Geek Patrol has done a benchmarking comparison between the Powerbook G4 and MacBook Pro using their own tool, Geekbench. The results are in line with Apple’s marketing numbers (3-4 times as fast), making me question the validity of the benchmark and how it’s written.
In particular, I have to wonder if the tests are particularly memory heavy, in which case the MacBook Pro would have a ridiculous advantage. Comparing 167 MHz to 667 MHz just isn’t fair and that was always my concern about PowerPCs until the G5s came out. It also makes me wonder how those kinds of comparisons would shake out. Like I always say, a benchmark is only as good as its reflection of actual usage.
The CPU benchmarks (where the MacBook Pro showed a 2-4x improvement over the PowerBook G4) aren't constrained by memory bandwidth; none of them are able to saturate the memory bus.
In other words, Intel Core Duo doesn't have an advantage thanks to having a better memory bus, it has an advantage because it's a better CPU.
Thanks for weighing in John. Any chance that you'll be releasing the source?
We're not sure how we'll license Geekbench once it's complete, but for the time being we don't have any plans to release the source code.