Download Embedded Systems - Lecture Slides - Operating Systems | CS 3733 and more Exams Operating Systems in PDF only on Docsity!
Summer 2007 CS 4953: SS-Embedded Systems 1
Welcome to CS 4953
SS: Embedded Systems
Instructor: Dr. Dakai Zhu
Classroom: SB 3.02.04A
General Course Information
a Required textbook: ¾ Embedded System Design: A Unified Hardware/Software Introduction (ESD) by Frank Vahid and Tony Givargis, Wiley, 2002; ISBN: 0471386782 ¾ Robotic Explorations: A Hands-on Introduction to Engineering (Robot) by Fred G. Martin, Prentice Hall, 2001; ISBN: 0130895687
a Prerequisites: ¾ CS 3733: Operating Systems (undergraduate), and ¾ CS 4753: Architecture (undergraduate), and ¾ CS 3343: Algorithms (undergraduate)
a Class Web: http://www.cs.utsa.edu/~dzhu/cs4953/
Summer 2007 CS 4953: SS – Embedded Systems 3
Grade Policy
a NO final exam! ☺
aTwo Mid-Term Exams ( 25% each )
¾ June 28 and July. 26 ( tentative ) ¾ Closed books, closed notes
aFive labs and one project ( 50%)
aAttendance to both lecture and labs is mandatory
¾ Miss 3 classes Æ one letter grade down ¾ Extra 5% based on class attendance & participation
aNo early or makeup exams
aFinal grade: >= 90% Æ A; >= 80% Æ B;
>= 70% Æ C; >= 60% Æ D; <60% Æ F
Class Environment and Lab Equipments
aClassroom is the ‘ Security Lab ’
¾ No ‘touch’ on other equipments except the ones for us ¾ Fail to follow this rule Æ remove from the class !!!
aLab equipments: LEGO set & Handyboard
¾ More than $500 for each set Æ reuse for future classes
aLabs/Project: team work (2 or 3 students)
¾ All teammates are responsible for the equipments ¾ If broken, will be asked to replace them at your cost ¾ Labs/project will be conducted in the classroom
Summer 2007 CS 4953: SS-Embedded Systems 7
Introduction to
Embedded Systems
Lecture Outline
aWhat is an embedded system?
aWhy do we care embedded systems?
aDesign challenges & metrics
aConstraints and design tradeoffs: cost, performance,
time etc.
aDesign technologies and approaches
aCourse objectives and topics
Summer 2007 CS 4953: SS – Embedded Systems 9
General Computing Systems
aPersonal Computers (PCs)
¾ Desktops ¾ Laptops
aMainframes
aServers
Embedded Computing Systems: Examples
aComputing systems embedded
within electronic devices
aComputing system other than a
desktop computer
aPerhaps 50 per household and
per automobile
aBillions of units produced yearly,
versus millions of desktop units
Summer 2007 CS 4953: SS – Embedded Systems 13
Digital Camera: An Example
Microcontroller
A2D CCD preprocessor^ Pixel coprocessor D2A
JPEG codec DMA controller
Memory controller ISA bus interface UART LCD ctrl
Display ctrl
Multiplier/Accum
Digital camera chip
lens
CCD
a Single-functioned -- always a digital camera a Tightly-constrained -- Low cost, low power, small, fast a Reactive and real-time -- only to a small extent
Definition: An Embedded System
aEmploys a combination of hardware & software to
perform a single/specific function
aPart of a larger system (not a “computer”)
aWorks in a reactive and time-constrained
environment.
aSoftware provides features and flexibility
aHardware = {Processors, ASICs, Memory,...} is used
for performance (& sometimes security)
Summer 2007 CS 4953: SS – Embedded Systems 15
Four Categories of Embedded Systems
a General Computing ¾ Applications similar to desktop computing, but in an embedded package ¾ Video games, set top boxes, wearable computers, automatic tellers
a Control Systems ¾ Closed loop feedback control of real time system ¾ Vehicle engines, chemical processes, nuclear power, flight control
a Signal Processing ¾ Computations involving large data streams ¾ Radar, Sonar, video compression
a Communication & Networking ¾ Switching and information transmission ¾ Telephone system, Internet
Why are Embedded Systems important?
a Market reasons ¾ The embedded systems market is also in billions of $ ¾ 90% of processors go into “ non-computers ”, only 10% in “computers” ¾ In year 2000, about $2,700 of every car goes to electronics
a Engineering reasons ¾ Why does a satellite need a Windows prompt? ¾ Does the McDonald’s POS (point-of-sale) terminal need MacOS? ¾ Any device that needs to be controlled can be controlled by a microprocessor
a Embedded system designers are often ¾ Jackofmanytrades ¾ Need to know hardware, software, and some combination of networking, control theory and signal processing ¾ business models
Summer 2007 CS 4953: SS – Embedded Systems 19
Optimization: Tradeoffs
a Improving one metric may worsen others a Expertise with both software and hardware is needed to optimize design metrics a Various technologies in order to choose the best for a given application and constraints
Performance Size
Power
NRE cost
Microcontroller
A2D CCD preprocessor^ Pixel coprocessor D2A JPEG codec DMA controller
Memory controller ISA bus interface UART LCD ctrl
Display ctrl
Multiplier/Accum
Digital camera chip
lens
CCD Hardware
Software
Time-to-market: a demanding metric
aTime required to
develop a product to the
point it can be sold to
customers
aMarket window
¾ Period during which the product would have highest sales
aAverage time-to-market
constraint is about 8
months
aDelays can be costly
Revenues ($)
Time (months)
Summer 2007 CS 4953: SS – Embedded Systems 21
Losses due to delayed market entry
a Simplified revenue model ¾ Product life = 2W, peak at W ¾ Triangle area equals revenue a Area = 1/2 * base * height ¾ On-time = 1/2 * 2W * W ¾ Delayed = 1/2 * (W-D+W)(W-D) a Loss: difference between the on- time and delayed triangle areas a Percentage revenue loss = (D(3W-D)/2W 2 )100% a Lifetime 2W=52 wks, ¾ delay D=4 wks: (4(326 – 4)/226^2) = 22% ¾ delay D=10 wks: (10(326 – 10)/226^2) = 50%
On-time Delayed entry entry
Peak revenue
Peak revenue from delayed entry
Market rise Market fall
W 2W Time
D
On-time
Delayed
Revenues ($)
NRE and unit cost metrics vs. unit #
a Costs: ¾ Unit cost: the monetary cost of manufacturing each copy of the system, excluding NRE cost ¾ NRE cost (Non-Recurring Engineering cost): The one-time monetary cost of designing the system ¾ total cost = NRE cost + unit cost * # of units ¾ per-product cost = total cost / # of units = (NRE cost / # of units) + unit cost a Example ¾ NRE=$2000, unit=$ ¾ For 10 units 9 total cost = $2000 + 10*$100 = $ 9 per-product cost = $2000/10 + $100 = $
Amortizing NRE cost over all the units
Summer 2007 CS 4953: SS – Embedded Systems 25
Trends in Embedded Systems
aIncreasing code size
¾ average code size: 16-64KB in 1992, 64K-512KB in 1996 ¾ migration from hand (assembly) coding to high-level languages
aReuse of hardware and software components
¾ processors (micro-controllers, DSPs) ¾ software components (drivers)
aIncreasing integration and system complexity
¾ integration of RF, DSP, network interfaces ¾ 32-bit processors, IO processors (I2O)
Embedded System Requirements
aR1 Functional requirements
aR2 Temporal requirements
aR3 Dependability requirements
Summer 2007 CS 4953: SS – Embedded Systems 27
Functional requirements
aData Collection
¾ Sensor requirements ¾ Signal conditioning ¾ Alarm monitoring
aDirect Digital Control
¾ Actuators
aMan-Machine Interaction
¾ informs the operator of the current state of the controlled object ¾ assists the operator in controlling the system
Temporal requirements
aTasks may have deadlines
aMinimal latency jitter
aMinimal error detection latency
aTiming requirements due to tight software control
loops
aHuman interface requirements.
Summer 2007 CS 4953: SS – Embedded Systems 31
Different Processor Technologies
aThe architecture of the computation engine used to
implement a system’s desired functionality
Application-specific
Registers Custom ALU
Controller Datapath
Program memory Assembly code for: total = 0 for i =1 to …
Control logic and State register
Data memory
IR PC
Single-purpose (“hardware”)
Controller Datapath Control logic State register
Data memory
index total
IR PC
Register file General ALU
Controller Datapath
Program memory Assembly code for: total = 0 for i =1 to …
Control logic and State register
Data memory
General-purpose (“software”)
Different Processor Technologies (cont.)
aGeneral-purpose processors
¾ Low time-to-market and NRE costs ¾ High flexibility
aSingle-purpose processors: digital circuit designed
to execute exactly one program
¾ Contains only the components needed to execute a single program Æ fast, low power and small size ¾ High cost
aApplication-specific processors: optimized for a
particular class of applications
¾ Some flexibility, good performance, size and power
Summer 2007 CS 4953: SS – Embedded Systems 33
IC (integrated circuit) Technologies
aThe manner in which a digital (gate-level)
implementation is mapped onto an IC
aFull-custom/VLSI
aSemi-custom ASIC (gate array and standard cell)
aPLD (Programmable Logic Device)
source drain cha nnel
oxide
gate
Silicon substrate
IC package IC
Full-custom/VLSI
aAll layers are optimized for an embedded system’s
particular digital implementation
¾ Placing transistors ¾ Sizing transistors ¾ Routing wires
aBenefits
¾ Excellent performance, small size, low power
aDrawbacks
¾ High NRE cost (e.g., $300k), long time-to-market
Summer 2007 CS 4953: SS – Embedded Systems 37
Processor & IC technologies Relation
a Basic tradeoff ¾ General vs. custom ¾ With respect to processor technology or IC technology ¾ The two technologies are independent
General- purpose processor
ASIP
Single- purpose processor
PLD Semi-custom Full-custom
General, providing improved:
Customized, providing improved:
Power efficiency Performance Size Cost (high volume)
Flexibility Maintainability NRE cost Time- to-prototype Time-to-market Cost (low volume)
Design Technology
aThe manner in which we convert our concept of
desired system functionality into an implementation
Libraries/IP: Incorporates pre- designed implementation from lower abstraction level into higher level.
System specification
Behavioral specification
RT specification
Logic specification To final implementation
Compilation/Synthesis: Automates exploration and insertion of implementation details for lower level.
Test/Verification: Ensures correct functionality at each level, thus reducing costly iterations between levels.
Compilation/ Synthesis
Libraries/ IP
Test/ Verification System synthesis
Behavior synthesis
RT synthesis
Logic synthesis
Hw/Sw/ OS
Cores
RT components
Gates/ Cells
Model simulat./ checkers
Hw-Sw cosimulators
HDL simulators
Gate simulators
Summer 2007 CS 4953: SS – Embedded Systems 39
The co-design ladder
a In the past: ¾ Hardware and software design technologies were very different ¾ Recent maturation of synthesis and CAD tools enables a unified view of hardware and software a Hardware/software “codesign” Implementation
Assembly instructions
Machine instructions
Register transfers
Compilers (1960's,1970's)
Assemblers, linkers (1950's, 1960's)
Behavioral synthesis (1990's)
RT synthesis (1980's, 1990's) Logic synthesis (1970's, 1980's)
Microprocessor plus program bits: “software”
VLSI, ASIC, or PLD implementation: “hardware”
Logic gates
Logic equations / FSM's
Sequential program code (e.g., C, VHDL)
The choice of hardware versus software for a particular function is simply a tradeoff among various design metrics, like performance, power, size, NRE cost, and especially flexibility; there is no fundamental difference between what hardware or software can implement.
Moore’s Law: Transistors on a Chip
1
10
100
1000
10000
100000
1000000
10000000
100000000
1000000000
19681970197219741976197819801982198419861988199019921994199619982000
Transistors 4004
8008 8080
8086
286 386
(^486) P-I P-II
P-III P-IV
Moore’s Law
IC transistor capacity has doubled roughly
every 18 months for the past several decades
