Announcement

Collapse
No announcement yet.

7-Way Linux Distribution Benchmarks To Kick Off September

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

  • 7-Way Linux Distribution Benchmarks To Kick Off September

    Phoronix: 7-Way Linux Distribution Benchmarks To Kick Off September

    In testing out a new Broadwell-EP system as well as for final validation of the new Phoronix Test Suite 6.6, I carried out a fresh Linux OS distribution comparison last week. Here are those results from Ubuntu, Clear Linux, Scientific Linux, openSUSE Tumbleweed, Fedora, Antergos, and Sabayon Linux.

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    OpenMP isn't some panacea for parallel programming. In fact, the folks at Blender have realized this and are rewriting their parallel code and leveraging CUDA/OpenCL more expansively to improve their product.

    I'm betting in three years no one will give much credence to it at all.

    Comment


    • #3
      Originally posted by Marc Driftmeyer View Post
      OpenMP isn't some panacea for parallel programming. In fact, the folks at Blender have realized this and are rewriting their parallel code and leveraging CUDA/OpenCL more expansively to improve their product.

      I'm betting in three years no one will give much credence to it at all.
      Lol... from the same idiot who pretends he used to work for Steve Jobs at Next, which was a total failure of a company.

      Comment


      • #4
        Interesting results. but why was Ubuntu run with the 4.8x kernel? I thought 16.04 is still on 4.4x out-of-the-box.

        Comment


        • #5
          Originally posted by chuckula View Post

          Lol... from the same idiot who pretends he used to work for Steve Jobs at Next, which was a total failure of a company.
          Per his LinkedIn profile......
          • ObjectLine Engineer

            Apple Computer Inc

            January 1997 – May 1998 (1 year 5 months)

            Worked for Apple Enterprise Software delivering Professional Service support for the NeXT Enterprise Suite of Tools and Applications.


          • Technical Support

            Apple

            December 1996 – May 1998 (1 year 6 months)

            Liason between Device Driver Engineering and Enterprise Professional Services. Job was to coordinate client device driver needs to support Openstep deployments.
          • SQA Engineer

            NeXT Software Inc

            May 1996 – December 1997 (1 year 8 months)

            Worked within the Software Quality Assurance team testing Openstep Operating System and Development Tools. Also, I coordinated third party companies testing progress against their own applications.

          Comment


          • #6
            In one test, "PostgreSQL pgbench v9.4.3 Normal Load" ClearLinux is much slower than the rest. Why is this? Should it be "Less is better"?

            Comment


            • #7
              Originally posted by Marc Driftmeyer View Post
              OpenMP isn't some panacea for parallel programming. In fact, the folks at Blender have realized this and are rewriting their parallel code and leveraging CUDA/OpenCL more expansively to improve their product.

              I'm betting in three years no one will give much credence to it at all.
              I don't believe this at all, at least not on the CPU. On the CPU using intrinsics for SIMD, manual unrolling for ILP, and OpenMP for threading I am confident my results would never be worse than OpenCL and in most cases my results would be better. I wrote a real-time ray tracer in OpenCL so I have some experience with it. I don't have any experience of OpenMP on the GPU so that's another issue.
              Stack Overflow | The World’s Largest Online Community for Developers

              Comment


              • #8
                This benchmarks mean nothing if the distros has not the same kernel version installed or at least we do not know which one are in use. Whatever it is. We canot know what can be kernel version or kernel configuration performance. I miss RT kernel benchmarks, and also I miss Debian Mint Manjaro Solus and Kaos far more popular than Scientific or Clear Linux.

                In my opinion Debian Ubuntu, Fedora Suse. Antergos (for archs) and Sabayon (for gentoos) must be there ALWAYS and they are 6 3 debs, 2 RPMS 1 arch 1 gentoo, and you can add slack. And some other used as Mint, Manjaro, Kaos and Solus and if there are specially good and unused Scientific and Clear or similar.

                Comment


                • #9
                  Originally posted by chuckula View Post

                  Lol... from the same idiot who pretends he used to work for Steve Jobs at Next, which was a total failure of a company.
                  Well Next lived 10 years and Apple purchased it in 1997 for $429 million... that is not such a failure after all

                  Comment


                  • #10
                    Ya know Sabayon is a pretty good "It just works with no BS" Linux distro. It's nice to see it show up here. I'd say it and Tumbleweed are the best rolling release distros out there atm.

                    Comment

                    Working...
                    X