By Katie Marsal

Published: 10:35 AM EST

Performance: objective tests

Readers are encouraged to read our detailed processor comparison to learn all of what's new in Apple's new NVIDIA-based architecture and an explanation of what determines performance, though there are key differences between the Pro and the new 13" models.

Curiously, Apple has willingly taken a step back in clock speed for the sake of technology. Spending $1,299 now only gets you a 2GHz processor -- and it now costs $1,599 for the 2.4GHz that once occupied the $1,299 price point. It's not quite a downgrade: both chips now have a faster 1067MHz system bus and consume just 25W of thermal peak power versus 35W for the older chips. That translates to cooler and more energy-efficient processors as well as a slight gain in performance per clock.

NVIDIA's GeForce 9400M hardware also has a much bigger role to play in the lower-cost MacBook. The MacBook Pro simply uses it to save power; in the MacBook, it replaces Intel's aging GMA X3100 integrated graphics and is claimed to boost performance by roughly 5X in games and other graphics-intensive apps. It also has the side benefit of reducing the footprint of the MacBook's internal components by consolidating the integrated graphics, memory controller and peripheral bus into one monolithic chip.

This two steps forward, one step back approach bears out in synthetic tests as well as less formal, hands-on experience. Testing using Primate Labs' Geekbench, which chiefly focuses on processor performance, shows a slight boost in performance at the same 2.4GHz clock speed in all categories of processor performance, but only a miniscule one. We don't have a 2GHz model to test but expect similar results versus the 2.1GHz chip that now occupies Apple's $999, previous-generation machine. We've also included the new 2.53GHz MacBook Pro with 4GB of RAM from our MacBook Pro review to compare performance with Apple's new flagship.



If we were to base performance solely on these benchmarks, the new MacBooks would almost be considered a failure: users are actually paying between $200 to $300 more for roughly the same processor.

Things start to change, however, when you toss hardware 3D speed into the equation. A test with Maxon's CINEBENCH R10 shows that the new MacBook is actually slightly slower than the previous model in pure CPU (software-based) rendering but utterly decimates the older system when it leans on its 9400M graphics for rendering a scene using the OpenGL video standard. The difference is so stark that it puts the 13" MacBook within spitting distance of the MacBook Pro, which has the benefits of a much faster GeForce 9600M GT, more memory and a faster core.



With speeds like these, the MacBook now effectively has mid-range graphics in a thin-and-light system. We wouldn't go so far as to call a $1,299 or even $1,599 system low-end, but it now saves at least $400 for those who want the headroom for heavy-duty production in a small Mac. Color questions aside, we wouldn't be surprised if Apple opened the doors to officially supporting Final Cut Studio on a MacBook for the first time.

Performance: games and general video performance

At the same time as we've quantified the MacBook's performance on the professional end, we've also run the system through its paces and monitored frame rates during play for some games as a way to test Apple's performance claims.

On the earlier MacBook, gaming for all but 2D or particularly old titles is virtually impossible. In Quake 4, frame rates are so low as to be extremely choppy at 1024x768. An older-still game like Unreal Tournament 2004 does play within a few frames per second of a TV-like 30 frames per second at the same resolution, but it also has much less detail. And some games are entirely beyond the GMA X3100's scope; we wouldn't want to play truly modern games like Call of Duty 4 on Intel's stock hardware at all.

By contrast, the GeForce 9400M is an out-and-out powerhouse. Quake 4 does stutter at points -- particularly those with several characters onscreen in complex indoor environments -- but holds much closer and sometimes surpasses the 30 frames per second mark with medium-level detail that would crush the Intel video. Unreal Tournament 2004 also runs much more quickly; we never once saw the frame rate dip below 30 frames per second even at the 13" screen's maximum 1280x800, and in-play speed was often significantly higher.

Excluding gameplay, the video speed is also noticeably improved just in general tasks. The previous-generation MacBook would frequently drop a few frames during Dashboard or Expose transitions, especially in heavy use. On the aluminum system's 9400M, those drops never occur. We can also vouch for reduced CPU usage during QuickTime movies thanks to the 9400M's hardware acceleration. When playing a 1080p trailer from Apple's website, processor usage never climbed above 28.6 percent in Activity Monitor where it would at times climb much higher on the earlier MacBook, often hovering between 70 and 90 percent.