The VP of Intel’s digital enterprises group told the INQ this morning that the Itanium microprocessor is starting to make waves in the corporate enterprise market. Kirk Skaugen, of the servers platform group, showed a slide which claimed the Itanium processor was eating into Sun and IBM Power shares, based on “customer revenues”. He also said that since the fourth quarter of 2003, applications for the Itanium family had grown to 5900 by the end of 2005.
I don’t think Itanium has put a significant dent in Sun’s market. Opteron and other x86-64 chips (including those from Intel) have. But the Itanic has not lived up to Intel’s expectations, clearly.
Opterons and Xeons are far more widespread and I have yet to read of a company that dumped thier servers running the above mentioned processors and get on the Itanium bandwagon. I think Intel is beating a dead horse. They should take what they learned apply it to their new procesors and thats it.
Opterons and Xeons are far more widespread
Actualy, Opterons would have much more of a market if Dell and some other vendors wouldn’t push Intel only sales based on the bussiness relations.
People often ask such resellers about Opterons, and then after receiving negative answer, they buy Xeon. Those customers buy that machine solely for Dell or that specific vendor support not for Intel value. I often talk to some friends of mine who work for various vendors that resell Intel only, and all say the same. There is a lot of asking about AMD, even though company policy is already known.
On the other side, it is not a joke that people call Xeon with the name Itanic. As I hate to remember it was still the worst machine I ever had on test.
I think Intel is beating a dead horse.
Has ever been anything different? With any /*not Intel specific, all companies do that*/ company PR bull they post?
PR is like statistics. Lies, lies and just more god damn lies. Except that PR lies are wrapped in paper spelling only “WE RULE, WE, THE BEST”
People often ask such resellers about Opterons, and then after receiving negative answer, they buy Xeon. Those customers buy that machine solely for Dell or that specific vendor support not for Intel value. I often talk to some friends of mine who work for various vendors that resell Intel only, and all say the same. There is a lot of asking about AMD, even though company policy is already known.
It seems that there’s a bit of a disconnect – it looks like the vast majority of self-built/mom-and-pop-built PCs use AMD chips these days, while the majority of big OEM PCs use Intel.
I think Intel is beating a dead horse. They should take what they learned apply it to their new procesors and thats it.
And what they learned is that there are quite severe limits to what compilers can do, and that out-of-order execution isn’t the waste of space that some people thought it was. Oh, and “Very Large Instruction Words” really are as bad as they sound.
VLIW was the transmeta name, while intel dubbed it “EPIC” (Explicitly Parallel Instruction Computing) – there, even the acronym sounds bad
on the serverside intel managed to kill off alpha, pa-risc and mips.
and i’m sure if they somehow can get rid of ppc and sparc they will get a whooping 10% marketshare with itanic
I never really thought of it that way. Especially since HP has killed Alpha and PA-RISC. Those two alone would give you a good share of the market on new purchases.
based on “customer revenues”
Given the price of the Itanium, could this just mean that people spend more on it than the competition?
Given the price of the Itanium, could this just mean that people spend more on it than the competition?
I was thinking the same thing. Also: isn’t Itanium the chip in which the most money has been sunk?
The poor sods that used to buy PA-RISC-based servers from HP, MUST now chose Itanium servers or… go to IBM, SUN and others, that will offer them better value. It appears that most have, in fact, defected, but the ones who decided to stay and buy into Itanics, have increased it´s marketshare.
One thing, however, that is important to mention here, and thatvery few are aware of: for HP, the PA-RISC servers were a loss. I talked with people from HP, and they couldn’t quite explain what is thereason for HP sticking with it. So, lack of HP-UX servers based on Itanics does not mean less profit for HP. If anything, it should have a positive effect :o)
Both Dell (mostly Xeon) and other Intel vendors (i.e. SGI->Itanium) are loosing big in the high performance computing market. Intel currently has simply no processor to match against Opteron clusters.
Companies like Dell don’t care… For SGI it means their market is eroding away.
Sun is winning; they noways offer Opterons instead of Sparcs in most HPC bids.
At the not for profit I work for, a bank donated to us two Itanium systems. When I went to get them I asked the IT there, besides the tax write off, why he was getting rid of them and he said, “Opterons were cheaper, faster and better.”
Aren’t these also the same people who said “Don’t worry if the P3 smacks the P4 at similar clockspeeds, the P4 will make up for it by being able to scale way way past the P3’s range”?
That is correct, then they turn around and now use the P3 core again.
I know people like to bash Intel for the P4, but that particular attack is unfair. Consider the context in which the P4 was designed, and the time at which Intel made that statement. During the history of the PII/PIII, the same basic core scaled from 266MHz to over 1GHz (about a 4x jump) over a period of about 3 years and 2 process shrinks (350nm to 250nm to 180nm). The Athlon scaled similarly, about 3.5x over 3 years and two process shrinks (250nm to 180nm to 130nm).
Even assuming a slower growth rate of doubling every 2 years, one would be entirely justified in expecting the P4 to have hit 6GHz+ in the 5.5 years and four process shrinks since its release (180nm to 130nm to 90nm to 65nm). Yet, the fastest P4 is still only 3.8 GHz, a scaling of only 2.5x over nearly 6 years. Nobody, save perhaps for AMD*, foresaw the clock-rate wall that the industry would hit. Who would have expected getting only a 2.5x scaling from four process shrinks, when previously a 3x-4x scaling had been had from only two?
*) I’d argue that AMD did not forsee the frequency wall so much as it just got lucky. Intel has consistently led AMD in process technology, so AMD probably figured that in the long run, it was not a good idea to try to compete with Intel by using clockrate (which is process dependent). Meanwhile, Intel probably assumed its lead in process technology would allow them to compete with clockrate. Both assumptions were resonable in the context of their time, although in hindsight Intel’s was doomed to failure.
Edited 2006-01-19 04:26
Nobody, save perhaps for AMD*, foresaw the clock-rate wall that the industry would hit. Who would have expected getting only a 2.5x scaling from four process shrinks, when previously a 3x-4x scaling had been had from only two?
To be precise, it wasn’t the clock-rate wall as such, but the power wall. Clock rates used to be limited by transistor switching speeds, but with the P4 they’re limited by power consumption. The P4 could probably run at 6GHz right now, but the cooling for that would just be too expensive and loud.
Intel must have foreseen the increase in power consumption, but they probably expected process shrinks to make up for that. But that hasn’t happened.
The unfulfilled 3GHz predictions and the liquid cooling suggest that the G5 with its 16-stage pipeline has hit the power wall to.
AMD’s processors with their 12-stage pipeline don’t have such big power problems, which means they’re probably still hitting the switching speed wall first.
I guess for maximum performance you’d want to hit both walls at the same time.
Intel must have foreseen the increase in power consumption, but they probably expected process shrinks to make up for that. But that hasn’t happened.
Quote from a New Scientist article titled Has Moore’s Law had its Chips? (vol 182, issue 2448, p26, 22nd May 2004).
<blockquote>”Somewhere between 130 nanometers and 90 nanometers the whole system fell apart. Scaling stopped working and nobody seemed to notice.” – Bernie Meyerson, Chief Technologist, IBM.</blockquote>
People did predict the power increase, but they didn’t predict all the problems caused by the process shrink. Have a look at a paper from ISSCC 2001 Microprocessors for the New Millenium: Challenges, Opportunites, and New Frontiers:
http://ieeexplore.ieee.org/iel5/7284/19686/00912412.pdf
At one point Gelsinger writes about power consumption:
<blockquote>… predicted powers are excessive and prohibitiveley large for any practical application, and it is clear total power consumption will become a limiting factor.</blockquote>
But he concludes the paper with:
<blockquote>if is boldly predicted that the microprocessor of the year 2010 will have 1B transistors on a die, operate at 20 to 30GHz, and perform over 1T operations per second.</blockquote>
Even though everyone expected problems they had 20 years of experience of solving such problems and were confident that they could solve this one.
When are we going to finally see that “massive migration from Xeon to Itanium” that Intel keeps going on about?
I guess its like Hypertransport competitor, aka, CSI – missing action.
As for ‘killing competition – nothing is killing RISC, the only thing that is being killed is the large margins that these machines used to attract for unjustified claims of performance superiority.
Right now, SUN is selling more SPARC machines they they ever have – be it with lower revenues (the revenue hasn’t increased at the same rate as margins have decreased), total POWER systems have increased – and with the costs associated with maintaining the POWER chip spread accross more markets and customers via the XBox and P3 contracts, I wouldn’t be surprised if in the future, seeing IBM offering some low cost PowerPC workstations (low cost being a subjective term).
On the topic of “Opteron kills everyone and my grandmother”:
It’s quite amazing how everyone’s behind AMD these days. And Linux. And Google.
I find it interesting how people are naturally pushing another single dominant player, ushering it to power.
Look at how Microsoft (and to some degree Intel) started out, and how the PC fan base ignored better products, like the Macintosh and the Amiga. (I’m talking 80’s, early 90’s here)
The underdogs turn into monsters.
Edited 2006-01-19 08:29
Your historical analysis is flawed. Microsoft and Intel were never the underdogs. Microsot/Intel started out with less marketshare than the Macintosh and Amiga, but they never had a “fan base” per se. Rather, the rapid adoption of IBM compatibles by businesses, and the resulting percolation of IBM compatibiles into the home market, drove Microsoft + Intel to dominance.
The interesting issue here is that Intel and HP have effectivly shot themselves in the foot. As a previously pure Digital Tru64 (Alpha RISC) environment we have gone through several vendor changes in recent years. First from Digital (DEC) then to Compaq then over to HP. Through that time the support has gone from excellent to good to horrid. Through that time we had assurances that they had planned out the migration from Tru64 to HPUX on Itanium. What they forgot to mention or even think about was if the customers where going to go down that route. We have now invested heavily in moving any new business away from HP, not only because the HP migration plans and support contracts where dead weight but because they tried to push us onto a Platform that our software would not work on and the vendor had no intentions to write an itanium port due to what they saw as “a difficulty developing for that arch”. Through the entire process i constantly have felt that the HP buyout of Compaq was a knive aimed at the RISC market rather than for any real benefit to HP. I dont know if that is how it really was intended behind the scenes but its the way its come across to us as customers and speaking to other Compaq/Dec Alpha owners I have got similar comments.
Looks like Xeon is doing better than Itanium. I think Itanium was a disaster and this is proving it. AMD is kicking Intel’s butt.
In all fairness though, from a marketing prespective, Itanium is a flop and Intel should immediately pull out of Itanium and concentrate on getting its Xeon processors in a far better position than AMD’s own. I think it’s possible but realistically, it might not happen.
Edited 2006-01-19 20:00