Itanium Will Only Be Partly Supported by Longhorn

ver.sicher.ungsvergleich writes "Although stopping short of pulling the plug entirely on Itanium, MS has said that Longhorn will only be able to work for a limited number of higher-end jobs. On the positive side, Microsoft does see a future for the chip, but that 'big iron' slot is not exactly what Chipzilla envisioned as Itanium's future."

Comments

I think it's time to pull the plug

[roman_mir]

on Longhorn.

Where the hell is Netcraft when you need it?

Is this really a big deal?

[CTho9305]

Sure, Intel may have originally hoped to migrate the world to IA64, but given the wild success of AMD64 in bringing 64-bit to the x86 world, it doesn't look like that's happening. The Itanium chips Intel is releasing are obviously not aimed at tasks that could be handled by a 386 with some SCSI drives ("fax server"? a file server?)... who is going to use a multi-thousand-dollar CPU for anything other than database|web|high-end server anyway?

Re:Is this really a big deal?

[jiushao]

Are there reasons other than poor support from Micorsoft for Itanics massive failure? Is it a poor arcitecture?

In the world of good compilers originally envisioned by HP/Intel: No.
In practice: Yes.

Why? Because compilers aren't nearly as good as HP/Intel hoped but state of the art Out-of-Order processors are great. There is only so much theorethically possible ILP to extract in regular code, and good OoO chips extract most of it in an automatic fashion from existing code. So the hardware guys did a better job here than the software guys, and the Itanium bet on software.

To clarify on OoO processors doing most of the possible work in extracting ILP: Even if the instruction window was increased to infinity (that is, all ILP is always found) it would still not yield dramatically much better performance (I have seen estimates of about 25% best-case). So even with a perfect compiler there is just not much to gain, and we do not have perfect compilers. This very high level of performance in extracting ILP is what is forcing the new shift to TLP with architectures like the Sun Niagara.

I don't think we should be pleased that the Itanium failed. As I have often discussed in the past I think Intel really deserves a lot of credit, they are the undisputed top dog in the market, and despite that they are also one of the companies that consistently attempt new different approaches in high-profile products. Neither the Itanium nor the Netburst (which really is interesting and innovative technology) worked out well, but it is trying things that makes technology move forward.

That's not to say that AMD is a bad company, they managed to make the best x86 implementation yet, which is great (though I still consider the K7 to have been the golden age since their pricing structure truly was incredible then).

New Design Getting Flushed Away

[SumDog]

It is rare when someone comes up with an entirely new architecture and instruction set. The IA64 was a complete break and had it been pushed correctly, AMD would be rushing to make IA64 clones instead of Intel supporting AMDs 64-bit extension.

If I remember correctly, the IA64 has 128 general purpose registers and 128 floating point registers. It's a load/store machine and it's pretty close to a RISC arch (really it's an "very long word" instruction set, but lets not get picky).

It was a chance to make a clean break from the old 32-bit legacy chips, however the price was and is too high and AMDs are cheaper and still very powerful.

I really hope this chip doesn't die off. At least with limited support in the new Windows, it will still have a strong server market, but I think a lot of companies are going to be afraid to buy because of running into compatibility problems. I know at where I work, we'd like to have servers that can do anything/general purpose. You put a limit on what the OS can do and then you're afraid of old legacy or propriety software not working correctly

But hey as long as you use Linux, the IA64 is fairly well supported, and it will be better supported in Linux than in Windows!

Sumdog

Re:New Design Getting Flushed Away

[sean23007]

Actually AMD probably wouldn't be rushing to make IA64 clones. They wouldn't be allowed to. Decades ago, Intel was forced to license the x86 technology to another manufacturer to prevent a 100% monopoly in the general purpose consumer chip market. Obviously, Intel doesn't like this at all, since it basically means they can't beat AMD as long as the world is still on x86. (They can still hold the lion's share of the market and make a metric shit ton of money, but they can't win, because AMD has to be there.)

One of the lesser known reasons for Intel's plan to develop and push the Itanium was that it would be a clean break with x86, which means that AMD would not be allowed to make them. Intel would be the only supplier allowed to make the chip. Then they'd get sued for it, and would settle by giving rights to manufacture them to some small company with one fab that's a generation or two behind. AMD would have been stuck with x86, and Intel would have won. (Bear in mind that if the switch had been successful, Itanium would have been adopted long before x86-64 and the Opteron were developed.)

Frankly, I'm glad the Itanium failed. Even though it's a pretty cool chip with an interesting design, I'd rather have Opterons available than not.

Wouldn't it be funny...

[mosel-saar-ruwer]

...or at least ironic, if the only operating system that gained a foothold on Itanium proved to be...

[drumroll please]

...VMS?!?

I mean, talk about a soap opera:

1) David Cutler leaves DEC for M$FT, where he unveils a VMS++, dubbed "Windows NT".

2) DEC invents Alpha, but the only way they can garner any OS support for it is by porting their own VMS and positioning Alpha as a replacement for VAX.

3) DEC sues M$FT to port "Windows NT/VMS++" to Alpha.

4) Intel & HP enter into a partnership to build a next generation super-chip.

5) Compaq purchases DEC.

6) Compaq sues Intel for theft of much of the intellectual property that went into Alpha.

7) Intel settles with Compaq by purchasing the manufacturing rights to Alpha.

8) HP purchases Compaq.

9) HP cancels Alpha and announces that the new upgrade path for VMS customers is Itanium.

10) M$FT announces an end to support for "Windows NT/VMS++" on Itanium, but then backtracks, and agrees to partial support.

Who knows what the moral of this story is?

Maybe: Hardware comes and hardware goes, but software is forever?

Re:Fortran programmers don't need (or want) Window

[slavemowgli]

I think you fail to realise an important thing: change typically happens by evolution, not revolution, and that's even more true when there's a multi-billion dollar industry involved. Do you honestly believe that everyone's just going to throw every system they have away?

It's not gonna happen. The industry likes migration/upgrade paths, and in 90% of all cases, a design that extends is gonna win over one that outright replaces.

Intel seems to have been unwilling to face that fact, but what they failed to realise is that their monopoly is not big enough to simply force change on people - rather, their move just gave AMD etc. an opportunity to slowly but steadily chip away at that monopoly.

From a market perspective, that's a good thing, of course - but if I was an Intel shareholder, I'd demand that heads roll for this gross mismanagement in the top executive floor.

Re:Fortran programmers don't need (or want) Window

[TheRaven64]

The Itanium was the only realistic chance we had to get away from the x86 for the forseeable future, and the designers blew it.

PowerPC/POWER is still viable, and IBM may have another go at putting them in consumer machines if an OS that runs on PPC becomes popular in the desktop space.

ARM-derived chips are still going strong. At IDF there was an XScale chip demo'd that ran at 1.25GHz - probably fast enough for 90% of users.

Alpha remains my all time favourite architecture - pure 64-bit, and the PAL code concept is remarkably elegant.