Announcement

Collapse
No announcement yet.

OpenSUSE Leap 15.1 Beta Is Running Well - Benchmarks On AMD EPYC Workstation

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

  • OpenSUSE Leap 15.1 Beta Is Running Well - Benchmarks On AMD EPYC Workstation

    Phoronix: OpenSUSE Leap 15.1 Beta Is Running Well - Benchmarks On AMD EPYC Workstation

    With openSUSE Leap 15.1 reaching beta this week I decided to take it for a quick spin of this Linux distribution derived from the same sources as SUSE Linux Enterprise 15 SP1. Here are some quick benchmarks compared to Leap 15.0 as well as the latest rolling-release openSUSE Tumbleweed.

    http://www.phoronix.com/vr.php?view=27569

  • #2
    I don't know about Leap, but Tumbleweed has a CPU governor running by default.

    You should switch to performance mode prior to testing -
    > sudo cpupower frequency-set -g performance

    To check info on the CPU, and see if you have any other preset governor settings:
    > sudo cpupower frequency-info

    Comment


    • #3
      Using KDE Leap on my desktop AMD, it's an excellent distribution, conservative as it should be a fix release, but it is also thanks to this conservative policy, that Leap has never abandoned me, it simply works, I've never had any problems .

      Comment


      • #4
        Originally posted by Charlie68 View Post
        Using KDE Leap on my desktop AMD, it's an excellent distribution, conservative as it should be a fix release, but it is also thanks to this conservative policy, that Leap has never abandoned me, it simply works, I've never had any problems .
        That conservative policy killed OpenSUSE's popularity. It used to be a cutting (not bleeding) edge distro with an 8-month release cycle that competed squarely with Fedora. Then SUSE's new ownership stepped in, their Richard Brown became chairperson and they proceeded to destroy OpenSUSE. The community had already voted on and approved a mission statement where we favored cutting edge but not over stability, etc. This essentially got torn up. They changed the dev cycle to one year and SUSE pulled their employees' assistance in readying the distribution. Then the "solution" was proffered to use the SLES kernel. Essentially they created a crisis and then offered to fix the problem they created. Suddenly OpenSUSE becomes an old fuddy-duddy distro with an antique kernel. They let SUSE Studio, the best distro customizing app their ever was, rot until it stops working. Brown becomes a holy terror on Reddit, attacking anyone who criticizes OpenSUSE. He tells me that the community mission statement "wasn't binding" and that there were "no more people willing to work on the kernel". When I asked where all the OpenSUSE contributors disappeared to, he tells me "Tumbleweed".

        OpenSUSE occupied position 4 or 5 on Distrowatch for over 10 years. Bryan Lunduke, formerly of Linux Action Show, gave a talk saying how OpenSUSE was 4 or 5 for ten years before and would be the same ten years from now. It was a remarkably consistent distro. Where is it today? #9. OpenSUSE stopped making cool innovations, stopped supporting a lot of its former cool ideas (SUSE Studio, WebYAST for administering a system over web UI, etc.), and its release schedule became too slow. It changed its number 13.x to 42 and now is back to 15, confusing everyone. Its new mission statement talks about making significant changes only ONCE EVERY THREE YEARS, a fact most OpenSUSE users don't even realize. It's wasting its time working on SUSE Enterprise crap like containerized this and that (and starting all over again by rejecting kubernetes, which dominates the niche, because Brown doesn't like it).

        I say now that OpenSUSE used to compete with Fedora; now it competes with Debian Stable. (In one of his latest interviews Brown used the word "Stability" over six times).

        Fortunately the true OpenSUSE fans are still working on Tumbleweed, which to me is the real OpenSUSE. Still, we need to throw Brown out and fork if necessary to go back to the days of being an enthusiast desktop OS rather than some junior version of SUSE Enterprise. We've dropped the ball and other distros have picked it up. OpenSUSE needs to get back to being aggressive and creative again. We don't need conservative stability. We need reliable innovation.

        Comment


        • #5
          Originally posted by alcalde View Post
          Fortunately the true OpenSUSE fans are still working on Tumbleweed, which to me is the real OpenSUSE.
          And Tumbleweed, which is "bleeding edge", is far more stable than openSUSE was in the days of 42.1, 42.2, 42.3, etc.

          Originally posted by alcalde View Post
          FBrown becomes a holy terror on Reddit, attacking anyone who criticizes OpenSUSE. He tells me that the community mission statement "wasn't binding" and that there were "no more people willing to work on the kernel". When I asked where all the OpenSUSE contributors disappeared to, he tells me "Tumbleweed".
          I did not know about that. That is some odd behavior. Funny that he views Tumbleweed as a competing project.

          Originally posted by alcalde View Post
          OpenSUSE occupied position 4 or 5 on Distrowatch for over 10 years.
          Even Distrowatch has acknowledged publicly for many years that their page hit numbers are meaningless. For example, Ubuntu has been the most used distro worldwide almost since its introduction in 2004, and it hasn't been at the top of DW's page hit rankings for the past 8 years. That's a weird popularity contest in which some distros have even cheated their numbers to get ahead.

          Comment


          • #6
            Can someone explain to me what the point is of back-porting patches and drivers to a kernel is instead of simply upgrading to the latest kernel? Near as I can tell, taking an older kernel and back-porting drivers and patches to it from the newer kernel effectively upgrades it to the newer kernel while keeping the older version number. What is the benefit of this? It just seems like it's extra work for no benefit.

            As for this beta, I downloaded the live-usb and gave it a test drive and while it does look like it will be a solid offering the sad truth is that Ubuntu is the premiere Linux distro and has been for years.

            Comment


            • #7
              Originally posted by andyprough View Post
              I did not know about that. That is some odd behavior. Funny that he views Tumbleweed as a competing project.
              /me sees Phoronix, ignores it After so many years of dealing with their nonsense I no longer consider any benchmark hosted on that site to be...

              Comment


              • #8
                Originally posted by sophisticles View Post
                Can someone explain to me what the point is of back-porting patches and drivers to a kernel is instead of simply upgrading to the latest kernel?
                Not breaking out-of-tree drivers as the kernel interface does not change, and not introducing a whole butt-load of new kernel code (especially in drivers) that can introduce regressions you don't want in a LTS.

                Comment


                • #9
                  Originally posted by opensuselover View Post
                  link to reddit
                  It mostly boils down to the fact that phoronix tests only default settings so any result is hard to compare (and have therefore limited value). For example most of the OpenSUSE slowdowns are because it's using btrfs by default. mostly the methodology
                  comparing distributions in the way they do is like to comparing a race car, a saloon car, and a 40ton truck, in a drag race and then saying which is the best
                  And then next week they'll compare the same 3 in a road haulage challenge..
                  And then next week they'll configure them all differently again so there's absolutely no way of knowing if the different results are a result of the distribution or the other changes made...
                  Benchmarking isn't a toy to be used for some Top Gear like public spectacle, but an assistance for real engineers to do real analysis on real problems that effect the goals of the Project and the users of that project.
                  If I counted the amount of time I have wasted answer questions or chasing wild geese over Phoronix nonsense, I can assure you it has cost the Project far more in terms of motivation and contribution than the occasional nugget of useful information it has ever popped up
                  So, while I do not dismiss every 3rd party benchmark everywhere, Phoronix have well earned their lonely place on my list of "benchmarks I will outright dismiss with prejudice"




                  Comment


                  • #10
                    Originally posted by alcalde View Post
                    That conservative policy killed OpenSUSE's popularity. It used to be a cutting (not bleeding) edge distro with an 8-month release cycle that competed squarely with Fedora.
                    ^- That's the point where you're comparing apples-to-oranges.

                    The oldschool "openSUSE Linux" was split into two.

                    The nowadays "openSUSE Leap" is closest to compare to CentOS : it's an enterprise-y like thingy, but free (unlike RedHat and SUSE's SLES which are for corporate custommers), that is somewhat stable-ish with 8-month cycles. (Thus, still more cutting edge than CentOS).
                    It's something that you'll run on a server, while wanting to still have somewhat not-to-old packages (basically, for when you DO NOT want Debian Stable, but something more "current century"). And still benefit from the giant openSUSE collection of 2nd and 3rd party repositories (open Build service), to get more recentish version of package.


                    Post split, it's "openSUSE Tumbleweed" that has continued the niche of "cutting edge" version. Except now they have moved to a rolling-reƩease distro. While at the same time putting the same quality assurance as old school openSUSE Linux.
                    Basically you have the advantages of Gentoo / Debian Sid (always having the latest version of everything) with their drawbacks (constant breakage from using always "experimental" packages).

                    Originally posted by alcalde View Post
                    Fortunately the true OpenSUSE fans are still working on Tumbleweed, which to me is the real OpenSUSE. Still, we need to throw Brown out and fork if necessary to go back to the days of being an enthusiast desktop OS rather than some junior version of SUSE Enterprise. We've dropped the ball and other distros have picked it up. OpenSUSE needs to get back to being aggressive and creative again. We don't need conservative stability. We need reliable innovation.
                    Post-split, Tumbleweed *is for all purpose* the "real openSUSE". That's the message that SUSE has tried to pass on every conference where they have been present.
                    Basically, they have sped up the cycles.
                    And Yes, it now suddenly a *rolling release* distro. But it's not your garden variety rolling distro (it's not Gentoo nor Debian Sid), it's a rolling distro with all the quality and stability assurance that you're used from "real openSUSE". Except that instead of having new version numbers every ~7 months, it has a new tagged release *every week*.
                    You have the quality of a whole new version of old "real openSUSE" coming to you every single fucking week.

                    openSUSE Leap is *specifically* marketed as a junion version of SUSE Enterprise. Litteraly. It's described as a free version of SUSE Enterprise, with a few more modern package migrated from Tumbleweed.
                    Don't use it on your laptop, that's not what it is for.

                    Tumbleweed is aggressive and creative enough. That's were you'll find the innovation you crave for and that's how openSUSE has tried to market themselves.


                    Comment

                    Working...
                    X