No announcement yet.

What Linux Distribution Should Be Benchmarked The Most?

  • Filter
  • Time
  • Show
Clear All
new posts

  • #31
    Debian stable could be old. Stable is recommended for servers, but for desktop the recommended is Testing. I think on Debian testing or even unstable.


    • #32
      For the love of everything that is good in this world, Xubuntu!

      I think Xubuntu is the "gamer" flavour of Ubuntu, so I'd say it's wise to compare that against other platforms.


      • #33
        Originally posted by peppercats View Post
        Kubuntu is probably the single most buggiest KDE distro and is regularly frowned upon when giving advice on which KDE distro to use.
        Nearly every KDE-related bug I've helped people diagnose was related to Kubuntu and not KDE.
        That's certainly not true. I was using Arch Linux in the past and there were bugs which weren't present in Kubuntu. From my experience Kubuntu is one of the most polished Linux distributions. OpenSUSE is maybe better, but Fedora is the worst.


        • #34
          openSUSE. It has the Open Build Service, and as mentioned SUSE Studio allows for easy replication of an image (although it doesn't have a "copy the official LiveCD packages" option that I'm aware of, only the base KDE/GNOME packages are selected by default, which makes setting up a LiveCD a bit longer task than it would strictly be necessary). Also, like you mentioned in the article:
          Aside from the debug build situation during the development cycle, another benefit of Ubuntu is their "daily mainline kernel PPA" where Debian packages are available daily (and tagged) releases of the mainline Linux kernel. When I maintain dozens of systems in my office/labs and am routinely reformatting them, it's very convenient to simply download the very latest Linux kernel code (important for Phoronix tests) without having to build the kernel from source. Beyond saving a few minutes each time, the mainline kernel PPA is very easy to cite Phoronix readers to when asking about how to install the very latest Linux kernel, what the kernel configuration was like as found in a Phoronix article, etc (enhancing the reproducibility of Phoronix tests). I'm not aware of any other Linux distribution offering easy, daily-updated kernel binary packages.
          openSUSE has that, it's either Kernel:Stable if you want the upstream stable versions, or Kernel:HEAD for the latest kernel git code. You can use them via SUSEStudio, too, to make a snapshot of a certain kernel version for others to reproduce.

          The second option for me would be Fedora, due to its corporate backing and bleeding edge status, although I see how the debug builds could be problematic.

          Originally posted by hadrons123 View Post
          Arch linux or Gentoo ~unstable would be more appropriate since they have the latest packages.
          No. Arch and Gentoo don't have defaults, and Phoronix is benchmarking defaults. Also, the way PTS works is not exactly compatible with Gentoo, because unlike Portage-managed packages, the tests are by default not compiled with -march=native and thus would show poorer results than if you installed the test packages through Portage.

          Originally posted by iniudan View Post
          Fedora, OpenSUSE, Debian, Ubuntu, CentOS are the primary one I would like to see. But would also like one of Mandriva fork (Mageia or PCLinuxOS most likely) and Crunchbang, this last one simply because it an out of the box Openbox Debian, so could could same time in setting up has to have a lightweight distro, mostly for comparaison of DE impact on the benchmark result, has you will only have to deal with a windows manager, at least for the test where it could be relevant.
          Debian and CentOS are way too stable to be interesting. Same goes for Ubuntu LTS. They are good as a reference platform, but benchmarking them doesn't show any interesting data by itself. Mageia is also a bit on the stable side, plus the latest benchmarks show that there was some performance issue that they had which would also influence the test results in a bad way.

          Originally posted by Kirurgs View Post
          I don't think list should be larger, but one thing that could actually matter is using other type of kernels, like the one with BFQ disk scheduler, TuxOnIce, maybe BFS. So it's -pf kernel or -lqx.
          In the end the there will be quite more work if You do include those, but that's what I'm interested in, just ubuntu and only the latest one, well not for me.
          Speaking of which, now that we have latency tests, some benchmarks of the realtime and low-jitter kernels should be made. It would be pretty interesting to see.


          • #35
            Originally posted by MrTheSoulz View Post
            You guys and your distro wars did it...
            i guess this where we say good bye Phoronix
            since i do not care about any other distro phoronix will not be news site for me anymore i guess wich makes me VERY sad becuase i used to check in at least 5 times a day here to see the news :C

            For the trolls: Be carefull what u wish for, if ubuntu would to fall it wont fall alone...
            people dont like change and they would probably be like me and just give up on linux and go back to windows.
            Go head and hate me for saying the truth that you ignore because u are in love with other distro.

            Thank you for the years of fun Phoronix, will miss you.
            MrTheSoulz, we are carefully wishing that Canonical would just take their Androidified Debian Phone Distro and move far away from what has been GNU/Linux for decades. They don't add one iota to the thing we most care about. The only thing Canonical is doing is taking what they need and occasionally they drop what they wrote for themselves (and which no other Linux project is interested in) on a public website. It may be under (sometimes unfavourable) FOSS licenses, but it is squarely Ubuntu only.

            If Canonical/Ubuntu disappears or cuts the ties officially with the GNU/Linux communities, we won't lose anything. Canonical doesn't write infrastructure (the few bits they do, no one else wants to use). So codewise they can be missed. We'll lose "the marketing that Canonical adds to GNU/Linux", but do they add it to GNU/Linux or are they just pushing their own increasingly non-GNU/Linux Mobi-buntu product?

            When it comes to the alledgedly 20 million Ubuntu users in Canonicals userbase, they are an unpricipled bunch who follow Canonical around and they accept whatever that company deems fit to give them. On top of that, these users, with their millions in numbers, are not enough to keep Canonical profitable. Tell me, what do those consumers add to GNU/Linux that we won't have without them? It isn't money flowing in the ecosystem, it isn't tecnical expertise, it isn't broader GNU/Linux advocacy, so what do you *buntu users add? Clout? In the decade that Canonical exists, these numbers have not brought any major software titles over from the Windows and Mac world. Don't take credit for Steam. we owe thanks for that to Microsoft, with their plan to turn Windows into a walled, application store controlled, phablet OS. Valve needed an escape.

            So, why should the GNU/Linux communities care if *buntu users use a *buntu or Windows or Mac? You certainly aren't using GNU/Linux and you do nothing to advance it.


            • #36
              Debian or Mint.
              Since Debian is the mother of the two most popular distros (Ubuntu, Mint), and Mint is the most userfriendly and popular.


              • #37
                I am personally an Arch user but I really do not care which distribution that gets used for most testings. As Michael pointed out, Ubuntu got a big bonus by having the versioned PPAs for kernel testing, so any distro (Xubuntu, Bodhi, Mint...) being able to use those should out of convenience be used. With some regular testing against other distributions to make sure that the distribution of choice does not deviate from the rest, one should be able to conclude that the results are representative.

                Fanboyism does not help anyone. Better to go for convenience.


                • #38
                  +1, that's what i wanted to say


                  • #39
                    I'm a Gentoo user at home and a Fedora user at work. I agree that Gentoo would be totally wrong for this. Far too much variation and it's not exactly a system you can just set up and tear down in a hurry. If the debug build issue can be resolved then Fedora seems ideal. It doesn't stray too far from the norm but does keep close to the edge. It's also commonly used.


                    • #40

                      Kubuntu provides ubuntu plus wayland plus a full and complete desktop. And it's not so buggy as before.