Well, maybe I am jumping the gun too fast, so forgive me if I do it. I honestly think AMD is playing dumb about the hammer, letting some specs go out but keeping the inners a secret. Maybe I am watching too much spy shows on TV, but there are things that make me scratch my head.
1) If you have read the patents rankings, AMD has generated more than intel for the last 3 years.... in 2000 and 2001 AMD generated more than 1000 patents each year. It was said also that 80% of those patents applied to the K8. Some of those patents were about dual cores and stuff like that. I honestly don't believe that a design wich has generated over 1000 patents will be just a souped up Athlon with SSE2, integrated memory controller and hypertransport.
2) When Compaq sold Alpha to Intel, AMD didn't really care. Some years ago, when DEC was sold AMD hired almost complete departments. That is how Dirk Meyer and Jim Keller (K7 & K8 architects) came to AMD. I think that if AMD this time didn't have interest in the Alpha guys is because AMd is really trusting its own cPU that could compete with the alphas.... that is the sledgehammer.
3) Some preeliminary benchmark of the clawhamer appeared on a german site. It showed the 800 MHz K8 beating a 1.6 P4 willamette in Q3A (the favorite benchmark of intel lovers). Despite those results, you shoud recall that the AMD management expressed disapointment for the "poor results of the hammer"... I wouldn;t consider that as poor performance. That means they expect much more of the chip.
4) Even if the clawhammer is as has been rumored, just an SSE2 AtThlon with PC2700 controller and hypertransport, that would be a killer budget machine. Flame me if you want, but with the glueless MP capabilities of the hammer thanks to the HT tunnels, any HIGH END machine will be a dual clawhammer, or even better, sledge hammer. A lot has been rumored about prescott and granite bay, prescott is still in the makes, and granite bay will be aimed at the high end.... If I were AMD, kill a dual-channel DDR competitor (granite bay) with a dual clawhammer configuration.
5) Granted, Intel has an edge in manufacturing and process, but I don't think that the clawhammer will debut at 1.6 GHz because the process doesn't allow it, it will debut at 1. 6 GHz because that will plenty to beat the competition. SOI and pure isotopic Si are also in the plans.
6) Related to #2, the lack of interest in keeping the Athlon as the best x86 CPU. Back in February, we saw the XP 2100+ ramping up pretty quickly after the XP 2000+. I thought a Palomino 2300+ was in the plans. Obviosly, AMD didn't think about it. The 166 MHz FSb is another one, a 166 MHz FSB xp 2200+ in a good Kt333 keeps up the pace with the P4 2.53.... but AMND says it doesn't matter.
As I said, maybe I am being too optimistic. The patents are the ones that really have been bugging me. Flame me if you want.