Diapositive 1 - BigBozoid
Download
Report
Transcript Diapositive 1 - BigBozoid
Big Blue
70% market share from 1950’s
onward
Any new entrant had to:
1) compete with their hardware
2) compete with their software
3) find a niche IBM did not serve already
Just after IBM introduced the
vacuum-tube 709 (1957)
Philco tried to compete with its surfacebarrier transistors
IBM countered the next year
with the 7090 and Philco
failed by 1964
IBM 709
Data processing system
Able to mete out technology at a pace that did
not render installed machines obsolete too
quickly
Kept punch cards alive from the 1930’s
thought the 1960’s and beyond
Only company to make a profit making big
machines in the 1950’s
The Counterbalancing
Influence of the
American Government
Military support also important even as far
back as ENIAC
After Korean War – large increases in
spending for basic research
Before - military decided what research
they want
After - researchers themselves decided
Dynamic Random Access
Memory
one-transistor memory cells
store each single bit of information as an
electrical charge in an electronic circuit
major increases in memory density
widely adopted throughout the industry
in widespread use today
DRAM
1966
Robert H. Dennard
The Minicomputer
Born - Chippewa Falls,
Wisconsin
Adolescent hobby –
electronics
Fought in WWII in Europe
and the Pacific
Returned home to study
electrical engineering and
mathematics
Seymour Cray
1925 - 1996
Started work for ERA and
then UNIVAC
Cray’s early contribution
the CDC 160
Navy Tactical Data System
(NTDS) one of the first
transistorized machines
1960 - Designed the model
160 for CDC to handle
input/output for their 1604
It had 12 bit word length
Able to access primary memory of 8000 words
6.4 micro-second clock cycle
What Cray had invented was,
in fact, a minicomputer.
Minicomputer not a direct
competitor to mainframes
Mainframe
Minicomputer
Operated on 36 bits
Operated on 12 bits
Other registers handled
addressing
indexing
extra digits generated
Instruction codes more
complex
Too big and expensive to
be used for one purpose
only
Small enough and cheap
enough to be bought for
one purpose
Equipment Corporation
Harlan
Anderson
Kenneth Olsen
Old woolen mill in Maynard, Massachusetts
Innovative architectural features
Not just machines with transistors replacing tubes
No channels
I/O proceeded directly to the core memory itself
Corporate culture
180º opposites
Only IBM had the right to modify their leased machines
Digital sold and encouraged user modification
As a result, the PDP series found a multitude of uses
to control and monitor
factories
transportation systems
nuclear power plants
etc
PDP- 8 controlling potato picker
User Manuals
Flip Chip Module
Printed manuals on inexpensive paper
Gave away users manuals for free to anyone
Published detailed specifications of their
products
Conjuncture of performance,
storage, packaging and
price
Word length – 12 bits
6 kb expandable to 48 kb
Used indirect addressing to
increase functional word
size
35,000 additions per second
PDP- 8
Logic modules mounted
on two towers
Discrete circuits not
integrated
$18,000
Over 50,000 sold
PDP-8
www.pdp8.net has a running
PDP8 that anyone can control
through a Java applet, plus a
webcam to show the results
Teletype Corporation
Model 33 ASR
(automatic send and receive)
Input/output device for early
mini-computers
Functioned as a type-writer
Printed onto a roll of
continuous paper
Sent a code indicating what
key was pressed directly into
the computer
6 to 10 characters/second
ESC and CTRL keys
Used ASCII-American Standard Code for Information Interchange
Snow White and the
Seven Dwarves
Or
The Bunch
–General Electric & RCA
Time sharing and
System/360
Each user had the illusion that a
complete machine and its software were
at his or her disposal
Happened in the milliseconds between
the typist’s keystrokes and ability of the
computer to fetch and execute simple
instructions
OS/360
Family of 3 control programs
PCP (Primary Control Program)
processed jobs sequentially
MFT (Multiprogramming with a Fixed
multitasking, fixed
number of concurrent tasks,
each w/ a preset memory
allocation
number of Tasks)
MVT (Multiprogramming with a Variable
varying numbers
of tasks, memory size could
change dynamically
number of Tasks)
JCL (Job Control Language) batchscripting language
State of the art computer graphic rendering in 1964
Digital Mona Lisa
H. Philip Peterson of Control Data Corporation
CDC 3200 computer and a "flying-spot" scanner
The production process - 14 hours to complete
Contained 100,000 pixels plotted using numerals,
sometimes overprinted for density.
Printed on sepia tone paper w/ India ink.
A first of its kind scanning process
1965 Fuzzy logic
to process approximate data
- such as « about 100 »
1965 -fuzzy sets
1973 -analysis of complex
systems and decision
processes
1979 -possibility theory
and soft data analysis.
Lofti Zadeh
University of California, Berkeley
Soft Computing
fuzzy logic
neural network theory
probabilistic reasoning
belief networks
evolutionary computing
DNA computing
chaos theory
parts of learning theory
Achievements of BISC
Berkeley Initiative Soft Computing
fuzzy reasoning (set and logic)
new soft computing algorithms for intelligent, semiunsupervised use of large quantities of complex data
uncertainty analysis
perception-based decision analysis and decision
support systems for risk analysis and management
computing with words
computational theory of perception
precisiated natural language
БЭСМ USSR 1965
BESM stands for "Большая Электронно-Счётная
Машина"
(Bolshaja Elektronno-Schetnaja Mashina) in
Russian,
"Large Electronic-Computing Machine" or simply
"Large Computer".
1965
BASIC (Beginners All Purpose Symbolic Instruction Code)
programming language by Thomas Kurtz
& John Kemeny, Dartmouth College
Packet switching, funded by ARPA was
developed
The first supercomputer, the Control
Data CDC 6600, was developed
The integrated circuit
"The future of integrated electronics is the future
of electronics itself.
The advantages of integration
will bring about a proliferation
of electronics, pushing this
science into many new areas.
Gordon Moore 1965
Integrated circuits will lead to such wonders as
home computers ..."
Co-founded
1968
Cray Supercomputers
Question:
What company makes
the fastest computer?
Answer:
Wherever Seymour
Cray is working now.
Cray I 1976
More supercomputers