeBSD against a default setup for benchmarks that are compatibles...

          Comment


          • #6
            Originally posted by Linuxxx View Post
            The usual & expected outcome:

            1.
            CentOS in the lead because it uses the performance CPU governor & always enables transparent hugepages, which can increase computational performance but also lead to higher (i.e. worse) latency, which is usually not a problem on a server system setup.

            2.
            As always, the various *BSDs are not even worthy of consideration when serious work needs to be accomplished!
            (When again was the last time any BSD showed up in the Top500 list?)
            Low Quality Bait

            Comment


            • #7
              Thanks Michael for the *BSD benchmarks! Appreciate them greatly. Would love to see OpenBSD and NetBSD get thrown in too, maybe on an older more supported platform like Skylake or Haswell.

              Comment


              • #8
                Why didn’t you include Hurd in this?????

                Comment


                • #9
                  I also want to know about Ubuntu over OpenZFS versus FreeBSD over OpenZFS, maybe same filesystem makes benchmark more valuable

                  Comment

                  Working...
                  X