Announcement

Collapse
No announcement yet.

Ubuntu 7.04 to 8.10 Benchmarks: Is Ubuntu Getting Slower?

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

  • #61
    Originally posted by glasen View Post
    A user at the german forum "www.ubuntuusers.de" did some tests with his Lenovo T60
    that would be me

    Originally posted by Michael View Post
    There's a few possible reasons... Do you have a link to his post / information?
    I'll run a few tests again (unfortunately i don't have "more than 100 hours" to run all the tests) and post the graphics/results here later this day. My T60 is the 2007-FVG Model.

    All i can tell you for now is that i get up to two times better results at all your audio-encoding tests -actually very close to Ubuntu 7.04/7.10

    Comment


    • #62
      This may have something to do with the fact that your Ubuntu 8.10 may be your first partition... Could it?

      Comment


      • #63
        OK I just ran a quick Encode-mp3 test on my system which runs Ubuntu 8.04 and my results do not jive with Phoronix's.

        My system is a 1.7GHz Pentium 4(first generation) with 768MB(RIMM) memory, so should be a little less than twice as slow as a T60 dual 1.83GHz Pentium M cores and 1GB memory. A Single core Pentium M 1.83GHz would perform as well as a 3 GHz Pentium 4 first generation, So since lame doesn't use multithreading a comparison can be made as the second processor in the T60 will remain mostly unused, and still the T60 should be about 44% faster. (my hardrive is very old 10 Gig drive, so the T60's should be the same or better)

        My results were 123.14s compared, to 116.67s on the T60, so yes my system is just slightly slower than the T60, which is only about 5% faster, where it should be about 44% faster (which is what you see the T60 do on the test under 7.04).

        I would also add, it is a really really really bad idea to do a OS version speed comparison, and only use one machine, two different machines is the bare minimum to be able come to any conclusion and write an article, and I would do probably five machines with different architectures, also changing system memory amounts to see any differences as well.
        Last edited by pataphysician; 29 October 2008, 10:17 AM.

        Comment


        • #64
          instant fail, you can't compare like that. Compare different ubuntu versions on your box.

          Comment


          • #65
            I can't do a clear comparison, but there is no way possible that my 8.04 machine should be anywhere close to a T60, I have done enough cross architecture testing in the past to know this. The values shouldn't be anywhere this close and since the T60 performs right where it should to a 1.7 pentium 4 first gen in the 7.04 and 7.10 tests, I would have to believe there is another problem.

            Besides Phoronix test is instant fail as they tested only one fucking machine to do a OS speed comparison, that is called amateur.

            Comment


            • #66
              oh boohuuu. How many sites out there test hardware or games on more than one system?

              Doing a comparism on one piece of hardware is fine. Comparing something completly different to that results and then conclude that the test was bogus is not.

              Comment


              • #67
                I didn't say the test was bogus, just that there must be something up on the t60 under ubuntu 8.04+, other than newer ubuntu just being naturally slower.

                I'm not a big ubuntu user and this older machine is the only one I have with it installed, and I don't really have the time to put on 7.04, though if I was running a benchmarking site that wanted to be taken seriously I probably would. It used to have debian lenny on it, but I decided I would take a look at ubuntu, I didn't notice any real performance difference between the two.

                There are many sites that are amateur benchmarking sites, it's unfortunate that Phoronix seems to be one of them.

                Yes most serious hardware benchmarking sites, for example for gamers, benchmark games on a huge host of hardware to do comparisons. There were comparisons done for example on gaming on XP versus Vista, and serious benchmarking sites compared across multiple machines and architectures and memory configurations, it just goes without saying. I don't see how this is so hard, anyone in IT usually has multiple machines they can test on, just in their own personal possession, let alone if they are running a benchmarking site. I have about 10 different machines in my house alone, most of which I got cheap from clients or jobs or bought used. So testing on at least two machines to weed out a possible hardware specific problems is a no brainer.

                Comment


                • #68
                  they stopped using xp or doing comparisms a long time ago - and when they test hardware, they only change ONE (!) piece at a time.

                  Comment


                  • #69
                    @energyman :

                    Have you ever heard the term "proof of plausibility".

                    Comment


                    • #70
                      Originally posted by energyman View Post
                      How many sites out there test hardware or games on more than one system?
                      How many readers judge hardware based on ONE review? I hope you get the point.

                      Originally posted by energyman View Post
                      that the test was bogus
                      If not then the results should be reproducible.

                      Encoded 52.6MB WAV to MP3 on one machine with one version of lame (3.98.2):

                      31.7 sec. on Feisty
                      31.1 sec. on Hardy
                      30.9 sec. on Intrepid RC

                      That doesn't correlate with Phoronix' encoding time increase of ~100% from Feisty to Intrepid and I haven't seen any other results that do.

                      I'm not a fanboy - I'd just like to see benchmarks that are worth reading and I know that's what Phoronix want too.
                      Last edited by LinuxAffenMann; 29 October 2008, 12:50 PM.

                      Comment

                      Working...
                      X