Announcement

Collapse
No announcement yet.

AMD Ryzen 9 7950X3D Linux Performance

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

  • #11
    One series of tests that might be interesting, what if you take the 7950x and use /sys/devices/system/cpu/cpu(x)/online to disable half the cores,
    then on the 7950x3d, use it to disable the cores on the CCD with the smaller cache, that should in theory see see how much performance the cache offers for things, without the scheduler getting in the way

    For even more science, it would be cool to see the 7950x3d with the cores on the CCD with the larger cache disabled

    Although I don't know if it's possible both ways, as I know I can't turn off CPU0



    But the few seconds off of code compiliation but using a lot less power already looks promising

    Comment


    • #12
      Originally posted by nerdopolis View Post
      One series of tests that might be interesting, what if you take the 7950x and use /sys/devices/system/cpu/cpu(x)/online to disable half the cores,
      then on the 7950x3d, use it to disable the cores on the CCD with the smaller cache, that should in theory see see how much performance the cache offers for things, without the scheduler getting in the way

      For even more science, it would be cool to see the 7950x3d with the cores on the CCD with the larger cache disabled

      Although I don't know if it's possible both ways, as I know I can't turn off CPU0



      But the few seconds off of code compiliation but using a lot less power already looks promising
      You could stuff the benchmark into a container with only the desired cores being available in that container.

      ETA: This benchmark run is using Windows and tests both the BIOS priorities for cache or frequency. Quite interesting. For my uses frequency is generally better:
      The Ryzen 9 7950X3D is the spearhead of the AMD Zen 4 X3D lineup. In our performance review we test AMD's new 16-core flagship with dozens of applications and 14 games at up to 4K, to see whether AMD can take the performance crown from Intel, how power consumption is affected, and what can be expected from overclocking.
      Last edited by unwind-protect; 27 February 2023, 11:55 AM.

      Comment


      • #13
        Originally posted by Volta View Post
        Awesome Ryzen performance. Not even a single reason to consider Intel as an alternative.
        Uh, there's plenty of reasons. There's more to a platform overall. I'm not dissing the Ryzen, just pointing out the errors in the statement.

        Comment


        • #14
          No Total War test results anymore ? Warhammer 3 is at least still pretty relevant/modern

          Comment


          • #15
            Originally posted by joshx1 View Post
            I only care about code compiling etc, and workstation use, gaming is ¯\_(ツ)_/¯ for me .

            Slightly slower in some benchmarks but MUCH lower thermals... interesting. Probably some half-decent overclocking to come for even better performance?​
            I assume it's because 7950X wasn't used in 105W eco mode. That would have been an interesting comparison with 7950X3D.

            Many pointed out, that gains in performance in 170W mode for 7950X vs 105W mode are pretty small, while power usage and thermals cost is disproportionately higher.

            I doubt 7950X3D has overclocking room.
            Last edited by shmerl; 27 February 2023, 12:04 PM.

            Comment


            • #16
              Solid CPU, good power consumption. Of course like others mentioned the challenges are around application scheduling. This is super hard any workload is different. Since this CPU is so specific, this is actually an argument for that eBPF based scheduler, which was recently proposed. The heuristics are so custom and maybe even user specific, this would be a perfect use case. In a way that's comparable go what AMD is doing on Windows by having external software do some monitoring and guiding the kernel.

              Comment


              • #17
                According to Techpowerup, to properly enable the new functions on Windows requires installing 8 special AMD drivers from Xbox Game Bar.
                Originally posted by nerdopolis View Post
                One series of tests that might be interesting, what if you take the 7950x and use /sys/devices/system/cpu/cpu(x)/online to disable half the cores,
                then on the 7950x3d, use it to disable the cores on the CCD with the smaller cache, that should in theory see see how much performance the cache offers for things, without the scheduler getting in the way

                For even more science, it would be cool to see the 7950x3d with the cores on the CCD with the larger cache disabled

                Although I don't know if it's possible both ways, as I know I can't turn off CPU0



                But the few seconds off of code compiliation but using a lot less power already looks promising
                Techpowerup tried this too, and published their findings here

                Comment


                • #18
                  This is more cache than my first computer had disk space! XD

                  Comment


                  • #19
                    This image is broken for me: https://www.phoronix.net/image.php?i...en7950x3d7_med

                    Interesting to see it did well with gaming performance. I was worried about how the scheduler would interact.

                    Comment


                    • #20
                      Originally posted by cjcox View Post
                      Uh, there's plenty of reasons. There's more to a platform overall. I'm not dissing the Ryzen, just pointing out the errors in the statement.
                      Such as?

                      Comment

                      Working...
                      X