I can't seem to locate a downloadable SPECcpu2000 program (and I'm not willing to put a huge amount of time into it). A quick download of PCMark and the Dell shows an aggregate processor score of 4500, or about 60% that of the typical 3ghz P4 score of 7800. Scaling the
apple measured SPEC scores from
http://www.apple.com/g5processor/ by the same amount (which is crap, but easy), gives fp:400 and int:513, which in turn about compares to a 1ghz G5 on float or a 1.3ghz G5 on integer.
Using the spec.org numbers for the 3ghz P4 to scale gives fp:709 and int:664. These map to approximately a 1.65ghz G5 chip. To be fair, I'll use IBMs numbers for the 970, with fp:1051 and int:937 at 1.8ghz (with a decent compiler). Using these to compare, the PM compares to about a 1.38ghz G5.
Is this very scientific? Nah, not in the slightest. It doesn't include the debatably higher build quality of the Powerbooks (solid, but prone to defects such as warping), or the benefits (nicer, cleaner) of OSX ... or the drawbacks (specific applications, etc) of OSX. Its just some musings on performance.
Either way, anyone still thinking that Apple's laptops still rule the roost is sadly mistaken. Again, the numbers above are for the slowest Dell 600M currently offered, that I personally bought for $1300 (seen 'em cheaper). The newer 1.7ghz models are available and 30% faster. Don't confuse the PM with the P4M -- these are very different chips.
So, unless I've made a serious mistake, you can expect the PM and the G5 to compete pretty strongly on a clock-for-clock basis. The powerbooks will probably feel faster, especially with factors like QE offloading jobs to the GPU where XP does them with the CPU, but for serious business-application number-crunching, I don't expect to see much of a difference.