LOL, any defense of that particular benchmark article should pretty much prove who signs your paycheck.
People have been OCing Phenoms to 3.0 - 3.5 Ghz, why can't we have an OCed Phenom comparison? and Penryn was NOT just a die shrink, there were plenty of differences, they just did not amount to any substantial performance gains.
Intel always said that AMDs features (Hypertransport, etc...) didn't amount to any significant advantage over Intels dated FSB platform architecture, what if they were right? AMDs IMC may have been a limiting factor in their inability to clock higher, the IMC takes up a lot of silicon, and typically will be running at full speed during heavy loads(re: high power consumption/heat dissipation), it's entirely possible that a die-shrink will allow them to catch up this go round.