First Look: From the Lab: XP-on-Mac benchmarks
You’ll forgive me if I seem a little confused these days. Despite the fact that I work in a lab full of Macs, I spend more of my time lately staring at Windows XP than at Mac OS X.
It’s not that I’m mulling over a platform switch—rather, this XP overload is part of Macworld Lab’s efforts to see how Microsoft’s operating system performs on Apple hardware now that software exists that enables you to boot into XP on an Intel-based Mac. With the help from our sister publication, PC World, we’ve been running the WorldBench 5 real-word benchmark suite on all of our Intel systems to gauge cross-platform performance.
Our testing efforts began shortly after hackers came up with a way to get Intel-based Macs to run Windows XP. The hacked method takes a lot of effort install and configure —and just as much effort to run benchmark tests, as it turns out. WorldBench automatically restarts the computer many, many times during the testing process, forcing us to manually select which OS—XP or OS X—to boot into after each restart.
That’s why we greeted last week’s arrival of Boot Camp with a mixture of relief and sadness. Developed by Apple to let Intel-based Macs run Windows XP, Boot Camp allows you to set Windows as the default startup system; that meant we no longer had to babysit the machines during testing. At the same time, we were also a little peeved about how many hours and days we had wasted benchmarking the hacked method.
Even with Boot Camp, it still takes a couple of hours to set these systems up. But once running, they’ve been very stable. Here are some WorldBench 5 results, compared to three computers recently tested by PC World.
Windows XP Testing
|Test System||Processor||WorldBench 5 Score||Mulititasking Test||Windows Media Encoder 9||Roxio VideoWave||Adobe Photoshop 7.0.1||Microsoft Office 2002 SP-2|
|Apple iMac||2.0GHz Core Duo T2500||96||423||294||267||320||541|
|Apple MacBook Pro||2.16GHz Core Duo T2600||98||419||280||259||305||535|
|Dell Inspiron E1705 (portable)||2.0GHz Core Duo T2500||97||498||305||271||325||549|
|HP Compaq nx9420 (portable)||2.16GHz Core Duo T2600||101||444||279||259||311||575|
|HP Pavilion a1250n Media Center Desktop PC||2.0GHz Athlon 64 X2 3800+||94||521||321||290||367||563|
Best results in bold. All individual test results in seconds.
As you can see, the Macs running Windows gave these PCs a run for their money, with the 2.16GHz MacBook Pro turning in the fastest scores on three of the five individual tests. The build-to-order MacBook configuration also tied the 2.16GHz HP Compaq in the sixth test, involving Roxio VideoWave.
Noticeably absent from the table is the Mac mini Core Duo which had trouble completing the test suite (the multitasking test in particular). Also, one of the 12 applications in the WorldBench 5 suite—3D Studio Max—won’t run on a Boot Camp system; luckily, you can still get valid WorldBench scores without the two tests which use that app. The problem isn’t a matter of horsepower, but one of copy protection. Hopefully, that issue will resolved soon.
We’ll keep trying to get the Mac mini to run WorldBench, and we’ll report those scores when we do. We’ve also started running the tests on our Intel Macs running Windows via Parallel’s new virtualization software. So check back soon for more results.