They should have called this monster Tukzilla!

Gives all those former DEC Alpha engineers a reason to retain their Intel Principle Engineer Baller Status.

Itanium has always had an advantage in performance-per-clock. 
But Itanium is now being blown away by x86 in overall performance simply because the x86 clocks higher - Itanium 2 9xxx caps out at 1.6GHz with a 533MHz FSB, so it's not too surprising that a 3.0GHz Xeon 53xx with 1333MHz FSB is beating it (the multiprocessor 7000 series Xeons don't scale that high yet). 
The only advantage Itanium has right now is that the caches are enormous, but even here x86 is catching up.

The trouble is that Itanium isn't mass-market at all, so the return on investment of ramping up clockspeed is pretty poor. x86 has been eating everyone else's lunch for years. 
The only other viable architecture seems to be ARM because of x86's poor power characteristics for mobile applications.

The only other advantage I'm aware of is that Itanium servers like HP's Integrity SuperDome scale out to 64 or even 128 processors (Windows can't handle more than 64 logical cores on one box at present, though), while the largest x64 servers I'm aware of stop at 32. IBM's latest x3850 M2 scales to 16 sockets but supports quad-core Xeons, so you do get 64 cores.

It might be better to be looking at some form of compute clustering or data partitioning rather than a single huge box, though.

I don't understand why Intel keeps pouring money into this technology. It's a complete failure both commercially and technologically.

Intel is at high margins of profit as usual, so I guess they can at least afford a little burn-the-money adventure. 

And they will keep trying just for the sake of teasing the technology itself. ;)

did the Alpha have to die for this hunk of junk? 

HP should be ashamed of itself for killing off one of the best processors of the past 20 years.

i was prairie doggin it after decoding the headline.

good thing i read the Inq in the nude, otherwise, i'd have soiled myself from
laughter.