Joe, Re: "This is my point. The biggest competitor of Itanium is and will be Xeon. And Xeon can't cede the top end to Itanium, because it is facing competition from Athlon. If there were no AMD and no Hammer, Intel could just limit the resources devoted to Xeon (server chipsets - # of CPUs, bandwidth etc.) just to leave the high end to Itanium. But because of AMD, Intel is forced to make the Xeon line as good as it can possibly be, to keep AMD away."
I recognize your point, Joe. I agree that this is exactly what has happened. I have always been a fan of IA-64, but only recently have I caught a glimpse of performance, and I have to say that I am impressed. Itanium 2 is able to achieve almost 30% performance over the fastest .18u generation Xeon MP - with Hyperthreading. And this is in TPC, a very intensive test involving the CPU, memory, and I/O. That Itanium 2 is able to achieve such an increase in performance (especially in a benchmark that scales rather poorly with the CPU) speaks volumes about the IA-64 architecture, its scalability, and its prospects for the future (think IA-64 with Hyperthreading...). Corporations are more than willing to spend proportionally more money for more performance, especially in the transaction processing market. Itanium 2 seems to offer quite a bit more to the table, with a 4 processor score close to 70K TpmC. This kind of performance should give Itanium 2 the credibility of a "high end" CPU, and then it will be up to Intel to migrate the industry towards 64-bit software.
The RISC market isn't going away overnight. For the first couple years, Intel can get Itanium 2 entrenched in this market with solid infrastructure, before being forced to move it down to higher volume markets. I think that's the way that Intel will have to handle it, if they want it to be successful. It's gone from a downhill battle to more of an uphill one, but I wouldn't count it out of the race, yet.
wbmw |
