et's just say that the chips scale so well that we see an overal 15 % clock for clock advantage for phenom over penryn, that means that intel would need a 3.45 ghz penryn to match phenom.
Since we're doing Phenom vs. Penryn, not Barcelona vs. Clovertown, I'd say there is no chance in hell Phenom will beat, or even match Penryn at same clock speed. Did you not see the benchmarks of Barcelona vs. Dual Opteron desktop app benchmarks from Anand?? Purely looking at that alone, it looks like Phenom will be fair bit behind
Conroe(in single thread)/Kentsfield(in multi-thread), let alone Penryn.
Even worse, someone has pointed out that Quad FX setups are sometimes quite a bit slower than the Athlon X2 setups at the same clock. Barcelona showed 21% advantage in Oblivion against Dual Opteron. Look on the QX6850 review:
http://www.anandtech.com/cpuch...owdoc.aspx?i=3038&p=15
Athlon FX 3.0 is 17% faster than Quad FX 3.0. If same holds true for Dual Opteron as it did with Quad FX, then we would see Barcelona is only less than 4% faster than Opteron per clock. Tech report results show that in workstation apps, which Barcelona should be comparatively better, has lots of situations where its slower per clock than Kentsfield.
Since Penryn is 5-10% faster than Conroe, the deficit will be in AMD's side, meaning they would need 3.4GHz Phenom to equal 3.0GHz Penryn. Since Penryn is coming at 3.33GHz with EE, AMD would need 3.7-3.8. Not likely.
On Intel platforms where performance is more chipset bound than AMD's since Intel platforms don't use integrated memory controller, the performance difference between best and worst variest at most 5%, averaging 2-3%. Phenom not only has integrated IMC to negate much of the differences that variations of memory controllers have, it even has L3 cache and a superior IMC than Athlon X2, meaning it'll be even less sensitive in theory.