Announcement

Collapse
No announcement yet.

Clear Linux vs. Ubuntu 19.10 Video Encoder Performance On The Core i9 9900K

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

  • Clear Linux vs. Ubuntu 19.10 Video Encoder Performance On The Core i9 9900K

    Phoronix: Clear Linux vs. Ubuntu 19.10 Video Encoder Performance On The Core i9 9900K

    Often when doing cross-distribution benchmarks, readers often comment on the performance of Clear Linux particularly for video encoding use-cases as surprisingly different from other distributions. Some argue that it's just over the default CPU frequency scaling governor or compiler flag defaults, so here is a look at that with Ubuntu 19.10 daily benchmarked against Clear Linux...

    http://www.phoronix.com/scan.php?pag...buntu-19.10-VE

  • #2
    Unless you called "x86_energy_perf_policy -c 0-7 performance" on Ubuntu before running the tests, I think the results are meaningless.
    Default kernels set this intel-only knob to "balanced", while the clear linux kernel sets it to "performance". And it does make a noticeable difference.

    Comment


    • #3
      It might be interesting to use lxd to install Ubuntu userspace inside Clear, and Clear userspace into Ubuntu, and see how they bench under each other. That will make it pretty obvious whether the kernel and sysconfig are contributing much to performance differences.

      Comment


      • #4
        I can't be the only one tired of hearing about Clear Linux. Yeah, I get it, you use it as your main system. It's not that great. You could compile a kernel with their tweaks anywhere and have a much more usable operating system with repos actually populated with packages and not flatpaks. It's cool if you wanna use it, but does it really need to be a part of like every other news article you write here? It definitely gets more coverage than any other distro lately...

        Comment


        • #5
          Originally posted by linuxgeex View Post
          It might be interesting to use lxd to install Ubuntu userspace inside Clear, and Clear userspace into Ubuntu, and see how they bench under each other. That will make it pretty obvious whether the kernel and sysconfig are contributing much to performance differences.
          Yes, please.

          LXC, KVM and/ or whatever.

          Comment


          • #6
            Originally posted by Syretia View Post
            I can't be the only one tired of hearing about Clear Linux. Yeah, I get it, you use it as your main system. It's not that great. You could compile a kernel with their tweaks anywhere and have a much more usable operating system with repos actually populated with packages and not flatpaks. It's cool if you wanna use it, but does it really need to be a part of like every other news article you write here? It definitely gets more coverage than any other distro lately...
            Since one of the main goals of Phoronix and the associated test suite is to evaluate performance, it would make sense that Clear Linux get more column inches because it is consistently outperforming so many of the mainstream Linux distros.

            Also by making repeated notes about it, it places pressure on the other distros to pull up on their performance skills and apply them to their spins.

            The fact Phoronix took many of Clear Linux known enhancements and tried to apply them to Ubuntu was to show if it is just settings or is it something else?

            But by testing, it also brings out overseen or overlooked options (like by mlau ) that others can look and compare.

            I have learned a lot about certain linux kernel behaviors just by reading Clear Linux test reports. I find them educational.

            Comment


            • #7
              Originally posted by Syretia View Post
              I can't be the only one tired of hearing about Clear Linux. Yeah, I get it, you use it as your main system. It's not that great. You could compile a kernel with their tweaks anywhere and have a much more usable operating system with repos actually populated with packages and not flatpaks. It's cool if you wanna use it, but does it really need to be a part of like every other news article you write here? It definitely gets more coverage than any other distro lately...
              Don't listen to him, Michael. People forget they are not forced to read every article, I just skip the ones I'm not interested in. Because of Phoronix, I've been following Clear Linux very closely, and if Ubuntu goes ahead and discards APT, Clear Linux is at the top of my list.

              If Clear Linux has considerable performance advantages over other distros, I don't see any reason not to keep talking about it.

              Comment


              • #8
                Originally posted by mlau View Post
                Unless you called "x86_energy_perf_policy -c 0-7 performance" on Ubuntu before running the tests, I think the results are meaningless.
                Default kernels set this intel-only knob to "balanced", while the clear linux kernel sets it to "performance". And it does make a noticeable difference.
                My man!

                That's one of the things I keep writing about here, yet the majority seems to enjoy staying clueless!

                Anyway, one can also set this value to 0 (zero) for maximum performance-bias.

                Could you please provide the output of:
                Code:
                sudo cpupower -c all info
                This way we can see the actual numerical value "x86_energy_perf_policy" is setting.

                Thanks again for bringing this up!

                Comment


                • #9
                  I'm about to add a filter to the RSS feed for our Discord guild to get rid of the Clear Linux related articles as they present no new info. We know that Clear Linux's kernel is faster. We know that other distros aren't bringing in Clear Linux's changes. This is just getting old, honestly. On top of that, with all of the coverage it gets, we have people thinking they should switch to it when it has next to no actual packages in its repositories and leans heavily on flatpaks.

                  Comment


                  • #10
                    Originally posted by timofonic View Post

                    Yes, please.

                    LXC, KVM and/ or whatever.
                    LXC/LXD are fancy chroots and as such they will do the job I requested. KVM would defeat the purpose entirely.

                    Comment

                    Working...
                    X