Features

The hard drives fitted as standard are only 5400rpm. This could be a real handicap for professional users, and while you have the option to upgrade to a 7200rpm drive (100GB versions with the 15in models for an extra £140 or £80, and down in capacity to a 100GB 7200rpm model for the 17in with no change in the price), the disk tests we carried out were really disappointing.

One last thing: the aluminium case sure looks fantastic, but because it conducts heat much more readily than the plastic shell of the MacBook, the MacBook Pros feel much hotter when running. Our testing regime obviously puts a lot of stress on the processor, and the area between the keyboard and the screen got so hot we couldn't actually touch it. However, the underside and the wrist rest became merely a little unpleasantly warm.

Which one should you go for? As we often find, it's the mid-range model that offers the best balance of features. It packs the same processor and graphics hardware as the 17in MacBook Pro, and is eminently more portable. Considerably better graphics performance compared with the MacBook range is complemented with some extra features such as the ambient light sensor and ability to power a 30in Cinema Display, but just make sure you really need all this before you go for the Pro range.

How we tested

MacUser's old benchmarking regime served us well, but the switch to Intel meant many of our tests were no longer appropriate. The new suite consists of five

ADVERTISEMENT

main tests: Xbench, Photoshop, iMovie, Quake and Cinebench.

We use Xbench (xbench.com) mainly to test hard disk performance; its other tests can be a bit abstract and aren't helpful in telling us how well a Mac will perform in the real world. That said, we do record its scores for CPU and memory benchmarking to give us some raw number-crunching figures.

Our Photoshop test consists of two Actions, each run three times, quitting and restarting between each, with the best time taken. The creative test includes applying complex blurs, making selections, using clipping paths and more. The second test simulates Photoshop being used as an image processor, creating a 300dpi, RGB contact sheet from a folder of 10 high-resolution CMYK images and one EPS file. Photoshop's preferences are set to use the 100% of available RAM. Remember, Photoshop is running under Rosetta.

The iMovie test creates a blank, default iMovie project and then times how long it takes to import 3GB of movie files from the hard disk - one DV stream, one AVI, and one 3GB clip from a cameraphone.

To test gaming, we use version 1.2 of Quake 4 (it's a Universal Binary) and set the video quality to high and the resolution to 1024 x 768 pixels. Finally, Cinebench tests mathematical ability and helps analyse how well the Macs cope with 3D.

All Macs are switched to 1024 x 768 pixels, Energy Saver preferences are set to Best Performance, and there are no optical media in the drives or attached peripherals. AirPort and Bluetooth are turned off. All Macs are tested in the configurations advertised by Apple. Our benchmark machine is a default-configuration 2.1GHz iMac G5 with iSight, the last G5-based iMac.

The last word

There's a phenomenon used in marketing to get people to do the job of a salesman to themselves: if you offer someone three different options and call them something like bronze, silver and gold, almost no one will pick bronze. We just don't like to think of ourselves as 'bronze' people, as a rule.