Transcript Chapter 2 – 1956 to 1964

Chapter 2 – 1956 to 1964
Computing Comes
of Age
IBM 1130
1
Introduction
 Clerks in offices performed many
“busy” tasks- Comptometer (Pg.48)
 Common Problem: Needed to store/
retrieve large amounts of dataquickly and
easily
2
Core Memory – a radical innovation
 Small, donut shaped materials
threaded together with fine wires
 See Description- Pg. 49
 Hysteresis – from Germany after WWII
 Advantages
 Small – non-volatile
 Random Access
 Began to install in existing
computers, e.g. Whirlwind
3
Core Development




German fire-control systems
Aiken’s Mark IV, 1952 - An Wang
ENIAC, 1952 – Burroughs Corp, 2D
Whirlwind, 1952 – Jay Forrester, 3D
 Made it the
“fastest”
4
Core Memory
5
Air Force SAGE
 Semi-Automatic Ground Environment
 Early called “Whirlwind II” – similarity
 Core Memory 8,192 – 32-bit words
 55,000 vacuum tubes per system
 Radar+Aircraft+Telephone+Radio+Ships
 To detect & identify enemy aircraft
 IBM won contract
 Delivered Prototype 1955; 30 more
 Each system = 2 identical computers
 Needed hundreds of thousands high-quality core
6
IBM and SAGE
 ½ Billion in revenue for IBM
 Began producing own core
 1956: IBM passed UNIVAC in
Installations of large systems
7
In the Meantime…..
 While IBM and UNIVAC were
leading, others did get in the
game
 Honeywell
 General Electric (GE)
 RCA
8
Honeywell
 Raytheon failed to deliver late 1940’s
government bid
 Joined with Honeywell, 1955
 1957- Datamatic 1000
 Immediately Obsolete
 Used Tubes, not transistors
 Withdrew; re-entered in 1960’s
9
General Electric (GE)
 1955- leading electronics firm
 $3 billion in sales
 200,000 employees
 1953- OARAC- USAF
 Sr. Management decided not to
market
 Why? IBM was GE’s largest
customer of vacuum tubes
10
GE (continued)
 Late 1950’s –ERMA
 Electronic Recording Machine Accounting
 “1-time project”; transistors + MICR
 1958 - Bank of America & Ronald
Reagan unveiling
 Research excellent but Mgmt. never
committed to computer industry
 1970- sold to Honeywell- $200 million
11
RCA
 $940 million sales; 98,000 employees
 BIZMAC, 1955 (Arnold Spielberg, engineer)





1 full system, few smaller ones
Specialized architecture
Several Hundred tape drives
Specialized processors; sort & search
Failure- behind improvements (tube to
transistors)
 Another specialized failure: UNIVAC File
12
Architecture -- Read Pg. 58-64
 By end of 1960, approx. 6,000 G.P.
computers installed in the U.S.
 Word Length: Prior to core memory,
fetch 1 Bit
 7-12 decimal digits; 30-50 bits
 Long words costly & complex
 Soon various lengths;
 Variable vs. Fixed
 1954: IBM 704-36 bit word length
 Today - not totally standard
13
Architecture Cont.
 Registers: Sets of circuits- 1950’s
 Accumulator; program counter; index
register (pg. 60)
 1956 – British, 7 GP registers, 1 PC
 Addresses
 Single address instructions heavily used
 Then 2 & 3 address schemes
 0 address - Stack architecture
 Later in calculators
14
Architecture Cont.
 I/O Channels - processor
 UNIVAC innovations
 Buffer: to help slow I/O
 Interrupt: I/O when necessary
 Channel: separate processor for I/O
 “Becoming” 2- processor system
 I/O Channel became defining characteristic of
mainframe
 Expensive but necessary
15
Architecture Cont.
 Floating Point Arithmetic
 Hardware (expensive) vs. Software (slow)
 Scientific vs. Commercial – parallel dev.
 IBM 360 – combined both components
 1st Computers in 1940’s had FP
Hardware (Zuse, Bell Labs)
 Co-Processors; incorporated into the
486 chip
16
Transistor
 Bell Labs- early 1950’s
 Replacement for tubes, not reliable for core
 Regulated Monopoly, telephone only
 Released transistor information
(small fee)
 Philco-surface barrier transistor
 Mass produced & reliable
 Leader
 SOLO: 1st general purpose, transistorized
computer in U.S. (for NSA)- 1956 to 1958
17
Transistor (cont.)




TRANSAC; S-2000 (1960)
UNIVAC- Solid State 80
Began Second Generation
1962- Ford bought , Philco out of
computer business
 ** Second Generation
 1962- Ford bought Philco
 Dropped computer business
18
Inventors of Transistor
Shockley (seated),
Bardeen (glasses),
Brattain, in 1946
Nobel Prize, 1956
19
Early Transistors
20
IBM
 By 1960, dominated computer industry
 1952- Justice Dept. alleged anti-trust
violations in punch card business
 1956- Consent Decree
 Must SELL and rent its computers
 Third-party vendors bought & leased IBM
 Stock soared, in spite of critics
 Combination: marketing, manufacturing,
& technical innovations
21
IBM cont’d
 Criticism
 Took innovations from smaller
companies
 704: core, floating-point, FORTRAN was
superior to UNIVAC
 Sales Force + Manufacturing Techniques
+ Field Service  success
22
IBM (cont.)
 Model 305 Disk






Announced 1956; marketed 1957
Pack of 50, 24’’ platters
1200 RPM
5 M characters- Random Access
“Boundary Layer”- air
RAMAC – Random Access Memory Accounting Machine
 1st United Airlines for reservations
 Watson, Jr. “greatest product day…”
23
IBM’s 7094 (early 1960’s)





709 tubes 7090 transistors (USAF)
Mainframe: floor, climate
36-bit word, 150 Kb core
Console – detailed control
Typical Process (p. 73)
 Batch Processing
 Separate 1401 for printing
 $1.6 million - $30,000 month
24
IBM 1401 & 1620 (Late 1950’s)
 Low-end, compact (sold 10K 1401)
 Made possible
by transistors
 Stored program
 Core
 1403 Printer
 Fastest of its time – 600 lpm
25
Conclusion






Second Generation
Transition from tubes to transistors
Core Memories
Disks
Business computing applications
IBM success
26