Slashdot Mirror
Intel Dumps Iitanium's x86 Hardware Compatibility
Spinlock_1977 writes "C|Net is running a story that Intel is going back to software x86 emulation on Itanium in order to reclaim chip real estate. (room for another 9MB of cache?)
One notable quote about x86 emulation: 'Basically, no one ever used hardware-based IA-32 execution, so better to use the silicon for something else,' said Illuminata analyst Gordon Haff. 'Of course, basically no one uses software-based emulation either, but at least that doesn't cost chip real estate.'"
I think the days of it mattering what the exact instruction set is are pretty much over.
This is very old news. Various sources and die photos have showed this for more than a year... ...and no one cares.
The die space reclaimed was somewhat significant, and the software emulation is faster than the hardware emulation.
-- Have you ever imagined a world with no hypothetical situations?
Yes, there is a whole family, and I think they sold like 2 units last year. I kid, I kid. However:
For 2004: Intel misses Itanium sales mark by $26.6bn
And this is biased, but more current: Reality Check: Itanium - A Sound Bet for the Future?
From that article's sidebar:
Granted, Sun is pretty well biased, but itanic looks like it's long since sunk to me.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Although it may not run as cool and will use around 100W of peak power (+/- 10%), Montecito will be dual-core and run at around 1.6-1.8GHz at launch. 100W is less power than the current high-end Xeon MP and just over the Sun US-IV+ processors, but each of the two cores gets 12MB of L3 cache. Compare that to the ~120-130W power envelope of the mid/high-end Itanium processors available right now.
Granted, the Itanium is not the fastest enterprise-focused processor out there, but at least they are trying to reduce the overall power consumption and heat generation of the next-gen Itaniums.
For the workload I deal with everyday, the Opteron and US-T1 are better suited.
Actually it was.
THe Itanium has an impressive FPU set to make it fast in certain situations like scientific apps but other than that its been subpar and expensive compared to cheaper Xeon's and other risc processors like IBM's Power.
The Mercedes was supposed to be the new Xeon for servers and workstations and NT with its portable HAL was supposed to eventually migrate to the chip and would overtake the desktop market after teh apps appeared. It was all teh rage at Zdnet computer magazines and was termed the next big thing by Intel.
Mercedes (Itanium) was Intel's kneejerk reaction the Powerpc which was a threat in the early to mid 1990's.
Mercedes was supposed to be here by 1997 and it still hasn't delivered its promise as the next platform out of x86. I get modded down every time I talk about the Itanium but the engineering specs and things that just went wrong are stunning. The first version never came out because it was too slow which delayed it for another 2 years for its next version. The next one(first publically released) had to be overclocked and require a 1 pound heatsink with a fan that sounded like a jet engine just to be nominal. Intel loaded it with huge cache to make it go faster in certain benchmarks which brought up the price and size of the chip. HP killed teh alpha next to make it look like the Itanium wasn't as slow in comparison.
Carly Fiona did alot of strange things in terms of arm struggling Intel and forcing the cancellation of the Alpha in favor of the ITanium because she lacked the concept of "sunkin costs" or bad investments. Itanium does not make HP or Intel really any money. I am not talking about its technolical abilities but from a business standpoint.
Switching to an alpha would have been better and cheaper with stronger performance. Windows2k was out in beta3 on it at the same time as x86 and Linux and BSD support was already strong. Not to mention HP already had VMS ported to it.
The premise behind VLIW was that as chip says limits things you can do with hardware there needs to be a shift to software and leave the fast ram (cache) on the chip. Turns out huge improvements in fabrication made this argument false and somethings like branch predictions just can't be done in software. Fast dedicated hardware is faster than software. Who came up with this idea of moving optimization to software?
If I were Intel I would can Itanium and start over. Transmeta had something interesting and the new PentiumM's are rumored to be designed by a small Isreali firm bought by Intel with similiar technology. I think that is the next big thing.
Of course a newer Alpha would rock. Sigh
http://saveie6.com/
I find is odd that Intel keeps backtracking to its 20 year old Pentium Pro design. Both of their recent high-budget designs, the P4 and the Itanium proved to be a flop to some extent, while the P6/Pentium Pro/PII/PIII/Centrino/Banias architecture has scaled amazingly well since its humble 200 MHz beginnings.
Was there a generation change at the design offices? What else could have caused the most prominent chip design firm to lose its ability to do solid engineering? Granted even the golden boys created a dead end (i960) architecture, it wasn't quite as expensive a mistake as Itanium...
I remember that in the nineties new chip generations would be popping up left and right, each of them offering some really unique and cool innovation in terms of memory management, execution streamlining or heat management. But Transmeta was the last memorable innovation, and since then everyone seems to be exclusively focused on cache megabytes and transistor sizes. I would love to see real experimentation and innovation reintroduced in the CPU arena...