Announcement

Collapse
No announcement yet.

AMD EPYC 7302 / 7402 / 7502 / 7742 Linux Performance Benchmarks

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

  • AMD EPYC 7302 / 7402 / 7502 / 7742 Linux Performance Benchmarks

    Phoronix: AMD EPYC 7302 / 7402 / 7502 / 7742 Linux Performance Benchmarks

    Last month we provided launch-day benchmarks of the AMD EPYC 7502 and 7742 under Linux in both 1P and 2P configurations for these exciting "Rome" Zen 2 server processors. For your viewing pleasure today is a fresh look at not only the EPYC 7502 and 7742 processors under the latest Linux 5.3 kernel but we've also expanded it to looking at the EPYC 7302 and EPYC 7402 processors as well with those processors recently being sent over by AMD. Under Ubuntu 19.04 with Linux 5.3, these four different AMD EPYC 7002 series SKUs were benchmarked along with some of the older AMD Naples processors and Intel Xeon Gold/Platinum processors for a fresh look at the Linux server performance.

    http://www.phoronix.com/vr.php?view=28257

  • #2
    such a fine processor I could use my self compiling t2 all day, ... ;-) https://www.youtube.com/watch?v=TBC9C4L1J-k

    Comment


    • #3
      I guess by now it's pc to say the Intel solutions are screaming "Help! I'm getting raped!" come the perf*dollar charts...

      Comment


      • #4
        Will you also get any of the not-yet-public SKUs with low core count and all 256MB of L3 enabled? They should be very sweet for all memory bound application out there ...

        Comment


        • #5
          Originally posted by pegasus View Post
          Will you also get any of the not-yet-public SKUs with low core count and all 256MB of L3 enabled? They should be very sweet for all memory bound application out there ...
          We'll see... I invite benchmarking on any of the processors, I am sure they will likely send whatever new SKUs out that may come. Worst case, someone else likely running PTS and uploading them to OB.
          Michael Larabel
          http://www.michaellarabel.com/

          Comment


          • #6
            Is the cost comparisons not a bit skewed by not including some base system costs?
            e.g. a 1P base server costs eg. $2000 and a 2P base server costs e.g. $3000. What does a Epyc server cost without the cpu? What does a Xeon server cost without the cpu? Or is the calculation cost too much work?

            Comment


            • #7
              Nice benchmarks!
              I just got a server with 2x Epyc 7402, and have also been doing some benchmarks involving MKL (using the quantum chemistry code ADF).
              As on all non-intel chips, MKL will not automagically select the correct code path for fastest exectuion. The same holds true for multi-code-path (fat) binaries generated by the intel compilers (ifort/icc). However, for MKL there is a trick to work around this problem by setting an undocumented environment variable: MKL_DEBUG_CPU_TYPE=N, where N is an integer that corresponds to a vectorization set. N=5 should force MKL to use AVX2 instructions, instead of the default SSE2 on non-intel chips.

              It would be really interesting to see the MKL benchmarks after doing "export MKL_DEBUG_CPU_TYPE=5"

              Comment


              • #8
                Originally posted by grigi View Post
                Is the cost comparisons not a bit skewed by not including some base system costs?
                e.g. a 1P base server costs eg. $2000 and a 2P base server costs e.g. $3000. What does a Epyc server cost without the cpu? What does a Xeon server cost without the cpu? Or is the calculation cost too much work?
                Unfortunately too many factors to work out, because of the dozens of different server motherboards / barebones / servers... you can find EPYC and Xeon motherboards at similar price points, so wouldn't really impact the outcome of the graphs anyhow. At the low-end, I can find a 1P EPYC board for ~$500 or 2P for ~$600 so if going by that it doesn't factor out much for multi-thousand dollar CPUs.
                Michael Larabel
                http://www.michaellarabel.com/

                Comment


                • #9
                  Originally posted by grigi View Post
                  What does a Epyc server cost without the cpu? What does a Xeon server cost without the cpu?
                  As long as you stay with modest amounts of memory, cost difference is negligible and cpu is the main cost factor. Once you get into multi-terabytes memory territory, then memory becomes your largest price chunk and cpu doesn't play that much of a role. Still, it's cheaper with epyc since it can drive 8ch memory with up to 2TB per socket.

                  Comment


                  • #10
                    I see. That makes sense.

                    Comment

                    Working...
                    X