Announcement

Collapse
No announcement yet.

Debian 8.5 vs. Debian Testing Benchmarks - July 2016

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

  • Debian 8.5 vs. Debian Testing Benchmarks - July 2016

    Phoronix: Debian 8.5 vs. Debian Testing Benchmarks - July 2016

    Here is the latest look at the performance of Debian GNU/Linux 8.5 vs. Debian Testing on the same system for showing how the performance is looking for Debian 9 "Stretch" ahead of its release next year...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    7 vs 14.

    Comment


    • #3
      You could get the same kernel from Debian backports for jessie. This is always a good idea with new hardware, please add some benchmarks with that.

      Comment


      • #4
        I think debian+xfce is the best combination, especially when installed from testing! If only someone could comment on that... (just kidding)

        Kano I am running Debian stable, with both kernel (4.6) and mesa (11.1 I think) from backports. It is a great idea .

        Btw, there is an interesting discussion going on about making a "jessie and a half" release (https://lists.debian.org/debian-back...7/threads.html).
        I think they still need to backport xserver and a few other packages.

        Thanks Michael for the tests!
        Last edited by franglais125; 08 July 2016, 03:09 PM.

        Comment


        • #5
          Originally posted by phoronix View Post
          Debian GNU/Linux
          Nice try, Stallman. Hacking Phoronix these days, are we?

          Comment


          • #6
            Originally posted by franglais125 View Post
            I am running Debian stable, with both kernel (4.6) and mesa (11.1 I think) from backports. It is a great idea .
            Great idea, but unusable... at least mesa from my radeonsi POV. That mesa backport is 7 month old compiled against 2 years old llvm... It is better to not use that

            If someone really wants stable, should compile recent (or one older) stable mesa branch against recent (or one older) stable llvm or branch, git2git of stable branches, take2... if not that, but one like to track point releases of mesa then llvm release diff should never be more then 6-9 months, etc...
            Last edited by dungeon; 08 July 2016, 06:06 PM.

            Comment


            • #7
              Originally posted by debianxfce View Post
              Debian should drop stable distribution and backports because it is a waste of resources.Experimental-unstable-testing cycle is enough and using testing allows you to use unstable and experimental packages and you will have fresh software and can make Debian better by testing it. Debian testing is a rolling release distribution and that is modern computing. Both stable and testing do have about 400 bugs, so the stable distribution is a myth.
              Stable, testing or unstable - your choice. First is release, second and third are development branches.

              Testing is not modern distribution, it is development branch of Debian possibly more stable then unstable but sometims isn't . Once Freeze period happen it is not so modern isn't it, for about 6 month of that period it does not roll. It is not fixed period, but sort of about 18 month development branches rolling and then you have 6 months of Freeze there.

              In that Freeze period, 'unstable' start to serve to the 'next stable' process so from rolling became semi rolling and 'testing' does not roll at all So who like to roll should use 'unstable' really, and alpha/beta testers who like to try next debian stable early, should use 'testing' and specialy in that Freeze period
              Last edited by dungeon; 09 July 2016, 04:08 AM.

              Comment


              • #8
                Originally posted by debianxfce View Post

                Debian should drop stable distribution and backports because it is a waste of resources.Experimental-unstable-testing cycle is enough and using testing allows you to use unstable and experimental packages and you will have fresh software and can make Debian better by testing it. Debian testing is a rolling release distribution and that is modern computing. Both stable and testing do have about 400 bugs, so the stable distribution is a myth.
                You are wrong. Debian has implication far behind just desktop users. From proprietary NAS setups to webservers to routers, quite a good number of devices run Debian. Discontinuing debian stable would be a great loss.

                Comment


                • #9
                  Originally posted by dungeon View Post
                  If someone really wants stable, should compile recent (or one older) stable mesa branch against recent (or one older) stable llvm or branch, git2git of stable branches, take2... if not that, but one like to track point releases of mesa then llvm release diff should never be more then 6-9 months, etc...
                  is llvm so fragile?

                  Comment


                  • #10
                    Originally posted by trek View Post
                    is llvm so fragile?
                    Well, yes and no If you want very latest 'radeonsi' checkout you should use current mesa main git compiled against llvm main git too.

                    But if you want latest stable releases, one should match releases where llvm isn't recommended so shouldn't be 1 or even 2 years old like in this backport case, that might work but you will only see old bugs there regardless of mesa being more recent, etc...

                    If not releases one might like to track recent stable git of both, like mesa 12 git against 3.8 git, there are stable things... let me find links... So let say, this against that:


                    Project moved to: https://github.com/llvm/llvm-project - GitHub - llvm-mirror/llvm at release_38


                    So if you want stable track that, then after about 3-4 months change to 12.1 against 3.9, etc...
                    Last edited by dungeon; 09 July 2016, 05:35 AM.

                    Comment

                    Working...
                    X