Remember Arthur, later RISC OS, on the Acorn Archimedes?

I consider a move away from x86 to be a good thing, personally. Many people like to complain about compatibility, but the simple fact remains that, these days, software tends to be compiled - not written in assembler. So the processor architecture shouldn't matter anymore.

How many variants of hardware does Linux work on? Loads. Not because it is a better OS, but simply because a decision was made, early on, that the source code should be readily available. The same is true for BSD (NetBSD is arguably the most portable - I remember first running it on my Amiga 1200, back in the day.)

For the server end, open source tends to go with the territory, these days - so unless you are running something really proprietory, compatibility should not be an issue for the vast majority of people. After all, when was the last time you cared about processor architecture when you compiled Apache?

For all those who lament the loss of many RISC architectures caused by Itanium, lets not forget Arm. Its one of the best RISC architectures ever created. Although it has not yet made its way into desktop PC's, it might yet still do so one day.

Honestly, x86 isn't x86 anymore is it.

Since the AMD k5 and the pentium ii (or was it the pentium pro?) most x86 chips have been RISC cores on the inside, with instruction translation on the outside.

QUOTE:
because, unlike most in this industry, I expect more, and ideologically I am not going to settle for anything less.
END QUOTE.

You should dig up the old bob Colewell video after he was kicked from intel for going against their ideology of ever higher frequencies on the P4. He has some choice words for ideology driving chip design, as was the case with both the P4 and the Itanium.

Shouldn't it be iTanium? It is overpriced. It runs like crap. And there is no reason to buy it other than some almost masochist snobbish satisfaction.

Sounds like an Apple product to me...

Andy, you couldn't have missed my point more.

The point is that there's an unnecessary and unwise aversion to non-x86 architectures.

Tell me.. would you, personally, be happy with a hugely kludged-up rendition of x86 doing the bulk of our computing 25+ years from now? I sure wouldn't.. because, unlike most in this industry, I expect more, and ideologically I am not going to settle for anything less.

Yes, HP did agree to kill PA-RISC and work with Intel. The point is that the move left it 100% dependent on Intel, whether x86 or Itanic.

Compaq acquired DEC in 1998, but it wasn't until mid-2001, simultaneous with Merced delivery, that it effectively killed Alpha by sending the first group of Alpha engineers to Intel. HP then acquired Compaq later that year and sealed the deal.

SGI was on the way down when it made the deal with Intel, but the MIPS roadmap was substantially pulled back at that time and the delays in Itanic rollout further damaged SGI's competitiveness.

As to AMD, Intel saw to it through dirty tricks that AMD64 never established a foothold in low-end enterprise, but more to my original point AMD never had a high-end enterprise technology solution to rival Itanic or POWER.

Intel thus effectively froze the low-end enterprise market to wait until it transitioned to 64-bit x86 processors, having blown its published plan to deliver Itanics to this segment.

Intel says it doubles performance. Well from 2 to 4 cores thats clear. On the other side this implies: no frequency increase, no internal speed-ups either.
Summary: Itanic will share the fate of the Intel iapx432 - if someone with memory remembers...

DEC was bought by HP-Compaq and then killed.

Itanium didn't kill the Alpha, HP-Compaq did.

HP killing PA-Risc was a no brainer as they had a large hand in developing Itanium. As for Alpha and MIPS, they were dead before Itanium development even started. AMD did indeed have an answer. AMD64. True, it was still x86, but an answer none the less. An answer that Intel later had to copy and license from AMD.

If Itanium was such a success, why did HP drop it and leave Intel with the bill?

We for one welcome our new "kitchen sink" overlord. We're sure he will be quite the "phone book killer". No need to launch a rocket scientist to the Moon, now you've found the Itanium, a "meta-cognitive processor".

Tick Tock, on the clock
But the party don’t stop
NO
Woah-oh oh oh
Woah-oh oh oh

And POWER 7 is coming and the architecture base of it is still alive and kicking.

Itanic was a stunning success. It convinced HP to kill PA-RISC, DEC to kill Alpha, and SGI to kill MIPS. And it left AMD without an answer. Never mind that Nehalem performance renders Itanic irrelevant. All in all, a great investment by Intel.

QUOTE:

Yes, God forbid someone try to make something work *other* than x86 without anyone immediately poo-pooing it.

END QUOTE:

The Itanium platform has been around for 9-years and still no-one wants it for anything other than the big-tin market.

Intel cocked-up totally on this venture.

Its not x86 so everything has to be re-coded or runs in emulation.

Its not any faster than its existing competitors and needs everything to be re-coded.

Desktop and "lesser" 64-bit x86 CPU's stole a huge amount of the thunder (thanks AMD).

Cluster computing took off.

The Itanium platform was doomed by all of the above amonst others from the moment it was launched, hardly a surprise that it was called the "Itanic" for the best part of a decade.

Yes, God forbid someone try to make something work *other* than x86 without anyone immediately poo-pooing it.

[sarcasm] Because we all know x86 is the be-all, end-all of server computing. [/sarcasm]