Allies pledge $10 billion to boost Itanium

SAN FRANCISCO--Intel, Hewlett-Packard and seven other server companies will spend $10 billion through 2010 to try to increase adoption of the Itanium processor.

The money is coming from Intel and HP--Itanium's co-developers and top backers--as well as from Unisys, Silicon Graphics Inc., NEC, Hitachi, Fujitsu, Fujitsu-Siemens and Groupe Bull. The companies said Thursday that they will spend the money on research and development, marketing, and work to help software companies support the high-end processor.

"Itanium has been taking share from both IBM power and Sun Sparc. We're on the right trajectory, but we want to go faster," Tom Kilroy, general manager of Intel's Digital Enterprise Group, said at a press event here. "The $10 billion investment is a statement that we want to accelerate as a unified body."

The companies are members of the Itanium Solutions Alliance, which includes Microsoft, Red Hat, Novell, Oracle and other software companies. The alliance also announced its Itanium Solutions Catalog, which describes various combinations of hardware and software for specific tasks.

Itanium, a high-end processor, was once expected to sweep the server world. But because of delays, initial performance issues and software incompatibilities, Intel has had major difficulties getting Itanium to catch on. Most recently, the first dual-core model, code-named Montecito, was delayed from 2005 until mid-2006.

Kilroy made it clear why Intel is pushing so hard to give Itanium a stronger future.

"This is a $140 billion opportunity on hardware. It's dwarfed by the opportunity in software and services on top of that," Kilroy said. "There's a reason there's $10 billion of investment in play."

SGI, a relatively small company that concentrates on the high-end technical computing systems, expects to benefit from the strength-in-numbers argument behind the alliance.

"The Itanium Solutions Alliance is useful for us," said Greg Estes, vice president of global marketing, in an interview. "If you're a company like ours, you try to stake out ground to be the high-performance player in the marketplace. If you can do that with a robust ecosystem underneath that, that doesn't feel esoteric--high-performance systems with something that feels like it has some weight on it so you're not just betting on SGI alone--that's a better story for customers."

[irisfailsafe]

Is it really worth it?

I mean, many ultra high end apps (such as pro engineer) stop
the support of the processor. The thing is the Itanium has many
design flaws that would require a major redesign. Only custom
made apps seem to be using and I don't know if that market is
big enough

[jonnie savell]

Itanium is echhh.

quote:
"Itanium has been taking share from both IBM power and Sun Sparc. We're on the right trajectory, but we want to go faster," Tom Kilroy, general manager of Intel's Digital Enterprise Group, said at a press event here.

observation:
Insofar as at least one Itanium CPU was purchased, this statement is true. But, because not many of them were purchased, this statement sounds silly.

Itanium is echhh.

[zkysr]

Tom Kilroy has increased crack cocaines market share!

Itanium has been forced down the throat of HP customers and has been taking market share from PA-RISC, MIPS, and Alpha, all of which are going away. And Opteron has been taking more of that market share than Intel would like to admit, in addition to taking Xeon market share. Not even Dell who is Intels wife supports this garbage.

If I introduced a new chip today and sold one system to an idiot, I could claim I was taking market share from everyone.

According to IDC, the quarter-by-quarter unit volume for Itanium servers since the fourth quarter of 2003 goes as follows:

* Q403 7616
* Q104 8678
* Q204 8085
* Q304 8235
* Q404 8996
* Q105 8127
* Q205 8500
* Q305 8596

IBM has surged from similar Power volumes in 2001 to shipping more than 100,000 servers per year, and Sun is up to more than 300,000 SPARC servers shipped per year.

[libertyforall1776]

Leave the itanic sunk

Why not just leave the itanic sunk...

There is a reason Apple did not choose it for their Mac OS X
transition... Why sink more $$$ into this tremendous flop?

See: http://www.theregister.co.uk/2004/02/17/who_sank_itanic/

[JTowner]

Real reason - Intel can't win against AMD

So Intel just admitted that it can't compete with AMD in the high end market. AMD gave us x86-64 and with that we can achieve the higher performance we want, without the additional Itanium costs.

[Intelrules]

: Real reason - Intel can't win against AMD

What are you smoking?

[fred dunn]

$10 Billion..Yeah Right. If you believe that...

Then I have a bridge I want to sell you. Just think about this. The Itanium processor is already 4 to 5 years old and the "Alliance" would have us believ that this processor class is still going to be around in 2010, ridiculous.
This is just a ploy by interested companies to "send a signal" to the IT community that the Itanium processor is alive and well and will be for many years, this too is ridiculous.

Sounds like these companies are making a last ditch stand to keep Itanium alive but remember this:
When you buy products or services from any of the companies within the Alliance you will be paying a part of that $10B. So then the smart thing would be to steer clear of all companies investing in Itanium so you don't pay the "Itanium Premium".

Fred Dunn

So who's going to pay the $10B?

You mister customer! Thats whos going to bankroll this science
project. With the dismal flatlining volumes of Itanium, and Opteron,
SPARC and Power outperforming Itanium which idiots will purchase
Itanium regardless of how much money is thrown at it? As an Intel
investor, I am frustrated at the billions they have spent on this lost
cause, with no ROI to speak of! Why not return this investment to
the customers in form of a rebate for wasting our times even
talking about this. My vote is cut the cord and take it off life
support.

[Seaspray0]

Cost vs Performance

It's not that the Itanium chip is a bad performer - and it's actually the Itanium 2 chip and not the orginal Itanium - but when you look at the bang for the buck, it doesn't make sense. The Itanium series is a 64 bit chip that will also run 32 bit code. The original Itanium was not a good performer when it came to running 32 bit code, hence the Itanium 2. The Itanium 2 does a better job running 32 bit code but still not giving a bang compared against a traditional 32 bit dual core processor. It does best when running straight 64 bit code.

So why is 32 bit processing so important for a 64 bit chip? It's because not all the applications out there are yet designed for 64 bit and it will take years before the need to do 32 bit processing is finished. If you look back at the transition from 16 bit to 32 bit software/hardware you can make a comparison of how things are expected to go today. It has been the Itanium's poor performance on 32 bit code that has hampered it all these years.

With the software side now catching up to the hardware, a pure 64 bit environment is now more feasable and the Itanium is now a more promissing choice but recent advancements by AMD has provided a serious threat. While not as established in 64 bit processing, AMD is making inroads quickly due to their lower cost for performance. Their timing was perfect to jump into 64 bit processing before Itanium became the default choice.

As of today, the developers have dumped billions of dollars into the Itanium series and have almost no return on their investment. HP and Intel are the biggest losers and you can be sure the 10 billion more is a drop in the bucket compared to what they've lost and possibly their only chance to get anything back before the chip gets pushed aside which is something they've tried to avoid at all costs. They know they're taking a beating on this one so I really don't expect them to try and recoup their 10 billion from increasing the price. I would more expect them to drop the price to a very low profit margin in the attempt for market share.

[freq]

I bet they'll use Rambus....

its called recooperating