Overshadowed by all the iPhone hoopla was the release late last week of a software update for some other Apple hardware. The details of MacBook Pro Software Update 1.0 remain sketchy--"This update provides important bug fixes, and is recommended for all 2.2/2.4GHz MacBook Pro models," is all the release notes offer. But the update appears to address a performance issue we found while running benchmark tests on the just-released MacBook Pro models.
The MacBook Pro models released in June sport new nVidia GeForce 8600M GT graphics cards. Based on the tech specs, we expected the nVidia cards would deliver higher frame rates than the ATI graphics found in the previous versions of the MacBook Pro. And while this was the case in our tests with Doom and Quake, Unreal Tournament scores were disappointing.
The test we run on all systems is the 1,024-by-768 Antalus Botmatch test. The results for this particular test showed the older ATI-driven 2.33GHz MacBook Pro pushing through 80 frames per second; the best-performing 2.4GHz MacBook Pro posted a score of 58 frames per second -- that's 28 percent fewer frames than the older, supposedly slower model. The same test with higher resolution settings showed the newer models with a slight advantage, maintaining their frame rates in the more demanding test while the 2.33GHz Mac Book Pro's frame rates dropped to just below the 2.4GHz's speeds.
After updating the new MacBook Pros with Apple's software update, we found that our new MacBook Pro models, while still a touch slower than the older systems at the 1,024-x-768 Unreal Tournament test, made up a lot of ground, with the 17-inch 2.4GHz model able to display 26 percent more frames per second than it did before the update. The updated systems were also faster in the higher-resolution Unreal Tournament test, increasing the 2.4GHz models' modest lead to a more impressive 20 per cent.
Speedmark 4.5 scores are relative to those of a 1.25GHz Mac mini, which is assigned a score of 100. All systems were running Mac OS X 10.4.9 with 2GB of RAM. We used Unreal Tournament 2004’s Antalus and Asbestos Botmatch average-frames-per-second scores; we tested at resolutions of 1024-x-768 pixels as well as 1440-x-900 pixels at the Maximum setting with both audio and graphics enabled. Doom 3 was set to use high video setting, 1024-x-768 resolution, V-sync No, Antialiasing Off and all other settings set to Yes. Quake 4 was set to 1024 x 768 resolution, high quality video, full screen, multiprocessor on , V-sync No, antialias Off and all other advanced settings set to Yes.
The 1,024-x-768 Unreal Tournament test is included in our Speedmark test suite, so the slow frame rates were holding down the overall score of the new MacBook Pros. With last week's software update, the significant improvement in the Unreal Tournament score also increased the Speedmark score.
The three new MacBook Pro models received a three to five point boost in Speedmark scores, pushing the new 17-inch MBP to a 6-percent faster Speedmark score than the old top of the line 17-inch 2.33GHz model--that's up from a 4-percent margin. Quake and Doom scores didn't change much, up or down a frame or two. We also spot checked the rest of the Speedmark tasks to see if the update improved or broke anything else, but those tasks appear to be unaffected by this update.
Though this type of mystery pops up quit often in my line of work, I don't like when systems don't behave the way they should. I like it less when I'm the only person seeing it. Now that this little problem has been ironed out, I can move onto the next.