John F., about the WSJ article:
It seems to me the article is reasonably accurate.
While the use of the word "copying" is not precisely correct, the article
does give a feel for relative technology positioning.
Unless you have never used anything except a Wintel box, you are aware that
Intel has not been a technology leader in the high-performance processor
category.
Until fairly recently, comparisons of high-end workstations to typical
Intel-based boxes has been a bad joke. There are Unix-based scientific
workstations available that are more powerful, more reliable, and can be
configured with considerably greater memory capacity, than any Intel-based
system. This has been true for some time.
The most telling anecdote I've heard is that Intel tried to coerce
engineers on the pentium design team into using Intel-boxes. The engineers
rebelled and the design work was instead done on IBM RS/6000s. The new
Intel chips of course were to compete with RS/6000 descendants (PowerPC).
More irony followed :-)
The research lab I am in has a DEC Alpha workstation, 300MHz processor, 1Gb(!) of RAM. For scientific applications it is something like 2-4 times
faster than a top-of-the-line P6-200 system is supposed to be. And just
try to find a single-processor P6 system configurable with >1Gb of memory.
This machine is over a year old and is not the most powerful currently
available. Now, how much did it cost? Please don't ask.
But the situation is changing fast. Intel is going to be playing with the
big boys in the next processor generations -- maybe P6, definitely P7. No
surprise here, convergence of PC and workstation has been on the horizon
for a long time.
As far as "long-term, original" research, Intel certainly has never been in
the same league as IBM, or AT&T in its heyday, nor are they likely to be in
the future.
But I have seen some anecdotal evidence that Intel is getting more
aggressive in recruiting top research people, in a variety of application
areas (not just chip technology).
What are the competitive implications?
If through open software platforms or whatever Intel loses its de facto
mass-market monopoly, and has to compete on technology alone, they are
going to see severe margin pressure. Lots of people can make good fast
chips. Many of them are making chips that are faster then Intel's.
OTOH, if Intel can retain this monopoly, their size-of-market and cost
advantages could effectively put Sun, Dec, HP, IBM out of the
workstation/server business entirely.
From what rumors I hear about P7 development, Intel is well
on top of the situation from a technology standpoint.
As is usual for the WSJ, the information contained in the article is more
historical than anything. It's not news.
--joel |
