Announcement

Collapse
No announcement yet.

A Big Operating System Benchmark Comparison

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

  • #41
    Originally posted by Qaridarium
    you also can use hackintosh ... .
    That would be illegal and put Micheal in the range of Apple's Ninja lawyers.

    Comment


    • #42
      Originally posted by Apopas View Post
      Yup, also the new OSX will be pure 64 bit as they say.
      Well yes and no. OS X when ran on a 64-bit system will be pure 64. If ran on one of the early intel Macs that only had a 32-bit intel it will still run as 32-bit.

      Comment


      • #43
        You need to give -ftree-vectorize as compiler parameter to get vectorization from gcc.

        And for old gcc version that didn't have autodetection capability you need also give -msse2.

        Comment


        • #44
          Originally posted by suokko View Post
          You need to give -ftree-vectorize as compiler parameter to get vectorization from gcc.

          And for old gcc version that didn't have autodetection capability you need also give -msse2.
          Automatic Vectorization for SSE/SSE2 has been added since GCC 4. Mearly adding in -O3 should kick it in.
          Last edited by deanjo; 15 August 2009, 09:47 AM.

          Comment


          • #45
            Originally posted by Qaridarium
            no thats a good argument becourse:
            debian-i386=486!!!!
            debian-amd64=64bit+sse2

            you are a laier if you say you can use sse3 on the atom with an defauld desktop linux distri!

            "save on manufacturing costs"

            only becourse they must pay lizence per cpu to amd for the 64bit part!
            so don't use debian - or do a source install, either way, saying the processor is completely worthless for everyone because one distro doesn't take full advantage of it is even worse then your last arguement.

            and no - it actually costs them more to make 64bit chips. more transistors required and such. the liscence fee is only part of that expense, not that that is really relevant. bottom line - it was cheaper, more energy efficient, and those were the important properties needed for the intended platform - why is that too hard to grasp?

            Comment


            • #46
              Hi!

              I would like to see:
              1. Fedora 11, Mandriva (whatever is the lastes), OpenSuses (whatever is the lastest), Bubuntu (whatever is the latest) - because these are popular and everyone use them
              2. ArchLinux - because it's rolling the bleediest of the bleeding edge (construct Your OS Yourself style), gentoo style with installation and configuration in couple of mins (actually I'm using it daily)
              3. SLES, RHEL, CentOS, Solaris - because these are latest, supported from respective vendors, etc., just wanna see how this compares to others
              4. OpenSolaris, Free[Open]BSD - just the results, no specific reason

              regards
              Kirurgs

              Comment


              • #47
                I would like to see SLED/SLES and RHEL in the benchmark. Since both are made for enterprise. Also they do have support and have a community built operating system (openSUSE and Fedora).

                Having to many Linux based operating systems would be meaningless I think, and hardware wouldn't matter as long it is the same for all. Maybe Windows 7 should be in to.

                also two test that would be intressting are SunSpider webbrowser test on the default browser of the operating system.
                The second are GeekBench since it works on the most platforms.

                Comment


                • #48
                  Originally posted by Qaridarium
                  you are a laier if you say you can use sse3 on the atom with an defauld desktop linux distri!
                  That's a choice of your distribution.
                  Moblin uses sse3 in 32 bit for everything.
                  Fedora 12 will use sse2 in their 32 bit install.
                  ..
                  ..

                  It's a choice. If you don't like the choice your distribution makes, either talk to them to change it, or change distributions.

                  Comment


                  • #49
                    Originally posted by tesuki View Post
                    also two test that would be intressting are SunSpider webbrowser test on the default browser of the operating system.
                    Well you can already run PeaceMaker which is a bit more comprehensive benchmark of browsers already online. They have comparitive results there. It would be a duplicated effort.

                    The second are GeekBench since it works on the most platforms.
                    Geekbench won't run on 64-bit unless you pay for it and because it's a precompiled blob it is extremely hard extrapolate results on processor architecture. We have no idea what has been optimized and for what. Since it's not a real-world use benchmark results would be speculative and inconclusive as to how it would pertain to real world use.

                    Comment


                    • #50
                      Originally posted by L33F3R View Post
                      Just because you use arch doesnt make it an ideal testing platform. I use linux mint but you dont see me asking for that to be tested, i can already assure you its bloated. If the man had the time i would say go for it but he and hopefully others here have lives to attend .
                      I use gentoo... Never tried Arch!
                      And because I respect other's lives I say to make the benchmark even more worthy

                      Comment

                      Working...
                      X