bb@tetons.UUCP (Bob Blau) (02/25/89)
This is the continuing saga of a startup company and its quest for
breakthroughs.
Background: The head of Imaginary Computing Corporation's architecture
department asked the research labs to concentrate on
achieving short term breakthroughs in chip density, speed,
and low power consumption (in that order.)
They came back and said that manufacturing will be able to produce
chips with 1.5 million transistors using a .75 micron CMOS process in
1991. The chips will be capable of a 60 MHz clock rate, with 3.5 watts
power consumption and a 250 pin package.
Not quite a breakthrough, but not too shabby!
Imagine that you're the president of Imaginary Computers Corp., and
must decide what product and architecture you will build with this
technology. Your task is to explain to us, the board of directors,
what your decision is.
First you ask your marketing people for advice. They all agree that
compatibility with a big software base is the most important goal.
But they disagree on which base.
The first marketeer suggests, "We could be the first to sell a Vax
plug compatible system!"
"No, no," another replies, "if you want to go after the fastest
growing market segment you should build a Crayette on a chip."
"You people don't know what big is," retorts a third marketeer. "If
you want a BIG installed base you should implement the 370/ESA
architecture on a chip."
"Well," a fourth says definitively, "you could gain access to the
biggest market by far if you implement a 386, floating point unit,
super VGA controller, floppy and hard disk controller, and bus logic
on a chip."
The discussion then degenerates into a wild argument about volumes,
margins, and competition.
So next you go to the architecture department. They all agree that
putting a RISC core with a memory management unit, floating point
unit and instruction and data caches on the chip is the only way
to go. However, they strongly disagree on what to do with all of the
leftover real estate.
The first engineer suggests "The best use of the extra area would be
to put in really large caches."
"No, no," another replies, "a better use would be to implement lots
of parallel functional units to get the cycles per instruction (CPI)
below 1."
"Au contraire," shouts a third engineer, "the most efficient usage
would be to squeeze two RISC systems on the chip in a multiprocessing
configuration."
"If you could do that," sneers a fourth, "then you'd be better off
halving the chip area and making cheaper RISC chips."
The discussion then degenerates into a RISC war about miss ratios, CPI,
and process granularity.
So finally you go to the research labs. They all agree that you
could do some really neat stuff with the technology. But they disagree
about what's best.
The first scientist suggests, "Hey, you could implement a 64 bit
architecture, with 64 bit integers and 128 bit floating point
hardware!"
"No, no," the second replies, "a fast scalar RISC with an attached
vector unit on chip would be perfect."
"Hogwash," the third scientist scoffs, "a VLIW on a chip would be a
screamin' demon."
"If you really want performance," the fourth proclaims, "you'd go to
massive parallelism, putting a 64 processor (with memory) connection
machine on a chip."
The discussion then degenerates into vigorous hand waving about the
best way to implement various scientific computing methods.
You go back to your office, think about it for awhile, and write
your report. What do you decide the company should make? Why?
What will give the most performance?
What will make the most money?
What will be the most fun?
This is an imaginary scenario. The situation, characters and plot are
entirely fictional.
Apply: Standard disclaimers
--
Bob Blau Amdahl Corporation 143 N. 2 E., Rexburg, Idaho 83440
UUCP:{ames,decwrl,sun,uunet}!amdahl!tetons!bb (208) 356-8915
INTERNET: bb@tetons.idaho.amdahl.comlearn@igloo.Scum.COM (william vajk) (02/25/89)
In article <745@tetons.UUCP>, bb@tetons.UUCP (Bob Blau) writes: > What do you decide the company should make? Why? > What will give the most performance? > What will make the most money? > What will be the most fun? Make no mistake, this is a marketing exercise. Items 1 and 3 are the only ones worth considering. Bill Vajk | A person of quality is never intimidated. learn@igloo |