Hewlett-Packard, one of the most aggressive promoters of Intel’s Itanium family of processors, is 86ing its line of workstations that use the chips. The decision by the Palo Alto, Calif.-based computing giant will likely be interpreted as a large symbolic blow to Itanium’s fortunes.
HP discontinues its Itanium workstations
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
maybe, HP will start to use AMD-64 bit processors.
I Think that Intel deserves what happened to its workstation market at HP. I Think that they need to make Itanium I at least the next Pentium V if they want to resurrect this chip again at the workstation market.
Intel would just introduce an Itanium for desktop use.
An interesting move from HP
As the article points out, the Itanium requires native i64 code, meaning any application has to be ported and recompiled. While if you got it with HP-UX or Linux that wasn’t a huge deal, the vast majority of the sales for it were as WinXP 64 bit edition – which long term with zero backwards compatability to 32 bit code nuetered the software base. As a (allegedly) high end workstation the ability to run multiple OS in both 64 bit and 32 bit modes allows a higher range of end testing… lacking that ability makes the Itanium less desirable to the current alternative – Opteron. The Opterons ability to run 32 bit code at native speed while offering native 64 bit execution as well goes above and beyond, and many (mostly developers) are taking notice of it. The Itaniums days are numbered not from it being a bad or slow chip, but from intel not taking the time to make it backwards compatable with the x86 line. There is a LOT of code optimized for the x86, and as many (if not more) binaries.
Oh, by the way l33t h4xx0r, eighty-sixed is a very common term for showing something the door, and abbreviating it 86’ed or to say you are 86ing it most certainly is NOT resorting to l33t. Next time, learn some common slang BEFORE you knee-jerk.
Oh, and as someone with a sense of humor I don’t think I could have avoided using the double pun.
http://dictionary.reference.com/search?q=86
Intel would just introduce an Itanium for desktop use.
Why? Is there a reason that Itanium is better than x86 for the desktop?
I am not being an idiot, I really would like to know why it is better.
Why? Is there a reason that Itanium is better than x86 for the desktop?
I am not being an idiot, I really would like to know why it is better.
Getting rid of that god for saken BIOS once and for all and replace it with something superior, namely EFI. Maybe then instead of having plug ‘n and pray, we’ll have a situation where the operating system doesn’t have to jump through 3,000 hoops, hoping between modes, just to load itself.
Why isn’t EFI being used for x86? buggered if I know, Intel has made it available for all and sundry, it would simply be a metter for the BIOS producers to adopt – IMHO, they reason why they won’t stop using their junk is because there is a under the table agreement between the two largest producers to ensure that superior solution never gets to the table.
Death to intel for:
-plug n pray(Also deat to microsoft for that)
-rambus
-motherboard recalls
-the pentium aborted 1.13GHz introduction LOL(AMD jealousy really starting to show there) and PS btw thanks to the hardware review sites for their honesty in that matter
-being second to 700, 800, and 1000MHz(maybe others)
-the snobbiness of their adherants
-the pentium III outperforming the pentium IV when it was first introduced, not to mention the Athlon
-and not even having a desktop 64bit processor.
In the K6 days, p5/p6 was outperforming AMD and Cyrix/Winchip in games, until Quake2 with 3DNow! came out, and Voodoo drivers were 3DNow! optimized. The FPU was why the K6, etc performed worse than the p5/p6 generally. AMD then introduced the Athlon with 3 FPU and started to kick intel around some. Chipset drivers were the main thing that held total benchmarking down for AMD/Via or SIS or Ali combos compared to intel ones.
Then, the pIV came out, with lots of bandwith for good performance. AMD counters with the integrated memory controller. In many areas AMD is exceeding the competition. I would like to see them and the chipset makers come up with a PC architecture with good features from PPC(oops, they have PCI and AGP and USB… tainted by intel there, sort of)… never mind.
I remember when I still mostly used winblows and had to uninstall and reinstall cards and drivers to make plug and pray work, but my DOS, OS/2, and Linux partitions had no such problems whatsoever. I would almost be tempted to get an old Mac, but I eventually overcame the crappiness that is wintel, and was happy again. Anyway, I could almost feel sorry for them and all their misfortunes, but competition is good for everybody.
>> Oh, by the way l33t h4xx0r, eighty-sixed is a very
>> common term for showing something the door, and
>> abbreviating it 86’ed or to say you are 86ing it most
>> certainly is NOT resorting to l33t. Next time, learn some
>> common slang BEFORE you knee-jerk.
It might be a common slang for you but, as you have probably already seen, English isn’t my native language. I don’t and can’t know everything about the slang and other unusual things used in other languages.
It’s like this “it’s raining cats and dogs” idiom, for example. How should a foreign speaker know what that means? Yes, I know it means that “het pijpenstelen regent” when they say that, but at first it sounds pretty weird, right?
(“pijpenstelen” is Dutch for pipe stems or pipe sticks)
Thanks for both explanations though (you and the other anonymous user).
I’m sorry to hear of Itanium’s problems. I worked on the chip in linux development years ago at Intel. However after working with the AMD64 chips, I can see why Itanic suffered.
It was a doomed chip to begin with. Constant failures and problems occured and it way too slow for today’s envrionment.
Time to move on Intel. Produce the Pentium V with 32/64 bit capabilites for desktops. Produce a Pentium V XEON 32/64 bit and make it work faster than the current Xeon chips. Intel, you had the chance and you blew it.
It’s no wonder that your CEO wants major changes.
>> Oh, by the way l33t h4xx0r, eighty-sixed is a very
>> common term for showing something the door, and
>> abbreviating it 86’ed or to say you are 86ing it most
>> certainly is NOT resorting to l33t. Next time, learn some
>> common slang BEFORE you knee-jerk.
It might be a common slang for you but, as you have probably already seen, English isn’t my native language. I don’t and can’t know everything about the slang and other unusual things used in other languages.
Seems to be an American slang term, can’t say I’ve ever heard it either – and I am a native English speaker.
The 3MB L3 version of the IA64 is slow, Intel got around it by adding another 3MB of L3 (total of 6MB) on chip, and it is now a bit faster!
But… just look at the P4EE, when you add tons of cache, any chip goes fast too.
And now, the SPEC FP king is IBM Power5, and the TPC-C king is the IBM Power4 (for non-clustered results) with using half of the processors than the IA64. And for low end servers and high end workstations, the AMD64 chips are great.
Remember Itanium was co-developed by HP when they decided to kill HP-PA risc processors.
They killed the Alpha and ported OpneVMS to Itanium. Itanium suffers the x86 “we must be compatible” credo.
Ludovic
—
http://perso.hirlimann.net/~ludo/blog/
86 defined:
http://www.urbandictionary.com/define.php?term=86
I hope they regret stopping development of the alpha by now. I wonder, could it be worth it to resurrect it?
I can’t say I’m very sad to hear this. I have never been able to see the Itanium as anything other than a product of NIH syndrome. That Intel wanted to move away from the kludgy x86 architecture is commendable; that they chose to do so by cooking up another 64 bit RISC architecture (besides Alpha, UltraSPARC, etc) is lamentable. And when they turned out not to have done a very good job at it, either, I just wished it would die. Unfortunately, HP killed Alpha first. A desperate move to save Itanium, perhaps?
Now there is ppc64 and x86-64 trying to make a niche for themselves. I think they will fly, and I am happy that open-source is flourishing today – at least we can move applications from platform to platform, and use whichever architecture leads or wins.
“Getting rid of that god for saken BIOS once and for all and replace it with something superior, namely EFI.”
There are interesting paralells between EFI and Itanium.
Where Itanium is Intel’s attempt to escape from x86 and make a 64 bit RISC CPU, EFI is an attempt to get with the times on boot firmware. Where IA64 was an imitation of (mostly) Alpha, EFI serves the same purpose, but is incompatible with Open Firmware.
Itanium had difficulty conquering the market because other RISC architectures were already established, and IA64 is compatible with neither them, nor with x86 (emulation is sloooow). EFI might encounter similar difficulties, although its chances are a lot better than Itanium’s: it’s cheap (Itanium was not), and it’s not as defining of an architecture as the instruction set, making it easier to support.
As far is I am concerned, EFI is yet another symptom of the NIH disease at Intel, and I wish it the same fate as Itanium. To the designers, who no doubt put hard work and long hours in it, I extend my sympathy. Perhaps you could solve a problem that has not been solved before, next time?
Well, they already make Opteron servers, so it’s only a matter of time.
Though I never contemplated buying one of these, cost being the huge factor as is for most everyone, it is too bad these seem to be on the way out. I remember when they were being introduced the break through feature seemed to be their EPIC. That being that the parallelism was decided and dealt with at compile-time, relieving the CPU of that task. Anandtech has an old article on that – http://www.anandtech.com/showdoc.aspx?i=274
The huge weakness here though were the compilers, which Intel didn’t always have control over.
I wonder how long it is going to take Intel to realize that Itanic was a bad investment and kill off the whole thing. HP’s shipments of Itanic powered servers are likely to look any less miserable as they are at the moment and nobody else seems to be planning to join the Itanic club. I give it maybe 2 years.
Seems to be an American slang term, can’t say I’ve ever heard it either – and I am a native English speaker.
I’m in the southern U.S. and rarely hear “86” except in movies. It’s commonly used when someone drinks too much and gets temporarily kicked out of a bar.
“axed” is more common, symbolic of an “axe” chopping off a project or other.
Now if Intel would just 86 the x86…
HP’s 4-way I2 with 8GB RAM & 4x73GB hard drives $73,780.00
IBM 4-way POWER5 with 8GB RAM & 4x72GB hard drives $21,049.00
IBM 4-way I2 with 2GB RAM & 2x148GB hard drives $58,795.00
Why are I2 systems so expensive. Also, the above pricing is for 1.5GHz processors. Benchmarks show the POWER5 to be almost 2x faster then the I2; so the POWER5 system is not only the cheapest but also the fastest.
For an extra $3,000.00 you can get the IBM configured with 1.65GHz processors.
Note: the above IBM pricing is for the new OpenPower servers which are designed to run Linux. Red Hat & SUSE both offer support for about $2,000.00 per year.
IA64 is not an imitation of the Alpha, and it’s not RISC. It would probably have been better if it would have been an imitation of the Alpha, but they chose a more radical design. The architecture does have a lot more potential than people think though, and it will be interesting to see what happens around 2007 if they can unify the Xeon/Itanium infrastructure.
intel took over the chip design and compiler teams of hpaq.
thats the only link between alpha/itanium.
i hope that hp goes back to alpha. at least they are on par with intel (after 2 years without alpha-improving)
I think we have to accept that Alpha is dead by now. Itanium is more likely to become a huge success than Alpha being resurrected. It was the fastest processor for quite some time, but it takes more than that to succeed unfortunately.
As far is I am concerned, EFI is yet another symptom of the NIH disease at Intel, and I wish it the same fate as Itanium. To the designers, who no doubt put hard work and long hours in it, I extend my sympathy. Perhaps you could solve a problem that has not been solved before, next time?
Well, EFI is about the *only* good thing that came out of the whole Itanium debarcle, but then again, if it were me, I would have gone OpenBoot.
Opteron CPU, OpenBoot firmware, would be a great marriage of technologies; but as you said, Intel suffer from the NIH sydrom, however, with that being said, nothing was ever stopping AMD from coming out and saying, “OpenBoot is the new firmware for the Opteron platform!”; Microsoft would have supported it, SUN would find it alot easier to provide good hardware detection since they’re already used to handling OpenBoot, and as for Linux, well, life for the kernel and boot strap developers would be made alot easier since there is no longer the jumping between the different modes just to get the damn thing booting.
It seems that SUN afterall made a good decision when keeping SPARC close at heart even when everybody said that they where !doomed! with the release of Itanium.
Just a thought
kaiwai,
At least for now, AMD has limited resources. It’s nothing short of a miracle that they managed to push the x86-64 standard and didn’t financially sink in the process.
EFI (or some sort of it) will replace the current BIOS real mode bootstap… sadly enough it’ll be set by Microsoft (in the longhorn) and will include numerous loopholes that will be used by Microsoft to further strengthen their hold on the OS market. (Let alone the Viruses that will flourish…)
Gilboa
At least for now, AMD has limited resources. It’s nothing short of a miracle that they managed to push the x86-64 standard and didn’t financially sink in the process.
Not entirely true, I mean, even if you take the 64bit factor out of the equation, you’re still left with a very fast processor; I think the 64bitness is more the icing on the cake rather than the actual cake itself.
The fact is, SUN and AMD have a VERY close relationship, whose to say that SUN couldn’t license OpenBoot source code to AMD? from what I understand, only a very small amount of the OpenBoot code is actually achitecture dependent, if AMD and SUN worked together and licensed their firmware to motherboard manufactures are a cheaper price than BIOS producers, then you’d find the technology would catch on.
As for Microsoft, anything that would make their product easier for the end user from the point of view of hardware/operating system integration, they’ll support.
EFI (or some sort of it) will replace the current BIOS real mode bootstap… sadly enough it’ll be set by Microsoft (in the longhorn) and will include numerous loopholes that will be used by Microsoft to further strengthen their hold on the OS market. (Let alone the Viruses that will flourish…)
Not necessarily, EFI is an openstandard, the source is open etc. You’ve actually got a greater risk of that occuring now in regards to the BIOS than if there was a huge shift to EFI and OpenBoot.
Yet more etymologies of ’86’:
http://www.chungdong.or.kr/middroom/luxellen/others_1.htm#Eighty-Si…