Announcement

Collapse
No announcement yet.

A Look At The Windows vs. Linux Scaling Performance Up To 64 Threads With The AMD 2990WX

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

  • #11
    Originally posted by Space Heater View Post

    Michael I'd be interested in hearing this too. It's one of the biggest critiques people have of these benchmarks currently, along with reports that the Windows binaries are not current and not compiled with visual studio.
    Each OS was at its defaults unless otherwise noted.

    The Windows binaries are using the official project binaries for each program. In regards to 'not current', someone pointed out a newer 7-zip was available but not aware of any other of these programs being outdated besides maybe a point/patch release?
    Michael Larabel
    https://www.michaellarabel.com/

    Comment


    • #12
      Ok, so you are running these tests with the balanced performance profile? I'd say that has the possibility of hurting these results, as most are running Windows Server with the high performance profile set.

      Comment


      • #13
        Originally posted by schmidtbag View Post
        Pretty interesting results - I sure wasn't expecting 64 threads to have such a major performance regression over 32 in so many tests.
        Remember that "32 to 64" is also "SMT off to SMT on" so it's not like the earlier steps where additional cores were brought online.
        Test signature

        Comment


        • #14
          Originally posted by F.Ultra View Post

          In exactly what way would GCC provide optimized binaries tuned especially for Linux? And AFAIK the pthreads library in gcc/msys2 is just a wrapper library around the native win32 threading library so it should have no impact on scaling except the small overhead of an extra function call when you call locking primitives since the library is implemented as:
          And most of the tests were calculating things (Blender, IndigoBench, ...), not stressing the sync/threading or filesystem infrastructure.

          Comment


          • #15
            So, this is the reason we never see Windows at netcraft's Top 10 most reliable hosting sites:

            News, updates and resources from Netcraft to detect, disrupt and take down phishing and cybercrime.

            Comment


            • #16
              Originally posted by Space Heater View Post
              Ok, so you are running these tests with the balanced performance profile? I'd say that has the possibility of hurting these results, as most are running Windows Server with the high performance profile set.
              I’d be interesting to see such a comparison, on the other hand if the balanced power plan really causes such performance issue on Windows I’d still say that this is a serious Windows defect.

              Anyway, would anybody chip in for a benchmark of various CPU governors and its effect on the TR2? I definitely would.

              Comment


              • #17
                Originally posted by MadCatX View Post
                I’d be interesting to see such a comparison, on the other hand if the balanced power plan really causes such performance issue on Windows I’d still say that this is a serious Windows defect.

                Anyway, would anybody chip in for a benchmark of various CPU governors and its effect on the TR2? I definitely would.
                It seems pretty reasonable to compare "balanced power" with the default Linux power management settings since it's directly equivalent, Ubuntu doesn't default to "performance" mode either.

                Comment


                • #18
                  These results are interesting especially when taking into account benchmarks etc as reported at eg Anandtech.

                  They couldn't recommend 2990WX because of problems with 64 threads. This shows that that may be a more common issue on Windows than on Linux.

                  Comment


                  • #19
                    Originally posted by Pawlerson View Post
                    So, this is the reason we never see Windows at netcraft's Top 10 most reliable hosting sites:

                    https://news.netcraft.com/
                    What Netcraft measures is not the same as what Phoronix is measuring. The two don't correlate.

                    Comment


                    • #20
                      Originally posted by You- View Post
                      These results are interesting especially when taking into account benchmarks etc as reported at eg Anandtech.

                      They couldn't recommend 2990WX because of problems with 64 threads. This shows that that may be a more common issue on Windows than on Linux.
                      Hm, I think that Anandtech, lost some of its glory the day when Anand "retired" from the site. The reviews are not really that good, look at the assumptions they make about the power usage of the infinity fabric. They conclude that its power hungry based from ways of measuring power that quite a few of the knowledgeably people in the forum say is plain wrong. The way phoronix do the testing is more scientific, easier to replicate and shows a broader variety depending on the wishes of the users.

                      Kind regards
                      Brutalix

                      Comment

                      Working...
                      X