Windows Server 2008 R2 to Phase Out Itanium

[Katanai]

I'm purely speculating, of course!

Of course you are, like most people do. I always have to fight against this misconception that Itanium as a whole is some sort of failure. Far from it my friend. You, for example, are probably interacting with an Itanium server every day without even knowing it. You know, a big part of telecommunications servers are Itanium based, so when you make a call on your mobile phone there's a very big chance that it's routed through an Itanium system. Or when you get money out of an ATM, there's a chance that the main server of that bank is Itanium based. Even now when you posted this message, chances are that a node somewhere was Itanium. I configure and help sell servers every day and I can tell you that, yes x86 sells more but the Itanium architecture is far far from being a failure and sells on a constant basis.

[Mats]

Of course you are, like most people do. I always have to fight against this misconception that Itanium as a whole is some sort of failure. Far from it my friend. You, for example, are probably interacting with an Itanium server every day without even knowing it. You know, a big part of telecommunications servers are Itanium based, so when you make a call on your mobile phone there's a very big chance that it's routed through an Itanium system. Or when you get money out of an ATM, there's a chance that the main server of that bank is Itanium based. Even now when you posted this message, chances are that a node somewhere was Itanium. I configure and help sell servers every day and I can tell you that, yes x86 sells more but the Itanium architecture is far far from being a failure and sells on a constant basis.

Did you even read my post???? I never said it was a failure, you made that up.

My only point was that when IA-64 showed up, the plans for it regarding what it would be used for was much wider,
not only for what you describe, and that it's supporting Windows is a relic from that time.
Intel even planned to use IA-64 instead of x86-64.

It's not really much speculation from my side anyway, here's what Craig Barrett said about it:

At the time, the best computer architects from Hewlett-Packard and Intel got together to create the next-generation big iron architecture, the Itanium family. The thought was it would cascade down to low-end computers because it was believed the 32-bit x86 architecture would run out of gas. You remember we went through the PowerPC alliance against Intel. And x86 kept growing upwards. We beat back PowerPC. We beat back the ACE consortium. We fooled ourselves a little bit because x86 beat back the Itanium challenge. The x86 Xeon servers and Opteron grew up faster into the server space than anyone thought they would. Itanium is profitable today. It’s growing market share in big iron architecture. But it certainly didn’t move down into the PC market as we had anticipated in the early 1990s. It made life difficult for Sun Microsystems. It did replace Precision Architecture at HP. So it is the big iron architecture, still battling with IBM at the very high end.

[Olivon]

Intel: "Itanium Will Survive Even Without Microsoft’s Support" - X-Bit Labs

Windows represents less than 6% of current Itanium sales according to IDC's Q3 2009 server tracker report. Most Itanium users run Unix, specifically HP-UX

At present Intel’s most powerful Itanium processor is model 9300 previously known as Tukwila that has four cores and is made using relatively old 65nm fabrication process. It is projected that in 2012 the world’s largest maker of chips will release 32nm eight-core IA64 chip code-named Poulson, which will also feature micro-architectural enhancements, including new instructions. In 2014 Intel plans to release code-named Kittson chip. Unfortunately, Intel does not make official claims about the future of Itanium at this point.