Never an exact science...but XBENCH does not show any improvement at all...
Am i crazy in suggesting that XBench would actually be expected not to show any improvement?
One side effect of this utility is that it causes the Mac clock to speed up. The clock probably increases in speed proportionately with the increase in proc speed. If XBench relies on the Mac clock to calculate results, those results will remain the same as one of the constants used in those calculations has now become a variable. Calculation screwed!!!