Grand overview
Ah, the time has come at last... Intel's long-awaited Response.
So, based on IDF comments, AMD owns the high-end desktop well into 2005. I don't think anyone can say they didn't use their one-year head start in server chips about as well as it could have been done. If you compare Opteron to Athlon-MP, there's a world of difference in mindshare, buzz, OEM acceptance, etc, etc. The position that Opteron has carved out for itself is strong enough that it will not be displaced *simply* because of the existence of an Intel clone. And that is quite an accomplishment for Ruiz in just one year.
Then there's AMDs technical advantage. Given the performance and power/heat issues with the p4 architecture, it looks like AMD will be able to compete for the forseeable future. Until Intel switches to Centrino-architecture and includes an on-chip memory controller, I see AMD having a nice advantage. The current p4 is so overwrought and unimpressive (and rare) that Elmer might call it unmanufacturable. The point is that there are very good reasons for OEMs to offer the Opteron, despite the attack of the Intel clones.
On the whole, it looks pretty good for AMD. Not as good as if Intel hadn't entered the ring, but better than any alternative I can think of.
The question is how can they maintain their superior position in the future? That will be a good fight, no doubt. I only hope, though, that AMD is being very attentive to its partners and customers in the design of k9. In particular, AMD might take advantage of Intel skittishness in regards to Itanium. E.g. enterprise-class cache coherency (which is supposed to be in K9)is something that Intel would not incorporate into Xeon, ever, so long as Itanium exists. Also, I think that there are alot of possible fruits of the AMD-IBM relationship that have yet to germinate (if I have the botany right).
