INTRODUCTION 

In this chapterwe set the stage for our study of embedded computing systemdesign. 

In order to understand the design processes,we ?rst need to understand how and 

why microprocessors are used for control, user interface, signal processing, and 

many other tasks. The microprocessor has become so common that it is easy to 

forget how hard some things are to do without it. 

We ?rst review the various uses of microprocessors and then review the major 

reasons why microprocessors are used in system design?delivering complex behav- 

iors, fast design turnaround, and so on. Next, in Section 1.2, we walk through the 

design of an example system to understand the major steps in designing a system. 

Section 1.3 includes an in-depth look at techniques for specifying embedded sys- 

tems?we use these speci?cation techniques throughout the book. In Section 1.4, 

we use a model train controller as an example for applying the speci?cation tech- 

niques introduced in Section1.3 that we use throughout the rest of the book. 

Section 1.5 provides a chapter-by-chapter tour of the book.

1.1 COMPLEX SYSTEMS AND MICROPROCESSORS

What is an embedded computer system? Loosely de?ned, it is any device that 

includes a programmable computer but is not itself intended to be a general-purpose 

computer.Thus,a PC is not itself an embedded computing system,although PCs are 

often used to build embedded computing systems. But a fax machine or a clock 

built from a microprocessor is an embedded computing system. 

This means that embedded computing system design is a useful skill for many 

types of product design. Automobiles, cell phones, and even household appliances 

make extensive use of microprocessors. Designers in many ?elds must be able to 

identify where microprocessors can be used, design a hardware platform with I/O 

devices that can support the required tasks, and implement software that performs 

the required processing. Computer engineering, like mechanical design or thermo- 

dynamics, is a fundamental discipline that can be applied inmany different domains. 

But of course, embedded computing system design does not stand alone. Many of 

the challenges encountered in the design of an embedded computing system are 

not computer engineering?for example, they may be mechanical or analog electri- 

cal problems. In this book we are primarily interested in the embedded computer 

itself, so we will concentrate on the hardware and software that enable the desired 

functions in the ?nal product.

1.1.1 Embedding Computers

Computers have been embedded into applications since the earliest days of com- 

puting. One example is the Whirlwind, a computer designed at MIT in the late 

1940s and early 1950s.Whirlwind was also the ?rst computer designed to support 

real-time operation and was originally conceived as a mechanism for controlling 

an aircraft simulator. Even though it was extremely large physically compared to 

todayfs computers (e.g., it contained over 4,000 vacuum tubes), its complete design 

from components to system was attuned to the needs of real-time embedded com- 

puting.The utility of computers in replacing mechanical or human controllers was 

evident fromthe very beginning of the computer era?for example,computerswere 

proposed to control chemical processes in the late 1940s [Sto95]. 

A microprocessor is a single-chip CPU. Very large scale integration (VLSI) 

stet?the acronymis the name technology has allowed us to put a complete CPU on 

a single chip since 1970s, but those CPUs were very simple. The ?rst microproces- 

sor, the Intel 4004,was designed for an embedded application,namely, a calculator. 

The calculator was not a general-purpose computer?it merely provided basic 

arithmetic functions. However, Ted Hoff of Intel realized that a general-purpose 

computer programmed properly could implement the required function, and that 

the computer-on-a-chip could then be reprogrammed for use in other products 

as well. Since integrated circuit design was (and still is) an expensive and time- 

consuming process, the ability to reuse the hardware design by changing the 

software was a key breakthrough. The HP-35 was the ?rst handheld calculator to 

perform transcendental functions [Whi72]. It was introduced in 1972, so it used 

several chips to implement the CPU, rather than a single-chip microprocessor. How- 

ever, the ability to write programs to perform math rather than having to design 

digital circuits to perform operations like trigonometric functions was critical to 

the successful design of the calculator. 

Automobile designers started making use of the microprocessor soon after 

single-chip CPUs became available. The most important and sophisticated use of  

microprocessors in automobileswas to control the engine:determiningwhen spark 

plugs ?re, controlling the fuel/air mixture, and so on. There was a trend toward 

electronics in automobiles in general?electronic devices could be used to replace 

the mechanical distributor. But the big push toward microprocessor-based engine 

control came from two nearly simultaneous developments: The oil shock of the 

1970s caused consumers to place much higher value on fuel economy, and fears of 

pollution resulted in laws restricting automobile engine emissions. The combina- 

tion of low fuel consumption and low emissions is very dif?cult to achieve; to meet 

these goals without compromising engine performance, automobile manufacturers 

turned to sophisticated control algorithms that could be implemented only with 

microprocessors. 

Microprocessors come in many different levels of sophistication; they are usu- 

ally classi?ed by their word size.An 8-bit microcontroller is designed for low-cost 

applications and includes on-board memory and I/O devices; a 16-bit microcon- 

troller is often used for more sophisticated applications that may require either 

longer word lengths or off-chip I/O and memory;and a 32-bit RISC microprocessor 

offers very high performance for computation-intensive applications. 

Given thewide variety ofmicroprocessor types available,it should be no surprise 

that microprocessors are used in many ways. There are many household uses of 

microprocessors. The typical microwave oven has at least one microprocessor to 

control oven operation. Many houses have advanced thermostat systems, which 

change the temperature level at various times during the day.The modern camera is 

a prime example of the powerful features that can be added under microprocessor 

control. 

Digital television makes extensive use of embedded processors. In some cases, 

specialized CPUs are designed to execute important algorithms?an example is 

the CPU designed for audio processing in the SGS Thomson chip set for DirecTV 

[Lie98]. This processor is designed to ef?ciently implement programs for digital 

audio decoding. A programmable CPU was used rather than a hardwired unit for 

two reasons: First, it made the system easier to design and debug; and second, it 

allowed the possibility of upgrades and using the CPU for other purposes. 

A high-end automobile may have 100 microprocessors, but even inexpensive 

cars today use 40 microprocessors. Some of these microprocessors do very simple 

things such as detect whether seat belts are in use. Others control critical functions 

such as the ignition and braking systems. 

Application Example 1.1 describes some of the microprocessors used in the 

BMW 850i. 

stability control (ASC T) system intervenes with the engine during maneuvering to improve 

the carfs stability. These systems actively control critical systems of the car; as control systems, 

they require inputs from and output to the automobile. 

Letfs ?rst look at the ABS. The purpose of an ABS is to temporarily release the brake on 

a wheel when it rotates too slowly?when a wheel stops turning, the car starts skidding and 

becomes hard to control. It sits between the hydraulic pump, which provides power to the 

brakes, and the brakes themselves as seen in the following diagram. This hookup allows the 

ABS system to modulate the brakes in order to keep the wheels from locking. The ABS system 

uses sensors on each wheel to measure the speed of the wheel. The wheel speeds are used 

by the ABS system to determine how to vary the hydraulic ?uid pressure to prevent the wheels 

from skidding. 

Sensor 

Sensor Sensor 

Sensor 

Hydraulic 

pump 

Brake Brake 

Brake Brake ABS 

The ASC T systemfs job is to control the engine power and the brake to improve the 

carfs stability during maneuvers. The ASC T controls four different systems: throttle, ignition 

timing, differential brake, and (on automatic transmission cars) gear shifting. The ASC T 

can be turned off by the driver, which can be important when operating with tire snow chains. 

The ABS and ASC T must clearly communicate because the ASC T interacts with the 

brake system. Since the ABS was introduced several years earlier than the ASC T, it was 

important to be able to interface ASC T to the existing ABSmodule, as well as to other existing 

electronic modules. The engine and control management units include the electronically con- 

trolled throttle, digital engine management, and electronic transmission control. The ASC T 

control unit has two microprocessors on two printed circuit boards, one of which concentrates 

on logic-relevant components and the other on performance-speci?c components.

1.1.2 Characteristics of Embedded Computing Applications 

important in both general-purpose computing and embedded computing, but 

embedded applications must meet many other constraints as well. 

On the one hand, embedded computing systems have to provide sophisticated 

functionality: 

¡ Complex algorithms: The operations performed by the microprocessor may 

be very sophisticated. For example, the microprocessor that controls an 

automobile engine must perform complicated ?ltering functions to opti- 

mize the performance of the car while minimizing pollution and fuel 

utilization. 

¡ User interface: Microprocessors are frequently used to control complex user 

interfaces that may include multiple menus and many options. The moving 

maps in Global Positioning System (GPS) navigation are good examples of 

sophisticated user interfaces. 

To make things more dif?cult, embedded computing operations must often be 

performed to meet deadlines: 

¡ Real time: Many embedded computing systems have to performin real time? 

if the data is not ready by a certain deadline, the system breaks. In some cases, 

failure tomeet a deadline is unsafe and can even endanger lives. In other cases, 

missing a deadline does not create safety problems but does create unhappy 

customers?missed deadlines in printers, for example,can result in scrambled 

pages. 

¡ Multirate: Not only must operations be completed by deadlines, but many 

embedded computing systems have several real-time activities going on at 

the same time. They may simultaneously control some operations that run 

at slow rates and others that run at high rates. Multimedia applications are 

prime examples of multirate behavior. The audio and video portions of a 

multimedia stream run at very different rates, but they must remain closely 

synchronized. Failure to meet a deadline on either the audio or video portions 

spoils the perception of the entire presentation. 

Costs of various sorts are also very important: 

¡ Manufacturing cost: The total cost of building the system is very important in 

many cases. Manufacturing cost is determined by many factors, including the 

type of microprocessor used, the amount of memory required, and the types 

of I/O devices. 

¡ Power and energy: Power consumption directly affects the cost of the 

hardware, since a larger power supply may be necessary. Energy con- 

sumption affects battery life, which is important in many applications, 

as well as heat consumption, which can be important even in desktop 

applications.

Embedded computing is in many ways much more demanding than the sort of 

programs that you may have written for PCs or workstations. Functionality is 

Finally, most embedded computing systems are designed by small teams on 

tight deadlines. The use of small design teams for microprocessor-based systems 

is a self-ful?lling prophecy?the fact that systems can be built with microproces- 

sors by only a few people invariably encourages management to assume that all 

microprocessor-based systems can be built by small teams.Tight deadlines are facts 

of life in todayfs internationally competitive environment. However,building a prod- 

uct using embedded software makes a lot of sense:Hardware and software can be 

debugged somewhat independently and design revisions can be made much more 

quickly.

1.1.3 Why Use Microprocessors?

There are many ways to design a digital system: custom logic, ?eld-programmable 

gate arrays (FPGAs), and so on.Why use microprocessors?There are two answers: 

¡ Microprocessors are a very ef?cient way to implement digital systems. 

¡ Microprocessors make it easier to design families of products that can be built 

to provide various feature sets at different price points and can be extended 

to provide new features to keep up with rapidly changing markets. 

The paradox of digital design is that using a predesigned instruction set processor 

may in fact result in faster implementation of your application than designing your 

own custom logic. It is tempting to think that the overhead of fetching, decoding, 

and executing instructions is so high that it cannot be recouped. 

But there are two factors that work together to make microprocessor-based 

designs fast. First,microprocessors execute programs very ef?ciently. Modern RISC 

processors can execute one instruction per clock cycle most of the time, and high- 

performance processors can execute several instructions per cycle.While there is 

overhead that must be paid for interpreting instructions, it can often be hidden by 

clever utilization of parallelism within the CPU. 

Second, microprocessor manufacturers spend a great deal of money to make 

their CPUs run very fast. They hire large teams of designers to tweak every aspect 

of the microprocessor to make it run at the highest possible speed. Few products 

can justify the dozens or hundreds of computer architects and VLSI designers cus- 

tomarily employed in the design of a singlemicroprocessor;chips designed by small 

design teams are less likely to be as highly optimized for speed (or power) as are 

microprocessors.They also utilize the latest manufacturing technology. Just the use 

of the latest generation of VLSI fabrication technology, rather than one-generation- 

old technology, can make a huge difference in performance. Microprocessors gen- 

erally dominate new fabrication lines because they can be manufactured in large 

volume and are guaranteed to command high prices. Customers who wish to fab- 

ricate their own logic must often wait to make use of VLSI technology from the 

latest generation of microprocessors. Thus, even if logic you design avoids all the 

overhead of executing instructions, the fact that it is built from slower circuits often 

means that its performance advantage is small and perhaps nonexistent. 

It is also surprising but true that microprocessors are very ef?cient utilizers 

of logic. The generality of a microprocessor and the need for a separate memory 

may suggest that microprocessor-based designs are inherently much larger than 

custom logic designs. However, in many cases the microprocessor is smaller when 

size is measured in units of logic gates. When special-purpose logic is designed 

for a particular function, it cannot be used for other functions. A microprocessor, 

on the other hand, can be used for many different algorithms simply by changing 

the program it executes. Since so many modern systems make use of complex 

algorithms and user interfaces, we would generally have to design many different 

customlogic blocks to implement all the required functionality.Many of those blocks 

will often sit idle?for example,the processing logicmay sit idlewhen user interface 

functions are performed. Implementing several functions on a single processor often 

makes much better use of the available hardware budget. 

Given the small or nonexistent gains that can be had by avoiding the use ofmicro- 

processors, the fact that microprocessors provide substantial advantages makes 

them the best choice in a wide variety of systems. The programmability of micro- 

processors can be a substantial bene?t during the design process. It allows program 

design to be separated (at least to some extent) from design of the hardware on 

which programs will be run.While one team is designing the board that contains 

themicroprocessor, I/O devices,memory,and so on,others can be writing programs 

at the same time. Equally important,programmability makes it easier to design fam- 

ilies of products. In many cases,high-end products can be created simply by adding 

code without changing the hardware. This practice substantially reduces manufac- 

turing costs. Evenwhen hardwaremust be redesigned for next-generation products, 

it may be possible to reuse software, reducing development time and cost. 

Why not use PCs for all embedded computing? Put another way, how many 

different hardware platforms do we need for embedded computing systems? PCs 

arewidely used and provide a very ?exible programming environment. Components 

of PCs are, in fact, used in many embedded computing systems. But several factors 

keep us from using the stock PC as the universal embedded computing platform. 

First, real-time performance requirements often drive us to different architec- 

tures.As we will see later in the book, real-time performance is often best achieved 

by multiprocessors. 

Second, low power and low cost also drive us away from PC architectures and 

toward multiprocessors. Personal computers are designed to satisfy a broad mix 

of computing requirements and to be very ?exible. Those features increase the 

complexity and price of the components.They also cause the processor and other 

components to use more energy to perform a given function. Custom embedded 

systems that are designed for an application,such as a cell phone,burn several orders 

of magnitude less power than do PCs with equivalent computational performance, 

and they are considerably less expensive as well. 

The cell phone may, in fact, be the next computing platform. Since over one 

billion cell phones are sold each year, a great deal of effort is put into designing 

them. Cell phones operate on batteries, so they must be very power ef?cient.They 

must also perform huge amounts of computation in real time. Not only are cell 

phones taking over some PC-oriented tasks, such as e-mail andWeb browsing, but 

the components of the cell phone can also be used to build non-cell-phone systems 

that are very energy ef?cient for certain classes of applications. 

1.1.4 The Physics of Software 

Computing is a physical act.Although PCs have trained us to think about computers 

as purveyors of abstract information, those computers in fact do their work by 

moving electrons and doing work. This is the fundamental reason why programs 

take time to ?nish,why they consume energy, etc. 

A prime subject of this book is what we might think of as the physics of 

software. Software performance and energy consumption are very important prop- 

ertieswhenwe are connecting our embedded computers to the realworld.We need 

to understand the sources of performance and power consumption if we are to be 

able to design programs that meet our applicationfs goals. Luckily,we donft have to 

optimize our programs by pushing around electrons. In many cases,we can make 

very high-level decisions about the structure of our programs to greatly improve 

their real-time performance and power consumption.As much as possible,we want 

to make computing abstractions work for us as we work on the physics of our 

software systems. 

1.1.5 Challenges in Embedded Computing System Design 

External constraints are one important source of dif?culty in embedded system 

design. Letfs consider some important problems that must be taken into account in 

embedded system design. 

How much hardware do we need? 

We have a great deal of control over the amount of computing power we apply 

to our problem. We cannot only select the type of microprocessor used, but also 

select the amount ofmemory,the peripheral devices,andmore. Sincewe oftenmust 

meet both performance deadlines andmanufacturing cost constraints, the choice of 

hardware is important?too little hardware and the systemfails tomeet its deadlines, 

too much hardware and it becomes too expensive. 

How do we meet deadlines? 

The brute force way of meeting a deadline is to speed up the hardware so that 

the program runs faster. Of course, that makes the system more expensive. It is also 

entirely possible that increasing the CPU clock ratemay notmake enough difference 

to execution time,since the programfs speedmay be limited by thememory system. 

How do we minimize power consumption? 

In battery-powered applications,power consumption is extremely important. Even 

in nonbattery applications, excessive power consumption can increase heat dis- 

sipation. One way to make a digital system consume less power is to make it 

run more slowly, but naively slowing down the system can obviously lead to 

missed deadlines. Careful design is required to slow down the noncritical parts 

of the machine for power consumption while still meeting necessary performance 

goals. 

How do we design for upgradability? 

The hardware platform may be used over several product generations,or for several 

different versions of a product in the same generation, with few or no changes. 

However, we want to be able to add features by changing software. How can we 

design a machine that will provide the required performance for software that we 

havenft yet written? 

Does it really work? 

Reliability is always important when selling products?customers rightly expect 

that products they buy will work. Reliability is especially important in some appli- 

cations, such as safety-critical systems. If we wait until we have a running system 

and try to eliminate the bugs,we will be too late?we wonft ?nd enough bugs, it 

will be too expensive to ?x them, and it will take too long as well. Another set of 

challenges comes from the characteristics of the components and systems them- 

selves. If workstation programming is like assembling a machine on a bench, then 

embedded system design is often more like working on a car?cramped, delicate, 

and dif?cult. Letfs consider someways inwhich the nature of embedded computing 

machines makes their design more dif?cult. 

¡ Complex testing: Exercising an embedded system is generally more dif?cult 

than typing in some data. We may have to run a real machine in order to 

generate the proper data.The timing of data is often important,meaning that 

we cannot separate the testing of an embedded computer from the machine 

in which it is embedded. 

¡ Limited observability and controllability: Embedded computing systems 

usually do not comewith keyboards and screens.Thismakes itmore dif?cult to 

see what is going on and to affect the systemfs operation.We may be forced to 

watch the values of electrical signals on the microprocessor bus, for example, 

to know what is going on inside the system. Moreover, in real-time applica- 

tions we may not be able to easily stop the system to see what is going on 

inside. 

¡ Restricted development environments: The development environments for 

embedded systems (the tools used to develop software and hardware) are 

often much more limited than those available for PCs and workstations.We 

generally compile code on one type of machine, such as a PC, and download 

it onto the embedded system.To debug the code,we must usually rely on pro- 

grams that run on the PC or workstation and then look inside the embedded 

system.

1.1.6 Performance in Embedded Computing

When we talk about performance when writing programs for our PC, what do 

we really mean? Most programmers have a fairly vague notion of performance? 

they want their program to run gfast enoughh and they may be worried about 

the asympototic complexity of their program. Most general-purpose programmers 

use no tools that are designed to help them improve the performance of their 

programs. 

Embedded system designers, in contrast, have a very clear performance goal in 

mind?their programmustmeet its deadline.At the heart of embedded computing 

is real-time computing,which is the science and art of programming to deadlines. 

The program receives its input data; the deadline is the time at which a computation 

must be ?nished. If the program does not produce the required output by the 

deadline, then the program does not work, even if the output that it eventually 

produces is functionally correct. 

This notion of deadline-driven programming is at once simple and demanding. 

It is not easy to determine whether a large, complex program running on a sophis- 

ticated microprocessor will meet its deadline.We need tools to help us analyze the 

real-time performance of embedded systems;we also need to adopt programming 

disciplines and styles that make it possible to analyze these programs. 

In order to understand the real-time behavior of an embedded computing system, 

we have to analyze the system at several different levels of abstraction.As we move 

through this book,we will work our way up from the lowest layers that describe 

components of the systemup through the highest layers that describe the complete 

system.Those layers include: 

¡ CPU: The CPU clearly in?uences the behavior of the program, particularly 

when the CPU is a pipelined processor with a cache. 

¡ Platform: The platform includes the bus and I/O devices. The platform com- 

ponents that surround the CPU are responsible for feeding the CPU and can 

dramatically affect its performance. 

¡ Program: Programs are very large and the CPU sees only a small window of 

the program at a time.We must consider the structure of the entire program 

to determine its overall behavior. 

¡ Task: We generally run several programs simultaneously on a CPU, creating a 

multitasking system. The tasks interact with each other in ways that have 

profound implications for performance. 

¡ Multiprocessor: Many embedded systems have more than one processor? 

they may include multiple programmable CPUs as well as accelerators. Once 

again, the interaction between these processors adds yet more complexity to 

the analysis of overall system performance.