Whatever slim chance Itanium had of a meaningful future, appears to be over

[CHADBOGA]

http://www.intel.com/content/www/us/en/processors/itanium/itanium-kittson-update.html

Intel has updated the definition of the next generation Itanium® processor, code name “Kittson”. Kittson will be manufactured on Intel’s 32-nm process technology and will be socket compatible with the existing Intel Itanium 9300/9500 platforms, providing customers with performance improvements, investment protection, and a seamless upgrade path for existing systems. The modular development model, which converges on a common Intel® Xeon®/Intel Itanium socket and motherboard, will be evaluated for future implementation opportunities.
​

So it doesn't look like HP is prepared to pay Intel to develop a 22nm successor to Poulson.

It looks like Kittson will be to Poulson what Montvale was to Montecito, a meaningless new name for the same chip and a pitiful frequency increase.

Where this leaves HP's Itanium based customers in the future will be an interesting fallout to observe.

Presumably HP will want to move them to Xeon based systems, but I wonder how many customers they will lose in the process.

All this will probably come as no great surprise to anyone who has been following Itanium's rocky path for the last few years.

Can't help but think that the encroachment by ARM, simply meant that Intel decided to focus all its efforts on x86, and maybe this was an inevitable consequence anyway, perhaps sped up by 5 or so years.

I wonder when the official announcement of the cessation of Itanium based systems from HP will be made, and if this will have any outcome in the legal wranglings between HP and Oracle.

 

[ShintaiDK]

Do you know how far back EX is on x86? Today they still use Westmere-EX for example.

This is where reliability is more important than speed.

 

[Phynaz]

Itanium is a low volume part. The costs of moving it to a bleeding edge manufacturing processes wouldn't pay off.

 

[NTMBK]

Itanium is a low volume part. The costs of moving it to a bleeding edge manufacturing processes wouldn't pay off.

22nm products have been out for almost a year now, and this is talking about a product which is still a long way in the future- 22nm won't exactly be bleeding edge by that point

 

[Nemesis 1]

The modular development model, which converges on a common Intel® Xeon®/Intel Itanium socket and motherboard

Chad I didn't read the article. But what a quoted . I not sure what your saying .

 

[dave_the_nerd]

22nm products have been out for almost a year now, and this is talking about a product which is still a long way in the future- 22nm won't exactly be bleeding edge by that point

They still fab their chipsets on 65nm. :whiste:

 

[CHADBOGA]

Chad I didn't read the article. But what a quoted . I not sure what your saying .

I just copied that from Intel's site.

There is no article as such, just that notice which I copied and pasted in full, in my opening post.

 

[Arachnotronic]

Forget Itanium. Xeon E7 will have all the RAS features of Itanium soon, so spending the $ to do all that work on a slower, inferior platform is...silly.

 

[ShintaiDK]

Forget Itanium. Xeon E7 will have all the RAS features of Itanium soon, so spending the $ to do all that work on a slower, inferior platform is...silly.

Soon? Its not with IB-EX.

 

[pm]

Chad I didn't read the article. But what a quoted . I not sure what your saying .

The original design of Kittson was that it would be socket compatible with Xeon motherboards. So you could buy a Xeon motherboard and then either socket an Itanium or a Xeon part into it depending the usage model and workload. They are saying that this feature has been dropped and will be evaluated for future designs... in other words, we were doing it, we aren't now but we might in the future.

* As always, I'm not an Intel spokesperson *

 

[Edrick]

22nm products have been out for almost a year now, and this is talking about a product which is still a long way in the future- 22nm won't exactly be bleeding edge by that point

In the Big Iron world, that is bleeding edge.

 

[Acanthus]

They still fab their chipsets on 65nm. :whiste:

I've always wondered why that didn't hurt them back in the Nvidia / VIA / ALI days. I would imagine the competitors didn't have access to the smaller processes and Intel just always stayed a process node behind on their own fabs.

 

[Genx87]

Itanium meanginful future was killed by Intels Xeon line.

 

[Charles Kozierok]

Itanium is almost like Intel being HP's fab.

More than that, of course, but you know what I mean.

 

[Idontcare]

Itanium is almost like Intel being HP's fab.

More than that, of course, but you know what I mean.

Foundry, and in more than one way. Itanium is essentially an HP product that which is designed externally (so Intel is the IC Design Foundry there) as well as manufactured externally (Intel again).

AMD wants the same model, to be a design house foundry. That is their vision with the ARM license.

Makes sense too, not everyone wants to maintain a full-time fully staffed 2,000+ IC design team year around. Go to them and have them design an IC that meets your needs, and have them deal with getting it fabbed at TSMC or GloFo for you as well.

It is a model that has opportunity for lowering the barrier for new and small product designers. Not everyone can be an Apple that can write a check to buy a PA Semi so they can design their own custom ARM-based IC.

Maybe Intel will get into the business too. They have the experience.

 

[pyjujiop]

The sooner it dies, the better. Itanium killed off or marginalized every other non-x86 architecture out there as soon as it was announced. God only knows what we would have if a fraction of that R&D money that went into Itanium had gone into further development of the Alpha instead.

 

[Idontcare]

The sooner it dies, the better. Itanium killed off or marginalized every other non-x86 architecture out there as soon as it was announced. God only knows what we would have if a fraction of that R&D money that went into Itanium had gone into further development of the Alpha instead.

I don't get the vitriol against Itanium. DEC was dead before Itanium was released. As for the architectures that were swept aside, survival of the fittest.

Itanium didn't kill them, the people who made the decisions to deprecate those architectures are the one's who killed them. Take it up with them if you think they chose poorly.

 

[Nemesis 1]

The original design of Kittson was that it would be socket compatible with Xeon motherboards. So you could buy a Xeon motherboard and then either socket an Itanium or a Xeon part into it depending the usage model and workload. They are saying that this feature has been dropped and will be evaluated for future designs... in other words, we were doing it, we aren't now but we might in the future.

* As always, I'm not an Intel spokesperson *

Thanks on 2 counts . The reply and understanding where i was at with this .