YouTube Vid a 2.66 Mac Pro vs i7 64-bit windows PC

Discussion in 'Mac Pro' started by skyline r34, Jan 3, 2009.

  1. skyline r34 macrumors 6502

    skyline r34

    Joined:
    Oct 10, 2005
    Location:
    San Diego
    #1
  2. Slowstick macrumors 6502

    Joined:
    Dec 16, 2008
    #2
    Seems unfair. A Mac with two workstation processors versus one Core i7?
     
  3. Tallest Skil macrumors P6

    Tallest Skil

    Joined:
    Aug 13, 2006
    Location:
    1 Geostationary Tower Plaza
    #3
    Yep. And when the Nehalem Xeon Mac Pro comes out, further ownage will commence.
     
  4. Eidorian macrumors Penryn

    Eidorian

    Joined:
    Mar 23, 2005
    Location:
    Indianapolis
    #4
    How so? 8 threads.
     
  5. Slowstick macrumors 6502

    Joined:
    Dec 16, 2008
    #5
    That's true, I am sorry. However, it may still be unfair. Need to know more info about the specific processors.
     
  6. Eidorian macrumors Penryn

    Eidorian

    Joined:
    Mar 23, 2005
    Location:
    Indianapolis
    #6
    Keep in mind on a good day that a Core i7 starts off around US$250. Try getting a Xeon DP for that price.
     
  7. Slowstick macrumors 6502

    Joined:
    Dec 16, 2008
    #7
    Agreed again! :) Price is certainly a factor!
     
  8. wheezy macrumors 65816

    wheezy

    Joined:
    Apr 7, 2005
    Location:
    Alpine, UT
    #8
    Hmmm, I just ran Geekbench 32 bit on my MacPro:

    2.8Ghz Quad x 2
    6GB RAM
    ATI 2600 HD 256MB
    320GB, 500GB, 2 1TB

    My score was 9094
    :Integer - 8725
    :Floating Point - 15076
    :Memory - 2619
    :Stream - 2404

    Why is my 2.8 getting a better score than the OP's 3.2? Amount of Memory?
     
  9. Slowstick macrumors 6502

    Joined:
    Dec 16, 2008
    #9
    I think there is something we don't know about what went on in the OP's test. Anyone disagree?:confused:
     
  10. Eidorian macrumors Penryn

    Eidorian

    Joined:
    Mar 23, 2005
    Location:
    Indianapolis
    #10
    Hopefully they had enough sense to kill as many background processes as possible for an accurate benchmark and run it several times for an average result.
     
  11. Slowstick macrumors 6502

    Joined:
    Dec 16, 2008
    #11
    I hope so. But did they do it on both machines? Remember... the i7 is running Windows!!!:D:D This test can't be fully accepted because of the different OSs. OS X handles background processes a hell of a lot better.
     
  12. Eidorian macrumors Penryn

    Eidorian

    Joined:
    Mar 23, 2005
    Location:
    Indianapolis
    #12
    What's the point of a cross platform benchmark then?
     
  13. Slowstick macrumors 6502

    Joined:
    Dec 16, 2008
    #13
    General comparison only. For a complete and accurate test, the Mac Pro would have to be running Windows as well. Its only fair.
     
  14. Ramashalanka macrumors regular

    Joined:
    Dec 26, 2008
    Location:
    Lanka Ravi Shanka
    #14
    No. Eight threads on a quad-core (even with hyper-threading) will be slower than eight threads on an eight-core, all else remaining equal, so it isn't a fair test of the difference between the core 2 and core i7 technology. We'll see that comparison most directly when there is a Gainestown Mac Pro to compare to my current Harpertown.

    In the meantime, the geekbench results for an eight-core core i7 Hackintosh look promising at:

    http://browse.geekbench.ca/geekbench2/view/96514

    Though, as always, differences in clockspeeds etc cloud the issue.
     
  15. Slowstick macrumors 6502

    Joined:
    Dec 16, 2008
    #15
    And operating system...:D:D
     
  16. Ramashalanka macrumors regular

    Joined:
    Dec 26, 2008
    Location:
    Lanka Ravi Shanka
    #16
    Absolutely :), that was in my "etc". Amount and speed of ram are other big ones. As Eidorian correctly says, a big issue will be how careful the test has been with background processes and multiple trials:
    From the videos, it looks like the test results are from a single run.
     
  17. Slowstick macrumors 6502

    Joined:
    Dec 16, 2008
    #17
    And speed of hard drive...
     
  18. Eidorian macrumors Penryn

    Eidorian

    Joined:
    Mar 23, 2005
    Location:
    Indianapolis
    #18
    Geekbench is CPU only.
     
  19. Slowstick macrumors 6502

    Joined:
    Dec 16, 2008
    #19
    Really?! Are you sure? I didn't know that.

    Then all of the above postings would be irrelevant. If what you say is true, then any Geekbench testing would be hard to believe because when you are testing two machines, operating system, ram, hard drive speed, etc. are all important factors in overall speed. Do you see what I am saying? Would you agree?
     
  20. scottlinux macrumors 6502a

    scottlinux

    Joined:
    Sep 21, 2005
    #20
    Yes geekbench is cross-platform, and CPU only. The entire point is to make a fair benchmark no matter the OS, and one that does not even factor in hard drive speed.

    I've benchmarked Geekbench with Linux, OS X (10.5 hackintosh) and Windows XP on the same PC, and scores are a little different but fairly close. Memory scores are higher in OS X, and a lot worse in XP curiously enough.
     
  21. NoSmokingBandit macrumors 68000

    NoSmokingBandit

    Joined:
    Apr 13, 2008
    #22
    Agreed. If they wanted a windows vs osx test they should have installed them on the same computer. If they wanted a hardware test they should have installed the same OS. There are to many variables here to draw any logical conclusion.
    The whole test is a fail.
     
  22. Slowstick macrumors 6502

    Joined:
    Dec 16, 2008
    #23
    .

    My point exactly. It just doesn't make sense.
     
  23. Ramashalanka macrumors regular

    Joined:
    Dec 26, 2008
    Location:
    Lanka Ravi Shanka
    #24
    I agree that geekbench is pointless, but I don't see why you decide this due to your surprise that it's CPU only. Geekbench is not affected by hard drive speed, but is affected by operating system, RAM size, RAM speed, motherboard (and many other details) as well as CPU (I'm certain that is what Eidorian meant by "CPU only").

    The problem is that it is affected by so much (including about two percent variation between successive runs), not by too little (no hard drive speed). Also, the idea of doing a bunch of tests and assigning each of them some arbitrary importance to come up with one number is an oversimplification.
     
  24. Eidorian macrumors Penryn

    Eidorian

    Joined:
    Mar 23, 2005
    Location:
    Indianapolis
    #25
    Yeah that is what I meant. Just for raw hardware the CPU impacts integer and floating point performance and there are memory and stream tests as well.

    The best I can manage on my G31 board is about 5500 on Geekbench under Vista Ultimate 64.
     

Share This Page