Jarkko Teppo <jate_at_uwasa.fi> wrote:
> In my ongoing search for all things HP 9000 Series 500 I accidentally
> stumbled into HP Journal. HP has a database on the web with just
> the article names:
>
> http://www.hp.com/hpj/journal.html
How...sad. A couple or three years ago, HP stopped printing the
Journal on paper and went to web-only publication. Then they offered
a reprint service for old issues/articles. Now it seems that they've
stopped that entirely, removed the web-only articles, and outsourced
the reprint service.
> But there seems to be a bug of sorts (I think I ought to report
> it to HP). Try searching for HP 9000 Series 500, and you'll get one
> article:
You want to look at two issues: August 1983 and March 1984.
August 1983:
(0) front cover: photo of FOCUS memory, IOP, and CPU boards with
protective covers removed. p. 1.
(1) VLSI Technology Packs 32-Bit Computer System into a Small Package,
by Joseph W. Beyers, Eugene R. Zeller, and S. Dana Seccombe.
pp. 3-6.
(2) An 18-MHz, 32-Bit VLSI Microprocessor, by Kevin P. Burkhart, Mark
A. Forsyth, Mark E. Hammer, and Darius F. Tanksalvala. pp. 7-8,
10-11.
(3) Instruction Set for a Single-Chip 32-Bit Processor, by James
G. Fiasconaro. pp. 9-10.
(4) VLSI I/O Processor for a 32-Bit Computer System, by Fred J. Gross,
William S. Jaffe, and Donald R. Weiss. pp. 11-14.
(5) High-Performance VLSI Memory System, by Clifford G. Lob, Mark J.
Reed, Joseph P. Fucetola, and Mark A. Ludwig. pp. 14-20.
(6) 18-MHz Clock Distribution System, by Clifford G. Lob and Alexander
O. Elkins. p. 17.
(7) 128K-Bit NMOS Dynamic RAM with Redundancy, by John K. Wheeler,
John R. Spencer, Dale R. Beucler, and Charlie G. Kohlhardt.
pp. 20-24.
(8) Finstrate: A New Concept in VLSI Packaging, by Arun K. Malholtra,
Glen E. Leinbach, Jeffery J. Straw, and Guy R. Wagner. pp. 24-26.
(9) NMOS-III Process Technology, by James M. Mikkelson, Fung-Sun Fei,
Arun K. Maholtra, and S. Dana Seccombe. pp. 27-29.
(10) Two-Layer Refractory Metal IC Process, by James P. Roland, Norman
E. Hendrickson, Daniel D. Kessler, Donald E. Novy Jr., and David
W. Quint. pp. 30-33.
(11) NMOS-III Photolithography, by Howard E. Abraham, Keith
G. Bartlett, Garry L. Hillis, Mark Stolz, and Martin S. Wilson.
pp. 34-37.
March 1984:
(0) front cover: photo of a "solar system" made up of colored balls.
p. 1.
(1) A New 32-Bit VLSI Computer Family: Part II--Software, by Michael
V. Hetrick and Micheal L. Kolesar. pp. 3-6.
(2) HP-UX: Implementation of Unix on the HP 9000 Series 500 Computer
System, by Scoyy W. Y. Wang and Jeff B. Lindberg. pp. 7-15.
(3) An Interactive Run-Time Compiler for Enhanced BASIC Language
Performance, by David M. Landers, Timothy W. Tillson, Jack
D. Cooley, and Richard R. Rupp. pp. 15-21.
(4) A Local Area Network for the HP 9000 Series 500 Computers, by John
J. Balza, H. Michael Wenzel, and James L. Willits. pp. 22-23,
25-27.
(5) Data Communications for a 32-Bit Computer Workstation, by Vincent
C. Jones. pp. 24-25.
(6) A General-Purpose Operating System Kernel for a 32-Bit Computer
System, by Dennis D. Georg, Benjamin D. Osecky, and Stephan
D. Sheid. pp. 28-34.
(7) The Design of a General-Purpose Multiple-Processor System, by
Benjamin D. Osecky, Dennis D. Georg, and Robert J. Bury.
pp. 34-38.
(8) An I/O Subsystem for a 32-Bit Computer Operating System, by Robert M.
Lenk, Charles E. Mear, Jr., and Marcel E. Meier. pp. 38-41.
(9) Coping with Prior Invention, by Donald L. Hammond. p. 44 (rear
cover). About the HP 2225 ThinkJet and HP's decision to work with
Canon when they found Canon had already developed similar ink-jet
printing technology.
> Does anybody have the magazine (and be willing to sell it :) or if
> somebody has it could they give a list of the articles and
> some info on is it worth ordering as a reprint.
As you might guess, the answers are yes (and no), and see above.
These two issues are primarily about the HP 9000 Series 500 family.
The August 1983 issue is about the hardware and processor design and
packaging, and the March 1984 issue is about the software.
As Joe mentioned, the 9000 was originally conceived as a sort of
follow-on to the 9845 high-performance BASIC workstation. Somewhere
in there the marketeers got the idea that it should run this new UNIX
thing too, and so the OS developers got the idea of this lower-layer
OS called "SUN" (no relation to that other UNIX vendor) that provided
basic multiprocessor and multiprogramming support: multiple processes,
memory management, multiprocessor-safe synchronization primitives,
stuff like that. They then implemented a single-user BASIC atop that,
and a port of System III UNIX called HP-UX.
And, well, it's really a fascinating approach to performance from the
very late 1970s and early 1980s. That first issue is largely about
the design of a high-performance 32-bit CISC microprocessor based on
the then-current notions within HP: it's a 32-bit microcoded stack
processor, just like a big classic HP3000.
Speed meant power, and power meant heat, and they needed to do
something about the heat. Ever pull the CPU or memory cards out and
notice how heavy they are for their size? That's because the boards
have a 1mm solid copper core. The ICs are not in packages, they sit
in holes in the Teflon printed-circuit base and are epoxied directly
to the copper core, which dissipates the heat that the ICs generate
when running. Bond-out wires go directly to pads on the
printed-circuit board. That's why there are protective covers on the
boards, there's naked silicon and other fragile stuff underneath.
Astronomical notions are what drive the code names:
SUN: the OS
FOCUS: the processor chip family
MONOCLE: the boot loader ("puts the SUN into FOCUS")
DAWN: the 9020 (520)
CORONA: can't remember, was this the 9040 (540)?
That (especially the "SUN" name of the OS) is why the "solar system"
picture on the cover of the March 1984 issue.
At the time of the August 1983 issue, the family wasn't called the
"Series 500", it was just the "HP 9000" with models 20, 30, and 40
(9020, 9030, 9040). Somewhere between there and March 1984, the
98[123]6 and 9920 were renamed as the HP 9000 Series 200 family
(2[123]6 and 220, respectively) and the models 20, 30, and 40 became
the Series 520, 530, and 540. I think the Series 550 came after this.
-Frank McConnell
Received on Mon Sep 04 2000 - 12:12:46 BST