Announcement

Collapse
No announcement yet.

8-Way Linux Distribution Benchmarks On The AMD EPYC 7742 2P Server

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

  • 8-Way Linux Distribution Benchmarks On The AMD EPYC 7742 2P Server

    Phoronix: 8-Way Linux Distribution Benchmarks On The AMD EPYC 7742 2P Server

    A few days ago I provided some benchmarks showing how running Intel's open-source Clear Linux on AMD EPYC Rome can provide some significant speed-ups over Ubuntu Linux, but how do other Linux distributions compare on AMD's new Zen 2 server processors? Here is an eight-way benchmark comparison on the AMD EPYC 7742 2P Daytona server with its 128 cores / 256 threads.

    http://www.phoronix.com/vr.php?view=28218

  • #2
    Originally posted by phoronix View Post
    The fastest NetBench video performance was on Red Hat Enterprise Linux 8.0 followed by Clear Linux.
    Sorry? Isn't it "NeatBench"?

    Comment


    • #3
      Originally posted by tildearrow View Post

      Sorry? Isn't it "NeatBench"?
      Whoops, yep. Thanks.
      Michael Larabel
      http://www.michaellarabel.com/

      Comment


      • #4
        How much does clear linux pay for these fake benchmarks? H265 you add so many optimization commands to clear and none to SUSE...

        Comment


        • #5
          Originally posted by Chaython View Post
          How much does clear linux pay for these fake benchmarks? H265 you add so many optimization commands to clear and none to SUSE...
          Fake? Check it for yourself if you don't believe it.
          Clear Linux adds lots of optimization flags in order to be the fastest Linux distro ever.
          Other distros don't do that.

          Comment


          • #6
            Originally posted by Chaython View Post
            How much does clear linux pay for these fake benchmarks? H265 you add so many optimization commands to clear and none to SUSE...
            All compiler flags were at their default. Clear does have more aggressive compiler flag tuning as pointed out countless times in many articles but in the x265 tests for instance that really isn't relevant in the case of openSUSE. OpenSUSE had the same default CFLAGS as the other non-Clear Linux distributions yet performed so poorly (it just doesn't show the numa flag due to no space to render).
            Michael Larabel
            http://www.michaellarabel.com/

            Comment


            • #7
              Originally posted by Michael View Post

              All compiler flags were at their default. Clear does have more aggressive compiler flag tuning as pointed out countless times in many articles but in the x265 tests for instance that really isn't relevant in the case of openSUSE. OpenSUSE had the same default CFLAGS as the other non-Clear Linux distributions yet performed so poorly (it just doesn't show the numa flag due to no space to render).
              I have said this previously, but let me repeat:

              You compile the x265 binary. Make sure you have *all* required dependencies installed. Your recipe lists yasm, but x265 requires nasm:
              https://github.com/phoronix-test-sui...nition.xml#L20

              You should include the output of configure/cmake/meson/whatever for all tools you compile yourself, so results can be verified.

              For x265, it even reports the CPU extensions usable:
              Code:
              x265 --version
              x265 [info]: HEVC encoder version 3.1.2+1-76650bab70f9
              x265 [info]: build info [Linux][GCC 9.2.1][64 bit] 8bit+10bit+12bit
              x265 [info]: using cpu capabilities: MMX2 SSE2Fast LZCNT SSSE3 SSE4.2 AVX FMA3 BMI2 AVX2

              Comment


              • #8
                Originally posted by tildearrow View Post

                Fake? Check it for yourself if you don't believe it.
                Clear Linux adds lots of optimization flags in order to be the fastest Linux distro ever.
                Other distros don't do that.
                Fastest at benchmarks. They admitted through their own survey recently that hardly anyone uses it for actual work.

                Comment


                • #9
                  I know Michael is super busy, but I would love to see a BSD round up on a system with this many threads. FreeBSD vs DragonFly would be very interesting since they target high end high core count server/workstations like this but I would also like to see how OpenBSD and NetBSD perform on a system with this much horsepower especially since those projects haven't focused so much on extreme multithreading of the kernel as much as Free and DragonFly have. Might suggest using UFS as the file system in all tests so that ZFS and Hammer don't skew the results.

                  Comment


                  • #10
                    Originally posted by kylew77 View Post
                    I know Michael is super busy, but I would love to see a BSD round up on a system with this many threads. FreeBSD vs DragonFly would be very interesting since they target high end high core count server/workstations like this but I would also like to see how OpenBSD and NetBSD perform on a system with this much horsepower especially since those projects haven't focused so much on extreme multithreading of the kernel as much as Free and DragonFly have. Might suggest using UFS as the file system in all tests so that ZFS and Hammer don't skew the results.
                    FreeBSD 12 UEFI wasn't booting on the Daytona server so for now left it at that and didn't poke at the other BSDs with FreeBSD not booting.
                    Michael Larabel
                    http://www.michaellarabel.com/

                    Comment

                    Working...
                    X