After all the benchmarking of the Penryn, you can say the following:
1) Running stock, a Penryn is a little better than the equivalent C2D.
2) Overclocked, a Penryn is somewhat better than the equivalent C2D.
3) The only real performance bump comes from SSE4, but given all the enthusiasm the software writers have shown for 64-bit programming and multithreading, don’t hold your breath waiting for it to be used a lot.
4) A 3.0GHz Phehom wouldn’t look too bad against 3.0GHz Penryns for most things. It’s going to lose most of the time, but not by too much.
5) It’s a shame AMD can’t make 3GHz Phenoms. It looks like people will be able to overclock an early one to 3GHz, but then, you’ll be able to overclock an early Penryn to 4GHz without much muss or fuss.
6) Maybe AMD can make 3GHz CPUs by this summer, maybe you can OC those to 3.5GHz, but by then, the planned “speed bumped” Penryns will probably be doing 4.5GHz or a bit more than that OCd.
7) Intel will continue to own the OC- and the high-end market for at least another year, and AMD will continue to own the low-end for the same period, unless they run out of money doing so.
8) What’s AMD going to do in the meantime? Looks like their marketing will consist of selling Phenom along with a new mobo and video card and post gaming benchmarks against C2Ds and video cards they can beat. This will be the PenWho? campaign.
9) I know this won’t be popular, but there sure isn’t too software out there today takes much or any advantage of four cores, if the benchmarks mean anything. If I had to build a new system where money mattered, I’m looking at two cores, not four.
10) Obviously I don’t think Penryn is a whoop-di-doo chip, but then, it wasn’t meant to be. Nehalem is supposed to be the whoop-di-doo chip. Penryn is just Intel’s way to get the needed new technologies down pat with a proven core, then they can try the whoop-di-doo stuff.