Re: It seems that by your logic that you keep repeating daily, Penryn would be far ahead of any K10-based CPU.
It sounds unlikely, but Elmer's logic does have merit. Most recently, AMD guided towards a late-summer Barcelona launch, which would put it in the August/September time frame (mid to late-Q3). If AMD hits this, they should beat out Penryn, which Intel has most recently pulled into Q4 for some segments (Harpertown, for example, and maybe Yorkfield).
On the other hand, if you look at the way AMD has slowly pushed out the ATI R600 launch the closer we get to the originally proposed dates, I wouldn't be surprised if the same thing happens to Barcelona. AMD has basically taken the best case schedule for a very new and risky design, and proclaimed that date to the world. Most people would have looked at Barcelona's Sept 2006 tape out (me, for example), and concluded that it wouldn't likely ship until Sept 2007 at the earliest. Well, far be it from me to toot my own predictions.... <g>
But still, I think it's appropriate to treat AMD's schedule guidance with a grain of salt. If Barcelona does end up slipping by a few months into Q4, then Intel has a good chance of beating them with a 45nm launch. I don't expect Intel to be comprehensive in their 45nm launch (it will be a couple of segments, like I said), but neither will AMD's Barcelona launch be comprehensive (it will just be a quad core server and extreme edition desktop part, most likely). The interesting thing will be which one can ramp their new parts faster.
AMD's will be on a mature process, but it's a new architecture with large die size that will require several proliferations to ramp. On the other hand, Intel's will be on a new, unproven process with risky new features, but it will be using a tried and true micro-architecture with small die size and MCM technology instead of multiple proliferations.
