Slashdot Mirror
Microsoft Drops Windows XP for Itanium
MBCook writes "According to an article on The Register, Microsoft has canceled the version of Windows XP for Intel's Itanium processor. They will continue to sell Windows Server 2003 for the Itanium in the high-end server market, but 'For the mainstream server and workstation markets, however, we believe we can best serve our customers needs with Windows Server 2003 Standard x64 Edition, and Windows XP Professional x64 Edition, respectively.' So much for Itainum workstations running Windows, but then again the article notes that no major vendors actually sell Itanium workstations anymore."
you're over a week late with this "news""
Sound like its a good time to snag some bargain boxen!
Everyone is lowering or entirely dropping their level of support for the Itanium, and now with Intel's interest moving to a better 64-bit system, this is good for everyone except maybe Intel and those who bought Itanium's.
MS 1: No one is selling Itanium based desktops or workstations, only servers ... err ... wait, that's stupid
MS 2: Indeed
MS 1: Why are we trying to make a version of Windows XP for it then?
MS 2: Because
MS 1: Indeed, let's not bother with that
MS 2: Cool.
MS 1: Don't want to piss off Intel though, let's pretend that we'll keep Server 2003 running on Itanium and that we support it
MS 2: hehe yeah
Although this is old news I will say this move does make sense for Microsoft. The Itanium is a server based processor, Windows XP is a consumer and workstation based operating system. This move doesn't seem too horribly suprising.
Just one more giant ram into the hind end of Intel. Man, they took a beating last year, and here we are only 6 days into 2005, and Intel is shaping up to be the industry punching bag.
I hate to jump on the underdog bandwagon, but given the high price of Intel processors over the past couple decades, I'm glad to see it finally catching up to them, and in spades no less.
The sad thing is that AMD seems to be heading down the Intel road now and in another decade or two AMD will just be where Intel is now... offering overpriced processors, and we'll be rooting for whoever is eyeing AMD's chops at that point.
Why can't any company come in, clean up with good products at cheap prices and STAY THAT WAY? Why do they all have to get greedy in the end? This phenomenon is not constrained to the CPU market, of course, we see it every single day.
Microsoft was dropping Windows XP Period.
They've had it coming. Truth be told we're stuck with x86 the same way we're stuck with IPv4.
With Internet we at least have some reason to upgrade, however with x86 which is obsolete and has manydrawback, well it still works and we're still seeing some nice performance improvements. I think we might have to depart from x86 pretty soon as the clockspeeds can't be raised anymore. So now its either SMP or simply a different architecture, like G5 for instance.
I've been holding off on purchasing a system for years. Initially it sounded like Itinium (sp) would be the 64 bit standard. Hopefully there will be some type of 64 bit standard as there is a great need for 64 bit work stations. I am in the CAD/CAM business and ever since the demize of the Alpha we have been waiting on a good cheap 64 bit windows based platform.
I don't think this is a good sign for the Itanic, but I don't think anyone will be surprised. This may not be the end of the line for it, though. MS has only dropped their workstation version, not their server version.
The really interesting question is: will Linux be able to carry Itanic, now that MS is starting to leave it behind?
See what I've been reading.
And where the fuck is windows x86-64 edition?
___
If you think big enough, you'll never have to do it.
Yeah. Lost jobs, negative economic impact...
Hi-larious.
Trolls: The high-tech version of those morons that scrawl obscenities in public bathrooms.
Wasn't the Itanium project dropped as a whole? As far as I knew some important partners working with Intel pulled out and Itanium's were going to stop being produced.
Intel dropped support for Alpha and MIPS
What? I thought Alpha was made by DEC... What support did Intel have for Alpha? You probably meant that Microsoft dropped support for Alpha and MIPS.
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
After that version, Intel dropped support for Alpha and MIPS, and look what happened to them.
Actually, it was Microsoft that dropped Alpha, MIPS and PPC support after NT4.
So should we all be signing up for landlines, even when our cellphones are better? What about buggy whips? I don't need one, but we could employ a lot of people making them.
Intel may be hurting with their Itanium mistake, but AMD will be gaining sales, and hiring more people.
I don't see it as a big issue, Itanium has always been targeted as a special use chip for servers and clusters, thats the only place that they have been selling. MS is wise to drop a desktop version of the os that would have met a non existent marked.. Instead MS can now focus on the AMD64 / Intel 64 hybrid version which is where the high end desktop market is more-likely to see growth.
I think linux and win 2003 server will share the server market.
EA David Gardner -"... but the consumers have proven that actually what they want is fun."
Iff it had come out on time the world would have changed and colleges would be teaching itanium design, not OoO
it was freaking 15 months late, and it still kicked ass decently well, give them some credit
..is always sucky, because they make a damn mess, and have sharp edges.
Its got to happen at some point, this project has been a complete business failure for Intel...regardless of the pet project clusters and supercomputer projects, the number of shipped units is only a tiny tiny percentage of Intel's vision for this project, although I am sure many here will attempt to justify Itanium as a niche product.
There is still one company marketing high end services running Windows - Unisys IS the Windows Mainframe market. And while Unisys does suck balls, that doesn't mean that they don't command marketshare or respect
The really interesting question is: will Linux be able to carry Itanic, now that MS is starting to leave it behind?
If there's isn't one already, I'm sure there will be a NetBSD port for it eventually. The Penguin isn't the be-all-end-all of free [as in beer] software that works you know...
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
It may have escaped you (I am pretty sure it did), but I wasn't saying anything of the sort. I was responding to the notion that a failure of this scope by a corporation of Intel's size is amusing. And however much AMD gains from this, when a publicly traded company of this size makes a boo-boo, the whole nation goes ouch.
Trolls: The high-tech version of those morons that scrawl obscenities in public bathrooms.
I call bullshit.
It was not an attempt to drop backwards compatibility, but rather an attempt to produce a product vastly superior to an x86 based design.
Itanium was not designed for the desktop, or even the standard server market. It was designed for number crunching, which it works quite well at.
Is a Cray XT3 backwards compatible with a Cray1 or even a YMP?
NO.
Same thing goes here. In fact Itanium was designed to compete with the likes of Cray. It was never, ever, designed with desktop in mind.
-nB
whois gawk date unzip strip find touch finger mount join nice man top fsck grep eject more yes exit umount sleep dump
What's really sad is that Intel still insists that the Itanium really is a superior architecture, and fully plans to push it forward in the high-end computing space. They even point out that they've pushed out most of the other 64-bit competitors, leaving them #3 behind IBM's POWER and Sun's SPARC. Of course, this sounds pretty silly since the other competitors were HP's PA-RISC, which was dated and being phased out anyway, Alpha, which was being phased out intentionally in favor of Itanium because of HP's deal with Intel, and MIPS, which never did that well to begin with.
Personally, I think Intel is going to keep beating the Itanium dying horse as long as they can, while attempting to improve overall revenues by pushing into other markets such as cellphones, PDAs, and other mobile/low-power devices. I really don't see how Itanium can possibly succeed over IBM's POWER, though it may have a good chance against SPARC since Sun is floundering so badly.
I agree that x86 will be with us for a long time, but only insofar as we are tied to PCs. Game units have jumped on the IBM bandwagon, personal media players and other digital devices are probably still up for grabs. The Mac is of course non-x86 but it is not nor will it ever again be a mass-market product, nascent sub $500 headless units notwithstanding.
And Google is starting to become evil despite their self-proclaimed "don't be evil" policy. Their new Group "beta" is pure evil, a) for being a compulsary beta b) for making the search and thread browsing facilities very, very sucky indeed.
I have been dying for their supercomputer OS. Microsoft considers supercomputer OS . I really hope MS doesn't ditch that OS too. I have been speccing itanium clusters and they seem to fit my needs. I also can't wait for .Net to come to Itanium, I'm sure MS will write the best optimizing compiler.
"brxref
Actually to be clear:
-The alpha version limped along until Compaq stopped willing to help Microsoft support it.
-The MIPS version was discontinued after EVERY SINGLE MANUFACTURER of MIPS clone workstations stopped producing them (e.g. Netpower and friends). It's not particularly evil to stop producing software for nonexistent hardware. Don't bring up SGI - it never ran on SGI's.
yeah, i'm sure netbsd/ia64 will be enough to drive the entire market on its own.
Well, the Itanium still runs x86 instructions.
The problem with the Itanium is that it just requires too much compiler magic to make it work well.
It's kind of similar to the "software bloat" problem. Yeah, you could spend a couple of years optimizing a single piece of software, or you could just throw more hardware at it, for less money.
Need to fix an error: :blush:
=~s/Cray XT3/Cray X1/g
Should have compared the X1 not the XT3 for two reasons:
Cray X1 is a vector based unit (what the Cray1 wanted to be), while the XT3 is a MPP unit, effectively a cluster computer optimized for density to the extreme of having 4 CPUs per PE.
Other reason: XT3 uses opteron CPUs
My argument still holds, however, because the X1 is the arena where the Itanium competes better.
-nB
whois gawk date unzip strip find touch finger mount join nice man top fsck grep eject more yes exit umount sleep dump
The Alpha did well with people that needed it, ie those running VMS and Tru64, and customers still buy them. Compaq continued development on the Alpha as well. It was HP who killed Alpha, and is attempting to kill Tru64. VMS on the otherhand must be supported for at least 15 years if they want to keep all those big contracts, and they have said that they will. VMS is being ported to Itanium, last I heard it boots and will run a dir.
"I use a Mac because I'm just better than you are."
Companies are often started for other reasons, however when they become big, public and faceless, greed is the motivating factor. That's why monopoly prevention is as important. Capatalism works to leverege corperations greed against each other to benefit consumers. Only works if there's more than one player, however.
And however much AMD gains from this, when a publicly traded company of this size makes a boo-boo, the whole nation goes ouch.
If the nation goes ouch because of a single company's misstep, that nation needs to rethink how it runs its economy.
Intel is not that large a company, and the Itanium division is not the largest part of the company. Even if the whole Itanium project was cancelled, Intel has lots of other products (which are all probably more profitable than Itanium). A few people would lose their jobs, others would be redeployed within the company, and Intel would have to write off a big investment. All the companies that depend on Itanium would be in trouble... oh wait, they wouldn't because no one depends on Itanium.
Honestly, it isn't like Intel is going belly-up any time soon. It'd probably do Intel a lot of good financially if they dumped Itanium, cut the excess baggage, and started thinking about something better with which to compete against POWER.
There is still a fairly solid market for Itaniums. HP will move HPUX and MPE customers (and Tru64 too?) gradually to the platform. Having Intel made processors (as opposed to PA-RISC and Alpha) means that servers can be cheaper and HP can get out of the Microprocessor development business. It may not be everything that HP and Intel hoped for, but it will still give some fairly solid life to the processor.
Custom, hands-free Linux installs. Instalinux
Maybe because the software they want to run is only available for Windows? In which case x86-64 is the right choice.
And
FreeBSD kinda supports it too.
Although who knows if it'll ever reach Tier 1 status now...
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
This just seems to be another iceberg to hit the Itanic.
You have to remember that in many pro fields, the software is limited to a few, or even single platform. Also it's not as simple as "find an alternative" often one simply doesn't exist. A half assed, kinda working but missing half the features, thing won't cut it, has to be all there. Sometimes, something you consider to be just trivial will be make or break.
We run into this with engineering apps all the time. They'll demand specific enviroments (as in OS, version, access privlidges, etc) that aren't what we want to use, but there's just no choice. It's take it or leave it, and the app is the only one of its kind that does what is needed.
Yeah, stupid typo. Thanks for catching it.
See what I've been reading.
Bingo. The quarter-to-quarter mentality is a product of this myopia, which has been called the Cult of Shareholder Value.
Read the EFF's Fair Use FAQ
Why didn't you submit this news a week ago?
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Of course, you are right in that the final kill of Alpha came later on, but even Netcraft could see the way it seemed to be going.
Itanium was not designed for the desktop, or even the standard server market.
Yes it absolutely was. itanium was designed to replace ia32, totally. They wanted itanium on everything from desktop to supercomputers.
the grand master plan was for itanium to take over the world, ia32 would die a horrible death and everyone would live happily ever after with a new, elegant architecture and forget the monstrosity ia32 ever existed.
intel saw what apple managed to pull off with the 68k -> ppc architecture migration, and enviously hoped to emulate them.
intel's own current marketing literature even promotes itanium2 as an entry-level server and workstation processor! delusional at best.
What are you talking about? You can run any x86 application just fine on any Itanium box. In fact I am writing this on a browser running in a x86 chroot environment on an IA64 box so I can use flash and other similar binary crap.
Itanium was the in-the-future-better-technology-if-compilers-catch -up -and-everybody-ports-all-their-software-to-it ... maybe, that pretty much killed itself. The Register have been calling it the Itanic from pretty much day one. They are now entitled to a "I told you so".
Itanic was a good research project that they made the mistake of telling the marketing people about. It is very much like Intel's new Socket format (with the pins on the motherboard rather than the processor). It was designed to make intel's life easy at the expense of everybody else. Strangely everybody else didn't like this idea very much.
Nerd: Derogatory term typically directed at anybody with a lower Slashdot ID than you.
Exactly - Intel knows where the bulk of their processor market is, and Cray isn't it...
The SME, SOHO, end Consumer market is where they make a the most of their money IMO.
To invest in a platform only to be used in Crays is a tad bit foolish isn't it? Especially because the Enterprise computing market is pretty crowded as it is...
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
Itanium probably IS a superior architecture but that matters not at all if intel fails to market it well enough to get sufficient sales to develop higher-clockrate parts. Also, unless keeping the number of cores down is critical to your problem, the price-performance ratio of itanic is inferior to ANY of its 64 bit competitors, some by a wider margin than others.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Support for NT on Alpha and no support for archs other then ia32 for Windows 2000 was pulled because relativly no one was buying PPC, MIPS or Alpha machines for Windows by the time it came to decide what 2k would run on. Compaq had very lively development cycles for Alpha until HP walked in.
"I use a Mac because I'm just better than you are."
Maybe because the software they want to run is only available for Windows?
I realize that; that's why I suggested putting pressure on them to provide versions for other OSes.
Vendor-customer relationships are not a one-way street, though many people these days seem to think they are.
I highly doubt any of them will be throwing good money at Itanium, and they will probably just drop HP if they feel they can't get better options.
boxen /bok'sn/ (By analogy with VAXen) A fanciful plural of box
I wonder how one might use such a word
often encountered in the phrase "Unix boxen", used to describe
commodity Unix hardware.
OMG !!! just like the original post, it's a miracle. If it doesn't run windows anymore, that leave a unix family of OS. . . . idiot.
So, has Netcraft confirmed the death of Itanic yet? If not, everyone else is off and even the captain admits the ship has sunk...
Moll.
What you hear in the ear, preach from the rooftop Matthew 10.27b
Running Gentoo huh?
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
Actually, Intel did own Alpha at the end.m l
See, for example, http://slashdot.org/articles/01/06/25/1359207.sht
The problem with the Itanium is that it just requires too much compiler magic to make it work well.
thats only one of the myriad of problems with itanium.
how about
* no software (and even less as itanium partners continue to abandon ship)
* much too expensive - very poor price/performance.
> It was not an attempt to drop backwards compatibility, but rather an
> attempt to produce a product vastly superior to an x86 based design.
True enough, but after a decade of Intel's marketing department telling everyone that anything that wasn't x86 compatible was a dead end, too many potential customers got mental whiplash when Intel then tried to get them to buy a non x86 processor.
Poetic justice if you ask me.
Democrat delenda est
they didnt predict evolutions, they gambled on it.
that is to say, they desperately hoped compilers would evolve to support epic, and gambled the entire future of their technology on it.
the technology didn't pan out, and they blew their entire wad on the gamble -- they lost. bigtime. they now have a supremely expensive, fancy processor with nothing to drive it. it's an expensive boat anchor right now.
and the technology required to efficiently drive EPIC doesnt look to be coming any time soon, either.
Nope, Debian, where the minimum one needs to do to run their x86 applications on an Itanium box is apt-get install ia32-libs.
Perhaps, but the Alpha team is engineering at AMD now...
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
I'm sorry, but I simply haven't seen any evidence at all the Itanium is a superior architecture for real-world software. It entirely depends on the compiler optimizing everything for the CPU, but the problem as I understand it is that most software simply isn't written in such a way that the compiler can optimize it for this CPU; there's too much parallelism, and most software simply isn't written that way. Just a simple re-compile with Intel's compiler doesn't fix the problem; all the software must be written specifically for the processor's features.
This sounds like a very bad idea to me. To be successful, a processor needs to perform well with software that is commonly available for it.
Here's an analogy: suppose someone builds a really efficient car, that gets 100 mpg. However, it doesn't run on the same gasoline that everyone else's car uses. It runs on a special fuel, which requires a lot of time and energy to synthesize. Because of this, the price of the fuel is 10 times the price of gasoline. Obviously, the car isn't really efficient at all; the overall cost per mile of driving it will be much greater than regular cars, and there's no indication that the special fuel would get any cheaper. Who'd buy that car?
It may be true that certain highly specialized applications can take advantage of the Itanium's architecture, but Intel isn't going to make any money selling a tiny number of these processors to a tiny number of customers. Even worse, the economy of scale isn't there: if Intel doesn't make many of these CPUs, their price will be very high. The customers could do the same jobs using more of the less-efficient processors that the competitors sell; since these cost a small fraction of what the Itanium costs, they'll still come out ahead.
Highly specialized, highly priced products only succeed in the marketplace when there's no other, cheaper (though less efficient) way to do the job. Itanium is only good for highly parallel computing workloads, which can also be easily done with commodity processors, running in parallel. This is a formula for failure.
http://slashdot.org/submit.pl under my login:
2004-12-28 05:30:10 Microsoft drops Itanium Windows XP (IT,Microsoft) (rejected)
/. is irrelevant.
There is a 64-bit standard: x86-64 (used in Athlon 64 and Opteron). There's another one too: G5.
Not to mention DEC's Alpha, Sun's UltraSPARC, SGI's 64-bit MIPS stuff, IBM's Power line (not the same as PowerPC), and those are just the ones I can think of off the top of my head. Some of them were around well before PowerPC, let alone IA64 and X86-64.
dragonhawk@iname.microsoft.com
I do not like Microsoft. Remove them from my email address.
What about buggy whips?
I read that and thought, How do you introduce bugs into a whip?! New Year's resolution to cut down on the caffeine is hereby revoked.
What's really sad is that Intel still insists that the Itanium really is a superior architecture...
Well, it is, in theory. In practice, others were not successful in VLIW (Sun MAJC; isn't Transmeta VLIW, too), and POWER 5 is showing that RISC is still going very strong. Itanium is like a really great show dog that has skin rashes and farts too much, while everyone who knows what they are doing goes for the well-rounded mixed breed (aka, POWER, SPARC, and AMD64).
-- Microsoft is the most expensive commodity operating system and office suite vendor in the marketplace.
I call bullshit on you!
At least partially.
THe first round of epic (merced) was supposed to be a server processor, like the p6, with desktop parts trickling down, later. With the original plan, all current intel cpus should have been epic-based till now.
But the whole project was delayed and delayed, the compilers took ages to get running and AMD came rather unsuspected with the athlon, which resulted in the need quickly push the existing x86 design.
So the late epic designs werent significantly faster than x86 anymore, plus more expensive/higher power requiring. -> nobody wanted them.
If the itanium never was supposed to become a normal server&workstation processor, why do you think that they included a dedicated x86 processing core into the die?
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?
Xeon and Itanium will share chipsets and sockets in a few years, I think 2006 or 2007.
Sometimes they legitimately are one way streets. In this case, the customer/market demanded x86 compatibility, and was unwilling to see Intel half way.
As a business, Intel's responsibility is to tune into reality, whatever that is. Looks like AMD beat them to it, and their stock price will reflect it.
The hardware vendors have extremely shabby support for it. Know why? It's in BETA. They won't even get close to being serious until it's officially released.
Vendors have always dragged their feet on new OS releases - remember when Win95 first came out? There was all sorts of isues with CD-Writers, sound cards, printers (Canon's 650 driver is STILL officially a beta!).
"...Well, there's egg and bacon; egg sausage and bacon; egg and spam; egg bacon and spam; egg bacon sausage and spam..."
Remember Bill Gates' public vision a few yeas ago about Windows everywhere? It looks like it is coming true.
...except Itanium.
Oh, and Alpha.
Oh, and PowerPC.
Compatibility wasn't dropped. Even Itanic systems are still more then capabale of running good 'ol DOS. Granted, you'd have to be a little peculiar to want to run IA-32 apps at Pentium-2 type speeds on such expensive systems, but you can hardly say they dropped compatibility. Perhaps if Itanium hadn't taken so long to appear we'd all be complementing it for it's speedy IA-32 execution.
3 1803.pdf
Intel's got more information on Itanium IA-32 support at http://www.intel.com/design/itanium/downloads/254
HP was a big fan of the Itanic when HP was still HP, and dumped their PA-RISC roadmap in the 90s for it. the thought was that they'd get Intel to carry most of the cost of developing and fabbing, they could swing the HP-UX and (RSC? -- brain fade on the old mainframe codeset line) software teams over to a common hardware architecture, and save money there, too. so HP joined in the party, hearty.
they actually got stronger into it when they got compaq, even though Intel and Samsung were the Alpha partners at that point.
woe betide them all when the first Itanic came out, and the performance sucked against PA-RISC, and the second and third generation silicon didn't get them in the MIPS game, either, as seen in the roadmap.
For a lot of reasons, AMD Opteron looks better and better every day to the former Itanic fans. they should keep an eye out for Power5 and Power6 as well, the IAM coalition would like more customers. AMD didn't lose sight of what the users want.
if this is supposed to be a new economy, how come they still want my old fashioned money?
Netcraft confirms: Itanium is dying! - Where do you want to meme today?
so say we all
So, what exactly is the state of 64 bit Windows? There is still only beta support for AMD64, Itanium support is getting dropped.
Microsoft supposedly spends billions of dollars a year in R&D, and they are unable so support anything but the same chip architecture they have had for the past 15 years.
A company like Redhat or Suse, meanwhile, support virtually every major architecture available today, including the Mainframe and Power processors, and do it with a fraction of the resources and manpower available to Microsoft.
If Microsoft is unable to economically develop for other platforms, perhaps the company's cost structure is way out of line.
Come on, folks, you bash x86 architecture (and rigtfully so) and then you turn around and bash Intel for trying to break away from this architecture and do something wildly different and superior. Their only failure was that they haven't "bet the farm" on Itanium. If they did, we'd be running EPIC-architecture 64 bit systems by now. As things stand, the only two viable desktop choices are IBM/Motorola Power architecture (that's 64 bit from the ground up) and this old tired x86 architecture with 64-bit extensions duct taped to its side.
No, this just illustrates my point more. What I meant before is that many people these days act as if customers are beholden to the vendors, and have no choice but to go along with whatever the vendors want to do.
Intel saw Microsoft acting this way successfully, and they thought they could act the same way by pushing Itanium when no one really wanted it. They've learned the hard way that, when markets are working correctly, customers are the ones calling the shots, not vendors.
My point with these EDA programs is that if enough EDA software customers complained to these vendors, they'd release Linux versions of their software. But unlike the CPU situation where AMD gave people a choice, and they flocked to it, these customers need to stand up and speak out instead of waiting for things to happen.
Of course, apple provided pretty good 68k emulation in their PPC OS so that you could run the old apps, and provided a scheme for "fat" binaries so that you could have one (bloated) app that would run on both architectures, protecting the legacy users as well. Microsoft, intel's biggest bed partner, neglected to provide these items and thus essentially guaranteed that itanic would crash and burn (or sink...) Of course, PPC is a RISC processor and the 68k is very RISClike so the 68k emulation was relatively trivial, while itanium is VLIW where modern x86 processors are somewhere between RISC and CISC. Itanium only provides significant speed improvements with the most effective of compilers, and a JIT recompiler to run x86 code on itanium would be too complicated for microsoft to get right, you know they'd botch it somehow. Thus, by not putting x86 compatibility into the chip itself (which is highly impractical given the difference in architecture) intel basically ensured that itanium had no chance to survive as a desktop processor and shot itself in the foot in terms of world domination. The only way they could have succeeded is if AMD did not exist or was completely incompetent. Neither is the case, so we can now simply say goodbye to itanic on the desktop.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
"why do you think that they included a dedicated x86 processing core into the die?"
They didn't.
That's what caused all the fuss. The thing had to emulate x86 and as a result was dreadfully slow compared to a comparably clocked x86 die.
-nB
whois gawk date unzip strip find touch finger mount join nice man top fsck grep eject more yes exit umount sleep dump
Unfortunately Itanium suffers from a very very bad and unproven design. Moving optimizations to software was pure genious.... rolls eyes.
I think the intel engineers drank too much of the risc coolade and instead of promoting simplicity so they could add features like pipelines and branch predictions they instead moved those to software as well.
x86 has improved and their are worts all over it but it works.
Today's x86 processor just wrap x86 instructions to a risc core to execute.
IT works and it works quite well.
I would prefer to see Alpha re emerge but it aint going to happen. Its a shame really.
I am glad we are not stuck with epic and Itanium it would make writing assembly code and good compilers near impossible.
http://saveie6.com/
The really interesting question is: will Linux be able to carry Itanic, now that MS is starting to leave it behind?
Is there a compelling story for Itanic now? It's not x86 compatible in its native mode, right? So the Power 5/PowerPC would be a more appropriate platform for fast non-x86 linux users as its star is rising, while Itanic is fading away. There's a vibrant PPC Linux community.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
To be more specific, the 64-bit extensions are duct-taped to the 32-bit extensions that are duct-taped to the 16-bit instruction set. :->
Any compiler or I should say most will use lots of registers on the cpu itself and the Opteron has the memory controller chip in the cpu itself.
Things like branch prediction must be done in software on the Itanium which some computer scientists now say is impossible.
You can do that in hardware with risc processor but not Itanium if I recall.
I think it is a clossal failure but I am not a chip engineer. I do say I am glad I do not own one and a 1 pound heatsink with a fan that sounds like a jet engine to keep it cool because Intel/HP overclock it so it wont look as slow is not a good sign either.
New != supperior.
http://saveie6.com/
a JIT recompiler to run x86 code on itanium would be too complicated for microsoft to get right, you know they'd botch it somehow. Thus, by not putting x86 compatibility into the chip itself (which is highly impractical given the difference in architecture) intel basically ensured that itanium had no chance to survive
Right - what's Microsoft's value proposition for supporting Itanic with an ia32 JIT recompilation emulator when x86-64 is real? Apple got it done because they owned the hardware and the operating system, so they had motivation. I can't even see why Bill Gates should care if Intel goes out of business.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
But being 50% faster at more then 4 times the price is NOT a good deal.
Well, if you go out of Windows completely and cross into the Linux world, there's plently of software out for 64-bit architectures. Mainly the Athlon64, though. A good example is tha Fedora Core 3 is out for x86 and x86-64. There are RPMs to install the software specifically to the 64-bit kernel. Although, there is a 32-bit area so you can still install the x86 RPMs and compiled packages.
OT, but I really hate this "I call bullshit" meme.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
Oh, Linux is not good enough for that. HPUX/IA64 will carry the torch!
HPUX on IA64 always seemed like, to me, putting a fleet of barely functioning tanks on deck the Titanic. On top of the paying customers. (You decide which is which.)
I can explanate how to administrate your network. You must configurate and segmentate it, so it can computate.
I guess everyone knew what you meant, but NT4 was actually produced by the other very large and widely despised company (actually that's a little bit unfair to Intel).
"There is more worth loving than we have strength to love." - Brian Jay Stanley
It is so common to support multiple processors
that I don't see why Windows can't do it.
I would imagine that they would have the majority of their code written in a language like C or C++ and that they could just set a compiler flag for the different processor.
I am sure that the gnu compiler will do it.
What if they compile Windows with the Gnu compiler.
Why not do it?
I know why they don't, because they have their own. They won't use anyone elses.
Another reason why open source rocks!
Or they could use the Diab compiler.
Seriously, if you have a real compiler company or organization there isn't that much to do to recompile code to work with a new operating system or a new processor as long as you maintain proper build standards and adhere to very comman methodologies that have been around for at least 20 years.
This just shows how the Windows built on quicksand development tools are of poor quality even for them.
This is a case of where antitrust would have helped them actually do something good. But because they use a compiler that they make themselves they can't seem to make it work for all processors.
Pretty damn stupid of them, if you ask me.
It really is to bad that there arent any more Itanic workstations to price compare to the new G5 iMac.
..... the Apple is a whole lot cheaper.
I get tired of seeing that Macs are more than Dell or HP, when in fact if you price out either Itanium work stations (this is only fair) aganst the G5 iMac, well
Apples to Oranges?
Cheers
* Carthago Delenda Est *
It seems like the efforts of the R&D are used more to pioneer new markets of untapped revenue. The idea being to market novelty - not really perfect any technology. Once the novelty has worn off, let some other vendor move in and perfect what they pioneered. Well, the desktop and Office apps are good and very standard. But it appears that Linux is catching up here. I mean, there's most stuff you need in the OpenOffice spreadsheet - it's not far away from Excel. But on the server side, they sell the concept of a server, not a rigorously developed server OS. Aside from the desktop, which I do use as my primary desktop for professional type tasks like bookkeeping and other business apps that are windows apps. But I've now found Linux to be a more secure way to access the web. I look back and marvel at the fact that I used to put my banck records and other private stuff on a Windows box that was connected to the web.
http://www.softwareobjectz.com
You're sorta ignoring the fact that Windows NT was never something one would call "popular" Alpha or MIPS (or PPC).
Right, and HPaq owns the Alpha, which is why:
A) They're pulling out of Itanium
B) You can still buy Alphaservers.
Well, it's an improvement on "Bzzzt. Wrong!" at least.
To be fair, PPC64 is not pure 64-bit from the ground up, it too is 64bit extensions to a 32-bit architecture (the processors commonly known as the G3 and G4 were only 32-bit, for example, and the exact same binary MacOSX runs on G5). This is made painfully aware by a lot of linux distros on PPC64, where if you fail to explicitly install the 64bit development utilities (or specify to use them at compile time), you'll end up with PPC32 binaries by default, which aside from linking into 64bit code or trying to do it as a kernel module, you'd never know the difference without running file against it. It is very much similar to the x86_64 to x86 relationship, with the nice distinction that it did start life as a 32-bit platform and only has legacy dating back to then, unlike x86_64 which continues legacy from the intel 8-bit computing days, which means a lot more strange quirks that no longer are optimal.
As far as Intel trying to 'bet the farm' to move the world to IA64, I can guarantee it wouldn't have worked no matter how hard Intel tried. Assume hypothetically that Intel had completely ditched x86 and stopped development and production of IA32 chips. At the time Itanium was ready, AMD had already established itself as a pretty viable solution, not as well respected in business, but certainly on the radar. Now when faced with replacement/upgrade of hardware solutions, companies see the poster-child they've grown up to love, Intel, unable to run their existing applications, and therefore a huge cost to migrate in terms of development. Meanwhile, the suboptimal AMD offers fresh, fast x86 processors. Intel's reputation at that point wasn't enough to offset the huge cost of a platform shift. I remember PentiumPro facing harsh criticism and some market problems due to it's slower execution of 16-bit code, and that was when AMD and Cyrix had pretty equal, small, low-budget marketshare.
Besides, Itanium wasn't exactly pure gold. It had strong points (good High Performance Computing mainly), it had weak points (not good at general workstation use, high volume servers, essentially uses that involve widely varying, unpredictable execution paths).
XML is like violence. If it doesn't solve the problem, use more.
The really interesting question is: will Linux be able to carry Itanic, now that MS is starting to leave it behind?
Another question is: now that MS is dropping Windows for Itanium, will Intel contribute more free development tools for ia64 Linux and make more "investments" in Linux for Itanium? Since 1998, Intel has made many contributions to Linux for x86 and Itanium - I'm assuming much more than Alpha and MIPS have.
But even if they do, will enough Linux developers choose to develop for expensive Itanium workstations when cheap x86 workstations are "good enough." I think we'll see a lot more Itanium-specific Linux investments from Intel, but I don't know if Linux developers will invest more time in Itanium.
For the near future, it looks like Itanium workstations have Debian 3.0 and Red Hat Enterprise WS.
TO START
PRESS ANY KEY
Where's the 'ANY' key? I see Esk, Kitarl, and Pig-Up...
when they make a black & white robosapein with advanced, sentient AI, minigun side-arms and *INEXPENSIVE* add-on servomechanical sentry satellite-bots. (SAY, thats alot of "S" sounds and "-" marks!)
DO NOT DOUBT THE POWER OF DUCT TAPE!!!
;)
Seriously, it's amazing what that stuff can do...I just wish I could get my hands on the military grade stuff...the kind of tape that is meant for patching bullet holes and doing minor airframe repairs to the helicopeter
How about having no pins on either motherboard nor CPU but instead have a "CPU-sized plastic square with pins through it" right between them? If you bend a pin you'll just have to replace the between-thingie.
Oh god. I'm gonna make MILLIONS.
Free as in mason.
For example, "high performance" in the automotive world means big displacement, turbochargers, big valves, and the like
That's because the 2010s oil crunch hasn't happened yet. "Performance" will come to mean miles per gallon for a given payload, and generator-braked vehicles such as the Honda Civic Hybrid will dominate for passenger and grocery payloads. And as oil prices go up, electricity prices will probably go up as well, making instructions per kWh a valid measure of performance. So Freaking What(tm) if each individual core is slow if you can Beowulf the shit out of them?
Itanium owns 20% of the Risc server market.
Nice troll, HP still sells a lot more PA-RISC chips than they do Itaniums. Hows that for a ROTFLMAO!
-- Microsoft is the most expensive commodity operating system and office suite vendor in the marketplace.
Betamax had the same length tapes ( 180 minutes ) as everyone else.
A VHS T-160 tape has 8 hours in SLP mode. Record 20 episodes of your kids' animated series on one tape (provided you press pause when the commercial comes on). The Betamax L-750 tape only got up to 4.5 hours in its equivalent of SLP.
I really don't see how Itanium can possibly succeed over IBM's POWER, though it may have a good chance against SPARC since Sun is floundering so badly.
Interesting you say this, because Sun sells many more UltraSPARCs than IBM sells POWER. Add in SPARC 64 from Fujitsu, and that just increases the market share futher. Note that you didn't say PowerPC, which is a different numbers game, entirely.
-- Microsoft is the most expensive commodity operating system and office suite vendor in the marketplace.
Vendor-customer relationships are not a one-way street, though many people these days seem to think they are.
If the customer is a residential user, how can the customer negotiate with a Big Corporate Vendor(tm) other than just through take it or leave it?
Microsoft has its Windows and Office cash cows to prop up the rest of the company. Now that the foreign counterparts to the LZW patent have expired, does Unisys have a cash cow left to prop up the Windows Datacenter market?
You have it pretty much, 100% backwards. MS does not pioneer anything. They wait for some other company, standards group, or working body, to come up with pioneering ideas, implementations, business plans, and especially market share. MS then comes in, embraces and extends it; never getting anywhere near correct or perfect.
I never understood why MS claims they spend on R&D, but it's obvious that their idea of R&D is different from the rest of the tech world. I guess for MS, R&D = market surveys, web site development, and technology scouts to find other companies that actually do R&D, and have come up with new, innovative products. That way, they can buy the company, steal the IP, or simply force them out of the market.
We have an old Digital AlphaServer 1000A 4/233 we use as a file/print server in one of our offices, running Windows NT4 SP6a. I think it was bought in 1994 and hasn't been turned off since, with no downtime as far as I know. Say what you like about Digital, but they sure knew how to make machines that last. Jonathan
Richard Dawkins called, he wants his meme back. You know, the one about memes, called... "meme."
Intel sunk several billion into what is, effectively, and R&D project. Whether it will pay off or not remains to be seen and is to a large extent irrelevant. There seems to be a general feeling on Slashdot these days that we should mock people who try to innovate and elevate people who simply mass produce things that other people have invented.
I am TheRaven on Soylent News
On the technet workshop I've been this week their roadmap showed that when SP1 for 2003 server is released it will be followed by a generic 64 bit 2003 server release and a generic 64 bit release for xp. This means that itanium, emt64, amd64 -opteron- will be surported by the same install.
I don't think MIPS or PPC ever even got service packs.. Microsoft used to make some MIPS machines themselves too. Alpha support continued up to sp6a and the release-candidates for win2k.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
But getting linux working on itanic was relatively easy... Intel helped port the kernel, gcc and glibc, and it still cost them a lot less than they paid MS to port windows... After the kernel was done, most userspace apps build just fine, itanic is just another 64bit architecture to linux, just like PPC64, MIPS64 or Alpha that linux has been running on for years.. A vast majority of opensource apps that run on linux are 64bit clean and porting them to itanic was literally just a recompile.
You can't do this with closed source vendors, you have to pay/bribe/blackmail them to support your new architecture. They won't do it on their own until you already have a large enough user base, and with the current commercial-dominated software world, the user base won't grow unless the apps are ported. So, your screwed... Even Intel and HP combined didn't have the cash to bribe enough software vendors.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
the page you linked to does not give the plural of 'box' as 'boxes' anywhere.
my password really is 'stinkypants'
What the hell? The slashdot collective hive-mind doesn't speak with one voice?!?
The Slashdot user-base is made of bazillions of people (approximately). You can get MOST (but not all) of them to agree that "Linux is good". Anything beyond that, and all of the individual people have their own opinions.
Personally, I thought that the original Itanium concept was really nifty. Anybody who's been following it beyond the initial "The die is HOW large?" fiasco has been waiting for it to fail.
Forward, retransmit, or republish anything I say here. Just don't misquote me.
The 386 went from 16-bit to 32, taking all 16-bit apps with it, but there was also the very little known 80376, all 32-bit none of the 16-bit parts.
The same can be made of the Athlon64, simiar to the Itanium, being 64bit only. I know, that'd be a disaster, but now that we have binaries, linux binaries, and possibly windows, such a chip would be cheap, powerful, cool and welcome by some.
Would you buy a cheap laptop that will run AMD64 binaries real fast, but none of the 32-bit x86?
Err you don't know what you are talking about.
The Athlon64 is totally BACKWARD compatible with the AthlonXP and Pentium processors, and will run all 32 bit x86 binaries just fine. It is NOT a 64 bit only cpu, but can run in TWO modes (32 and 64 bit) as well as running 32 bit binaries INSIDE of a 64 bit OS. The Athlon64 is a desktop version of the Opteron. The major difference between these two cpu's is the number of processors that can be connected together in an SMP system. (Opteron's are SMP enabled, Athlon64's are for single cpu systems.) Next time RTFM and engage brain before putting mouth in motion.
"There seems to be a general feeling on Slashdot these days that we should mock people who try to innovate"
I'm not sure there was much innovative about Itanic: trying to eliminate hardware complexity by transferring it to the compiler has been done before with some of the early RISC chips, and it was a failure then, too. It's just way, way harder to figure out beforehand what the chip is going to do when running the program than it is for the chip to figure out what it's going to do while it's actually running the program.
Frankly, I think we should mock people who decide to spend billions trying to do something which failed before and, duh, it fails again.
Which was a very affordable & fast architecture. HP has converted most of its unix systems to itanic in hopes of aborting any further arch updates.
However i think they are once again improving the stable & faast PA risc systems since HPUX 11i itanium still stinks
from what i can tell SGI is still on the itanium bandwaggon
.. is the worst. How many millions of dollars has your company spent on M$?
How many hours, weeks, years, have companies lost messing with M$ software?
Got Windows? Got hacked.
Considering even the corporate leaders of most nations have finally discovered Windows is wasteful, expensive, and even damaging to thier enterprises why the frell would anyone bother with Windows ever again?
There are simply better solutions. Obviously.
They DID.
If you tell me that they were emulating it, you could also say that the athlon is emulating x86.
It used the same function units than the epic core, but the instruction decoding/dispatching was all done in hardware. (although very slow, because it was in order, single issue, ect)
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?
What he's saying is that perhaps there ought to be a 64 bit only Athlon64.
This question answers itself, though. Who remembers the 80376?
What we call folk wisdom is often no more than a kind of expedient stupidity.-Edward Abbey
Actually the i80376 was not meant for workstations and servers. It was targeted at the embedded market. The reason it would not work as a replacement for the i80386SX was that it lacked the integrated MMU. It is true that it only supported protected mode though, but it would not have been able to do the sort of unique segments with different protections for each app deal that a real 32-bit OS does.
The reason most people never heard of it is because it was not supposed to compete with the i80386SX/DX/SL chips. In its intended market for embedded systems it did, well... well enough for me to have heard about it at least, though for the systems from that vintage here the people before me went mainly went with 68K.
Bull. Intel spent 2 billion dollars on R&D for itanium. That's more money than cray made in profit in its entire history. There is no chance whatsoever that intel would dump that kind of money into exploiting a niche market which could never return their investment. When intel designed itanium they wanted to win the whole enchilada (commodity desktops & servers, where the money is) not just high-end cluster computers.
How does your thinking apply to the term "virii?" I've often thought that although "virus" applies both to computers and biological agents, they are really two very different things. Perhaps "virii," because it applies specifically to a computer concept as opposed to a biological disease, is a legitimate plural. Or is the use of "virus" strictly metaphorical, and only the accepted plural should be used? Perhaps its obsessive, but it can also be interesting and informative. The term "hacker," and how it gets used in the news and popular culture, for example.
And you didn't notice I used "Sound" where I should've put "It sounds?" If I knew this would get such a response, I would've thrown in the word "virii," too!
An "Economy" p275 with a single 1GHz POWER4+, 1GB RAM, 36G disk, 16 meg graphics adapter, and no optical drive is a mere $5,575 - bump it up to a 1.45GHz chip, 2 gigs of ram, a DVD drive, and a (sweet 1600x1200 20") TFT panel with a 128 meg card, double the price. Go for the dual-1.45 setup with 4 gigs of RAM and two HDDs and you're looking at a cool $15,993. Granted, that box will utterly spank just about anything else that you can fit under your desk, but it's just a little pricier than what Apple offers. Kind of like comparing an F2004 to an Enzo...both lust objects, but still totally different orders of magnitude.
Facts do not cease to exist because they are ignored. - Aldous Huxley
Well, that depends on how we prioritize the importance of our testicles relative to other things.
I don't know about you, but when I talk on a cellphone, I hold it to my head, not my crotch.
The reason it would not work as a replacement for the i80386SX was that it lacked the integrated MMU. It is true that it only supported protected mode though, but it would not have been able to do the sort of unique segments with different protections for each app deal that a real 32-bit OS does.
Actually the 376 DID have a MMU. But it ONLY supported a segmented memory layout. It lacked the paged mode memory managment. Intel ripped that one out to prevent you from running UNIX on the 376. The embedded project I used the 376 for DID have separate memory segments for protection. I recall during debug often hitting the segment protection violation trap! (Usually due to null pointers!)
Wow that's neat history to know, thanks! And actually it makes sense now that I think about it. You would like the segments at least for stack, heap, and code. The protections at the segment level are there already, so it would make sense to leave that in.
If I wanted to run code at 1/10th normal speed, I'd pick something out of a dumpster.
I actually said "Pioneer new markets of untapped revenue" - not "pioneer or develop new technology". They have a way of effectively packaging stuff that was invented by some obscure or out of favor group and delivering it via the greased channels that exist in whatever market is being worked. You are very astute in your observations. I still remember the documentary about the Xerox Palo Alto facility on PBS - it's played quite a bit. One of the lead developers or designers was ordered to show Steve Jobs everyhting - so he could use it in a new venture (Apple). History has shown it does not pay a software company to pioneer anything significant as it will simply be usurped by an organization with existing marketing channels to the core customer base. Doug Hettinger
http://www.softwareobjectz.com
IBM Power _architecture_ is 64 bit from the ground up. G3/4 processors were specifically castrated because the market wasn't ready for 64 bit. G5 is less castrated. Who knows, maybe sometime in the future we'll get the full blown multi-core IBM Power chips running in our Macs.
Which is duct-taped to 8-bit instruction set of Intel's 8088 processor.
The Reg actually has been calling it Itanic since 1999...
I actually said "Pioneer new markets of untapped revenue" - not "pioneer or develop new technology".
Opps. So you did! Sorry about that. Sounds like we on the same page.
No sweat, I'm glad you clarified what I was trying to say.
http://www.softwareobjectz.com
...and hasn't been for a long time.
I am very small, utmostly microscopic.
Itanic is EPIC. Explicitly parallel. You don't get much performance out of it with single threaded programs, which is why it sucks for most things.
I hate latin grammer! I was taught for years that splitting an infinitive in English, as in "to boldly go," was wrong. Much later I found out its because Victorian English scholars wanted English to follow Lain as much as possible, and the Latin infinitive doesn't have an equivalent of the word "to." A split infinitive is wrong only because its trickier for the pedants to translate into latin. So explain to me how latin plurals are appropriate, but Old English plaurals are only "antiquated?"