Oracle to stop all software development on the Itanium processor

[Idontcare]

Intel surely doesn't seem to mind to fan the flames of doubt regarding their views of the long term viability of Itanium.

http://www.xbitlabs.com/news/cpu/di...Reliability_and_Performance_than_Itanium.html

"We used to position Itanium up here: highest performance, highest reliability. In our ten years anniversary next year we are fully committed to the Itanium architecture. It is really now a choice of operating system, Xeon's reliability and performance is now equal and in some cases better than Itanium and they [processors] are going to leapfrog in performance over time. So, if you like HP-UX, OpenVMS, HP NonStop, other mainframe operating systems, we are fully going to support you on Itanium," said Kirk Skaugen, vice president of the Intel architecture group and general manager of Intel's data center group, during his keynote at the Intel Developer Forum (IDF) in Beijing, China.

The backpedaling has begin...looking more and more like Oracle was told/shown something concrete by top Intel execs as they claimed all along.

 

[jones377]

Either that or Intel has been forced to re-evaluate Itanium. Any guess what the rumored joint Intel/HP announcement on the 26th will be? A HP-UX port to x86 perhaps?

 

[VirtualLarry]

I wouldn't mind full-blown VMS for x86-64.

 

[Voo]

The backpedaling has begin...

That sounds less like backpedaling and more like fullblown reverse gear, or maybe I'm misunderstanding the quote. But that reads as "If you like itanium OSes like HP-UX don't worry we'll support you" and "well yeah x86 is way better than Itanium anyhow" in consecutive sentences.

 

[bryanW1995]

Uh oh; anybody working in the itanium dept might want to eye a transfer...

 

[taltamir]

That sounds less like backpedaling and more like fullblown reverse gear, or maybe I'm misunderstanding the quote. But that reads as "If you like itanium OSes like HP-UX don't worry we'll support you" and "well yeah x86 is way better than Itanium anyhow" in consecutive sentences.

Those sentences in no way contradict if you apply doublethink comrade.

 

[Idontcare]

That sounds less like backpedaling and more like fullblown reverse gear, or maybe I'm misunderstanding the quote. But that reads as "If you like itanium OSes like HP-UX don't worry we'll support you" and "well yeah x86 is way better than Itanium anyhow" in consecutive sentences.

I'll admit that it contained a lot less of the typical double-speak that tends to get interwoven into your standard "we deny we are doing xyz" when xyz is exactly what they are doing...so yeah, backpedaling may well be an understatement on my part.

It definitely speaks to a sentiment of "we do not foresee a viable growth strategy with this product, the ecosystem is shrinking, if it was 1910 and we were a huge supplier of horse-buggy whips then we'd be fools to stop selling them but at the same time we really think this horse-less carriage thing may well just eat our lunch so we are intending to transition to making rubber tires in the next 5 yrs...but for all you die-hard horse buggy owners out there, we got your back!".

 

[ydnas7]

there is no way Intel is going to admit anything is better than X86....
but Intel has always had some UBER post X86 processor in development from i432 to i860 and beyond. Now Itanium is a powerful and future oriented system, Intel investor's may doubt the Itanium strategy, but Intel needs a chip like itanium to bench X86 against and to develop. The lessons learnt on itanium are applied back to X86 (cache anyone) and the upcoming Itanium will have a very instructive take on SMT/CMT etc.

 

[Voo]

The lessons learnt on itanium are applied back to X86 (cache anyone) and the upcoming Itanium will have a very instructive take on SMT/CMT etc.

But there are many MUCH cheaper ways to research than to build a complete product, that can also be much more informative since you are much less constrained.

Also what was learnt from Itanium in regard to caches? Maybe I missed something essential there, but it seems more like they mostly integrated RAS features from Itanium back into their Xeons and there's absolutely no reason they can't integrate those into Xeons from the get-go.

@Idontcare: Is there anything in those quotes that denies the possibility of Intel helping HP port HP-UX to x86 and abandoning Itanium? Because that seems like an interesting approach (and in the long term much cheaper than developing an additional architecture), considering that Xeons could've feature parity with Itanium (at least RAS) in a few generations..

 

[Idontcare]

@Idontcare: Is there anything in those quotes that denies the possibility of Intel helping HP port HP-UX to x86 and abandoning Itanium? Because that seems like an interesting approach (and in the long term much cheaper than developing an additional architecture), considering that Xeons could've feature parity with Itanium (at least RAS) in a few generations..

Heh, Apple made the migration, and HP ported their OS once already (to itanium) so we know they've the tools to do it again if need be.

It'll be money in the end, both HP and Intel have plenty of it so that's not a limiting factor but profit margins will be.

I think it will be interesting to see what HP does with HP-UX and to see what Oracle does with Solaris hardware in the face of 10+ core x86 chips from both AMD and Intel.

The handwriting has been on the wall for big-iron since the downfall of Cray and then DEC, it's just going to take a decade longer than some people had anticipated.

I think of x86 as the automobile circa 1910. ARM is the motorcycle, big-iron is the steam locomotive.

Zoom forward in time and its obvious why the automobile is the high volume product. Motorcycles and locomotives have their place, they can never be fully displaced by road vehicles, but people like their cars.

And sure there are improvements in the steam locomotive since 1910, but even a nuclear powered 600mph bullet train is not going to displace the car as the preferred mode of transportation for the masses. Even in Japan the market TAM for automobiles vastly outstrips the TAM for bullet trains.

 

[podspi]

I think of x86 as the automobile circa 1910. ARM is the motorcycle, big-iron is the steam locomotive.

Zoom forward in time and its obvious why the automobile is the high volume product. Motorcycles and locomotives have their place, they can never be fully displaced by road vehicles, but people like their cars.

And sure there are improvements in the steam locomotive since 1910, but even a nuclear powered 600mph bullet train is not going to displace the car as the preferred mode of transportation for the masses. Even in Japan the market TAM for automobiles vastly outstrips the TAM for bullet trains.

So are the forthcoming high-end ARM chips something like these?

:biggrin: (Motorcycle with car-like chassis)

(More) on topic: I can't decide whether I am surprised or not by this announcement. On the one hand it was inevitable (they can't hold back x86 development lest they get overtaken by AMD and now ARM(!) ) but on the other hand it does sound to me like they are essentially acknowledging Itanium is no longer special in terms of performance or reliability.

Of course, they may want to keep Itanium going if only because if they force their partners to port their software to something else, they have no way of guaranteeing everyone will port their software to x86 :sneaky:

 

[Idontcare]

xbit has a blurb up about Kittson:

Intel Set to Detail Itanium “Kittson” Later This Month.

Intel Corp. and Hewlett Packard plan to reveal the future of Itanium platform and architecture on an event on the 26th of April, a source revealed. The event will be concentrated on the next-generation Itanium code-named Kittson.

http://www.xbitlabs.com/news/cpu/di...tail_Itanium_Kittson_in_Later_This_Month.html

For some reason this photo of Gelsinger inspires two thoughts in my mind:

1) Holy ESD, batman! I feel sorry for those chips being handled like that D:

2)

 

[Edrick]

Of course, they may want to keep Itanium going if only because if they force their partners to port their software to something else, they have no way of guaranteeing everyone will port their software to x86 :sneaky:

Where else would they port to? Once Itanium goes away, the only other big player in that market (other than x86) is IBM's Power platform. And I do not see HP porting their software to work on their biggest competitor's platform.

 

[podspi]

Where else would they port to? Once Itanium goes away, the only other big player in that markey (other than x86) is IBM's Power platform. And I do not see HP porting their software to work on their biggest competitor's platform.

ARM? If they're really serious about the server market who knows what we'll see in the 2~5 year timespan. I agree it is an unlikely long shot, though.

 

[Idontcare]

ARM? If they're really serious about the server market who knows what we'll see in the 2~5 year timespan. I agree it is an unlikely long shot, though.

SPARC is open too, isn't it?

May not be practical, and may be a truly bitter pill to swallow, but I think it is available for HP to take and do with as they please just as Fujistu has done.

 

[taltamir]

http://www.xbitlabs.com/news/cpu/di...anium_Engineers_to_Xeon_Projects_Sources.html

intel relocates engineers from itanium to xeon

 

[Lonbjerg]

Wow, if this is true...

I expected to see Itanium outlive everything - it's not like Intel to give up like that.

Itanium still makes more money for Intel in the server market than AMD does on x86 in the server market...so don't be so gloomy

 

[pm]

http://www.xbitlabs.com/news/cpu/di...anium_Engineers_to_Xeon_Projects_Sources.html

intel relocates engineers from itanium to xeon

It's weird to read this stuff when it's your stuff this anonymous source is commenting on. I guess I must be a member of the "small development team" that seems awfully similar to the same team that I've been working on for the rest of the Itanium projects.

 

[Idontcare]

It's weird to read this stuff when it's your stuff this anonymous source is commenting on. I guess I must be a member of the "small development team" that seems awfully similar to the same team that I've been working on for the rest of the Itanium projects.

It's all silver lining though, worst case is that you are destined for some x86 design team in your future, best case you get to work on post-Kitson stuff. Either way you got a job.

I remember when SUN cancelled their Millenium chip and layed off 5,000 employees. Now THATS the kind of bad news you don't want to be reading on S/A!

 

[jones377]

So the 26th came and went. Were there any major Itanium announcements?

 

[Idontcare]

So the 26th came and went. Were there any major Itanium announcements?

Just this:

Intel and HP May Postpone Itanium Roadmap Update amid Uncertainties – Sources.

Intel and HP planned to disclose post-Kittson (Itanium chip due in 2014 or 2015) in late April, 2011, according to unofficial sources. However, the event may be put on hold, some believe that because HP is unsure about its roadmap; other people suggested that Intel wanted to hold the information till its analyst day in May. In both cases the delay – which may actually last longer than for a couple of weeks, but for months – seems to be a result of Oracle’s decision to stop developing software for Itanium.

http://www.xbitlabs.com/news/cpu/di...oadmap_Update_amid_Uncertainties_Sources.html

It's long on speculation and short on facts.