Announcement

Collapse
No announcement yet.

Intel's Clear Linux Still Outperforming Other Distributions For Mid-2020

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

  • #21
    Beside Clear Linux, am I the only one to notice than Ubuntu 20.10 at mid schedule is coming along much faster than 20.04? Hopefully it stays that way.

    Comment


    • #22
      Originally posted by CTTY View Post
      What is Debian Buster Testing? Debian Buster is stable since a year.
      Good question. Does he mean buster stable, or the new testing bullseye?

      Comment


      • #23
        So many comments ignore that this is a comparison of many Linux operating systems. Of special interest are the "also-rans". Two outsiders are one RPM-based (openSUSE) and one Arch-based (Endeavour OS Rolling).
        There are several unreported variables in these tests. After installation of Linux operating systems, the internal components are "updated" to the latest versions, at the time of installation. Which actual updates are included can differ greatly from the first downloaded operating system. All the Debian-based & Ubuntu-based systems have authorized compilations of the latest Linux kernels. The significance of these updates seems never to be tested, in the comparative bench tests.
        Unusual are the slower performances of the RPM & Arch based systems. Both these minority systems are chosen by their users for reasons other than benchmark performance. The "daring" experimental nature of Ubuntu 2010, compared to LTS 2004 is expected. All the three test versions of the Ubuntu systems, released every six months, should perform better than the LTS released every two years.
        Clear Linux is designed mainly it seems for high performance, with little regard for other factors? If the Debian & Ubuntu based systems were adjusted for high performance, they should be using the official "low-latency" kernels officially released by Ubuntu. Ubuntu knows this to be supposedly true. Why does other not know this?
        Last edited by gregzeng; 07 August 2020, 04:34 AM.

        Comment


        • #24
          No, I didn't know that such a thing existed.
          Thank you very much!

          Comment


          • #25
            Originally posted by gregzeng View Post
            So many comments ignore that this is a comparison of many Linux operating systems. Of special interest are the "also-rans". Two outsiders are one RPM-based (openSUSE) and one Arch-based (Endeavour OS Rolling).
            There are several unreported variables in these tests. After installation of Linux operating systems, the internal components are "updated" to the latest versions, at the time of installation. Which actual updates are included can differ greatly from the first downloaded operating system. All the Debian-based & Ubuntu-based systems have authorized compilations of the latest Linux kernels. The significance of these updates seems never to be tested, in the comparative bench tests.
            Unusual are the slower performances of the RPM & Arch based systems. Both these minority systems are chosen by their users for reasons other than benchmark performance. The "daring" experimental nature of Ubuntu 2010, compared to LTS 2004 is expected. All the three test versions of the Ubuntu systems, released every six months, should perform better than the LTS released every two years.
            Clear Linux is designed mainly it seems for high performance, with little regard for other factors? If the Debian & Ubuntu based systems were adjusted for high performance, they should be using the official "low-latency" kernels officially released by Ubuntu. Ubuntu knows this to be supposedly true. Why does other not know this?
            Because Phoronix is simply testing these releases as a reflection of their performance in their current state at this moment of time.

            Phoronix does also test mods to certain modules but only against themselves.

            To try and tweak every OS to their ultimate performance condition before you test them would be a very time consuming and somewhat futile approach as everyone knows Linux has so many updates that come out daily.

            Therefore these tests should be considered in the context that they are a snapshot of the state of Linux at that particular time.

            What one deems "performant" depends on the personal use case of the user.

            Comment


            • #26
              Originally posted by gregzeng View Post
              So many comments ignore that this is a comparison of many Linux operating systems. Of special interest are the "also-rans". Two outsiders are one RPM-based (openSUSE) and one Arch-based (Endeavour OS Rolling).
              ...
              Unusual are the slower performances of the RPM & Arch based systems.
              Its just the usual "Michael throws a benchmark at a distribution" without even barely evaluating the results.

              The bad final results of openSUSE can be mostly attributed to the bad MariaDB results. As Michael does the DB setup with a homegrown script, he misses a very important point for the DB storage directory - disable CoW!.

              The benchmark setup does not resemble a real world setup. Most users would just install the RPM and be happy with it. They would also benefit from the 'No-CoW' flag already set for the MariaDB storage directory (same for PostgreSQL etc.).

              @Michael: Run
              Code:
              chattr +C $HOME/mysql/.data
              before
              Code:
              mysql_install_db

              Comment

              Working...
              X