Al Balmer <albal
...@att.net> writes:
> What timeframe are we talking about. There was an oregano frontend to
> spice in the late 90's.
misc. previous references to this series of trip reports
http://www.garlic.com/~lynn/2001l.html#61 MVS History (all parts)
http://www.garlic.com/~lynn/2006t.html#37 Are there more stupid people in IT than there used to be?
http://www.garlic.com/~lynn/2006n.html#56 AT&T Labs vs. Google Labs - R&D History
http://www.garlic.com/~lynn/2006t.html#37 Are there more stupid people in IT than there used to be?
SPICE reference from part of trip report to CMU 22jul81-24jul81
Shifting Towards Personal Computing
Most of what I said above about the advantages of small machines over
large machines is, I believe, also applicable when comparing personal
machines to small machines. That is, personal machines will probably
have as many advantages over small machines as small machines do over
large machines.
But before I go into that, let me define what I mean by *large*,
*small*, and *personal* machines. My primary criterion in
categorizing a machine as *large* or *small* is the number of
simultaneous users it typically supports. The following table is my
general feeling:
1 user - personal machine
2 to 10 users - very small machine
10 to 25 users - small machine
25 to 50 users - medium machine
50 to 150 users - large machine
150 to 300 users - very large machine
over 300 users - enormous machine
Of course, the larger machines are generally faster and have more
memory and storage as well as more users. But for a machine to be
effective, regardless of the number of users it supports, it must
deliver a certain amount of computing power to each user and there
must be available a certain amount of memory and storage per user. As
suggested above, there seems to be quite a bit of evidence to indicate
that doubling the MIPS, memory, storage, and number of users, results
in less of everything for everyone.
Several people at CMU seem to feel that the natural extension of this
is the personal computer where a fixed amount of compute power and
memory are dedicated to a single user and not subject to load
fluctuations or other user's hardware or software failures. The SPICE
people at CMU seem to feel that such a machine would need about 1 MIPS
of compute power and about 1 Mbyte of memory plus around 10 Mbyte of
DASD part of which should be on a removeable medium.
Current technology seems to fall just short of providing this
combination at an affordable price, but it's close and getting closer.
I tend to agree with the SPICE people that we should be heading toward
personal machines and should begin getting them for our Computer
Science researchers even if they cost 10 times what we'd like to pay
for them
The cost will come down dramatically and we need to get leading edge
experience with machines that will be cost effective 4-5 years from
now which means we'll have to pay much higher prices today. The
alternative is to work with today's cost effective technology and gain
very little experience on how to use tomorrow's.
...
The Computer Science Department
The goal of the computer science department computer facilities
is to optimize the productivity of researchers,
and to provide sufficient cycles for the various research projects.
They want to provide a minimally constrained solution space
for the researchers.
"Solve the problem, then specifiy the requirements"
The department is committed to doing research with equipment that
will be available in 3 to 5 years; thus it costs much more to simulate
that equipment with currently available gear. They view such capital
expenditures as leverage for their researchers. Try to choose
productivity of the users over throughput of the systems - minimize
administrative overhead and constraints.
The computer science department personnel are as follows:
* 40 faculty and researchers
* 80 graduate students
* 55 staff, including administrators, secretaries, programmers,
engineers, operators, technicians
Systems and Their Use
The computer science department has several different
kinds of machines:
* General purpose time sharing
* Project machines
* Personal machines
* research systems
* connectivity and networking
* special resources
...
The SPICE Project
For a complete description of the Spice project, see
"Research in Personal Computing at Carnegie-Mellon University,
Peter G. Hibbard, 25 November 1980, Spice Document S008"
Spice, *Scientific Personal Integrated Computing Environment* is a
major research project currently underway at CMU.
Spice is aimed at increasing user productivity. The environment will:
* Comprise at least 100 personal computers connected in a high
bandwidth network, providing facilities for scientific computing.
* Provide access to shared facilities such as printers and filing
systems.
* Provide a consistent style of user interaction for all the software
components.
* Provide tools to encourage modular extension and enhancement of the
software during its lifetime.
Work started on Spice during the summer of 1979. It is expected that a
version of the system will be available to users by 1983; and
essentially complete by 1985. The computer science department is
heavily committed to using Spice for its principal computing resources
until the 1990's.
The Spice machine is a 2 to 4 mips processor, 1 megabyte of memory, a
100 megabyte disk, a full page APA display, tablet, and a 10 megabit
ethernet connection. While CMU is using the Perq as a Spice
development machine, they feel that it is under powered, and does not
represent what will be needed in the second half of this decade.
The machines will be connnected via the 10 megabit ethernet, and will
share a common file system, with the local disk being managed as a
cache to the primary file server(s). Thus the system will give the
appearance of a large timesharing system, yet have all the advantages
of personal, separate machines.
|
