Announcement

Collapse
No announcement yet.

Arch-Based Manjaro Linux Runs Against Ubuntu 13.04

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

  • Arch-Based Manjaro Linux Runs Against Ubuntu 13.04

    Phoronix: Arch-Based Manjaro Linux Runs Against Ubuntu 13.04

    Here are our first benchmarks of the newly-released Manjaro Linux 0.8.5. Manjaro is a promising distribution based upon Arch Linux and has been benchmarked against Ubuntu 13.04...

    http://www.phoronix.com/vr.php?view=MTM1ODQ

  • #2
    So Manjaro is faster for bzip2, C-Ray, ffmpeg, and TSCP;
    equal for FS speed, prime seive, vpx, FLAC, lame;
    and slower for everything else--including GL, Dolfyn fluid dynamics, FFTE, compile tests, and surprisingly large margins on the apache benchmark.
    Which hypothetically means Manjaro is better for raycasting, chess, and playing videos, but worse for games and science.

    I wonder what's making Apache slower?
    It's not FS overhead, but is it a kernel regression, an optimization gone wrong in GCC, or something else?

    Comment


    • #3
      Originally posted by Ibidem View Post
      I wonder what's making Apache slower?
      It's not FS overhead, but is it a kernel regression, an optimization gone wrong in GCC, or something else?
      You should always ignore the Apache test here. Michael runs it in a weird configuration that doesn't actually test what a real Apache server is doing, in a way that is specifically warned against by the Apache devs. No one really knows what it's testing, it just seems to fluctuate randomly whenever a kernel gets upgraded.

      Comment


      • #4
        Originally posted by Ibidem View Post
        So Manjaro is faster for bzip2, C-Ray, ffmpeg, and TSCP;
        equal for FS speed, prime seive, vpx, FLAC, lame;
        and slower for everything else--including GL, Dolfyn fluid dynamics, FFTE, compile tests, and surprisingly large margins on the apache benchmark.
        Which hypothetically means Manjaro is better for raycasting, chess, and playing videos, but worse for games and science.

        I wonder what's making Apache slower?
        It's not FS overhead, but is it a kernel regression, an optimization gone wrong in GCC, or something else?
        I wonder if the Manjaro devs recompile their packages... the Arch devs purposefully pull out all debugging support, if the manjaro devs recompile the packages though they could be compiling back in debugging support which would make the packages slower than the stock arch packages

        Comment


        • #5
          Originally posted by Ericg View Post
          I wonder if the Manjaro devs recompile their packages... the Arch devs purposefully pull out all debugging support, if the manjaro devs recompile the packages though they could be compiling back in debugging support which would make the packages slower than the stock arch packages
          Debug symbols don't affect performance. They're only loaded if actually needed.

          Comment


          • #6
            C-ray must be pretty important:

            Comment


            • #7
              There's no reason Arch or Ubuntu should perform faster than one another. They're both compiled for an x86_64 baseline with stack smashing protection and -DFORTIFY_SOURCE. The benchmarks are likely just showing improvements and regressions in the kernel and compiler. The Arch packages are the latest stable releases and the Ubuntu ones are a frozen snapshot of older ones - it's the same software.

              Anyway, Manjaro is not Arch Linux. They build their own set of packages, with their own configuration options and they don't follow the same rolling release policy. Their repositories are organized a bit differently and there are packages included that aren't in Arch. It's not just a spin with different pre-installed packages. It's essentially the same relation Ubuntu has to Debian.

              Comment


              • #8
                I think Manjaro is faster

                I have been using manjaro for a while, and manjaro is faster than any ubuntu based distro. I really like Manjaro.

                Comment


                • #9
                  Originally posted by edoantonioco View Post
                  I have been using manjaro for a while, and manjaro is faster than any ubuntu based distro. I really like Manjaro.
                  There's no way it's faster at the kind of throughput benchmarks michael runs.

                  What it might be faster at is responsiveness, due to less background processes running and less RAM usage, particularly on older slower machines. Perhaps not so coincidentally, that's the type of improvement that normal users might feel - they don't use benchmarking tools to record the fps that their web browser renders at, but they can feel when an app starts up a little faster.

                  But the same exact code running on the same exact hardware isn't going to magically run faster just because you have a different logo on the desktop.
                  Last edited by smitty3268; 04-25-2013, 10:15 PM.

                  Comment


                  • #10
                    Originally posted by smitty3268 View Post
                    There's no way it's faster at the kind of throughput benchmarks michael runs.

                    What it might be faster at is responsiveness, due to less background processes running and less RAM usage, particularly on older slower machines. Perhaps not so coincidentally, that's the type of improvement that normal users might feel - they don't use benchmarking tools to record the fps that their web browser renders at, but they can feel when an app starts up a little faster.

                    But the same exact code running on the same exact hardware isn't going to magically run faster just because you have a different logo on the desktop.
                    ok, I have no trouble recognizing than you are right, I was just talking about my personal experience, possibly, for being a less cluttered distro with software than usually I do not use, it feels faster (mi PC is 4 RAM, not old).

                    Comment


                    • #11
                      Do you plan to include a Gentoo box in the multi-distro comparision? It would be interesting to see the effect of optimizing for the local processor with -march=native.

                      Though essentially Gentoo is not a single distro. Off the top of my hat I would guess that at (almost) full-system LTO, march=native and -O2 vs. -O3 might make a difference. And naturally the desktop environment (on a stock Gentoo you dont necessarily have X).

                      Comment


                      • #12
                        Originally posted by ArneBab View Post
                        Do you plan to include a Gentoo box in the multi-distro comparision? It would be interesting to see the effect of optimizing for the local processor with -march=native.

                        Though essentially Gentoo is not a single distro. Off the top of my hat I would guess that at (almost) full-system LTO, march=native and -O2 vs. -O3 might make a difference. And naturally the desktop environment… (on a stock Gentoo you don’t necessarily have X…).
                        I wonder why there still is no benchmark made for "full-system LTO, march=native and -O2 vs. -O3", not that i care about linux, its just that i saw hundreds of useless and freaky benchmarks made on phoronix by cocky linux fans, who thinks that linux is better (or will be after 1000 years) than any other os ( loool ), and to make this one should be equal to get nobel prize for achievements for linux, instead of getting a "Losers cap" So, who wants it ?

                        O.T.
                        I was watching shameless last 4 season, and there was a campaign made by a few people with some big words, and i instantly thought about linux community Oh boy, that was a lot of fun
                        Last edited by startzz; 04-29-2013, 10:49 AM.

                        Comment


                        • #13
                          Originally posted by startzz View Post
                          I wonder why there still is no benchmark made for "full-system LTO, march=native and -O2 vs. -O3", not that i care about linux, its just that i saw hundreds of useless and freaky benchmarks made on phoronix by cocky linux fans, who thinks that linux is better (or will be after 1000 years) than any other os ( loool ), and to make this one should be equal to get nobel prize for achievements for linux, instead of getting a "Losers cap" So, who wants it ?

                          O.T.
                          I was watching shameless last 4 season, and there was a campaign made by a few people with some big words, and i instantly thought about linux community Oh boy, that was a lot of fun
                          Come back when Windows runs on 80% of super computers, runs most of the infrastructure of the web, can be turned into special purpose version that is just 5 MB in size, runs more than a few phones or when you are able to fix a bug in your OS yourself.
                          If Windows is superior to Linux all that should be achieved right now already, but for some reason those poor losers that actually make money with using computers of do some real work on them prefer those loser OSes, which should tell even the dumbest Windows troll something.

                          Comment

                          Working...
                          X