ZDNet details Itanium struggles
ZDNet has published an excellent article on the history of the Itanium, beginning with HP's 1988 decision to develop a next-generation 64-bit chip, and ending with present-day observations and trends. Those of you who regularly follow Itanium news won't find any bombshells or new information, but if you're curious about Itanium, this is a good place to start.
One area the article covers particularly well is how Intel's failure to deliver what it promised in Itanium has impacted other companies. SGI, in particular, appears to have gotten the shaft—the company initially planned to abandon its MIPS-based processors and Irix OS in favor of Itanium and Linux. Not only did Intel's Merced problems force SGI to extend its CPU design another two generations, but the recently-announced Montecito delay has forced SGI to revamp its roadmaps and substitute Madison 9M as a solution. SGI isn't the only company affected by Itanium's problems; HP and Unisys have had to delay their own product deployment schedules, as well.
Intel itself has expressed nothing but stalwart support for the troubled IA64 design, and it's still possible that future generations of Itanium will improve the CPU's troubled legacy. At the moment, however, Itanium is confined to certain niche markets, and seems unlikely to break free from them in the near future. The promise of what Itanium seemed to offer (combined with the merger between HP and Compaq) was enough to drive Alpha and SGI's MIPS design from the market, and lead HP to gradually discontinue PA-RISC. The reality of Itanium performance and technological advancement, however, has utterly failed to live up to the vision Intel articulated prior to the architecture's launch, and may ultimately leave IA-64 pinned between the IBM POWER series, AMD's Opteron, and of course, Intel's own 64-bit Xeon.