Digital PDFs

XX-70D41-9B-1

September 1974

5 pages

Original

0.3MB

Document:	730903 KL15
Order Number:	XX-70D41-9B
Revision:	1
Pages:	5
Original Filename:	http://bitsavers.org/pdf/dec/pdp15/memos/730903_KL15.pdf

OCR Text

womDDmo INTEROFFICE MEMORANDUM
TO:

Distribution

September 3, 1973

FROM:

Dick Clayton

DEPT:

11/45 Marketing

EXT
SUBJ:

WILD IDEA FRO

C-.
$- .

DATE:

. 3638

$EP 61
973

For sane time now the PDP-15 team has been looking for bright ideas
that are low cost. high payoff. sell significantly more POP-15
oriented hardware and software. and don't appreciably interfere with
other segments of DEC's business. Some of these items included
using the 8K DEC standard memory. software enhancements. and of
course the very successful Unichannel project. The nature of most
of these projects has been aimed at six month payback and high margin •

In searching for what next. the concept of recoding the microprocessor of the KL10 came to tf:le surface. It is obvious to everyone
this idea could never fly. yet it isn't quite so cut and dried as one
might like (if you want life sim~le!) • .
I am personally enthused about the thought process and the idea appears
to demand a bit more digging. As a cold hearted businessman it will
probably by put on the shelf. As the manager of a group who is willing
to overturn any stone in their quest for good profit as part of a consistent spectrum of DEC products I am very proud of the attached proposal.· I am pushing it a few man weeks further to better understand if
anything might come of it. We are sending this memo around at this time
to let you know what one group is thinking about as well as again request
any help you can provide to expand our horizons or clarify...J)ur thinking •
. Thanks.
DC:vb
attachment

D,/)C C,

momo mD
TO:

SUBJ:

INTEROFFICE

Distribution

f'v1EI\/lDRANDUM

DATE:

August 30, 1973

FROM:

Gary Cole

DEPT:

15 Ma rketi ng

EXT :

4187

THE KL 15
The PDP-15 has been a fine, profitable machine which has developed
a superb set of software and excellent customer base. But it is
rapidly becomming technically obsolete and needs new electronics to
continue over the long haul. To engineer a new processor I/O system, etc. not only involves a several year lead time, but also at
least $2 million. Moreover, we would arrive at a machine with only
marginal appeal over the high end 11 family.
Out of this has arisen an idea which is so attractive that it demands
our collecticYe scrutiny.
.
Bmu1ate the PDP-15 on the KL10 processor.
Since the PDP-10 is a 36 bit machine the 18 bit PDP-15 data structures fit nicely onto ·the data paths.
There are several areas to consider:
1.

Technical - can it be done without redesign of the KL10
processor.
Mike McCarthy, Dave Rosenblum and I did a brief study of
the KL10 microprocessor and concluded that it definitely
could be done and would yield some· performance increase
over the existing PDP-15.

Marketing - Can it aid DEC in the medium scale market?
Can it be sold easily by the DEC sales force? Can it
develop into a large busines~?

First of all this will be a uniquely saleable computer, independent of
all other features, because it can be field upgraded to a PDP-10!
So, industrial concerns will buy KL15 1 s because it's good management
sense.
University departments will buy KL1S ' s because it represents such a
good long term investment.
Labs will buy KL15 1 s because it~s a way to get to a big computer by
the bac k door.

Page 2

KL15

.All this natural appeal." combined with gr.eat PDP-15 software and high
processor "{~ 2 )Jsec floating add} should make the KL15 a superbly saleable product.
Given DEC's sales volume 18 months from now a minimum volume projection
would be 500 systems (@ l20K)· per year ie., $60 million.
~

Cost - The KL10 is not an inexpensive device, hence the minimum sales
price will be at least $lOOK for a 64·K DOSjRSX system using PDP-l1 peripherals. Thus, we will stay definitely in the medium scale area.
PDP-10 Competition: The basic difference between PDP-10 and PDP-15
would become system size and software. If the performance of the KL15
was too high, we could always introduce delays in the microcode. The
real effect would seem to be more pro 10 than anti 10 since in the.end
there will be a lot more people having PDP-lOis.
PDP-l1j45 Competition: The KL15 will be capable of competing head-on
with current and anticipated "big midicomputers" such as the Datacraft
24·bit series and possible 32 bit Novas or Modcomps. This will save us
from having to do a usmall configuration" set of PDP-10 software (ie:
another RSX class monitor) which might otherwise be r~quired to compete.
Such a software development alone would cost several times more than the
.entire KL 15 project.
Corporate Benefits - Increased volume on KL10 processors will help production costs. Increased volume on KL10's will allow more field service
personnel and hence better customer service~ The KL15 will use all common
DEC components except software and a few ROM control boards so it should
help the entire prodcution organization. The KL15 will take some of the
pressure off the high end 11 family in the marketplace and immensely
strengthen DEC's competitive posture in this area.
The KL15 will bring in a large sales volume and should be highly profitable since its engineering cost will be low (both hardware and software)
as compared to any discrete cpu/system product. Because it is in the
high end of the midi marketplace it will be less susceptible to obsolence
than small machines and so have a long product life (5 years? 61). Moreover, the high income will help fund memory and peripheral development
which will benefit both 10 and 15.
.
Schedule - "The KL15 project should be phased about 6 months after the
KL10 to avoid interference with PDP-10 operations - this means that the
machine would be announoed about September 1974 for delivery March 1975.
As a first step, I propose the funding of a design project commencing
October 30, 1973 and ending March 31, 1974. This project would
1.

Prove technical feasibility by generating a preliminary
microcode for the machine using simulation.

DIGITAL

EQUIPMENT

CORPORATION

KL15

Page 3
2.

Provide detailed performance data for the finished product.

Develop a production schedule.

Develop detailed cost data both of production system cost
and development project cost.

Select the proper approach for I/O on this machine.

Decide what changes have to be made ( if any) in each price
of existing software and develop software project plans for
such changes •.

Develop system packaging and configuration for marketing
analysis and system pricing.

Build a technical team which can carry the project through
the next two years •

. 9.

Coordinate with all other groups in engineering, production,
programming and marketing which need to be involved.

10.

Establish the K£..15 design review committee and the project
schedul e.

This project team should consist of five people.
1.

Project Leader

Production Engineer

Electrical/Mechanical Engineer

4. Microcode Programmer
5.

PDP-15 System/Microcode Programmer

This project will cost approximately $75K and would finish about 1 year
before the first production units are·to be ready.
GC:vb

DIGITAL

EQUIPMENT

CORPORATION

Distribution:
Dick Devlin

Gene Smith

Gordon Bell

Phil Markell

Win Hindle

Bob Puffer

Ted Johnson

John Fisher

Pete Kaufman

Bill Thompson

Dick Clayton

Dave Knoll

larry Portner

John Holman

Claude Proteau

Dennis Burke

Mike McCarthy

Operations Committee

Ed Morri SOrf

Ed Schwartz

Alan Kotok

Jack Smith

Bill Kiesewetter
John Leng
Fred Wilhelm
PDP-15 Marketing
Gerry Moore
J. C•. Peterschmidt
Ed Kramer
Bill long
Brad Vachon
Bill Hogan
Julius Marcus
Jack Shields
Bob Lane
Larry Portner
Irwin Jacobs
DIGITAL

EQUIPMENT

CORPORATION