PT, Re: "This sounds like you see the IA-64 and X86-64 architectures in the same market? I have always thought of IA-64 more towards the big iron end of the spectrum (Power4 / Alpha) with high scalability?"
Effectively, they will be in different markets. Itanium 2 will be used almost exclusively by the OEMs (with almost full acceptance so far, and very little "whitebox" interest), while Opteron will have nearly all sales coming from the "whitebox" market (since zero OEMs have committed to the design at this point in time).
Therefore, I have a hard time seeing the two compete. But, like others have said, all it takes is a design win or two to open the door for Opteron, and AMD has expressed much interest in having the two platforms compete directly. Because of this, I think it will be necessary for Itanium 2 to show a compelling value proposition for customers to latch on to, or else the low cost Opteron solution may start sounding appealing.
I don't know how much of a market there is any more for fault tolerant systems. There is so much redundancy already, that failures are pretty rare, and even a low end system can "get by". If the high end market is indeed shrinking, then that's good news for AMD, since they will want to be there to offer their low cost alternatives.
Xeon will also be competing in the "low end", and I think it should have much continued success. Mostly, though, because it's so firmly rooted in the industry. Itanium 2 and Opteron will be starting nearly from scratch, so that's why I think it will come down to a race. Itanium 2 can offer all kinds of features, performance, and reliability, but if software is limited, customers will stick to the 32-bit solution, or with their RISC solutions (or in some cases, go with AMD). As AMD gains share, they will also gain credibility.
So I think it's still possible for Intel to snatch defeat from the jaws of victory. Vice-versa for AMD, but risk is "business as usual" for them.
wbmw |
