cs305 computer architecture autumn 2010 lecture 02br/courses/cs305-autumn2010/... · 2010. 7....
TRANSCRIPT
CS305 Computer ArchitectureAutumn 2010
Lecture 02
Bhaskaran RamanDepartment of CSE, IIT Bombay
http://www.cse.iitb.ac.in/~br/
http://www.cse.iitb.ac.in/synerg/doku.php?id=public:courses:cs305autumn10:start
Types of Computers
Q: What are the different computers you have seen or used or encountered or heard/read about?
Types of Computers: Desktops● Desktops or PCs: most well known class
Types of Computers: Servers● Servers: most powerful: individual PCs to supercomputers
Types of Computers:Embedded Computers
● Embedded computers: most specialized, vary in compute power, most prevalent but least noticed/known● Cellphones, robots, sensors, TV, fridge, washing
machine, etc.
Some Terms
● Processor or microprocessor: the main part of a computer
● Terms to be explained in subsequent lectures:● Instruction set architecture● nbit architecture
32/64bit Processors Sold
1998 1999 2000 2001 20020
200
400
600
800
1000
1200
The number of distinct processors sold between 1998 and 2002
Embedded Computer Desktops Servers
Mill
ions
of
com
pute
rs
Sale of Different MicroprocessorsSource: course text, Fig. 1.2
Source: course text, Fig 1.17
Growth of Processing Power
What is this Course About?
● What are the principles underlying the design of these various microprocessors?
● What support does the microprocessor provide for running applications?
● What are the components alongside a microprocessor?
● How to measure performance, and optimize for speed?
A Hierarchy of Languages
Algorithm
HighLevel Language (HLL)
Assembly Language
Machine Language
Digital Logic
Natural language, with some symbols
Concise; Structured, yet human readable
Not concise; Human readable representation of machine language
Binary numbers (0's and 1's)
Electrical signals
Mac
hin e
spe
cifi
cM
achi
n e
inde
pen d
ent
Programmer
Compiler
Assembler
Computer designer
An ExampleC code:a = b + c;d = e + f;
Machine code:...0...1......0...1......0...1......0...1......0...1......0...1......0...1......0...1...
Assembly code:lw $s1, 4($s0)lw $s2, 8($s0)add $s3, $s1, $s2sw ($s0), $s3lw $s3, 16($s0)lw $s4, 20($s0)add $s5, $s3, $s4sw 12($s0), $s5
Von Neumann Architecture
Memory=
Program+
Data
Controlpath
Datapath
Proc
esso
r
Input
Output
● The stored program concept: program (instructions) as well as data are stored in memory
● Processor fetches instructions from memory, and executes them on data (also fetched from/to memory)● Example from previous slide: LW and SW instructions
The Five Components● All computers have these five components: input,
output, memory, [data path + control path = processor]
● Underlined aspects: topics in this course
● Input: keyboard, mouse; also disk, network● Output: monitor; also disk, network● Memory: different kinds of memory● Data path + control path = processor
Inside a Computer...
Inside a Computer (continued...)
Magnetic Tape
Parts of a Computer● Integrated circuits, or chips:
● Flat and black● Processor (CPU), main memory, cache memory, etc.
● Motherboard:● Houses the various chips● Also has many I/O interfaces (PCI, USB, Serial, etc.)
● Secondary memory: nonvolatile● Magnetic disks, optical (CD/DVD), tape, flashbased
(e.g. USB pendrives, CF cards), floppies (obsolete)
Inside a Processor
Source: course text, Fig. 1.9
The IC Revolution behindProcessors and Memory
● IC's or Integrated Circuits● Precursors: transistors, vaccum tubes● Today: VLSI, ULSI
The IC Revolution in Numbers
Year Technology used in computers
Relative performance / unit cost
1951 Vacuum tube 11965 Transistor 351975 Integrated circuit 9001995 Very large scale integrated
circuit2400000
2005 Ultra large scale integrated circuit
6200000000
Source: course text, Chapter-1
Growth of Memory Capacity
1976 1981 1986 1991 1996 20011
10
100
1000
10000
100000
1000000
Growth of capacity per DRAM chip over time
Year of introduction
k bi
t c a
paci
ty (
k=2^
10=
1024
)
The IC Manufacturing Process
Source: course text, Fig. 1.14
IC CostCost per die=
Cost per waferDies per wafer×yield
Dies per wafer≈Area of waferArea of die
Yield=1
[1Defects per area×Area of die /2]2
Straightforward algebra
Approximation
From experience
● Unit cost of chip decreases with volume● Fixed costs amortised: design, masks in chip manufacture● Tuning to improve yield
Limits to IC density
● Fundamental physical dimension limits● Power consumption
● ~ 1 cm2 Pentium: 82 Watts!● Fan needed to sink the heat
● Observe: frequency scaling (Linux PC)
Upcoming topics
● Instruction set architecture● Assembly language programming