Announcement

Collapse
No announcement yet.

Benchmarks Of The Gentoo-based Sabayon

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

  • Benchmarks Of The Gentoo-based Sabayon

    Phoronix: Benchmarks Of The Gentoo-based Sabayon

    For those looking to experiment with a Gentoo-based Linux system but are not looking forward to the obstacles of installing Gentoo itself, an easier and quicker approach can be to use a distribution like Sabayon Linux. Sabayon uses pre-compiled x86 and x86_64 packages for installing the Linux distribution from its LiveDVD and uses their own Entropy system for package management, though these binary packages are compiled from Gentoo's Portage and using the Portage system is still available. The LiveDVD installer is also very easy to use and is just like using Ubuntu's Ubiquity or Red Hat's Anaconda. With all the benchmarking though of Ubuntu and Fedora as of late on Phoronix, we found it time to put out some benchmarks of Sabayon Linux. Up today are benchmarks from the recently released Sabayon 5.1 along with the older Sabayon 4.2 and for comparison is Kubuntu 9.10.

    http://www.phoronix.com/vr.php?view=14481

  • #2
    Nice comparision! Could you do some more testing with opensource graphic drivers on both Sabayon and Kubuntu systems? I'm most interrested in Radeon R300 driver benchmark.

    Edit: There's no information about which version of NVidia binary driver the tested distribution were using in benchmark but it would be very usefull, too.
    Last edited by xeros; 01-04-2010, 02:57 AM.

    Comment


    • #3
      Thank You Michael for looking at something not Ubuntu
      I'm a little bit disappointed that the article is kind of superficial.
      There is no real background, difference or any other useful info like installation "flavors". BTW, Sabayon uses Red Hat's Anaconda installer and not "like".
      I have to admit that article is about what was stated in title.
      IMHO, "real" articles are what we (readers) want.
      By "real" i mean not only 10 benchmark graphs.
      But also more useful info. After reading this article you probably will not try it.
      It's like "We know well Mazda (Ubuntu) and now we will learn about Toyota (Sabayon) by comparing max speaker volume and break lights intensity."
      If you need help, just ask. At least few readers will be happy to help.
      Why i'm not doing it by myself ? Because I can only help with this.

      Comment


      • #4
        Well you could test the latest official Kanotix Excalibur preview too and fix the pts gui issues the same time. No longer needed windows should be closed not stay on top.

        Comment


        • #5
          Meh! Sabayon is not as fast as I thought since it's Gentoo based. But that's normal since it uses prebuilt generic x86_64 binaries.

          Comment


          • #6
            Hmmm, Im surprised we got beaten by Kubuntu in a couple of things. I agree a more through analysis is needed however.

            Comment


            • #7
              Originally posted by Apopas View Post
              Meh! Sabayon is not as fast as I thought since it's Gentoo based. But that's normal since it uses prebuilt generic x86_64 binaries.
              Yup, they're prebuilt, some obviously with slightly different optimization
              parameters than others, considering some of the larger differences. It's
              great to see a non-Ubuntu distro for a change. Gentoo would be too much,
              ofcourse, since any gains are quickly offset by sheer compilation time.

              As a system for benchmarking x86 vs x64, though, Gentoo makes sense, cause
              you can have optimization parameters such as SSE1/2/3, -O2/3 and similar
              under control as opposed to Ubuntu/Debian where apparently some of these
              are being set only at the x64 level.

              Once more, credits to Michael for 'rocking the boat' a little bit
              Keep up the good work!

              Comment


              • #8
                Also i vote to set a filesystem standard for benchmarks. Use ext4 as filesystem even when it is not the default one. Then differences could also happen due to different default from one kernel to the next, but usually should be similar with the same kernel version (unless those defaults are patched or changed in the /etc/fstab).

                Comment


                • #9
                  Originally posted by clavko View Post
                  Yup, they're prebuilt, some obviously with slightly different optimization
                  parameters than others, considering some of the larger differences. It's
                  great to see a non-Ubuntu distro for a change. Gentoo would be too much,
                  ofcourse, since any gains are quickly offset by sheer compilation time.
                  As a Gentoo user i can tell you that "Sheer Compilation Time" is a myth
                  For example, it takes 56 seconds to install Firefox from source.
                  To recompile the whole thing with over 1K packages is less then 7 hours (I don't measure it usually and the "emerge -j 5 -vuNDe world" executed once something severely broken or something like GCC updated)
                  Again, the average package installation takes same time is precompiled binary.
                  And i'm talking about Q9300 with 4GB RAM... It's not fastest computer in any way. But you do need a lot of RAM. The trick is to use RAM disk (autofs) for portage work folder. But, RAM is cheap

                  Comment


                  • #10
                    I don't know how much does it take to compile Firefox on a quad core,
                    but i know how much it takes to do it on a single 2GHz Mobile Sempron
                    core, and let me tell you... it's bordering with 'not worth it".
                    (the usual dependencies of mozilla-firefox are about an hour or so).

                    Just yesterday it took me about 2-3 hours to compile Qt 4.6.1 and about
                    an hour and a half for recompiling KDE 4.3.4 afterwards. So in a way,
                    yes, the 'sheer compilation time' is a rather substantial variable here
                    at "clavko's". However, I'm an enthusiast and prefer to do it my own way.
                    I don't expect anyone with a less than dual core try and copy me.

                    As a one Gentoo sufferer to another, best regards

                    Comment


                    • #11
                      Originally posted by n0nsense View Post
                      As a Gentoo user i can tell you that "Sheer Compilation Time" is a myth
                      [...]
                      To recompile the whole thing with over 1K packages is less then 7 hours [...]
                      So, how many centuries would it take to take back those 7 CPU-hours lost in recompilation?

                      Think about your poor CPU... It could have spent those 7 hours playing its favorite music, watching movies, sleeping or just hanging out with its best pals, RAM and Southbridge (unfortunately, we lost Northbridge in an accident a few years back.)

                      Sad, sad CPU.

                      Edit: on topic, this articles shows there's no real performance advantage in favor of Sabayon or Kubuntu (4 wins on each side and several ties). I'd like to see a Gentoo benchmark, but I'm not sure it's worth the setup time...
                      Last edited by BlackStar; 01-04-2010, 09:57 AM.

                      Comment


                      • #12
                        I have a 5 years old 3000+ Athlon with 1.5 GB RAM and I compile Gentoo since 2004
                        The trick is to use as newer gcc as possible, don't compile with O3 coz it needs more time and for things like QT, compile them while you are sleeping
                        Btw with O2, Firefox with xulrunner needs about 40 minutes in my system.

                        Comment


                        • #13
                          Originally posted by BlackStar View Post
                          I'd like to see a Gentoo benchmark, but I'm not sure it's worth the setup time...
                          Well, there's something of a sort right... here.

                          Naturally, I'd like to see Gentoo trash Ubuntu some more, but I'll have to settle for this.

                          Comment


                          • #14
                            Don't run gentoo on anything less than a quad core. You'll regret it. I run arch on anything less.

                            Also with gentoo stay away from heavy weight vdesktop environs like gnome and especially kde. Anything using qt takes an extra compilation hit.

                            I will tell you a pair of 5520's (16 virtual cores) rips through a gentoo install in amazing time. With the added bonus of having system stability tested during install.

                            I would definitely say that using a "pre built" gentoo is really really dumb. That just makes it yet another average distro with average performance. I'd pick arch in a heartbeat over gentoo in this case...arch IMHO is a better system in general.

                            Comment


                            • #15
                              i dont see why compiling would be worth it in the end for the average user. For something like firefox the total time-use ratio would be stupid. firefox works just fine here in ubuntu and i cant complain about speed. I suppose you could compile everything while you sleep; But to compile everything you need you would need to sleep alot. Your doctors must be prescribing you guys some good shit.

                              I can see compiling games or other apps where every frame per second counts, as this bench demonstrates.

                              Comment

                              Working...
                              X