introduction to the design of embedded system - se...

Introduction to the Design of Embedded

SystemSystem2008 12 32008. 12. 3

Presented by Sang-Uk Jeon

ⓒ KAIST SE LAB 2008

Contents

Part I – What is Embedded System?Part II – Designing Embedded SystemPart III Software Engineering inPart III – Software Engineering in Embedded System Design

ⓒ KAIST SE LAB 2008 2/35

Part I. What is Part I. What is Embedded System?


Definition of Embedded System*Part I. What is Embedded System?

y

A combination of computer hardware and software, and perhaps additional mechanical or other partspDesigned to perform a dedicated functionIn some cases, part of a large system or productp

e.g.) Antilock Braking System(ABS) in a car

* Michael Barr "Embedded Systems Glossary” Netrino Technical Library* Michael Barr, "Embedded Systems Glossary”, Netrino Technical Library, Available at http://www.netrino.com/Embedded-Systems/Glossary


Characteristics of Embedded System*Part I. What is Embedded System?

y

Single-functionedExecutes a single program, repeatedly

Tightly-constrainedTightly-constrainedLow power, small, fast, etc.

Reactive and real-timeContinually reacts to changes in the system’sContinually reacts to changes in the system s environmentMust compute certain results in real-timeMust compute certain results in real-time without delay

* Frank Vahid and Tony Givargis Embedded System Design: A Unified* Frank Vahid and Tony Givargis, Embedded System Design: A Unified Hardware/Software Introduction, John Wiley & Sons, ISBN: 0471386782, 2002.


Categorization of Embedded SystemPart I. What is Embedded System?

g y

Computation orientedMP3 player, MPEG decoder, etc

Control orientedControl orientedHome appliances, industrial controller, safety-

iti l t llcritical controllerHybrid(computation + control)y ( )

Portable information devices• e g) Cellular phonee.g) Cellular phone

Networked multimedia applications


Part II. Designing Part II. Designing Embedded System


Issues in Embedded System DesignPart II. Designing Embedded System – Traditional Embedded System Design

y g

Top-priority design goalConstruct the system with desired functionality

Design issueDesign issueSimultaneously optimize numerous design

t i tconstraints• Size, performance, power, flexibility, etc.


Traditional Embedded System DesignPart II. Designing Embedded System – Traditional Embedded System Design

y g

System Specification

HW & SW P titi

•Designers partition the system into hardware and software early in the flow

HW & SW Partition

HW Development

Test not passed

•HW and SW engineers design their respective components in isolation•HW and SW engineers do not talk to each

hp

SW Development

other

•Integration problemsHi h t d l it ti

Prototype Test

•High cost and long iteration

Prototype

Test passed•Need new methodology!!

HW & SW co-designPlatform-based design…


HW & SW Co-DesignPart II. Designing Embedded System – HW & SW Co-Design

g

System Specification

Design Space ExplorationHW & SW Partition

HW Model SW Model* Synonym *ApplicationFunctionality

* Synonym *ArchitecturePlatform

Map HW & SW

E l ti f il d Evaluation*

SW SynthesisHW Synthesis

Evaluation failed

SW SynthesisHW Synthesis

HW & SW Integrationg*Evaluation subject : Performance, energy, etc.ⓒ KAIST SE LAB 2008 10/35

Design Space Exploration(DSE)*Part II. Designing Embedded System – HW & SW Co-Design

g p p ( )

Finding the optimal design of software and hardware

That satisfies given design objectivesThat satisfies given design objectivesProblem Space

(Characteristics of SW & HW)Solution Space

(Design objectives)

•Software functionality•Hardware parameters

Processor architecture

•Performance objective•Energy objective•…Processor architecture

Clock rateCache size…

Memory architecturePage replacement policy… * M. Gries. Methods for evaluating and

… covering the design space during early design development. Integr. VLSI J., 38(2):131–183, 2004.ⓒ KAIST SE LAB 2008 11/35

HW & SW PartitionPart II. Designing Embedded System – HW & SW Co-Design

Decides whether each functionality is implemented in HW or SW

Requirements : Summing up 100 valuesq g p

Software implementation

Alternative 1 Alternative 2

Software implementationADD v1, v2ADD v1, v3…

ADD v1, v2, …, v100

ADD v1, v100

ADD x1, x2 ADD x1, x2, …, x100Hardware support Hardware support

•More flexible software •Smaller software size•Better performance

pp pp


Software ModelPart II. Designing Embedded System – HW & SW Co-Design

Kahn process networksDirected Acyclic Graphs(DAG)Directed Acyclic Graphs(DAG)DAGs with periods and deadlinesSynchronous data flowyControl data flow graphs and dynamic data flowHigh-level programming languageCo-Design Finite State MachinesCommunication analysis graphsClick model of computationClick model of computationTransaction Level Modeling

SystemC

Hierarchical and heterogeneous models of computationPtolemy framework, Metropolis meta-model language, etc.


Hardware ModelPart II. Designing Embedded System – HW & SW Co-Design

Abstract modelsInstruction-accurate modelTask-accurate modelTask accurate modelNon-linear, accumulative service descriptions

E t bl d lExecutable modelsMicro-architecture templatespHardware description languageArchitecture description languageArchitecture description language


Mapping Software and HardwarePart II. Designing Embedded System – HW & SW Co-Design

pp g

Binding software tasks to hardware building blocks

May require additional jobs such asMay require additional jobs such as• Rewriting/adapting software code

– To link required interface of software and provided interface q pof hardware

• Compilation of the software onto the hardware

Software Tasks

ARM P1 ARM P2 DSP P3Hardware blocks


Evaluation (1/3)Part II. Designing Embedded System – HW & SW Co-Design

( / )

Simulation-based methodsCycle-accurate simulation

Software HardwareRead Memory

Cycle 1 (do Inst1)

Cycle 2 (do Inst2)

Done! (3 cycle)

Cycle 2 (do Inst2)

Cycle 3 (do Inst2)

System-level simulationSystem level simulationSoftware HardwareRead Memory

Done! (3 cycle)Read Memory :

Assume that it ( y )consumes 3 cycle



( / )

Analytical methodsStatic profiling

• Complexity analysis of algorithms, the dependency p y y g , p yanalysis of function call graph, etc.

High-level synthesisg y• Find an optimal mapping for the software tasks to

the hardware– Exact method such as Integer/mixed linear program

formulations, heuristics, evolutionary algorithms, etc.



( / )

Combination of simulation-based and analytical methods

Trace-based performance analysisTrace based performance analysisA trace contains all memory access

Cache hit & miss statistics…

Analytical models with initial, calibrating Initial simulation

yUse the trace to calculate performance

simulation

Cache miss rate : 10~20% P = A / B * …

Exhaustive simulation

Simulation results

Obt i f

Cache miss rate : 10 20%…

F d th i f ti

Q = C * 100 + …….

Exhaustive simulation Obtain ranges ofperformance factor

Feed the informationInto the analytical model


Frameworks for DSEPart II. Designing Embedded System – HW & SW Co-Design

System-level Micro-architecture centricMetropolisMescal

Mescal/TipiASIP-Meister / PEAS-III

StepNPSPADE

EXPRESSIONLisaTek

ArtemisMILAN

Chess / CheckersMaxCore & MaxSim

MESHSEASIncisive-SPWCoCentric System Studio


Example – Metropolis (1/4)Part II. Designing Embedded System – HW & SW Co-Design

p p ( / )

Functionality modelingProcess Src Medium S Process Sink

Process Src {port Out out;void thread() {

d(d )

Process Sink {port In in;void thread() {

i i (d )

Medium S {void send() {

…}out.send(data);

}void doSomething() {}

in.receive(data);}

}

}void receive() {

…}}

}}

}

Port Out {void send() {}

}

Port In {void receive() {}

}} }



p p ( / )

Architecture modelingTask 1

CPU

Task 2

P T k { M di CPU {

Task n

Process Task {port TaskToCPU taskToCPU;void thread() {}

Medium CPU {void read() {

// read from memory}}

void read() {…}

}void write() {

// write to memory}}

void write() {…}

}}

}void execute(int n) {

// Consume n CPU cycles}}

}



p p ( / )

MappingProcess Src Medium S Process Sink

Task 1

CPU

Task 2

Mapper {…constraint {constraint {

…Src.send => Task1.write;Src doSomething => Task1 execute(50);Src.doSomething => Task1.execute(50);Sink.receive => Task2.read;

…}}

}ⓒ KAIST SE LAB 2008 22/35


p p ( / )

A part of the output in “Simple case study” example


Platform-Based DesignPart II. Designing Embedded System – Platform-Based Design

g

Main ideaReusing & facilitating a common design to a variety of different applicationsy pp

Cell Phone Platform

Phone SW N 1

Phone SW N 2

Phone SW No nNo. 1 No. 2 No. n

Phone Product No. 1 Phone Product No. 2 Phone Product No. nⓒ KAIST SE LAB 2008 24/35

What is the Platform?* (1/2)Part II. Designing Embedded System – Platform-Based Design

( / )

PlatformProcessor

M

BusA library of components• To generate a design at certain level of

Memoryabstraction

Platform instanceA set of components

Processor P1-Clock : 300MhzA set of components

• Selected from the library(platform)P t t

Clock : 300Mhz-Cache size : 16Kb-…

• Parameters are setBus B1

-Bandwidth : 10Mb/s* A. Sangiovanni-Vincentelli, Quo Vadis, SLD? Reasoning About the T d d Ch ll f S t L l D i P di f th -…Trends and Challenges of System Level Design, Proceedings of the IEEE, Vol. 95, No. 3, March 2007.


What is the Platform? (2/2)Part II. Designing Embedded System – Platform-Based Design

( / )

Multiple abstraction levels of the platformEach platform instance in the abstraction levels can be reused

FunctionalityLevel 1



Processor ARM MIPS ARM920T R2000

Memory Bus

ARM

DDR2 SDRAM

ISA

MIPS

PCI

DDR SDRAMDDR-200

ARM926EJ-S R4600

8-bit ISADDR2 SDRAM PCI

PCI 2.2

DDR-333 16-bit ISA

DDR2-533

PCI 3.0DDR2-800ⓒ KAIST SE LAB 2008 26/35

Design ProcessPart II. Designing Embedded System – Platform-Based Design

g

Meet-in-the-middle processCombination of top-down and bottom-up approachpp


A Platform for Software ReusePart II. Designing Embedded System – Platform-Based Design

API Platform*

* A. Sangiovanni-Vincentelli and Grant Martin, Platform-Based Design and Software Design Methodology for Embedded Systems, IEEEMethodology for Embedded Systems, IEEE Design & Test, Vol. 18(6), November 2001.


Part III. Software Engineering in Embedded Engineering in Embedded

System Designy g


Challenge in Embedded System Designg y g

Consideration of flexible implementationTo address rapidly changing & increasing requirementsqSacrificing some degree of performance

Increasing usage of programmable elements instead of ASICs (Application Specific Integrated Circuits)

Increasing complexity of application(software)


Research Issues (1/3)Part III. Software Engineering in Embedded System Design

( / )

In the view of the software process*Embedded System Engineering

Issue 1 : Coordination of embedded software development process with other sub-processes

Embedded Software Engineering

Electrical Engineering

Mechanical EngineeringEngineering Engineering Engineering

Issue 2 : Specialized software development process for

* Bass Graaf Marco Lormans and Hans Toetenel “Embedded Software Engineering:

dealing with non-functional requirements (Memory, power, real-time requirements, etc.)

Bass Graaf, Marco Lormans and Hans Toetenel, Embedded Software Engineering: The State of the Practice,” IEEE Software, Nov.-Dec., 2003



( / )

In the view of the software modelHW & SW Partition

HW M d l SW M d l

Issue 1 : DSE in a more abstract levelHW Model

(Abstraction lv. 1)SW Model

(Abstraction lv. 1)abstract level

Issue 2 : Performance/ energy aware refactoring

Evaluation

Evaluationfailed

Map HW & SWenergy-aware refactoring

Issue 3 : Performance/ Evaluation

Issue 4 : Model-Driven

energy-related metric

HW Model(Abstraction lv. n)

SW Model(Abstraction lv. n)

Development

Map HW & SW Need to consider the underlying hardwareⓒ KAIST SE LAB 2008 32/35


( / )

In the view of the software code

HW & SW Partition

HW Model SW Model Usually SystemC code

Map HW & SWEvaluationfailed

Issue 1 : UML to SystemC

Evaluation*failed

Issue 2 : Application of code-level software engineering techniques

TestingClone detectionReverse engineeringMetric…


References

[1] Michael Barr, "Embedded Systems Glossary”, Netrino Technical Library, Available at http://www.netrino.com/Embedded-Systems/Glossarya ab e at ttp // et o co / bedded Syste s/G ossa y

[2] Frank Vahid and Tony Givargis, Embedded System Design: A Unified Hardware/Software Introduction, John Wiley & Sons, ISBN: 0471386782, 20022002.

[3] M. Gries. Methods for evaluating and covering the design space during early design development. Integr. VLSI J., 38(2):131–183, 2004.

[4] A. Sangiovanni-Vincentelli, Quo Vadis, SLD? Reasoning About the Trends and Challenges of System Level Design, Proceedings of the IEEE, Vol. 95, No. 3, March 2007.

[5] Bass Graaf, Marco Lormans, and Hans Toetenel, “Embedded Software Engineering: The State of the Practice,” IEEE Software, Nov.-Dec., 2003

[6] A Sangiovanni-Vincentelli and Grant Martin Platform-Based Design and[6] A. Sangiovanni Vincentelli and Grant Martin, Platform Based Design and Software Design Methodology for Embedded Systems, IEEE Design & Test, Vol. 18(6), November 2001.


Discussion

Q & A


introduction to the design of embedded system - se...

Documents