Speed : Parallels vs a "PC"

Discussion in 'Windows, Linux & Others on the Mac' started by SDDave2007, Jul 6, 2007.

  1. SDDave2007 macrumors regular

    Apr 12, 2007
    I recently purchased my first Mac [MacPro 2.66] and created a clone of my WinXP desktop using Parallels Transporter [awesome software by the way]..

    Anyways if anyone is interested in how much speed difference there is... I ran GEEKBENCH under Windows on both machines...

    Pentium 4 2.0ghz w/1gig RAM scored 837
    MacPro 2.66 w/Parallels 3.0 w/1gig Ram scored 2182!

    Now yeah the clock speed is higher... so if you want to compensate for that the Mac would be about 1640 [still TWICE as fast]

    And it looks like Geekbench only using ONE logical processor.. so the 4 cores did not skew the results

    MacMinis, MacBooks etc will be different.... would be interested to see others results
  2. miniConvert macrumors 68040


    Mar 4, 2006
    Kent, UK - the 'Garden of England'.
    There's no doubt that Windows runs very speedily inside Parallels - having other operating systems running alongside OS X really does seem to work well. I find your particular numbers quite surprising given how much better they are than on Windows running normally, regardless of the hardware, but if that's where we're at with virtualization technology then that's just great!

    If it wasn't for one piece of software I have that uses a dialup modem and just wont work on Parallels I'd be more than happy to find new uses for my remaining Windows machines.
  3. LeviG macrumors 65816

    Nov 6, 2006
    Norfolk, UK
    your benchmarks really are not true as you have forgotten about differences in architecture.

    The p4 is 2 architectures older so clock for clock is not a fair comparison.
    a p4 would need to be around 5Ghz (possibly more) to rival a core2duo 2.66Ghz.

    So working on your doubling for dual core you would also need to double it again to give a fair comparison.
    So (estimates) the p4 would be relatively around the 3300 mark and the c2d would be 2182. Allowing for a percentage of error on the p4 and make it 3000 so it appears the parallels is working ok but still can not rival windows without emulation.

    My Athlon Opteron X2 175 (2.4Ghz, 4GB ram) pulled 2191 in xp and the 2.66Ghz c2d is upto 1.5x more powerful than the x2.
  4. DaiKirai macrumors member

    Jan 17, 2005
    If someone has both Parallels and a Boot Camp partition on the same Mac, that'd be the ideal way to perform this test.
  5. psychofreak Retired


    May 16, 2006
    And, if you don't need the internet in Parallels, the VM will not be slowed down by anti-virus and spyware like a normal PC :)
  6. gnasher729 macrumors P6


    Nov 25, 2005
    First, the Core Duo and even more the Core 2 Duo processors in the newer Macs run much much faster at the same clock speed than a Pentium 4. 2 GHz Core Duo did beat the fastest P4s; so a single Core2 Duo processor at 30 percent higher clock speed than a P4 should give you about that speed advantage.

    There are two things to remember with Parallels: The Macintosh will be running Windows and MacOS X at the same time, and Parallels only uses one of the cores. That means on one hand, if you have a Windows application that is optimised for multiple processors, it will only run as fast as on a single processor machine with Parallels. On the other hand, if that Windows application uses 100 percent of that one CPU, you still have three CPUs in your MacPro that can run Mac apps at full speed (or you could run Windows XP, Windows Vista and two versions of Linux, all at full single-processor speed).

    You could, just for fun, run MacBench (which is useless as a benchmark) and Geekbench on Parallels at the same time. Geekbench should run at almost the same speed.
  7. SDDave2007 thread starter macrumors regular

    Apr 12, 2007
    In any case.... My MacPro running Windows XP is 2.7 times faster than my "real pc".... nuff said I guess
  8. LeviG macrumors 65816

    Nov 6, 2006
    Norfolk, UK
    which actually isnt very good, having had another read of the geekbench site the program is multi threaded meaning that it can take into account all 4+ cores of your mac pro. Now iirc parallels has an issue with utilising all 4 cores at present which would account for a lower than average score in some respects.

    If you were to run geekbench in bootcamp the score would be even higher.
  9. SDDave2007 thread starter macrumors regular

    Apr 12, 2007
    Well Geekbench said it used 1 logical and 1 physical processor core....

    Now I'm sorry I even brought the subject up... it seems that all anyone on these forums can do is attempt to bash any/all information.....

    So the heck with it.... My MacPro screams past my PC..... and thats good enough for me

Share This Page