Announcement

Collapse
No announcement yet.

Is Windows 7 Actually Faster Than Ubuntu 10.04?

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

  • #61
    Originally posted by BlackStar View Post
    Nice work, Michael!

    The most important results are without doubt those of XPlane, Unigine and SPECviewperf,
    Not sure how important SPECviewperf tests are when there is only one app in the tests that has a linux version still.

    Comment


    • #62
      I would recommend the author of this article to display the results differently once you have the MacOS results in. In this article the graphs where unreadable. There were too many lines and it became almost impossible to keep track of which color was associated with which system and OS, especially if you are partially colorblind.

      So I would suggest that you separate the line graphs by system in the next run. Yes it will result in many more graphs, but that also means more displayed ads which for you is a good thing.

      The tables were great and actually saved the article from being unreadable.

      Comment


      • #63
        Originally posted by deanjo View Post
        Ya, don't know what is up with the server. But going back the scheduler improvements found in Win 7
        Yup, 2.6.28 is very old as well and Linux scheduler is improved a lot for desktops since 2.6.32.
        The truth is one. Almost everyone (including me) believed till today that windows had the upper hand by a large marging in games. It has been proved that's not the case.

        Comment


        • #64
          Originally posted by deanjo View Post
          Not sure how important SPECviewperf tests are when there is only one app in the tests that has a linux version still.
          And games for Linux are rare as well, but that does not mean that the game tests don't have their importance.

          Comment


          • #65
            Originally posted by Apopas View Post
            The truth is one. Almost everyone (including me) believed till today that windows had the upper hand by a large marging in games. It has been proved that's not the case.
            That can't be truthfully said without test runs on DX as well. Like many other applications the ogl / DX performance can differ greatly especially in the unigine benches.

            Comment


            • #66
              Originally posted by deanjo View Post
              That can't be truthfully said without test runs on DX as well. Like many other applications the ogl / DX performance can differ greatly especially in the unigine benches.
              True, but still Linux can be tweaked a lot as well, so we will never be able to run a really objective test. Take for example that even Fedora (which is not as tweaked as Gentoo for example) was much much faster than Ubuntu with NVidia drivers. Over 50% if I recall correctly. Both me and you pointed that there is a problem with games in Ubuntu.
              Think it in another way also, even if DX runs games faster than GL (not necessarily true though), the difference won't be that high to say that windows is the undisputable leader in games performance.

              Comment


              • #67
                Damned 1 minute limit

                I believe, for example, that my Gentoo is gonna be faster than windows in the same machine, in almost every game. But Gentoo is not a desktop OS, so it doesn't count for many people. I use it as a desktop OS though.
                So what do we have if we remove every subjective point of view and stay with the sheer numbers?

                Comment


                • #68
                  Originally posted by Apopas View Post
                  True, but still Linux can be tweaked a lot as well, so we will never be able to run a really objective test. Take for example that even Fedora (which is not as tweaked as Gentoo for example) was much much faster than Ubuntu with NVidia drivers. Over 50% if I recall correctly. Both me and you pointed that there is a problem with games in Ubuntu.
                  Think it in another way also, even if DX runs games faster than GL (not necessarily true though), the difference won't be that high to say that windows is the undisputable leader in games performance.
                  That is all dependent on the implementation of course but one thing that has been shown over and over again for example is that unigine is a lot faster using DX. Also many apps are profiled and optimized in the windows drivers for their DX engines and may not have a ogl profile. 10 years ago you didn't see DX engines in professional CAD / Modelling / etc but now days pretty much all the major players default to a DX engine.

                  Comment


                  • #69
                    Looks like the Phenom Quad-Core @ 2.20GHz system with HD 5770 was CPU limited in most benchmarks both on Linux and Windows. So it got a rather poor showing compared to the HD 4670 .

                    Comment


                    • #70
                      Nice tests. Looking forward for the follow up.

                      For somebody looking for a medium PC configuration to run Linux on it, I found AX3 configuration to fit that role beautifully. It was on par with Windows on almost all gfx tests. The ATI Radeon HD 4670 seems to perform very nice and the drivers seems to be mature for this video card.

                      Comment


                      • #71
                        Well everyone would choose their own @#$%^ux intead of Ubuntu for testing, but they are minor of the minor which makes the result not reproducible without them being around for installation and configuration and shit. I know the world do not revolve around Ubuntu but it certain less likely to revolve around your little tiny ego distribution.

                        Comment


                        • #72
                          Thanks Phoronix/Michael, that was interesting. I agree with others that Linux (Ubuntu in this case) is just as viable as Windows for many people. When you already test closed source games, why not also run Doom3 or Quake4? Really hope you are right about the Valve Source Engine comming to Linux, and not just the new Postal game. Will be interesting to compare more modern games than always ioquake3...

                          Originally posted by deanjo View Post
                          Not sure how important SPECviewperf tests are when there is only one app in the tests that has a linux version still.
                          Well, most highend DCC apps runs on Linux because, you should know, all major CG/VFX studios use Linux. Maya is just one of the widely used ones. These workstation numbers are consistent with what I experience myself working in VFX using Linux compared to Windows.

                          And what is that bull**** about DCC apps moving to DX?? You must have heard this from a 3DSMAX user who thinks he's 1337... OpenGL is THE workstation graphics API, you can't take Direct3D serious in this space.

                          Comment


                          • #73
                            The title is of this article is misleading. Clearly you aren't suggesting that that Windows simply performs better than Linux. Nobody would be running it on their supercomputers if that were the case. Everyone already knows that graphics drivers on windows outperform Linux, and that's not likely to change any time soon nor is there anything we can really do about it until the open-source drivers are at least on par with proprietary ones. This is only a test of graphics drivers though and doesn't tell us much about the performance of the kernel and such.

                            Comment


                            • #74
                              Originally posted by Smorg View Post
                              The title is of this article is misleading. Clearly you aren't suggesting that that Windows simply performs better than Linux. Nobody would be running it on their supercomputers if that were the case. Everyone already knows that graphics drivers on windows outperform Linux, and that's not likely to change any time soon nor is there anything we can really do about it until the open-source drivers are at least on par with proprietary ones. This is only a test of graphics drivers though and doesn't tell us much about the performance of the kernel and such.
                              So because linux runs better HPC nodes means it's a better OS? I think you have to set criteria before you claim any thing is better then another. I'm pretty sure that most people here are not running linux for HPC purposes.

                              Comment


                              • #75
                                This test was about gaming, so it does not really matter for computing in general. With users of Ubuntu that I know, myself included, we all agree Ubuntu is faster, more stable and safer. This is what counts. For advanced and even professional gamers, Windows is the OS to use. Ubuntu and other Linux distros may close the gaming gap later though this is not that important. For us, Ubuntu has proved its worth in terms of speed, reliability and safety. I am very happy with 10.04.

                                Comment

                                Working...
                                X