Computer Architecture

Introduction

Architect

Is the art or science of building A methode or style of building (Webster Dictionary) Here the approprite analogy is perhaps with building architecture. An architect capabilities are greatly enhanced if he is also exposed to the design aspects of the computer system

Computer Architecture

Deals with the functional behaviour of acomputer system as viewed by a programmer. This view includes aspect such as:

The size of data types

Eg. Using 16 binary digit to represent an integer Like addition, substraction, & subroutine calls

The types of operations that are supported

Computer Architecture Is
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, and Brooks, 1964 SOFTWARE

Computer Architectures 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: Computer Architecture Course: Design of CPU, memory system, I/O system, Multiprocessors, Networks 2010s: Computer Architecture Course: Self adapting systems? Self organizing structures? DNA Systems/Quantum Computing?

Digital & Org.

Strong Prerequisite

Comp. Arch.
Why, Analysis, Evaluation

Parallel
Parallel Architectures, Languages, Systems

How to build it Implementation details Basic knowledge of the organization of a computer is assumed!

To Understand the design techniques, machine structures, technology factors, evaluation methods that will determine the form of computers in 21st Century
Technology Parallelism Programming Languages Interface Design (Inst. Set Arch.)

Applications
Computer Architecture: Instruction Set Design Organization Hardware

Operating Systems

Measurement & Evaluation

History

A Computer Architect concentrates on

1.

2.

Processor-memory-switch level at which an architect view the system. Its simply a discription of system component and their interconnections. Instruction Set level, at which level the function of each instruction is described. The emphasis of this description level is on the behaviour of the system rather than the hardware structure of the system

Computer Architecture Topics (1/2) and Storage Input/Output

Disks, WORM, Tape RAID Emerging Technologies Interleaving Bus protocols Coherence, Bandwidth, Latency DRAM

Memory Hierarchy

L2 Cache

L1 Cache
VLSI
Instruction Set Architecture

Addressing, Protection, Exception Handling Pipelining and Instruction Level Parallelism

Pipelining, Hazard Resolution, Superscalar, Reordering, Prediction, Speculation

Computer Architecture Topics (2/2)

P M P M

P M

P M

Interconnection Network

Processor-Memory-Switch

Multiprocessors Networks and Interconnections

Topologies, Routing, Bandwidth, Latency, Reliability

Throughout this text we will focus on optimizing machine cost per performance

Computer Architecture
Evolusi

Human Brain Comparative with the computer

Mechanical Era

Gear, tangkai dan katrol Tingkat kecerdasan: menghitung SCHICKHARD(1623): Kalkulator pertama PASCAL (1642): menjumlah dan mengurangi otomatis LEIBNIZ (1671):mengalikan dan membagi otomatis BARBAGE (1827):mesin differensial 1834: Mesin Analitik, komputasi Multi Purpose (program)

Zuse (1941) Z3 AIKEN (1944) Harvard Mark I

Komputasi Multi Prupose

Historical Generations

1st Generation: 194659, vacuum tubes, relays, mercury delay lines 2nd generation: 195964, discrete transistors and magnetic cores 3rd generation: 196475, small- and medium-scale integrated circuits 4th generation: 1975present, single-chip microcomputer Integration scale: components per chip

Small: 10100 Medium: 1001,000 Large: 100010,000 Very large: greater than 10,000

Early Computing
1946: ENIAC, us Army, 18,000 Vacuum Tubes

1949:
1954: 1957: 1958: 1964: 1969: 1970: 1981: 1986:

UNIVAC I, $250K, 48 systems sold

IBM 701, Core Memory Moving Head Disk Transistor, FORTRAN, ALGOL, CDC & DEC

Founded
IBM 360, CDC 6600, DEC PDP-8 UNIX FLOPPY DISK IBM PC, 1st Successful Portable (Osborne1) Connection Machine, MAX Headroom Debut

Underlying Technologies Year Logic Storage Prog. Lang.

54 58 60 64 66 67 71 73 75 78 80 84 87 89 92 Tubes core (8 ms) Transistor (10s) Hybrid (1s) IC (100ns) LSI (10ns) (8-bit P) (16-bit P) VLSI (10ns) (32-bit P) ULSI GAs (64-bit P) FORTRAN ALGOL, COBOL thin film (200ns) Lisp, APL, Basic PL/1, Simula,C 1k DRAM O.O.

O/S

Batch

Generation Evolutionary

Multiprog. V.M.

Parallelism

4k DRAM 16k DRAM 64k DRAM 256k DRAM 1M DRAM 4M DRAM 16M DRAM

Networks ADA C++ Fortran90

Context for Designing New Architectures

Application Area

Special Purpose (e.g., DSP) / General Purpose Scientific (FP intensive) / Commercial

Level of Software Compatibility

Object Code/Binary Compatible (cost HW vs. SW) Assembly Language (dream to be different from binary) Programming Language; Why not?

Context for Designing New Architectures

Operating System Requirements for General Purpose Applications

Size of Address Space Memory Management/Protection Context Switch Interrupts and Traps IEEE 754 Floating Point I/O Bus Networks Operating Systems / Programming Languages

Standards: Innovation vs. Competition

Predictions for the Late 1990s (1/2)

Technology

Very large dynamic RAM: 64 MBits and beyond Large fast Static RAM: 1 MB, 10ns

Complete systems on a chip

10+ Million Transistors

Superscalar, Superpipeline, Vector, Multiprocessors Processor Arrays

Parallelism

Predictions for the Late 1990s (2/2)

Low Power

50% of PCs portable by 1995 Performance per watt

Parallel I/O

Many applications I/O limited, not computation Computation scaling, but memory, I/O bandwidth not keeping pace New interface technologies Video, speech, handwriting, virtual reality,

Multimedia

Original

Big Fishes Eating Little Fishes

1988 Computer Food Chain

Mainframe Work- PC MiniMinistation computer supercomputer

Supercomputer

Massively Parallel Processors

1998 Computer Food Chain

Massively Parallel Processors MiniMinisupercomputer computer

Mainframe

Server
Supercomputer

Work- PC station

Now who is eating whom?

1985 Computer Food Chain Technologies

ECL TTL MOS

Mainframe Minicomputer Supercomputer Minisupercomputer Work- PC station

Why Such Change in 10 years? (1/2)

Function

Rise of networking/local interconnection technology Technology Advances CMOS VLSI dominates TTL, ECL in cost & performance Computer architecture advances improves low-end RISC, Superscalar, RAID,

Performance

Why Such Change in 10 years? (2/2)

Price: Lower costs due to

Simpler development CMOS VLSI: smaller systems, fewer components Higher volumes CMOS VLSI : same dev. cost 10,000 vs. 10,000,000 units Lower margins by class of computer, due to fewer services

Technology Trends: Microprocessor Capacity

100000000

Graduation Window
10000000

Moores Law
1000000

Pentium i80486

Transistors

i80386 100000 i80286

Alpha 21264: 15 million Pentium Pro: 5.5 million PowerPC 620: 6.9 million Alpha 21164: 9.3 million Sparc Ultra: 5.2 million
CMOS improvements: Die size: 2X every 3 yrs Line width: halve / 7 yrs

i8086 10000 i8080 i4004 1000 1970 1975 1980 1985 Year 1990 1995 2000

Memory Capacity (Single Chip DRAM)

size 1000000000

year

size(Mb)

cyc time

100000000

10000000

1000000

100000

10000

1980 1983 1986 1989 1992 1996 2000

1970 1975 1980 1985 Year 1990 1995 2000

0.0625 250 ns 0.25 220 ns 1 190 ns 4 165 ns 16 145 ns 64 120 ns 256 100 ns

Bits

1000

CMOS Improvements

Die 25 size 2X every 3 yrs Die size increase plus Line widths halve every 7 yrs
20 15
Transistor Count

transistor count increase

10

5
Line Width Improvement

0 1980

Die Size

1983

1986

1989

1992

Memory Size of Various Bytes Systems Over Time

8G
1G 128M

4K chips 512 chips 64 chips workstation 8 chips-PC

8M 1M 128K

1
8K 1970 1Kbit 4K 16K 1980 64K

chip
256K

1/8
1M

chip
1990 4M 16M

640K DOS limit

Time
64M 2000 256M

Technology Trends (Summary)

Capacity Logic DRAM Disk 2x in 3 years 4x in 3 years 4x in 3 years Speed (latency) 2x in 3 years 2x in 10 years 2x in 10 years

Processor Frequency Trend

10,000 100
Intel IBM Power PC DEC Gate delays/clock

Processor freq scales by 2X per generation

1,000

21264S 21164A 21264 Pentium(R) 21064A 21164 II 21066 MPC750 604 604+

10

100

Pentium Pro 601, 603 (R) Pentium(R)

486 386
10 1

1987

1989

1991

1993

1995

1997

1999

2001

2003

Frequency doubles each generation Number of gates/clock reduce by 25%

2005

Gate Delays/ Clock

Mhz

Processor Performance Trends

1000
Supercomputers

100

Mainframes

10
Minicomputers Microprocessors

0.1 1965 1970 1975 1980 1985 1990 1995 2000

Performance vs. Time 100

Performance (VAX 780s)
4K Mips (65 MHz)

10
8600 | MV10K

Mips 25 MHz

o 9000

o | MIPS (8 MHz)

uVAX 6K (CMOS)

1.0

780 5 MHz
68K

uVAX CMOS

Will RISC continue on a 60%, (x4 / 3 years)?

0.1 1980 1985

1990

1000

1200

Sun-4/260 MIPS M/2000 MIPS M/120 IBM RS/6000 HP 9000/750 DEC AXP/500 IBM POWER 100 DEC Alpha 4/266 DEC Alpha 5/300 DEC Alpha 5/500 DEC Alpha 21164/600

200

400

600

800

1.54X/yr

Processor Performance (1.35X before, 1.55X now)

87 88 89 90 91 92 93 94 95 96 97

Summary: Performance Trends

Workstation performance (measured in Spec Marks) improves roughly 50% per year (2X every 18 months) Improvement in cost performance estimated at 70% per year

Putting performance growth in perspective:

IBM POWER2 Workstation 1993 > 200 MIPS 140 MFLOPS $120,000 71.5 MHz 256 KB 512 MB Cray YMP Supercomputer 1988 < 50 MIPS 160 MFLOPS $1M ($1.6M in 1994$) 167 MHz 0.25 KB 256 MB

Processor Perspective

Year MIPS Linpack Cost Clock Cache Memory

1988 supercomputer in 1993 server!

Where Has This Performance Improvement Come From?

Technology? Organization? Instruction Set Architecture? Software? Some combination of all of the above?

Performance Trends Revisited (Architectural Innovation) 1000

100
Supercomputers Mainframes

10

Minicomputers Microprocessors

1
CISC/RISC

0.1
1965 1970 1975 1980 1985 1990 1995 2000

Year

100000000

Performance Trends Revisited (Microprocessor Organization)

Bit Level Parallelism Pipelining
r4400

10000000

Caches Instruction Level Parallelism Out-of-order Xeq Speculation ...

Trans istors

1000000 i80386 100000 i80286 r3010

r4000

i8086 10000 i8080 i4004 1000 1970 1975 1980 1985 Year 1990 1995 2000

What is Ahead?

Greater instruction level parallelism? Bigger caches? Multiple processors per chip? Complete systems on a chip? (Portable Systems) High performance LAN, Interface, and Interconnect

Hardware1980 Technology 1990

Memory chips Speed 5-1/4 in. disks Floppies .256 M 64 K 1-2 40 M 1.5 M 4M 20-40 1G

2000 256 M-1 G 400-1000 20 G

500-2,000 M

LAN (Switch)
Busses

2-10 Mbits

10 (100)

155-655 (ATM)

2-20 Mbytes 40-400

Software Technology
1980 1990 Languages C, FORTRAN C++, HPF Op. System proprietary +DUM* User I/F glass Teletype WIMP* stylus, audio,video, ?? Comp. Styles T/S, PC Client/Server New things PC & WS parallel proc. Capabilities WP, SS WP,SS, mail 2000 object stuff?? +DUM+NT voice, agents*mobile appliances video, ??

DUM = DOS, n-UNIX's, MAC WIMP = Windows, Icons, Mouse, Pull-down menus Agents = robots that work on information

Computing 2001 (1/2)

Continue quadrupling memory every 3 years

1K chip in 72 becomes 1 gigabit chip (128 Mbytes) in 2002

On-line 12-25 Gigabytes; $10 1-Gbyte floppies & CDs Micros increase at 60% per year ... parallelism Radio links for untethered computing

Computing 2001 (2/2)

Telephone, fax, radio, television, camera, house, ...Real personal (watch, wallet,notepad) computers We should be able to simulate:

Nearly everything we make and their factories Much of the universe from the nucleus to galaxies

Performance implies: voice and visual

Ease of use. Agents!

Applications: Unlimited Opportunities (1/2)

Office agents: phone/FAX/comm; files/paper handling Untethered computing: fully distributed offices ?? Integration of video, communication, and computing: desktop video publishing, conferencing, & mail Large, commercial transaction processing systems Encapsulate knowledge in a computer: scientific & engineering simulation (e.g.. planetarium, wind tunnel, ... )

Applications: Unlimited Opportunities (2/2)

Visualization & virtual reality Computational chemistry e.g. biochemistry and materials Mechanical engineering without prototypes Image/signal processing: medicine, maps, surveillance. Personal computers in 2001 are today's supercomputers Integration of the PC & TV => TC

Challenges for 1990s Platforms (1/2)

64-bit computers video, voice, communication, any really new apps? Increasingly large, complex systems and environments Usability? Plethora of non-portable, distributed, incompatible, non-interoperable computers: Usability? Scalable parallel computers can provide commodity supercomputing: Markets and trained users?

Challenges for 1990s Platforms (2/2)

Apps to fuel and support a chubby industry: communications, paper/office, and digital video The true portable, wireless communication computer Truly personal card, pencil, pocket, wallet computer Networks continue to limit: WAN, ISDN, and ATM?

TAMAT ELECTRONIC
E R A