Instead of running Crysis or benchmark suites, Kingsley-Hughes decided to go a different route and perform 23 common tasks on each of the three Windows versions (all 32bit), and then simply mark the winner, the runner-up, and number three. He performed the same 23 tests on these Windows versions on two different machines: an AMD Phenom 9700 2.4GHz system fitted with an ATI Radeon 3850 and 4GB of RAM, and an Intel Pentium Dual Core E2200 2.2GHz fitted with an NVIDIA GeForce 8400 GS and 1GB of RAM. While I personally don't have a system nearly as fast as either of those, they are hardly top-of-the-line configurations.
The results are telling. On the AMD machine, the Windows 7 beta beat both XP and Vista SP1 on 21 out of the 23 tests. The results were the same for the Intel box: 21 out of 23 tests went to Windows 7. While Windows 7 tripped on the installation of Office 2007 and DVD burning on the AMD box, it tripped over two file moving operations on the Intel machine.
There are, as always, issues with these tests and the data they produced. First of all, there are no absolute figures for any of the 23 tests; we don't know if Windows 7 beat its predecessors by a huge margin, or only by a barely statistically significant one. In addition, I found the tests to be a little vague, as the author doesn't supply the files he used for his tests - in other words, we can't reproduce his exact steps.
Despite these obvious deficiencies, these results confirm the general feeling you're getting from those who have already tested Windows 7: it seems like Microsoft is delivering on its promise to make Windows 7 no more resource demanding than Windows Vista - in fact, it seems to have overshot its goal by making Windows 7 perform better than its illustrious predecessor, which is nothing short of a miracle in Windows land.
Remember, though, that your mileage may vary.