It's worth noting that network performance is noticeably worse when running Windows XP on a Boot Camp partition or VM than under OS X. To show you the performance difference, I ran some makeshift network benchmarks, copying a 1.96GB file from a fileserver on a GigE switch to the MacBook Pro. I tested both wired and wireless connections, the results are below:

Wired Network Performance

Wireless Network Performance

The latest beta of Parallels' Workstation 2.1 improves network performance tremendously, but in the case of using the wired Ethernet port on the MacBook Pro Boot Camp is still noitceably quicker.

The final thing to note is that using Boot Camp, the MacBook Pro actually posts some not so great numbers for a Core Duo notebook. The charts below will put its performance into perspective for you:

General Performance - Business Winstone 2004

General Performance - MMCC Winstone 2004

While the Dell Inspiron E1705 performs pretty poorly given its CPU speed, the MacBook Pro is actually even worse. The E1705 holds about an 8% performance advantage over the MacBook Pro with an 8% faster CPU; however, neither Winstone test scales 1:1 with CPU speed increases so Dell's faster CPU is most likely only buying it another 3% performance advantage here. Obviously neither notebook comes anywhere close to the performance of the ASUS offerings, which continue to be the fastest I've ever encountered in a Core Duo notebook.

I don't really have a good explanation for the MacBook Pro's disappointing Windows XP performance, because all of its hardware is built out of the same major components that ASUS and Dell use for their notebooks. The only thing I can think of is that out of all these companies, ASUS is far more experienced with tweaking and tuning their motherboards for every last ounce of performance while honestly, Apple has never had to really care. Given that ASUS actually manufactures some of Apple's machines, it may be time to enlist its help in performance optimization as well.

Rosetta vs. Boot Camp vs. VT Problems in Mac-land
POST A COMMENT

48 Comments

View All Comments

  • Visual - Thursday, April 13, 2006 - link

    you have some borked page 18 after the "final thoughts" page showing up in the dropdown.
    in printarticle.aspx it shows up as a duplicate of page 17 for some reason

    as to the article itself - good job :)
    my guess as to why the vm is faster than the real thing is because the hdd emulation works somewhat like a ramdrive - its a file on the apple hdd but it probably gets cached up by osx or by the vmsoftware itself.
    Reply
  • plinden - Thursday, April 13, 2006 - link

    quote:

    my guess as to why the vm is faster than the real thing is because the hdd emulation works somewhat like a ramdrive - its a file on the apple hdd but it probably gets cached up by osx or by the vmsoftware itself.


    That could be - I have maxed out at 2GB RAM in my iMac, and I get wired RAM is close to the max and a hefty number of Page Outs (up to 210,000 last time I looked. before it setayed below 5000 even after being on for a week) while running Parallels VM.
    Reply
  • ibisbowti - Thursday, April 13, 2006 - link

    I been using the 1.83 Core Duo for about a week now. I think it is one of the latest builds according to the serial number. No problems at all, other than it does get pretty warm. Heat issue seems better after latest firmware update. I think the aluminum is designed to be a big heat sink! I thought the Front Row software would be a little gimmicky, but it is pretty cool, especially when sitting the unit on a coffee table and watching the HD movie trailers, IPhoto pics, etc with others. It's an awesome machine so far. Reply
  • artifex - Thursday, April 13, 2006 - link

    Since you say you ran the same tests as in your earlier review, I'd like to see graphs comparing the results of the Intel iMac vs. the MBP. and add in ones for the Intel Mini, if you can. I suspect we'll see iMac > MBP > Mini, but it would be nice to be sure.

    Also, if you could slap Parallels on the Mini and tell us how much of a hit the virtualization takes because the hardware virtualization is disabled for that line, that would be really interesting.

    Thanks.
    Reply
  • AppaYipYip - Thursday, April 13, 2006 - link

    "Apple quality control at it's best"

    That comment bothers me. Overall, there are no other manufacturers that come even close to Apples quality, design, and workmanship. Yet, you find one key that sits slightly off and suddenly feel the need to make such a blanket statement. If it bothers you so much, take it back and Apple will repair it for you, in record time.
    Reply
  • Calin - Thursday, April 13, 2006 - link

    The IBM thinkpads (before the Lenovo deal) were regarded as the best business laptops (or at least PC laptops) as quality and workmanship. Too bad they were designed with cramped keyboards (at least the models I saw) and no trackpad. Reply
  • Ryan Norton - Thursday, April 13, 2006 - link

    I'm a PC guy but I like Macs a lot and will probably buy a Macbook Pro, either now or when Merom ones come out. I figure Anand probably is too. Yet I work with Mac zealots who give me endless shit about the unequivocal superiority of Apple everything over PC (except for games, which they concede). So when someone like me finds a glaring flaw that seems like something that should have been spotted before it got to the end user, it's easy to take a cheap shot at Jobs =^) Reply
  • Sunrise089 - Thursday, April 13, 2006 - link

    If I had my dream review of this product, here's how I would have you test gaming performance:

    Test performance in Windows mode. Then compare it to other Core Duo notebooks. Then see if there is any game written natively for OS-X under Core Duo, and run the OS-X version and the Windows version to see the difference in OS on performance on the same machine.

    Other than no gaming info, terrific review.
    Reply
  • Calin - Thursday, April 13, 2006 - link

    This could be thanks to slower drivers in BootCamp Windows XP, or slower hard drive access/speed. Everything else is a disadvantage for VM: one more level of indirection in disk access, less memory, running the OS X behind the VM.
    Could you do some disk speed comparation between VM and native XP?
    Reply
  • BigLan - Thursday, April 13, 2006 - link

    I think it's going to be hard drive speed throwing off the benchmarks. The BootCamp partition is going to be at the outer edge of the disk, with much slower speeds than the VM client virtual drive which is on the faster Apple partition.

    I'm not sure if it's possible to assign the entire HD to a windows partition using Bootcamp, but that's about the only way i can think of to level the playing field.
    Reply

Log in

Don't have an account? Sign up now