No announcement yet.

What Shall We Benchmark Next? Let Us Know!

  • Filter
  • Time
  • Show
Clear All
new posts

  • #51
    Slightly off-topic, but I assume each of these benchmark systems would spend a significant amount of time idle?

    For suggestions such as benchmarking multiple distros/OSs, have you thought about having one machine that dual-boots (or triple-boots, etc) into multiple operating systems? I imagine it wouldn't be too difficult to arrange for this to happen automatically with GRUB. Each OS would update, run benchmarks etc, then reboot into the next.. Seems it could avoid the need for a lot of dedicated machines (and associated power requirements.)

    Just a thought, I don't know what your current set up is or how feasible this is


    • #52
      Originally posted by sifnt View Post
      While I don't know the specific benchmarks to use, anything that measures latency / responsiveness would be awesome.
      I agree that it would be very useful to add such tests to the suite before going on another graphics stack benchmark tracking bonanza.


      • #53
        Multithreading / Latency

        Michael, this is a great proposal of yours. I have been very impressed with the work you have already done in Benchmarking for Linux.

        I would like to see how the major Desktops and or Distros are performing in Responsiveness / Latency and how well they are using multicore cpus.

        It would be a great continuous benchmark, because we could see whether desktops are making best use of multicore cpus to produce the most responsive system.

        It could be done at the distro level (comparing responsiveness gains/losses as a new distro release evolves) or it could be done at the desktop level (considering whether KDE or GNOME on say the same Ubuntu are improving responsiveness.



        • #54
          Earlier in the thread, I asked for R200 2006 FGLRX vs modern OSS driver. As I did not noticed it was for phoromatic, then I will add a +1 for comparaison between closed and open driver in a vast array of cards. The GPU switcher could be used there with a MB with more than 1 PCI-E 16x slot.

          Nvidia 6600GT
          Nvidia 8800GT
          Nvidia 7300GT

          ATI 9200
          ATI x800
          ATI x1800
          ATI HD4***

          Should be tested on a weekly basis. Older ATI card will need a comparison with the last supported driver, so 6.08 and 9.03 respectively.


          • #55
            One twist I'd like to place on this is I personally would like to see the originating project having an interest in this.

            So if you run an OSS project that has an interesting in functional or performance tracking from build to build - post here and PM Michael and Me.



            • #56
              I can't suggest about what to review next, but there is one aspect that could help linux users better. PTS is already strong (or rather the de facto tool) with the benchmark suites, how about having a sub-section for overclocking guides/examples/reviews/tips/information. A sort of linux overclocking club, perhaps ? because if you see around the net, you won't find many official linux overclockers club or something along that line. Even if this means just adding a section in the forums, that should also be useful, shouldn't it ?


              • #57
                Originally posted by loonyphoenix View Post
                Wine is an interesting idea, I agree. +1

                I also like the idea of monitoring the performance of some non-Ubuntu distribution, especially one which is specifically designed to be a rolling release (Arch or Gentoo or even Debian Sid), with latest STABLE software, not the way Ubuntu fluctuates between buggy new untested software and old and therefore slow software.

                Arch would be awesome Except there is the slight problem that sometimes it needs user input and configuration to upgrade or continue running after an upgrade. I don't know anything about Gentoo, but I think this could be a good option. Sid would be too much like Ubuntu.

                These are my thoughts
                I am told that Gentoo has fewer problems than Arch in that regard. Seeing a Gentoo Linux performance tracker would be awesome, but I suggest that there be two trackers. One for the stable tree and one for the testing tree. I explained the difference between the two (and why there should be two) in my previous post.


                • #58

                  Would like to see a benchmark between the different *buntus.

                  Would also like to see a new benchmark of "Power & Memory Usage Of GNOME, KDE, LXDE & Xfce" using vanilla installation. Arch Linux would be a good distro to use.


                  • #59
                    Webkit, gecko and presto?


                    • #60
                      more hardcore benchmarks

                      It would be very interesting if you guys along with file system benchmarks, benchmarked core features of operating systems, such as scheduler latencies and responsivness, tcp/udp network stack latencies and throughput and such.