Computer System

Virendra Singh
Professor
Computer Architecture and Dependable Systems Lab
Department of Electrical Engineering
Indian Institute of Technology Bombay
http://www.ee.iitb.ac.in/~viren/
E-mail: [email protected]
EE-739: Processor Design
Lecture1 (11 Jan 2021) CADSL
 Historic Events
• 1623, 1642: Wilhelm Strickland/Blaise Pascal built a
mechanical counter with carry.

• 1823-34: Charles Babbage designed difference

engine. http://www.youtube.com/watch
v=0anIyVGeWOI&feature=related

11 Jan 2021 EE-739@IITB 2 CADSL

 Babbage’s Difference Engine
• Babbage Difference Engine
Ø Hand-cranked mechanical
computer.
Ø Computed polynomial
functions.
Ø Designed by Charles
Babbage in the early to mid
1800s.
²Arguably the world’s
first computer scientist,
lived 1791-1871.
Ø He wasn’t able to build it o His plans survived and this
because he lost his funding. working model was built.
n Includes a working printer!
http://www.computerhistory.org/babbage/

11 Jan 2021 EE-739@IITB 3 CADSL

 Historic Events
• 1943-44: John Mauchly (professor) and J. Presper
Eckert (graduate student) built ENIAC at U.
Pennsylvania.
• 1944: Howard Aiken used “separate data and
program memories” in MARK I – IV computers –
Harvard Architecture.
• 1945-52: John von Neumann proposed a “stored
program computer” EDVAC (Electronic Discrete
Variable Automatic Computer) – Von Neumann
Architecture – use the same memory for program and data.

11 Jan 2021 EE-739@IITB 4 CADSL

 Electronic Computer
First Computer ENIAC: made of huge number of vacuum tubes 1946
Big size, huge power, short life time filament

11 Jan 2021 EE-739@IITB 5 CADSL

 Most Influential Document
• “Preliminary Discussion of the Logical
Design of an Electronic Computing
Instrument,” 1946 report by A. W. Burks,
H. H. Holdstine and J. von Neumann.
Appears in Papers of John von Neumann,
W. Aspray and A. Burks (editors), MIT
Press, Cambridge, Mass., 1987, pp. 97-
146.

11 Jan 2021 EE-739@IITB 6 CADSL

 Theory of Computing
• Alan Turing (1912-1954) gave a model of
computing in 1936 – Turing Machine.
• Original paper: A. M. Turing, “On Computable
Numbers with an Application to the
Entscheidungsproblem*,” Proc. Royal Math.
Soc., ser. 2, vol. 42, pp. 230-265, 1936.
• Recent book: David Leavitt, The Man Who
Knew Too Much: Alan Turing and the
Invention of the Computer (Great
Discoveries), W. W. Norton & Co., 2005.

* The question of decidability, posed by

mathematician Hilbert.

11 Jan 2021 EE-739@IITB CADSL

 History Continues

• 1946-52: Von Neumann built the IAS computer at

the Institute of Advanced Studies, Princeton – A
prototype for most future computers.
• 1947-50: Eckert-Mauchly Computer Corp. built
UNIVAC I (Universal Automatic Computer), used in
the 1950 census.
• 1949: Maurice Wilkes built EDSAC (Electronic Delay
Storage Automatic Calculator), the first stored-
program computer.

11 Jan 2021 EE-739@IITB 8 CADSL

 What was Computing Like?

• A data processing application involved

passing decks of punched cards through
electromechanical “unit record” machines.

• Repetitive sort, calculate, collate,

and tabulate operations ...
– ... were programmed with hand-wired
plugboard control panels.

11 Jan 2021 EE-739@IITB 9 CADSL

 Plugboard Control Panel

IBM 407 Accounting Machine (1949)

11 Jan 2021 EE-739@IITB 10 CADSL

 Von Neumann Bottleneck
• Von Neumann architecture uses the same
memory for instructions (program) and data.
• The time spent in memory accesses can limit
the performance. This phenomenon is
referred to as von Neumann bottleneck.
• To avoid the bottleneck, later architectures
restrict most operands to registers (temporary
storage in processor).

Ref.: D. E. Comer, Essentials of Computer Architecture, Upper Saddle

River, NJ: Pearson Prentice-Hall, 2005, p. 87.

11 Jan 2021 EE-739@IITB 11 CADSL

John von Neumann (1903-1957)

11 Jan 2021 EE-739@IITB 12 CADSL

Second Generation Computers
• 1955 to 1964
• Transistor replaced vacuum tubes
• Magnetic core memories
• Floating-point arithmetic
• High-level languages used: ALGOL, COBOL
and FORTRAN
• System software: compilers, subroutine
libraries, batch processing
• Example: IBM 7094

11 Jan 2021 EE-739@IITB 13 CADSL

 Third Generation Computers
• Beyond 1965
• Integrated circuit (IC) technology
• Semiconductor memories
• Memory hierarchy, virtual memories and caches
• Time-sharing
• Parallel processing and pipelining
• Microprogramming
• Examples: IBM 360 and 370, CYBER, ILLIAC IV,
DEC PDP and VAX, Amdahl 470
11 Jan 2021 EE-739@IITB 14 CADSL
C Programming Language and
UNIX Operating System

1972 Now

CADSL
 The Current Generation
• Personal computers
• Laptops and Palmtops
• Networking and wireless
• SOC and MEMS technology
• And the future!
• Biological computing
• Molecular computing
• Optical computing
• Quantum computing

11 Jan 2021 EE-739@IITB 16 CADSL

 Technological Push: Moore’s Law
• In 1965, Gordon Moore predicted that the
number of transistors that can be integrated
on a die would double every 18 months (i.e.,
grow exponentially with time).
• Amazingly visionary – million transistor/chip
barrier was crossed in the 1980’s.
– 2300 transistors, 1 MHz clock (Intel 4004) - 1971
– 16 Million transistors (Ultra Sparc III)
– 42 Million, 2 GHz clock (Intel P4) – 2001
– 40 Million transistor (HP PA-8500)

11 Jan 2021 EE-739@IITB 17 CADSL

 J.L.Hoyt
MIT

11 Jan 2021 EE-739@IITB 18 CADSL

 Microprocessor Journey
Intel Pentium (IV)

Intel 4004

11 Jan 2021 EE-739@IITB 19 CADSL

 Technology Push
• Technology advances at varying rates
– E.g. DRAM capacity increases at 60%/year
– But DRAM speed only improves 10%/year
– Creates gap with processor frequency!
• Inflection points
– E.g. enough devices for multicore processor (2001)
• Current issues causing an “inflection point”
– Power consumption
– Reliability
– Variability

11 Jan 2021 EE-739@IITB 20 CADSL

 Application Pull
• Corollary to Moore’s Law:
Cost halves every two years
In a decade you can buy a computer for less than its sales
tax today. –Jim Gray

• Computers cost-effective for

– National security – weapons design
– Enterprise computing – banking
– Departmental computing – computer-aided design
– Personal computer – spreadsheets, email, web
– Mobile computing – GPS, location-aware, ubiquitous

11 Jan 2021 EE-739@IITB 21 CADSL

 Application Pull
• What about the future?
– E.g. weather forecasting computational demand

• Must dream up applications that are not cost-

effective today
– Virtual reality, telepresence
– Web agents, social networking
– Wireless, location-aware
– Proactive (beyond interactive) w/ sensors
– Recognition/Mining/Synthesis (RMS)
– ???

11 Jan 2021 EE-739@IITB 22 CADSL

Single Processor Performance
Move to multi-processor

RISC

EE-739@IITB CADSL
Needed: Energy Efficient Computing

Performance: 11PF
Power: 6 – 11 MW (idle to loaded)
10 MW = $10 M per year electricity
UIUC Blue Waters Supercomputer

Extreme Scale Computing: 100=1000 more capable for the same

power consumption and physical foot-print
• Exascale (1018 operations/cycle) datacentre: 20 MW
• Petascale (1015 operations/cycle) departmental server: 20 kW
• Terascale (1012 operations/cycle) portable devices: 20 W

11 Jan 2021 EE-739@IITB 24 CADSL

Mobile Devices Power Consumption

11 Jan 2021 EE-739@IITB 25 CADSL

Running Program on Processor
Time
Processor Performance = ---------------
Program

Instructions Time
= X
Program Instruction

(code size)

Architecture
Compiler Designer

11 Jan 2021 EE-739@IITB 26 CADSL

 Computer Architecture

• Instruction Set Architecture (IBM 360)

– … the attributes of a [computing] system as seen by the
programmer. I.e. the conceptual structure and
functional behavior, as distinct from the organization
of the data flows and controls, the logic design, and
the physical implementation. -- Amdahl, Blaaw, &
Brooks, 1964

11 Jan 2021 EE-739@IITB 27 CADSL

Running Program on Processor
Time
Processor Performance = ---------------
Program

Instructions Cycles Time

= X X
Program Instruction Cycle

(code size) (CPI)

Architecture --> Implementation

Compiler Designer Processor Designer

11 Jan 2021 EE-739@IITB 28 CADSL

Running Program on Processor
Time
Processor Performance = ---------------
Program

Instructions Cycles Time

= X X
Program Instruction Cycle

(code size) (CPI) (cycle time)

Architecture --> Implementation --> Realization

Compiler Designer Processor Designer Chip Designer

11 Jan 2021 EE-739@IITB 29 CADSL

 Iron Law
• Instructions/Program
Ø Instructions executed, not static code size
Ø Determined by algorithm, compiler, ISA
• Cycles/Instruction
Ø Determined by ISA and CPU organization
Ø Overlap among instructions reduces this term
• Time/cycle
Ø Determined by technology, organization, clever circuit
design

11 Jan 2021 EE-739@IITB 30 CADSL

 Computer Architecture’s
Changing Definition
• 1950s to 1960s:
Computer Architecture Course = Computer Arithmetic
• 1970s to mid 1980s:
Computer Architecture Course = Instruction Set
Design, especially ISA appropriate for compilers
• 1990s onwards:
Computer Architecture Course = Design of CPU
(Processor Microarchitecture), memory system, I/O
system, Multiprocessors

11 Jan 2021 EE-739@IITB 31 CADSL

 Instruction Set Architecture
(ISA)

software

instruction set

hardware

11 Jan 2021 EE-739@IITB 32 CADSL

 Computer Architecture
Firefox, MS Excel

Windows 7
Applications
Visual C++

x86 Machine Primitives Computer

Von Neumann Machine Architecture
Technology
Logic Gates & Memory

Transistors & Devices

Quantum Physics • Rely on abstraction layers to manage

complexity
11 Jan 2021 EE-739@IITB 33 CADSL
 Performance vs. Design Time
• Time to market is critically important
• E.g., a new design may take 3 years
– It will be 3 times faster
– But if technology improves 50%/year
– In 3 years 1.53 = 3.38
– So the new design is worse!
(unless it also employs new technology)

11 Jan 2021 EE-739@IITB 34 CADSL

Thank You

11 Jan 2021 EE-739@IITB 35 CADSL