HP Terminates Itanium Workstations

vincecate writes "The largest Itanium system maker, HP, has terminated its Itanium workstations. It seems their workstation customers have spoken in favor of x64. In related news, Intel expects to ship over 100,000 Itaniums in all of 2004 while AMD is estimating 1.5 to 2 million AMD64 chips in Q4."

Comments

The good news

[raider_red]

I've heard that HP actually sold both of the Itaniums they had in inventory, so there shouldn't be too much to write off.

How Ironic

[Anonymous Coward]

HP killed Alpha in favor of Itanium. Which in turn happenh to be dead at birth.

Makes me think about their technical vision ...

Re:How Ironic

[ackthpt]

HP killed Alpha in favor of Itanium. Which in turn happenh to be dead at birth.

Makes me think about their technical vision ...

Intel sued by DEC for stealing Alpha technology for Pentium

Intel agrees to buy production plant, pay undisclosed cash, continue to make Alphas for DEC

Merced goes on for years, uses lots of Alpha technology.

Revamped as Itanium

Sells for huge $$$$ when it hits the market

Still sells for $$$$

Intel gets clubbed like a baby harp seal by AMD x64

Seems somewhere in that long build up to the release of the Itanium they forgot how they made their money in the first place. Psst! Processors are a commodity.

Intel may have a lot of better technology than AMD, but AMD has clearly shown they've learned a lot about getting a product out there.

Re:How Ironic

[Mateito]

Makes me think about their technical vision ...

HP's current innovation strategy may be sumarized in the their unwritten Mission statement:

Carly Gets Paid.

Under Carly, the Calculator division has had the guts ripped out of it, the printer division has had the guys ripped out of it, the server division has had the guts ripped out of it.

Um.. what else does HP make?

And Carly gets her US$20m a year, despite the fact that none of her "innovations" have moved the company forward.

Re:How Ironic

[Rauser]

What else does HP make?

iPods... oh wait...

Re:How Ironic

[Mateito]

what about that MBA from that fluff business school, sloane?

Yes, and don't call me sloane.

Oh, right.

As an MBA in training, I can see exactly where MBA and technology diverge. MBAs are great for ideas on how to manage people, finances, suppliers, clients, to anticipate market trends etc etc... and a name school gets you great contacts (what I don't have.. but hey, its 1/10th the price of the Harvard BS course).

What it doesn't teach you is how to work R&D. The economics of R&D don't work the same way as everything else does. IBM get it. Xerox got it. AT&T may still get it. Sun hopefully will get it again.S

Stuff you do now may pay off for years. In some cases for IBM and AT&T, decades. MBAs don't think on those scales. Long term is 8 quarters... 2 years.

Carly might be great in charge of the Sales part of HP, the pure commerse stuff... but she doesn't have any idea about how to run and engineering firm because she's not an engineer.

Could it be?

[KingKire64]

I AMD has caught up to intel a couple of times in the desktop market only to fall back again. Could this be the time that they leapfrog over Intel and be far and away leader in a market? One could only hope. In a tech world of dominate players (Intel, MS) its nice to see the underdog win with a superior product.

Re:Could it be?

[RzUpAnmsCwrds]

"AMD chipsets support slower CPU to memory times than Intel (32bit or 64bit) counterparts. "

Bzzzt... wrong answer.

In AMD64 chips, the chipset doesn't have the memory controller - it's in the CPU.

AMD's CPU-to-DDR latency is much lower than Intel's.

Re:Could it be?

[Nazmun]

I can't believe this was modded up...

A64 FX's and Opterons support dual channel ddr and have much lower latencies then intel at the same mhz (400mhz ddr X 2). Usually the FX's and Opteron's win the memory bandwidth benchmarks.

As low power AMD has a line of mobile barton core processors that use as little power as 45 and even 35 watts. They can also be placed in a destop motherboard. The xp2400 35 watt is also under $100. But there is a good chance the pentium-m uses less power but they are only found on certain laptops (can't be bought).

AMD64 include consumer processors

[hattig]

AMD sold around 100,000 Opterons in Q2 however. This should increase to 200,000 in Q3 given recent products from HP, Sun, IBM etc, especially with the increase in 4P systems.

Of course, the ASP of Itanium is a lot higher, so Intel need to sell a lot fewer Itaniums to get the same money back as AMD. On the other hand, AMD haven't sunk $billions into K8!

Top 10

[AKAImBatman]

Top 10 Itanic jokes:

10. HP decided that they didn't want to go down with the Itanic
9. Hear that flushing sound? That's billions of dollars being invested into a lemon.
8. HP must of realized it was a 64-bit Pinto.
7. HP's just upset that they didn't get to sit on the bow and yell, "I'm the King of Computers!"
6. HP's Itanic line is sunk.
5. "The Itanic is the most advanced chip of her kind. She's practically unsinkable!"
4. HP didn't want to be compared to Leonardo Di Caprio
3. HP Execs suddenly realized that Di Caprio dies in the end
2. Intel assured HP that the Itanic was not sinking, despite being hit by a AMDBerg
1. "My clock wiiilllll, count on and on!"

Sorry, I just couldn't resist. :-D

Re:Top 10

[KingKire64]

Horrible, simply horrible.

Eck

A victory for 32 bit backwards compatibility

[celerityfm]

AMD deserves the win here for pushing 32 bit backwards compatibility, Intel had to and still is playing catch-up with them in this arena.

Good job AMD!

Re:A victory for 32 bit backwards compatibility

[ackthpt]

Backwards compatibility? Why? I mean you can just recompile, right?

Tell that to Microsoft.

Microsoft is running most of their software on AMD64 in 32bit, thanks to that backward compatibility, but you know they're sweating over getting full 64 out, since Linux has been 64.

Funny how Intel and Microsoft have to scramble to keep up with underdogs, isn't it?

Re:A victory for 32 bit backwards compatibility

[aristotle-dude]

Exactly. If you guys really hate MS and Intel, stop supporting them.

If you really like what Apple and IBM are doing with and for Open source, support them by buying their hardware and running whatever operating system you wish (be it linux for PPC, one of the BSD's or OSX).

I laugh when I see open source advocated saying how evil MS is and yet they probably helped put MSFT in the position they are now in by not buying Corel/Wordperfect products instead of MS Office and buying PC's bundled with Windows instead of now dead platforms like the BeBox, Commodore Amiga, Next Cube. Even if they had bought macs from Apple, MS would not have the power it now has in the industry and Corel/Wordperfect would still have a significant portion of the office market.

I also feel that Open Office should stop trying to closely emulate MS Office and try to produce something much better.

How the times change...

[Kazymyr]

I remember 5 or 6 years ago the new 64-bit chips from Intel were "hot" with everyone talking about them, and also supposedly right around the corner in terms of schedule. AMD surely stole their thunder on this.

O tempora...

Yeah, Itanium tanked... So what?

[stratjakt]

I guess because (for some moronic reason) AMD are "good guys" and Intel are "bad guys" we just have to get all giggly and rub their noses in it.

BFD. Sometimes you win, sometimes you lose. Some products take off, some don't.

Itanium looks like a good architecture for transaction processing, at least on paper. Turns out the market was more interested in backwards compatibility.

Re:Yeah, Itanium tanked... So what?

[roca]

1) When the IA-64 design first became public, it was clear that they'd made some incredibly poor decisions. For example, the architectural design was based on the assumption that the chip would not do out-of-order execution in hardware. Such deficiences were to be remedied by a god-like compiler that would emerge at some later date. Unsurprisingly, it never has.

2) These predictions were borne out by the fact that Itanium performance has always sucked, especially considering the enormous die size, cost and heat dissipation.

3) It looked like Itanium might win in the market despite its technical limitations, just because of Intel's vast marketing budget, its momentum, and its monopoly leverage forcing OEMs to stay away from technically superior alternatives like AMD64.

4) Thankfully this hasn't happened. The technically superior, open solution is winning. Thanks AMD.

Re:Yeah, Itanium tanked... So what?

[Animats]

Such deficiences were to be remedied by a god-like compiler that would emerge at some later date. Unsurprisingly, it never has.

Yeah. A few years ago, the compiler guys from HP came over to Stanford to speak about Itanium compilers. They didn't have a clue how to solve the problems they faced.

Re:Yeah, Itanium tanked... So what?

[rainman_bc]

No, we're all cheering for AMD for a number of reasons.

Their decision to support 32-bit mode in their x86 64 bit platform was a wise decisions and all of us knew that.

Furthermore, AMD keeps forcing Intel to innovate. As long as AMD is around, CPU's will get faster and better and do more per cycle.

Without AMD, we'd not have good competition, and Intel could comfortably cut their R&D costs to turn a bigger profit - their only rival would be PowerPC, and it's not a x86 platform. Let's not forget, Intel is first responsible to its shareholders.

Furthermore without competition, rest assured we'd already have DRM shoved down our throats too.

Re:Yeah, Itanium tanked... So what?

[roca]

Predication and explicit speculative loads were primarily added to the IA-64 architecture because they'd decided not to implement out-of-order execution (dynamic scheduling) and other dynamic techniques*. They aren't nearly as important on a modern superscalar processor.

(* Itanium2 doesn't even do next-line prefetching!)

Explicit speculative loads was a major mistake because in many kinds of code the compiler cannot place speculative loads far enough ahead of the actual use for it to pay off. Often the address to be loaded from is simply not known far in advance of the load (consider executing the C code "x = a->b->c"). So Itaniums spend a lot of time stalled waiting on memory accesses. That's why Intel spends so many transistors on gigantic on-chip caches, to try to reduce that pain. The architecture's pretty good for workloads with very regular and compiler-analyzable access patterns (regular number crunching, SpecFP) but it's bad for everything else (servers, user applications, irregular numeric codes).

Yes, IA-64 is a aggressive, radical, clean and somewhat novel design, so it's understandable that some geeks love it. However, it is not a good design.

If it was a good design, then with Intel engineering, 5x the transistor count, and no backward compatibility requirements, it would be absolutely crushing Opteron performance. Instead it is merely competitive.

BTW it is quite odd to consider IA-64 a small tweak over RISC chips. IA-64 is the most dissimilar of all viable architectures today.

bring back alpha

[Anonymous Coward]

Why doesn't Intel just get over the NIH syndrome and start fabbing the Alpha (proven design, existing software base, the geeks love it)... Don't they own the rights for it via some legal-fall out with Compaq?

- Friendly A.C.

Re:hp server

[ultrabot]

Let me put it this way. I would not buy a server from HP anyway.

I don't think they will care. Most people in the business of buying servers seem to do. Comp... er, HP Proliants are probably the most popular Linux servers at the moment.

So, question for the crowd...

[Featureless]

What was Intel thinking?

An architecture switch breaking x86 ISA compatibility (i.e. emulation is noticeably slower than the original item) would put it on a level playing field with other 64-bit workstation/server-class chips, yet they never seemed to offer either world-beating design improvements or substantial price benefits, or appear as though they would in the future.

This looked like a loser from the first minute I saw it, and I obviously wasn't the only one: I mean, the chip has been "The Itanic" in Register parlance for years now.

Intel, for all their flaws, is a smart company with a lot of smart people working for it. I must just not be seeing the whole picture. They must have had some good reason not to have flushed this project years ago, right?

Re:So, question for the crowd...

[HungSoLow]

Intel, for all their flaws, is a smart company with a lot of smart people working for it. I must just not be seeing the whole picture. They must have had some good reason not to have flushed this project years ago, right?

If there's one thing I've learned from working in high-tech, it's that no matter how smart and capable the grunts are (engineers, etc.) you always have a dim-witted marketing guy or manager steering projects in the wrong direction (and not listening to criticism).

Re:So, question for the crowd...

[Junks Jerzey]

What was Intel thinking?

An architecture switch breaking x86 ISA compatibility (i.e. emulation is noticeably slower than the original item) would put it on a level playing field with other 64-bit workstation/server-class chips, yet they never seemed to offer either world-beating design improvements or substantial price benefits, or appear as though they would in the future.

Intel decided to break with the past and start fresh, in hopes that they could make a large leap forward. That's a good goal. But what actually happened was a couple of things:

1. Their experiment failed, in that they didn't get the monstrous across-the-board benefits they expected.

2. They started this back in the days of the Pentium, when it looked like the x86 CPU architecture and instruction set were the big problems. The Itanium design team didn't forsee the crazy lengths that would be taken--by both Intel and AMD--in order to speed up the crappy x86 architecture.

Honestly, you can't fault Intel for trying. Where did chips like the ARM and MIPS come from (two of the most popular non-desktop processors)? From designing a new architecture. That's the same kind of thinking that resulted in the amazing GPUs from ATI and nVidia.

As a footnote, it's somewhat sad to see radical advances in CPUs come to a halt. I'd love to see someone set the industry on its ear.

Re:So, question for the crowd...

[e40]

They were not thinking. They were being arrogant.

I have a hypothesis: it was a power play to eliminate all competition. It would have been difficult for AMD and others to follow them down this IA64 road.

Corrolary: Intel wanted to establish compiler dominance. I work for a compiler company that produces every part of the source to machine translation for our compiler. Intel told us we would not be able to do an IA64 port all the way to machine code and that we'd have to use their assembler. This was shocking. Upon probing this, the Intel guy would not relent. He said it was near impossible for anyone but Intel to produce machine code for IA64. For over 20 years we've done countless ports, to some really weird hardware. Our expert said it would take 2 years to do the port. The most time we *ever* spent doing a port was a year and that was for a Cray (and a lot of that was for operating system interface issues).

Itanium will crush all... hardly

[cant_get_a_good_nick]

The interesting thing is 3 RISC chips were killed because of the threat of Intel - MIPS (well, at least in workstations, embedded lives on), Alpha, and PA-RISC. PA-RISC even had a technology that could be seen as the opposite of EPIC, instead of moving scheduling logic to the compiler, they actually moved some of the optimization the compiler could do to the chip itself, since it knew current state of the machine and the compiler couldn't. Just shows you what a bit of monopoly muscle can do I guess.

That's actually quite sad

[gsasha]

Architecturally, IA-64 is a very advanced architecture.
Ok, many people don't like it. And OK, it's complex. And OK, many people are making other quite good 64-bit processors.
If its competition was Power or MIPS, then OK, I'd say that the worse it is, let IA-64 die, but x86 (and x86-64 as well) is UGLY and laden with all kinds of OLD JUNK. Come on, it will be junked sooner or later. Granted, Intel can make high-performance x86s, but that at a price of devoting over 1/3 of the stages for decoding!
Or, let's put it that way. It is a Good Thing (TM) to have several different architectures. If all we'll be stuck with will be x86, it'll be quite sad.

Re:That's actually quite sad

[Waffle Iron]

but x86 (and x86-64 as well) is UGLY and laden with all kinds of OLD JUNK

The old junk is a constant overhead, but processor architectures keep getting bigger and more complex with or without the old junk. Processors are now so large that the old junk is a tiny percentage of the total logic.

All modern processors translate their user-visible instruction set on-the-fly into some other internal format anyway. The X86 ISA is just a kind of bytecode, and it's a relatively compact one at that. It's easier for compilers to generate than Itanium bytecodes, so it's not hard to see why X86 is still around.

I kind of doubt that X86 will ever get junked. Now that X86 has 64-bit addressing, there's little reason to create any new user-visible changes to the instruction set. Processors can continue to improve and change their internal architecture without bothering the users with silly implementation details.

Re:That's actually quite sad

[roca]

x64's 64-bit mode fixes quite a few of the problems of x86 as well as giving you 64-bit support. For example, a number of useless old instructions are no longer supported (they still work in x86 mode of course). It increases the number of general purpose registers from 8 to 16. Using SSE2 to do floating point, you get a reasonable floating-point instruction set with 16 registers. If you squint a bit it looks like a decent instruction set which just happens to have a weird instruction encoding.

Yes, the decode stages are a pain (though trace cache helps), but in return you get significantly higher instruction density than competing RISC chips which helps with your instruction cache.

OTOH the IA-64 architecture was designed around unfounded implementation assumptions like "we won't be doing out-of-order execution". Sorry, WRONG. Sometimes polishing up old junk gives better results than designing completely new and differently broken junk.

Intel outsider

[Doc Ruby]

Has AMD finally proven that the x86 "standard" can produce truly 100% compatible CPUs, without Intel IP, after decades of doges and ruses, including MMX?

In other news, Honda outsells Bentley.

[Anonymous Coward]

Amazing, isn't it, that a Honda Civic would outsell such a high end car?!?!!! It just boggles the mind.

The Opteron isn't in the same league as the Itanium, no matter how much AMDroids wish it were. AMD needs to be comparing Opteron/AMD64 sales to Xeon/Pentium4 sales. Itanium is a very high end processor and it's one of the best you can buy for certain high-end applications.

Not to say Intel didn't make a mistake in trying to push Itanium too early as a general purpose CPU - it's clearly not.

you're not the only one mocking the Itanium...

[exhilaration]

Check out this article: IBM mocks Itanium server sales - again, make sure you look at their very amusing graph of changing sales forecasts.

TFA?

[SuperQ]

Is it just me, or does the article gloss over the fact that "EM64T" is actual a clone of the AMD64 architecture? Are intel's market-droids trying to brainwash people, or are people really that clueless to the fact that INTEL IS MAKIGN A CLONE OF AN AMD CHIP?

Give credit where credit is due.. EM64T is clone crap, and is signifigantly slower than the AMD chips.

Just one little note...

[ltwally]

Just one little note that the author of this article fails to mention:

The Itanium is a high-end workstation/server chip. ONLY. -- While the AMD64 architecture is AMD's entire product line right now. It's their desktop chip; it's their workstation chip; it's their server chip; hell, it's even their notebook/laptop chip.

Whoever submitted this article seems to think that every AMD64 sold is going to be going into the high-end server market. Either that, or he thinks that home users are buying Itaniums. Funny... I don't seem to recall ever seeing a laptop with an Itanium in it.

A more honest comparison would be the 800 series Opterons vs. Itaniums, the 200 series Opterons vs. Xeons, and Athlon64's vs. Pentium 4's.

Re:What about servers?

[joib]

Of course, they're pledging to continue selling Itanium servers.

In the longer run, IMHO it sounds somewhat problematic, considering that all the engineers developing software will be running on x86-64. I.e. the software will first be available on x86-64, more tested etc.

So why should the customer shell out money for an Itanium server instead of an x86-64 server which has better bang-per-buck and runs the software more reliably? In the short run HP can probably contain x86-64 in low end servers, keeping high end stuff reserved for Itanic. But in the long run, they'll have to start providing higher end x86-64 gear too, or their customers will move to a competitor that will.

Re:Should have stuck with Alpha

[Kevin Burtch]

Exactly what I was thinking.
HP and Intel deserve this for killing off the two most powerful processor lines in history.

Back when PA-RISC and Alpha were in production, the gap between them and the next fastest CPU lines were staggering. I used to check the CPU Info Center at Berkeley every time a new one was released, just to see how badly it humiliated the competition (sadly, the CPU Info Center is no longer maintained).

The Athlon (before it was named such) uses the Alpha's bus... and the original slot-A design was compatible with both the Alpha and the Athlon, all you would need to sell a motherboard for the other one is a different BIOS. This was the selling point that convinced many motherboard manufacturers to actually make these boards. Unfortunately, only a tiny handful of companies actually marketed the resulting systems using the Alpha CPUs (mostly in Linux Journal & Linux Magazine as rackmount servers).

They could have done so much more... oh well.
My current favorites are UltraSPARC and PowerPC (with POWER close behind).

Funny shift in /. mindshare

[tji]

Around here, you used to find all kinds of people complaining about the old kludgy x86 architecture and how the backwards compatibility placed terrible limitations on the CPUs and on software that runs on it.

Now, everyone jumped on the bandwagon spouting "what were they thinking? Trying to define a new architecture.. dumb asses!"

So, which is it?? I learned architecture and assembly on a Motorola 68k processor. So, the x86 stuff has always seemed kludgy to me. Have the problems been overcome, or do people just not care anymore?