[geeks] more itanium floppyness...

[Peter L. Wargo]

Subject: FW: John Dvorak- "Itanium will turn out to be a flop"
Date: Thu, 7 Mar 2002 11:08:34 -0800 
MIME-Version: 1.0


Respected industry pundit John Dvorak says it bet on Itanium in his 
latest editorial. I interviewed with John live on the Silicon Spin TV 
show at the end of February.   About 1/4th of our interview was on 
Hammer vs. Itanium.

Happy reading.


Itanium Conundrum Dept.
March 5, 2002
By John C. Dvorak <http://www.pcmag.com/author_bio/0,3055,a=123,00.asp> 

Itanium Conundrum Dept.: Intel has to be concerned that its Itanium 
chip, long touted as the greatest thing since sliced bread, will turn out 
to be a flop, not unlike the once-idealized 432-MHz chip. The Itanium, 
which is the official name for the 64-bit chip codeveloped with Hewlett-
Packard, may have no appeal for the largest market segments. It's
beginning 
to look more like an expensive specialty chip for servers in a market
glut. 
This means it's possible that the mass market will eventually rock over 
toward the AMD x86-64 chip, which was designed both to compete with the 
Itanium and to be a fully compatible upgrade to the Athlon/Pentium 
architecture.  This situation has resulted in the emergence of numerous 
rumors that Intel is cloning the AMD 64-bit Hammer architecture and
readying 
its chip (code-named Yamhill) to roll into the marketplace -in case of an 
emergency. Cynics are referring to this as the Intel 64-bit Plan B
strategy. 
This would mark the first time that Intel has had to play performance 
catch-up and follow the AMD lead in design, which would be the most 
humiliating blow for Intel, ever.

How did Intel get into this pickle? I believe it began when the company 
was too aggressive in dealing with its competition and felt that there 
was no room in the market for Alpha, MIPS, or SPARC, not to mention AMD. 
Thus, Intel promised the be-all, end-all chip of the future some years
back. 
The chip was then dubbed the IA-64 architecture and code-named Merced, and 
nearly every manufacturer in the world lined up behind it because of the 
Intel hype. 

In all my years of following this industry, I have never seen such sheep
like behavior.  Not that I was surprised.   Sun Microsystems was even 
included for a little while. But Sun got smart fast and backed out of 
the herd.

The grand promises Intel made had the effect of ending HP's PA-WW chip 
development that was under way in the early 1990s and killing PA-RISC,
while pummeling any hopes for Alpha and MIPS.  By 1997, everyone had 
signed on, and in 1998, some people were even suggesting that Apple 
should drop the PowerPC in favor of the IA-64. Thus, it was easy for 
Intel to begin to believe its own publicity when the company saw it had 
the future market for servers and perhaps most of the computer world 
all to itself.
 
In the process, Intel may have become lax and taken its eye off the 
x86 ball. The irony here is that Intel built a patent blockade around 
the IA-64 architecture that experts say is impermeable. This literally 
forced AMD to continue on the old x86 treadmill, with faster x86
processors 
that add new features. Hence, AMD developed the AMD x86-64 Hammer designs, 
which is what the bulk of users want anyway. These designs could leave
Intel 
with a high-priced, who-needs-it architecture that will forever remain 
high-end and subject to competition from below-from AMD, from older Intel 
designs, and possibly from an Intel clone of the AMD processor. Wouldn't 
that be funny!

Genuinely Interesting Web Site Dept.: The pre-Internet era of IA-64 
hype was a time of weird assertions.  And since the business was heating
up, 
you had the "in a hurricane, even a pig can fly" phenomenon. Thus we saw 
Intel's "Pinky and The Brain" scheme to take over the world with the
IA-64. 
Enter a cool research tool called The Wayback Machine, at www.archive.org. 
The tool links to a database of archived Web sites and started in 1996.
Articles that were run on the Web in 1996 are quite revealing. For
example, 
Intel's then-CEO Andy Grove, while speaking at the 1996 Intel Internet 
Strategy Day, claimed that TV streamed to the PC would become a major
trend, 
and he expected average U.S. PC viewing hours to 
overtake average television viewing hours by the year 2000. These kinds of 
crackpot assertions, which became rampant during this period, helped 
contribute to the Internet bubble. Check out the site for some good
laughs.



-----
Peter L. Wargo
pwargo at basenji.com
Owner/operator of basenji.com.
[This space for rent]