Computer architecture fundamentals and

principles of computer design 2nd Edition Joseph

D. Dumas Ii download

https://textbookfull.com/product/computer-architecture-
fundamentals-and-principles-of-computer-design-2nd-edition-
joseph-d-dumas-ii/

Download full version ebook from https://textbookfull.com

We believe these products will be a great fit for you. Click
the link to download now, or visit textbookfull.com
to discover even more!

Fundamentals of Computer Architecture and Design Ahmet

Bindal

https://textbookfull.com/product/fundamentals-of-computer-
architecture-and-design-ahmet-bindal/

Logic Computer Design Fundamentals 5th Edition Mano

https://textbookfull.com/product/logic-computer-design-
fundamentals-5th-edition-mano/

Digital design and computer architecture. ARM Edition

David Harris

https://textbookfull.com/product/digital-design-and-computer-
architecture-arm-edition-david-harris/

Essentials of computer architecture Second Edition

Comer

https://textbookfull.com/product/essentials-of-computer-
architecture-second-edition-comer/
Fundamentals of Computer Graphics Steve Marschner

https://textbookfull.com/product/fundamentals-of-computer-
graphics-steve-marschner/

Computer Architecture John L. Hennessy

https://textbookfull.com/product/computer-architecture-john-l-
hennessy/

Computer Organization and Architecture 9th Edition

William Stallings

https://textbookfull.com/product/computer-organization-and-
architecture-9th-edition-william-stallings/

Fundamentals of Business Process Management Marlon

Dumas

https://textbookfull.com/product/fundamentals-of-business-
process-management-marlon-dumas/

Computer System Architecture M. Morris Mano

https://textbookfull.com/product/computer-system-architecture-m-
morris-mano/
SECOND EDITION

Computer
Architecture
Fundamentals and Principles
of Computer Design

Joseph D. Dumas II
U n i v e r s i t y o f Te n n e s s e e a t C h a t t a n o o g a
Chattanooga, TN, USA
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2017 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business

Version Date: 20160908

International Standard Book Number-13: 978-1-4987-7271-6 (Hardback)

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
 Contents

Preface.............................................................................................................. xiii

Chapter 1 Introduction to computer architecture..................................... 1

1.1 What is computer architecture?.............................................................. 1
1.1.1 Architecture versus implementation....................................... 2
1.2 Brief history of computer systems.......................................................... 4
1.2.1 The first generation.................................................................... 5
1.2.2 The second generation............................................................... 6
1.2.3 The third generation.................................................................. 8
1.2.4 The fourth generation.............................................................. 10
1.2.5 The fifth generation.................................................................. 13
1.2.6 Modern computing: the sixth generation............................. 16
1.3 Types of computer systems................................................................... 20
1.3.1 Single processor systems......................................................... 21
1.3.2 Parallel processing systems..................................................... 24
1.3.3 Special architectures................................................................ 25
1.4 Quality of computer systems................................................................ 25
1.4.1 Generality and applicability................................................... 26
1.4.2 Ease of use ................................................................................ 27
1.4.3 Expandability............................................................................ 27
1.4.4 Compatibility............................................................................ 28
1.4.5 Reliability���������������������������������������������������������������������������������� 30
1.5 Success and failure of computer architectures
and implementations............................................................................. 30
1.5.1 Quality and the perception of quality................................... 31
1.5.2 Cost issues�������������������������������������������������������������������������������� 31
1.5.3 Architectural openness, market timing, and other issues..... 33
1.6 Measures of performance...................................................................... 35
1.6.1 CPU performance..................................................................... 35
1.6.2 Memory system performance................................................. 37
1.6.3 I/O system performance.......................................................... 40
 1.6.4 Power performance.................................................................. 40
1.6.5 System benchmarks.................................................................. 43
1.7 Chapter wrap-up.................................................................................... 47

Chapter 2 Computer memory systems...................................................... 53

2.1 The memory hierarchy.......................................................................... 53
2.1.1 Characteristics of an ideal memory....................................... 53
2.1.2 Characteristics of real memory devices................................ 56
2.1.3 Hierarchical memory systems................................................ 60
2.2 Main memory interleaving................................................................... 62
2.2.1 High-order interleaving........................................................... 63
2.2.2 Low-order interleaving............................................................ 64
2.3 Logical organization of computer memory........................................ 69
2.3.1 Random access memories........................................................ 69
2.3.2 Sequential access memories.................................................... 71
2.3.3 Associative memories.............................................................. 73
2.4 Cache memory........................................................................................ 76
2.4.1 Locality of reference................................................................. 76
2.4.2 Hits, misses, and performance............................................... 78
2.4.3 Mapping strategies................................................................... 80
2.4.4 Cache write policies................................................................. 87
2.4.5 Cache replacement strategies.................................................. 89
2.4.6 Cache initialization.................................................................. 90
2.5 Memory management and virtual memory....................................... 91
2.5.1 Why virtual memory?.............................................................. 91
2.5.2 Virtual memory basics............................................................. 92
2.5.3 Paged virtual memory............................................................. 93
2.5.4 Segmented virtual memory.................................................... 97
2.5.5 Segmentation with paging.................................................... 100
2.5.6 The MMU and TLB................................................................ 101
2.5.7 Cache and virtual memory................................................... 102
2.6 Chapter wrap-up.................................................................................. 105

Chapter 3 Basics of the central processing unit.................................... 111

3.1 The instruction set.................................................................................111
3.1.1 Machine language instructions............................................ 112
3.1.2 Functional categories of instructions....................................114
3.1.3 Instruction addressing modes...............................................117
3.1.4 Number of operands per instruction................................... 121
3.1.5 Memory-register versus load-store architectures.............. 123
3.1.6 CISC and RISC instruction sets............................................ 126
3.2 The datapath.......................................................................................... 129
3.2.1 The register set........................................................................ 129
3.2.2 Integer arithmetic hardware................................................. 131
 3.2.2.1 Addition and subtraction...................................... 132
3.2.2.2 Multiplication and division...................................141
3.2.3 Arithmetic with real numbers.............................................. 151
3.2.3.1 Why use floating-point numbers?....................... 151
3.2.3.2 Floating-point representation.............................. 152
3.2.3.3 Floating-point arithmetic hardware.................... 158
3.3 The control unit..................................................................................... 160
3.3.1 A simple example machine....................................................161
3.3.2 Hardwired control unit......................................................... 166
3.3.3 Microprogrammed control unit........................................... 168
3.4 Chapter wrap-up...................................................................................176

Chapter 4 Enhancing CPU performance................................................ 183

4.1 Pipelining............................................................................................... 184
4.2 Arithmetic pipelines............................................................................ 189
4.3 Instruction unit pipelines.................................................................... 192
4.3.1 Basics of an instruction pipeline.......................................... 193
4.3.2 Control transfers and the branch penalty........................... 195
4.3.3 Branch prediction................................................................... 198
4.3.4 Delayed control transfers...................................................... 203
4.3.5 Memory accesses: delayed loads and stores....................... 206
4.3.6 Data dependencies and hazards.......................................... 207
4.3.7 Controlling instruction pipelines......................................... 210
4.4 Characteristics of RISC machines...................................................... 217
4.5 Enhancing the pipelined CPU............................................................ 222
4.5.1 Superpipelined architectures................................................ 223
4.5.2 Superscalar architectures...................................................... 224
4.5.3 Very long instruction word (VLIW) architectures............ 225
4.5.4 Multithreaded architectures................................................. 230
4.6 Chapter wrap-up.................................................................................. 234

Chapter 5 Exceptions, interrupts, and input/output systems............ 239

5.1 Exceptions.............................................................................................. 239
5.1.1 Hardware-related exceptions................................................ 240
5.1.1.1 Maskable interrupts............................................... 242
5.1.1.2 Nonmaskable interrupts....................................... 245
5.1.1.3 Watchdog timers and reset................................... 246
5.1.1.4 Nonvectored, vectored, and autovectored
interrupts................................................................ 247
5.1.2 Software-related exceptions.................................................. 250
5.2 Input and output device interfaces.................................................... 252
5.3 Program-controlled I/O....................................................................... 254
5.3.1 Memory-mapped I/O............................................................. 255
5.3.2 Separate I/O............................................................................. 257
 5.4 Interrupt-driven I/O............................................................................. 260
5.5 Direct memory access.......................................................................... 261
5.6 Input/output processors...................................................................... 266
5.7 Real-world I/O example: the universal serial bus........................... 267
5.8 Chapter wrap-up...................................................................................274

Chapter 6 Parallel and high-performance systems.............................. 279

6.1 Types of computer systems: Flynn’s taxonomy................................ 280
6.1.1 Vector and array processors.................................................. 282
6.1.2 GPU computing...................................................................... 288
6.1.3 Multiprocessor systems......................................................... 292
6.1.4 Multicomputer systems......................................................... 308
6.2 Interconnection networks for parallel systems.................................311
6.2.1 Purposes of interconnection networks.................................311
6.2.2 Interconnection network terms and concepts.................... 312
6.2.2.1 Master and slave nodes......................................... 312
6.2.2.2 Circuit switching versus packet switching........ 312
6.2.2.3 Static and dynamic networks............................... 315
6.2.2.4 Centralized control versus distributed control....316
6.2.2.5 Synchronous timing versus asynchronous
timing.......................................................................316
6.2.2.6 Node connection degree........................................318
6.2.2.7 Communication distance and diameter..............318
6.2.2.8 Cost, performance, expandability, and fault
tolerance.................................................................. 319
6.2.3 Bus-based interconnections.................................................. 322
6.3 Static interconnection networks......................................................... 325
6.3.1 Linear and ring topologies.................................................... 325
6.3.2 Star networks........................................................................... 326
6.3.3 Tree and fat tree networks..................................................... 327
6.3.4 Nearest-neighbor mesh.......................................................... 328
6.3.5 Torus and Illiac networks...................................................... 329
6.3.6 Hypercube networks.............................................................. 332
6.3.7 Routing in static networks.................................................... 334
6.4 Dynamic interconnection networks.................................................. 340
6.4.1 Crossbar switch....................................................................... 340
6.4.2 Recirculating networks.......................................................... 344
6.4.3 Multistage networks............................................................... 346
6.4.3.1 Blocking, nonblocking, and rearrangeable
networks.................................................................. 346
6.5 Chapter wrap-up.................................................................................. 351
Chapter 7 Special-purpose and future architectures.......................... 357
7.1 Dataflow machines............................................................................... 358
7.2 Artificial neural networks................................................................... 364
7.3 Fuzzy logic architectures.................................................................... 371
7.4 Quantum computing........................................................................... 379
7.5 Chapter wrap-up.................................................................................. 386

Appendix: reference and further reading materials with web links...... 391
Index..................................................................................................................411
 Preface

Digital electronic computer systems have gone through several genera-

tions and many changes since they were first built just before and during
World War II. Machines that were originally implemented with electro-
mechanical relays and vacuum tubes gave way to those constructed with
solid-state devices and, eventually, integrated circuits containing thou-
sands, millions, or billions of transistors. Systems that cost millions of dol-
lars and took up large rooms (or even whole floors of buildings) decreased
in price by orders of magnitude and shrank, in some cases, to single chips
less than the size of a postage stamp. CPU clock speeds increased from
kilohertz to megahertz to gigahertz, and computer storage capacity grew
from kilobytes to megabytes to gigabytes to terabytes and beyond.
Although most people have noticed the obvious changes in modern
computer system implementation, not everyone realizes how much has
remained the same, architecturally speaking. Many of the basic design
concepts and even the advanced techniques used to enhance perfor-
mance have not changed appreciably in 30, 40, even 50 years or longer.
Most modern computers still use the sequential, von Neumann program-
ming paradigm that dates to the 1940s; they accept hardware interrupts
that have been a standard system design feature since the 1950s; and
they store programs and data in hierarchical memories that are, at least
concep­tually, very similar to storage systems built in the 1960s. Although
computing professionals obviously need to stay abreast of today’s cutting-
edge architectural breakthroughs and the latest technical wizardry, it is
just as important that they study historical computer architectures—not
only because doing so gives a valuable appreciation for how things were
done in the past, but also because, in many cases, the same or similar tech-
niques are still being used in the present and may persist into the future.
Over many years of teaching the computer architecture course to an
audience of mostly undergraduate computer science (and a few computer
engineering) students, I have observed that few—if any—of the students
in a typical CS (or even electrical engineering) program ever go to work
designing microprocessors, memory devices, or other ICs, let alone com-
plete computer systems. Many, probably most, of them instead become
system administrators, programmer/analysts, technical managers, etc.
In these positions, one is not generally called upon to design hardware,
and so there is no need to know where each individual transistor goes
on a particular chip. However, it is quite likely that at some point almost
every computing professional will have to specify or purchase a computer
system to run a particular application(s). To do so, he or she must know
enough about computer architectures and implementation technologies
to be able to understand the characteristics of the machines under con-
sideration, see through manufacturer hype, intelligently compare system
performance, and ultimately select the best and most cost-effective system
for the job. A course designed around this textbook should prepare stu-
dents to do exactly that, without getting them lost in the myriad technical
details characteristic of other excellent, but lengthy and involved, texts.
My philosophy in developing—and now revising—this book was to
concentrate on the fundamental principles of computer design and per-
formance enhancement that have proven effective over time, and to show
how current trends in architecture and implementation rely on these prin-
ciples while in many cases expanding them or applying them in new ways.
Because specific computer designs tend to come and go quickly, this text
does not focus on one particular machine or family of machines. Instead,
important concepts and techniques are explained using examples drawn
from a number of different computer architectures and implementations,
both state-of-the-art and historical. In cases in which explanations based
on real machines would include too many “trees” for the reader to see the
“forest,” simpler examples have been created for pedagogical purposes. The
focus is not on understanding one particular architecture, but on under-
standing architectural and implementation features that are used across
a variety of computing platforms. The author’s underlying assumption is
that if students have a thorough grounding in what constitutes high perfor-
mance in computers, a good concept of how to measure it, and a thorough
familiarity with the fundamental principles involved in making systems
perform better, they will be able to understand and evaluate the many new
systems they will encounter in their (hopefully long) professional careers.
This book is primarily designed to be used in a one-semester upper
(usually senior)-level undergraduate course in computer architecture as
taught in most traditional computer science programs, including the one
at the University of Tennessee at Chattanooga where I have been on the
faculty since 1993. It is also suitable for use in an undergraduate elec-
trical engineering or computer engineering curriculum at the junior or
first semester senior level, with the idea that it would likely be followed
by an additional hardware-oriented course(s) covering topics such as
advanced microprocessor systems, embedded systems, VLSI design, etc.
(Refer to the chapter breakdown for suggestions on how to incorporate
the text into courses based on the quarter system.) This book can also be
effectively used (perhaps with some supplementary materials provided
by the instructor and/or an additional text chosen to meet specific needs)
in an introductory, master’s level graduate course. I believe it would be
particularly useful for what appears to me to be the increasing number of
students seeking a master’s degree in computer science after obtaining an
undergraduate degree in another field.
To get the maximum benefit from this text, students should have had
a previous course(s) covering introductory topics in digital logic and com-
puter organization. Although this is not a text for a programming course,
it is assumed that the reader is quite familiar with computer program-
ming concepts. Ideally, students should not only be well versed in at least
one high-level programming language, such as C, C++, or Java, but they
should also have had some exposure to machine-level programming in
assembly language (the specific architecture covered is not all that impor-
tant, but the concepts are highly relevant). Previous courses in operat-
ing systems and/or systems programming would be helpful, but are not
essential in following the material presented here.
To use computer architecture terminology, this textbook is a RISC
design. The idea is to help students achieve success in understanding the
essential concepts not by including a lot of extra “bells and whistles,” but
by providing an easily understood, almost conversational text illustrated
with simple, clear, and informative figures. Each chapter begins with an
introduction that briefly lays out the ideas to be covered and the reasons
why those topics are important to the study of modern computer systems.
The introduction is followed by several sections explaining the material in
detail and then a “chapter wrap-up.” This section briefly reviews the key
concepts that the student should have gleaned from reading the material
and participating in class, placing them in context to help reinforce the
student’s learning by helping him or her to see the “big picture” or broader
context into which the detailed material that has just been presented fits.
Finally, the end-of-chapter review questions include not only prob-
lems with concrete, numerical, right and wrong answers, but also “fill in
the blank” items that reinforce key concepts and terminology, as well as
open-ended short answer and essay-type questions. In years past, when
using other texts, my students have often complained about textbook
exercises that were either too trivial or too involved to be of much use
in preparing for exams (which, of course, is a major concern of students).
One of my goals in developing this book was to provide review questions
that would be sufficiently thought provoking to challenge students, but
manageable within a reasonable time frame (as they would have to be to
serve as viable testing items). To this end, many of the review questions
are drawn from actual exam questions used over the years I have taught
the computer architecture course. By working out answers to a variety of
problems and questions comparable to items that their instructor would
be likely to include on a test, students should not only master the material
thoroughly, but also (of equal if not greater importance to them) be pre-
pared to demonstrate their mastery when called upon to do so.
Chapter 1 provides a background for the topics in the rest of the book
by discussing the difference between architecture and implementation
and the ways in which they influence each other. It also includes a brief
history of computing machines from the earliest, primitive computers up
to the present day. (To understand where the field is now and where it
may go in the future, it is important to know where it has been.) The vari-
ous types of single-processor and parallel, general- and special-purpose
systems that will be covered in following chapters are introduced. Then
the reader, who will likely at some point in the future be responsible for
specifying and selecting computer systems, is introduced to some con-
cepts of architectural quality and other factors that may cause particular
systems to succeed or fail in the marketplace. Last, but certainly not least,
methods for quantifying and measuring the performance of computers
and their major subsystems are discussed.
Chapters 2 through 5 are the heart of the text; they cover, in appro-
priate detail, the architecture of traditional, single-processor computer
systems. (Even though multicore chips and other parallel systems have
become commonplace, understanding the operation of a uniprocessor
machine is still fundamental.) Chapter 2 deals with the important topic
of memory systems. It begins by explaining the characteristics of an ideal
memory system (which of course does not exist) and then discusses how
various memory technologies approximate ideal behavior in some ways
but not others. This naturally leads to an explanation of how a hierarchi-
cal storage system may be used to maximize the benefits of each type of
device while hiding their less desirable characteristics, thus enabling the
overall memory system to keep up with the demands of the processor
(and the needs of the programmer).
Because CPU architecture and implementation are so complex and so
critical to the performance of a system, both Chapters 3 and 4 are devoted
to this topic. Chapter 3 explains the basics, including the design of CISC
and RISC machine instruction sets, the datapath hardware that is used
to carry out those machine instructions by operating on integer and real
number values, and the control unit, which develops the control signals
that operate the datapath as well as the rest of the machine using either a
microprogram or hardwired logic. Chapter 4 then discusses techniques
that can be used to enhance the performance of the basic CPU design, with
particular emphasis on pipelining. Both arithmetic and instruction-unit
pipelines are covered; because so many modern microprocessors make
extensive use of pipelined implementation, we pay particular attention
to RISC, superpipelined, superscalar, VLIW, and multithreaded designs.
Chapter 5 completes the coverage of single-processor system design
considerations by discussing I/O related topics, including basic inter-
facing approaches, exceptions and interrupts, and the use of DMA and
input/output processors to offload I/O-related tasks from the main sys-
tem processor. These concepts are further illustrated via a discussion of
how the ubiquitous Universal Serial Bus (USB) interface works.
The final two chapters deal with approaches that may be adopted
when even the most advanced conventional, single-processor systems do
not provide the desired performance or are not well suited to the intended
application. Chapter 6 covers the most common types of high-performance
systems, most of which are parallel to varying degrees. Flynn’s venera-
ble taxonomy of computer architectures is discussed in relation to vector
and array processors, graphics processing unit (GPU) computing, shared-
memory multiprocessors, and message-passing cluster (multicomputer)
systems. Because communication between processors is critical to the per-
formance of parallel systems, static and dynamic interconnection networks
are discussed at some length. Finally, Chapter 7 goes beyond Flynn’s clas-
sifications to explore the characteristics of unconventional architectures of
the past, present, and future. From dataflow machines to artificial neu-
ral networks to fuzzy systems to quantum computers, students will see
that there are other approaches, in addition to von Neumann’s, that can be
taken to solve particular types of computational problems.
As mentioned earlier, most instructors and their students should be
able to cover the entire contents of this book in a one-semester course, given
satisfactory completion of the prerequisites suggested above. If it is desired
to spend more time covering conventional architectures at a leisurely pace,
Chapter 7 may be omitted or left for outside reading. At institutions using
the quarter system, it is suggested that Chapters 1 through 5 might be cov-
ered in one quarter while Chapters 6 and 7, plus perhaps some additional,
advanced topics added by the instructor—or a substantial research or
design project—could be reserved for a second, quarter-long course.
Writing (and updating) a textbook is a challenging task, and no mat-
ter how hard one tries, it is impossible to avoid error or to please every
reader. I have, however, done my best, and I hope you enjoy the result.
All known errors in the first edition have been corrected, and (I hope)
not too many new ones have been introduced. I would like to hear from
instructors and students alike about how you found this book useful and,
conversely, of any mistakes or misconceptions you may have encountered,
or any parts of the text that were less than clear. Please feel free to suggest
any improvements that I might make in future revisions of the text. I wel-
come your comments via e-mail at [email protected].

Joe Dumas
Signal Mountain, Tennessee
chapter one

Introduction to computer
architecture
“Computer architecture” is not the use of computers to design buildings
(although that is one of many useful applications of computers). Rather,
computer architecture is the design of computer systems, including all of
their major subsystems: the central processing unit (CPU), the memory
system, and the input/output (I/O) system. In this introductory chapter,
we take a brief look at the history of computers and consider some general
topics applicable to the study of computer architectures. In subsequent
chapters, we examine in more detail the function and design of specific
parts of a typical modern computer system. If your goal is to be a designer
of computer systems, this book provides an essential introduction to gen-
eral design principles that can be expanded upon with more advanced
study of particular topics. If (as is perhaps more likely) your career path
involves programming, systems analysis or administration, technical
management, or some other position in the computer or information
technology field, this book provides you with the knowledge required to
understand, compare, specify, select, and get the best performance out of
computer systems for years to come. No one can be a true computer pro-
fessional without at least a basic understanding of computer architecture
concepts. So let’s get underway!

1.1 What is computer architecture?

Computer architecture is the design of computer systems, including all
major subsystems, including the CPU and the memory and I/O systems.
All of these parts play a major role in the operation and performance
of the overall system, so we will spend some time studying each. CPU
design starts with the design of the instruction set that the processor will
execute and includes the design of the arithmetic and logic hardware that
performs computations; the register set that holds operands for computa-
tions; the control unit that carries out the execution of instructions (using
the other components to do the work); and the internal buses, or con-
nections, that allow these components to communicate with each other.
Memory system design uses a variety of components with differing char-
acteristics to form an overall system (including main, or primary, memory
2 Computer Architecture

and secondary memory) that is affordable while having sufficient storage

capacity for the intended application and being fast enough to keep up
with the CPU’s demand for instructions and data.
I/O system design is concerned with getting programs and data into
the memory (and ultimately the CPU) and communicating the computa-
tional results to the user (or another computing system) as quickly and
efficiently as possible. None of these subsystems of a modern computer
is designed in a vacuum; each of them affects, and is affected by, the
characteristics of the others. All subsystems must be well matched and
well suited to the intended application in order for the overall system to
perform well and be successful in the marketplace. A system that is well
designed and well built is a powerful tool that multiplies the productivity
of its users; a system that is poorly designed, or a poor implementation of
a good design, makes an excellent paperweight, doorstop, or boat anchor.

1.1.1 Architecture versus implementation

It is important to distinguish between the design (or architecture) of a
system and the implementation of that system. This distinction is eas-
ily understood through an analogy between computer systems and
buildings.
Computer architecture, like building architecture, involves first of all
a conceptual design and overall plan. Architects ask, “What is this build-
ing (or computer) going to be used for, and what is the general approach
we will take to fulfill the requirements of that application?” Once these
general decisions are made, the architect has to come up with a more spe-
cific design, often expressed in terms of drawings and other specifications
showing the general composition and layout of all the parts of the build-
ing or computer. This tells how everything will fit together at the higher
levels.
When the design gets down to the level of specification of the actual
components, the building or computer architect needs to have engineer-
ing knowledge (or enlist the help of a construction or computer engineer)
in order to make sure the paper design is feasible given available materi-
als. A construction engineer needs to make sure the building founda-
tion and beams will carry the required loads, that the heating and air
conditioning units have sufficient capacity to maintain the temperature
of the building, and so on. A computer engineer must make sure the
electrical, mechanical, thermal, timing, control, and other characteris-
tics of each component are sufficient to the job and compatible with the
other components of the system. The result of all this architectural and
engineering effort is a design specification for a building or for a com-
puter system. However, this specification exists only on paper or, more
Chapter one: Introduction to computer architecture 3

likely, as computer-aided design (CAD) files containing the drawings

and specifications.
If one wants to have an actual building to occupy or a computer sys-
tem to use, it must be built in the physical world. This, of course, is what
we mean by implementation. In the case of a building, the design docu-
ments prepared by the architect and engineer are given to a contractor
who uses components made of various materials (steel, wood, plastic,
concrete, glass, brick, etc.) to construct (implement) an actual building.
Likewise, the design produced by a computer architect or engineer must
be put into production and built using various electrical and mechani-
cal components. The end result—the implementation—is a working com-
puter system.
It should be obvious that, for the end product to be a properly func-
tioning building or computer, both the architectural plan and the physical
implementation must be done well. No amount of attention to quality in
construction will turn an inadequate design into a building or computer
that meets the requirements of the intended application. However, even
the best and most well–thought out design can be ruined by using sub-
standard components or shoddy workmanship. For a good end result, all
aspects of design and implementation must come together.
We should also recognize that although a clear distinction should
be made between architecture and implementation, they are intimately
interrelated. Neither architecture nor implementation exists in a vacuum;
they are like two sides of a coin. Architectural vision affects the type of
technologies chosen for implementation, and new implementation tech-
nologies that are developed can broaden the scope of architectural design.
Taking the example of building architecture again, in the distant past all
buildings were made of materials such as stone and wood. When iron
was discovered and first used as a building material, it allowed architects
to design new types of buildings that had previously been impossible to
build. The advent of steel and concrete enabled the design and construc-
tion of skyscrapers that could never have existed before the invention of
those materials. Modern materials, such as polymers, continue to expand
the possibilities available to building architects. Sometimes the desire to
include certain architectural features in buildings has led to the devel-
opment of new construction materials and techniques or at least to new
applications for existing materials.
Building architecture and implementation have progressed hand in
hand over the course of human civilization, and the vision of computer
architects and the available implementation technologies have likewise
(over a much shorter time but at a much faster pace) moved forward hand
in hand to create the computers we used in the past, the ones we use today,
and the ones we will use tomorrow.
4 Computer Architecture

1.2 Brief history of computer systems

Computing devices of one type or another have existed for hundreds of
years. The ancient Greeks and Romans used counting boards to facilitate
mathematical calculations; the abacus was introduced in China around
a.d. 1200. In the 17th century, Schickard, Pascal, and Leibniz devised
mechanical calculators. The first design for a programmable digital com-
puter was the analytical engine proposed by Charles Babbage in 1837.
Given the technology available at the time, the analytical engine was
designed as a mechanical rather than electronic computer; it presaged
many of the concepts and techniques used in modern machines but was
never built due to a lack of funding. Ada Augusta, Countess of Lovelace,
developed “cards” (programmed instructions) to demonstrate the opera-
tion of the proposed analytical engine and, to this day, is revered as the
first computer programmer. Although Babbage’s machine was never com-
mercially successful, mechanical punched-card data processing machines
were later developed by Herman Hollerith and used to tabulate the results
of the 1890 U.S. census. Hollerith’s company later merged with another to
become International Business Machines (IBM).
While Babbage’s design was based on the decimal (base 10) system
of numbering used by most cultures, some of his contemporaries, such
as George Boole and Augustus DeMorgan, were developing a system of
logical algebra (today known as Boolean algebra) that provides the theo-
retical underpinnings for modern computers that use a two-valued, or
binary, system for representing logical states as well as numbers. Boolean
algebra was an intellectual curiosity without practical application until
a young mathematician named Claude Shannon recognized (in his 1937
master’s thesis at the Massachusetts Institute of Technology) that it could
be used in designing telephone switching networks and, later, comput-
ing machines. Boolean algebra is the logical basis for virtually all modern
digital computer design.
In addition to Babbage’s engine and Hollerith’s punched-card machine,
another type of “computer” predated the existence of modern digital com-
puters. From the early 20th century, mechanical and electrical “analog
computers” were used to solve certain types of problems that could be
expressed as systems of differential equations with time as the indepen-
dent variable. The term computer is in quotes because analog computers
are more properly called analog simulators. They do not actually perform
discrete computations on numbers, but rather operations such as addition,
subtraction, integration, and differentiation of analog signals, usually rep-
resented by electrical voltages. Analog simulators are continuous rather
than discrete in their operation, and they operate on real values; num-
bers are measured rather than counted. Thus, there is no limit, other than
the tolerances of their components and the resolution of the measuring
Chapter one: Introduction to computer architecture 5

apparatus, to the precision of the results obtained. During the 1930s,

1940s, 1950s, and even into the 1960s, analog simulators were widely used
to simulate real-world systems, aiding research in such fields as power
plant control, aircraft design, space flight, weather modeling, and so on.
Eventually, however, analog simulators were rendered obsolete by digital
computers as they became more powerful, more reliable, and easier to
program. Thus, in the rest of our discussion of the history of comput-
ing devices we will restrict ourselves to the electronic digital computing
devices of the 20th and now the 21st centuries.

1.2.1 The first generation

During the late 1930s and early 1940s, mostly as part of the Allied effort to
win World War II, digital computers as we know them today got their start.
The first generation (approximately late 1930s to early 1950s) of computer
systems were one-of-a-kind machines, each custom built for a particular
purpose. Computers of the early 1940s, such as the Mark-I (also known as
the IBM Automatic Sequence Controlled Calculator or ASCC) and Mark-II
machines built by Howard Aiken at Harvard University, were typically
built using electromagnetic relays as the switching elements. This made
them very slow. Later machines were built using vacuum tubes for switch-
ing; these were somewhat faster but not very reliable. (Tubes, like incan-
descent light bulbs, have a nasty habit of burning out after a few hundred
or a few thousand hours of use.) Two of the first electronic computers
built using vacuum tube technology were the Atanasoff-Berry Computer
(ABC) developed at Iowa State University and the Electronic Numerical
Integrator and Calculator (ENIAC) built by John Mauchly and J. Presper
Eckert at the University of Pennsylvania for the U.S. Army Ordnance
Department’s Ballistic Research Laboratories. ENIAC, which was used
to calculate bomb trajectories and later to help develop the hydrogen
bomb, was more similar to today’s pocket calculators than to our general-
purpose computers as it was not a stored-program machine. Connections
had to be rewired by hand in order to program different calculations.
The first modern computer designed to run software (program
instructions stored in memory that can be modified to make the machine
perform different tasks) was the Electronic Discrete Variable Computer
(EDVAC) designed by Mauchly, Eckert, and John von Neumann of
Princeton University. EDVAC was designed to perform sequential pro-
cessing of instructions that were stored in memory along with the data,
characteristics of what has become known as the von Neumann archi-
tecture (to be discussed further in Section 1.3.1). It is debatable whether
von Neumann deserves primary credit for the idea—the report outlining
EDVAC’s design mysteriously bore only his name although several other
6 Computer Architecture

researchers worked on the project—but what is certain is that the stored-

program concept was a major step forward in computer design, making
general-purpose machines feasible. To this day, most computers are still
basically von Neumann machines with some enhancements.
Although the original EDVAC design was never built (it was even-
tually modified and built as Princeton’s Institute for Advanced Studies
[IAS] machine), its concepts were used in many other machines. Maurice
Wilkes, who worked on EDVAC, built the Electronic Delay Storage
Automatic Calculator (EDSAC), which became the first operational stored-
program computer using the von Neumann architecture. In 1951, the
first commercially available computer was produced by the Remington-
Rand Corporation. Based on the stored-program designs of the EDVAC
and EDSAC, this computer was known as the UNIVAC I (for Universal
Automatic Computer). The first of 46 of these machines were delivered to
the U.S. Census Bureau in 1951; the following year, another UNIVAC found
the spotlight as it was used to predict the outcome of the Eisenhower–
Stevenson presidential race on election night. With only 7% of the vote
counted, the machine projected a win for Eisenhower with 438 electoral
votes (he ultimately received 442).

1.2.2 The second generation

The second generation (approximately mid-1950s to early 1960s) of digital
computer systems included the first machines to make use of the new
solid-state transistor technology. The transistor, invented in 1947 by John
Bardeen, Walter Brattain, and William Shockley of Bell Laboratories, was
a major improvement over vacuum tubes in terms of size, power con-
sumption, and reliability. This new implementation technology paved the
way for many architectural enhancements, mainly by allowing the total
number of switching elements in machines to increase.
Vacuum tube-based computers could never have more than a few
thousand switching elements because they were constantly malfunc-
tioning due to tube failures. The mean time between failures (MTBF), or
average lifetime, of vacuum tubes was only about 5000 hours; a system
containing 5000 tubes could thus be expected to have a hardware-related
crash about once per hour on average, and a system with 10,000 of the
same type of tubes would average one failure every half hour of opera-
tion. If the machine were to run long enough to do any meaningful calcu-
lations, it could only contain a limited number of tubes and (because all
logic required switching elements) a very limited set of features.
Then, along came the transistor with its much longer life span. Even
the earliest transistors had a typical MTBF of hundreds of thousands or
millions of hours—one or two orders of magnitude better than vacuum
tubes. They were also much smaller and generated less heat, allowing
Chapter one: Introduction to computer architecture 7

components to be more densely packaged. The result of these factors was

that second-generation computers could be more complex and have more
features than their predecessors without being as large, expensive, or
power-hungry and without breaking down as often. Some architectural
features that were not present in first-generation machines but were added
to second-generation computers (and are still used in modern machines)
include hardware representation of floating-point numbers (introduced
on the IBM 704 in 1954), hardware interrupts (used in the Univac 1103 in
1954), general-purpose registers used for arithmetic or addressing (used
in the Ferranti Pegasus in 1956), and virtual memory (introduced on the
University of Manchester’s Atlas machine in 1959). The IBM 709, released
in 1958, featured asynchronous I/O controlled by independent, parallel
processors as well as indirect addressing and hardware interrupts.
Another technology that came into use with the second generation of
computer systems was magnetic core memory. Core memory stored binary
information as the magnetized states of thousands of tiny, doughnut-
shaped ferrite rings or “cores.” This technology reduced the space and
power required for large amounts of storage. Although typical first-
generation computers had only 1 to 4 kilobytes (KB) of main memory in
the form of delay lines, vacuum tubes, or other primitive storage tech-
nologies, second-generation computers could have comparatively huge
main memories of, for example, 16 KB up to 1 megabyte (MB) of core. This
increase in the size of main memory affected the types of instructions
provided, the addressing modes used to access memory, and so on.
Despite technological advances, second-generation machines were
still very bulky and expensive, often taking up an entire large room and
costing millions of dollars. The relatively small IBM 650, for example,
weighed about a ton (not counting its 3000-pound power supply) and cost
$500,000 in 1954. Of course, that would be several millions of today’s dol-
lars! Core memory cost on the order of a dollar per byte, so the memory
system alone for a large computer could have a cost in the seven-figure
range.
With such large, expensive systems being built, no organization could
afford to run just one program, get an answer, then manually load the
next program and run it (as was done with the first-generation machines).
Thus, during this time came the advent of the first batch-processing and
multiprogramming operating systems. Batch processing meant that pro-
grams were loaded and executed automatically by the system instead
of the operator loading them manually. Multiprogramming meant that
more than one program was resident in memory at the same time (this,
of course, was made possible by the larger main memory space). By keep-
ing multiple programs ready to run and by allowing the system to switch
between them automatically, the expensive CPU could be kept busy
instead of being idle, awaiting human intervention. Programs were still
8 Computer Architecture

executed one at a time, but when one program was completed or encoun-
tered an I/O operation that would take some time to complete, another
program would be started to occupy the CPU’s time and get useful work
done.
The first attempts to make human programmers more productive also
occurred during this time. Assembly language, as a shorthand or mne-
monic form of machine language, was first developed in the early 1950s.
The first high-level languages, Formula Translation (Fortran), Algorithmic
Language (Algol), and Common Business-Oriented Language (COBOL),
came along a few years later. It is interesting to note that computer pio-
neer John von Neumann opposed the development of assemblers and
high-level language compilers. He preferred to employ legions of human
programmers (mostly low-paid graduate students) to hand-assemble code
into machine language. “It is a waste of a valuable scientific computing
instrument,” von Neumann reportedly said, “to use it to do clerical work.”
Fortunately, his point of view did not win out!

1.2.3 The third generation

The third generation (approximately mid-1960s to early 1970s) marked
the first use of integrated circuits in computers, replacing discrete (indi-
vidually packaged) transistors. Integrated circuits (ICs) are semiconduc-
tor chips containing multiple transistors, starting with one or two dozen
transistors comprising a few gates or flip-flops (small scale integration
[SSI]) and later increasing to a few hundred transistors (medium scale
integration [MSI]), which allowed an entire register, adder, or even an
arithmetic logic unit (ALU) or small memory to be fabricated in one pack-
age. Core memory was still common in third-generation computers, but
by the 1970s, core was beginning to be replaced by cheaper, faster semi-
conductor IC memories. Even machines that used magnetic core for main
storage often had a semiconductor cache memory (see Section 2.4) to help
improve performance.
Integrated circuits continued the trend of computers becoming smaller,
faster, and less expensive. All first- and second-generation machines were
what would come to be called mainframe computers. The third generation
saw the development of the first minicomputers, led by Digital Equipment
Corporation’s (DEC) PDP-8, which was introduced in 1964 and cost
$16,000. DEC followed it with the more powerful PDP-11 in 1970. Data
General brought the Nova minicomputer to market in 1969 and sold 50,000
machines at $8000 each. The availability of these small and relatively inex-
pensive machines meant that rather than an organization having a single
central computer shared among a large number of users, small depart-
ments or even individual workers could have their own machines. Thus,
the concept of a workstation computer was born.
Chapter one: Introduction to computer architecture 9

Also during the third generation, the first of what came to be called
supercomputers were designed. Control Data Corporation introduced the
CDC 6600, widely considered to be the first supercomputer, in 1964. A key
contributor to the machine’s design was Seymour Cray, who later left CDC
to form his own very famous supercomputer company. The 6600 was the
fastest computer in the world at the time—roughly three times the speed
of the IBM Stretch machine, which previously held that title. The main
processor operated at a 10 MHz clock frequency (unheard of at the time)
and was supported by 10 parallel execution units. It could execute a then-
astonishing 3 million instructions per second and cost about $7 million in
mid-1960s dollars.
The 6600 was followed in 1968 by the even more powerful CDC 7600,
which topped out at 15 million instructions per second. The 7600 had a
heavily pipelined architecture and is considered the first vector proces-
sor (Mr. Cray would go on to design many more). Other third-generation
vector computers included the CDC Star-100 and Texas Instruments’
Advanced Scientific Computer (TI-ASC), both announced in 1972. The
third generation also saw the development of some of the first high-
performance parallel processing machines, including Westinghouse
Corporation’s SOLOMON prototype and later the ILLIAC IV (a joint ven-
ture between the University of Illinois, the U.S. Department of Defense,
and Burroughs Corporation).
The third generation also gave us the first example of a “family” of
computers: the IBM System/360 machines. IBM offered multiple models
of the 360, from low-end machines intended for business applications
to high-performance models aimed at the scientific computing market.
Although their implementation details, performance characteristics, and
price tags varied widely, all of the IBM 360 models could run the same
software. These machines proved very popular, and IBM sold thousands
of them. Ever since then, it has become common practice for computer
manufacturers to offer entire lines of compatible machines at various
price and performance points.
Software-related developments during the third generation of com-
puting included the advent of timesharing operating systems (including
the first versions of UNIX). Virtual memory became commonly used,
and new, more efficient computer languages were developed. Although
third-generation hardware was becoming more complex, computer lan-
guages were being simplified. Combined Programming Language (CPL),
developed circa 1964 at Cambridge University and the University of
London, was an attempt to streamline the complicated Algol language
by incorporating only its most important features. Ken Thompson of Bell
Laboratories continued the trend of simplifying computer languages (and
their names), introducing the B language in 1970. Move to the head of the
class if you can guess which language he helped develop next!
10 Computer Architecture

1.2.4 The fourth generation

The fourth generation (approximately mid-1970s to 1990) saw the con-
tinuing development of large-scale integration (LSI) and very large-scale
integration (VLSI) circuits containing tens or hundreds of thousands, and
eventually millions, of transistors. For the first time, an entire CPU could
be fabricated on one semiconductor microcircuit. The first microprocessor,
or “CPU on a chip,” was Intel’s 4-bit 4004 processor, which debuted in
1971. It was too primitive to be of much use in general-purpose machines,
but useful 8-, 16-, and even 32-bit microprocessors followed within a few
years; soon essentially all computers had a single-chip microprocessor
“brain.” Semiconductor main memories made of VLSI RAM and ROM
devices became standard, too. (Although core memory became extinct,
its legacy lives on in the term core dump, which refers to the contents of
main memory logged for diagnostic purposes when a crash occurs.) As
VLSI components became widely used, computers continued to become
smaller, faster, cheaper, and more reliable. (The more components that are
fabricated onto a single chip, the fewer chips that must be used and the less
wiring that is required. External wiring is more expensive and more eas-
ily broken than on-chip connections and also tends to reduce speeds.) The
Intel 486 CPU, introduced in 1989, was the first million-transistor micro-
processor. It featured an on-chip floating-point unit and cache memory
and was in many ways the culmination of fourth-generation computer
technology.
The invention of the microprocessor led to what was probably the
most important development of the fourth generation: a new class of com-
puter system known as the microcomputer. Continuing the trend toward
smaller and less expensive machines begun by the minicomputers of the
1960s and early 1970s, the advent of microcomputers meant that almost
anyone could have a computer in his or her home or small business. The
first microcomputers were produced in the mid- to late 1970s and were
based on 8-bit microprocessors. The Altair computer kit, introduced in
1975, was based on the Intel 8080 microprocessor. More than 10,000 Altairs
were shipped to enthusiastic hobbyists, and the microcomputer revolution
was underway. (Bill Gates and Paul Allen of Microsoft got their start in
the microcomputer software business by developing a BASIC interpreter
for the Altair.) Californians Steve Wozniak and Steve Jobs quickly joined
the wave, developing the Apple I computer in Wozniak’s garage using a
Mostek 6502 as the CPU. They refined their design and created the Apple
II in 1977; it outsold its competitors (the 6502-based Commodore Pet and
the Radio Shack TRS-80, based on a Zilog Z80 processor) and became a
huge success.
IBM saw the success of the Apple II and decided to enter the micro-
computer market as well. The IBM Personal Computer (or PC) was an
Chapter one: Introduction to computer architecture 11

immediate hit, prompting other manufacturers to create compatible “PC

clones” using Intel’s 16-bit 8086 processor. (The IBM PC and PC/XT actu-
ally used the cheaper 8088 chip, which was architecturally identical to the
8086 but had an 8-bit external interface.) The availability of compatible
machines from competing manufacturers helped bring down the price of
hardware and make PCs a mass-market commodity.
While IBM was enjoying the success of the original PC and its suc-
cessor the PC/AT (which was based on Intel’s faster 80286 CPU), Wozniak
and Jobs were developing new systems around Motorola’s 16-/32-bit 68000
family of microprocessors. Their ambitious Apple Lisa (the first microcom-
puter with a graphical user interface) cost too much (about $10,000) to gain
wide acceptance. However, its successor, the Apple Macintosh (launched
in 1984), incorporated many of Lisa’s features at about one-fourth the price
and gained a large following that continues to the present day.
Acceptance of microcomputers was greatly increased by the develop-
ment of office applications software. Electric Pencil, written by Michael
Shrayer for the Altair, was the first microcomputer word processing pro-
gram. Electric Pencil was not a big commercial success, but it was fol-
lowed in 1979 by WordStar, which gained widespread acceptance. dBase,
the database management package, and VisiCalc, the first microcomputer
spreadsheet program (originally developed for the Apple II) also appeared
in 1979. In particular, the spreadsheet program VisiCalc and its successor
Lotus 1-2-3 (developed for the IBM PC) helped promote the use of micro-
computers in business. Microsoft, which got its start with Altair BASIC,
won the contract to develop the PC-DOS (later generically marketed as
MS-DOS) operating system and a BASIC interpreter for the IBM PC. Soon
Microsoft branched out into applications as well, introducing Microsoft
Works (a combined word processor, spreadsheet, database, graphics, and
communication program) in 1987. The Windows operating system and
Microsoft Office (the successor to Works) gained massive popularity in
the 1990s, and the rest, as the saying goes, is history.
While microcomputers were becoming smaller, less expensive, and
more accessible to the masses, supercomputers were becoming more pow-
erful and more widely used. The first new supercomputer of the fourth
generation, the Cray-1, was introduced in 1976 by Seymour Cray, who had
left CDC to form his own company, Cray Research, Inc. The Cray-1 cost
about $8 million and could execute more than 80 million floating-point
operations per second (MFLOPS). Within a few years, Cray followed this
machine with the X-MP (Cray’s first multiprocessor supercomputer) in
1982, the Cray-2 in 1985, and the Y-MP in 1988. The eight-processor Y-MP
had a peak calculation rate of more than 2600 MFLOPS—about 30 times
the performance of the Cray-1.
Cray Research was not the only company developing fourth-generation
supercomputers. Most of Cray’s competition in the area of high-end vector
12 Computer Architecture

machines came from Japanese manufacturers. Some important Japanese

supercomputers of the 1980s included the Nippon Electric Company (NEC)
SX-1 and SX-2 systems, introduced in 1983; Fujitsu’s VP-100 and VP-200
machines (1983), followed by the VP-400 in 1986; and Hitachi’s S820/80
supercomputer released in 1987. Although the dominant supercomputers
of this period were mostly pipelined vector machines, the fourth genera-
tion also saw the debut of highly parallel supercomputers. The Massively
Parallel Processor (MPP), first proposed in 1977 and delivered in 1983 by
Goodyear Aerospace Corporation to NASA Goddard Space Flight Center,
was a one-of-a-kind machine constructed from 16,384 1-bit processors. A
few years later (1986–1987), Thinking Machines Corporation entered the
supercomputer market with its Connection Machine CM-1 and CM-2 sys-
tems. These machines contained 65,536 processors each; with more than
70 installations, they were the first commercially successful massively
parallel supercomputers.
Although most of the new developments of the fourth generation of
computers occurred at the high and low ends of the market, traditional
mainframes and minicomputers were still in widespread use. IBM con-
tinued to dominate the mainframe computer market, introducing a series
of upgrades (more evolutionary than revolutionary) to its System/370 line
of machines, which had replaced the System/360 in 1970. Among IBM’s
workhorse fourth-generation mainframes were the 3030 series (1977–
1981), the 3080 series (1980–1984), the 3090 series (1985–1989), and the 4300
series (1979–1989). All of these machines saw extensive use in a variety of
medium-to-large business applications.
The dominant minicomputer of the fourth generation was Digital
Equipment Corporation’s VAX series, which was a successor to the
popular PDP-11. DEC’s VAX 11/780, released in 1977, was the first mini-
computer with a 32-bit architecture, allowing large amounts of memory
to be addressed at once by the programmer. It was also the first mini-
computer to execute one million instructions per second. (The VAX
acronym stood for Virtual Address eXtension, and the operating sys-
tem, VMS, stood for Virtual Memory System.) To address compatibility
concerns, early VAX models incorporated a PDP-11 emulation mode to
ease migration to the newer system. The 11/780 was followed by mod-
els 11/750 and 11/730, which had close to the same performance but
were smaller and cheaper. Higher performance needs were addressed
by the dual-processor 11/782 and the 11/785. Largely due to the success
of these computers, by 1982, DEC was the number two computer com-
pany in the world, behind only IBM. DEC remained a major force in the
market as it continued to expand the VAX line through the 1980s, devel-
oping the higher performance 8x00 series machines and the microVAX
line of microcomputers that were compatible with DEC’s larger and
more expensive minicomputers.
Chapter one: Introduction to computer architecture 13

Characteristics of fourth-generation machines of all descriptions

include direct support for high-level languages either in hardware (as in
traditional Complex Instruction Set Computer [CISC] architectures) or
by using optimizing compilers (characteristic of Reduced Instruction Set
Computer [RISC] architectures, which were developed during the fourth
generation). The MIPS and SPARC architectures gave rise to the first RISC
microprocessors in the mid-1980s; they quickly rose to challenge and, in
many cases, replace dominant CISC microprocessors such as the Motorola
680x0 and Intel x86 CPUs as well as IBM’s mainframes and DEC’s mini-
computers. We study these competing schools of architectural thought
and their implications for computer system design in Chapters 3 and 4.
With the advent of hardware memory management units, time-
sharing operating systems and the use of virtual memory (previously
available only on mainframes and minicomputers) became standard on
microcomputers by the end of the fourth generation. By giving each pro-
gram the illusion that it had exclusive use of the machine, these techniques
made programming much simpler. Compilers were considerably improved
during the 1970s and 1980s, and a number of new programming lan-
guages came into use. BASIC, a language simple enough to be interpreted
rather than compiled, became popular for use with microcomputers. In
1984, Borland’s inexpensive Turbo Pascal compiler brought high-level lan-
guage programming to the personal computer mainstream. Meanwhile,
C (developed in 1974 by Dennis Ritchie of Bell Laboratories as a refine-
ment of Ken Thompson’s B) became the dominant language for systems
(especially UNIX) programming. Fortran 77 was widely used for scientific
applications, and Ada (named for the Countess of Lovelace) was adopted
by the U.S. military in 1983 for mission-critical applications. Finally, object-
oriented programming got its start during the fourth generation. C++ was
developed by Bjarne Stroustrup at Bell Laboratories during the early 1980s;
the first version of the new OO language was released in 1985.

1.2.5 The fifth generation

The fifth generation (approximately 1990 to 2005) of computing systems
can arguably be termed more evolutionary than revolutionary, at least in
terms of architectural features. Machines of this era used VLSI and ultra
large-scale integration (ULSI) chips with tens (eventually hundreds) of
millions of transistors to perform many complex functions, such as graph-
ics and multimedia operations, in hardware. Processors became more
internally parallel (using techniques such as superscalar and superpipe-
lined design to execute more instructions per clock cycle), and parallelism
using multiple processors, once found only in mainframes and supercom-
puters, started becoming more common even in home and small business
systems. Processor clock frequencies that reached the tens of megahertz
14 Computer Architecture

in fourth-generation machines increased to hundreds and thousands of

megahertz (1 GHz = 1000 MHz). Intel’s Pentium microprocessor, intro-
duced in 1993, had two pipelined execution units and was capable of exe-
cuting 100 million instructions per second. Ten years later, its successor,
the Pentium 4, reached clock speeds of 3.2 GHz on a much more highly
concurrent internal microarchitecture, implying a microprocessor perfor-
mance improvement of roughly two orders of magnitude (a hundredfold)
over that time span.
With microprocessors becoming ever more powerful, microcomput-
ers were the big story of computing in the fifth generation. Personal com-
puters and scientific and engineering workstations powered by single or
multiple microprocessors, often coupled with large amounts of memory
and high-speed but low-cost graphics cards, took over many of the jobs
once performed by minicomputers (which effectively became extinct) and
even mainframes such as IBM’s zSeries (which were mostly relegated to
important but “behind the scenes” applications in business transaction
processing). Even the supercomputers of this generation increasingly
made use of standard microprocessor chips instead of custom-designed,
special-purpose processors.
Supercomputing experienced a transformation during the fifth gen-
eration. The high-speed, pipelined vector processors of the late 1970s and
1980s only proved cost-effective in a limited number of applications and
started to fall out of favor for all but extremely numerically intensive
computations done by well-funded government agencies. By the 1990s,
Fujitsu and other manufacturers quit making vector computers, leaving
NEC and Cray as the only vendors of this type of system. Cray Computer
Corporation was spun off from Cray Research, Inc., to develop the Cray-3,
but only one machine was delivered before the company filed for Chapter
11 bankruptcy in 1995; plans for the Cray-4 were abandoned. Tragically,
supercomputer pioneer Seymour Cray met the same fate as Cray Computer
Corporation, perishing in an automobile accident in 1996.
Cray Research, Inc., began to move toward nonvector, massively par-
allel, microprocessor-based systems in the mid-1990s, producing its T3D
and T3E machines before being bought by Silicon Graphics, Inc. (SGI), in
1996. SGI’s Cray Research division produced the SV1 vector supercom-
puter in 1998 but was sold in March 2000 to Tera Computer Company
and renamed Cray, Inc. Cray introduced the X1 (formerly codenamed the
SV2) in 2002. Meanwhile, in 2001 NEC signed an agreement with Cray,
Inc., to market its SX-5 and SX-6 supercomputers, effectively combining
all remaining high-end vector machines into a single vendor’s product
line.
By June 2005, only 18 of the top 500 supercomputers were vector
processors (including nine Crays and seven NEC machines). The other
482 were highly parallel scalar machines, mostly built by IBM (259),
Another Random Scribd Document
with Unrelated Content
 vivacious.”
+ Outlook. 81: 382. O. 14, ‘05. 90w.
—
* “He has a sense of atmosphere, his point of view is
individual, and he is not without that kindly humour which
laughs while it sympathises. But he is terribly sentimental.”
+ Sat. R. 100: sup. 6. D. 9, ‘05. 170w.
—

* Long, John Luther. Miss Cherry Blossom of Tokyo. † $2.50.

Lippincott.
A reincarnation of this Japanese romance, in which wide
margins, Japanese flowers and fancies which wander across
the text, and various full page illustrations, some of which are
in color, lend to the interesting story of Sakura-San and the
“excellent barbarians” from England and America who play at
cross purposes thruout its pages a new and subtle charm.
* “In this and ‘Madame Butterfly’ he is seen at his best.”
+ Critic. 47: 583. D. ‘05. 20w.
+

* + Dial. 39: 449. D. 16, ‘05. 110w.

* “The author has contrasted Oriental and Occidental traits in
his well-known style.”
+ Ind. 59: 1378. D. 14, ‘05. 30w.

* Nation. 81: 381. N. 9, ‘05. 80w.

* + N. Y. Times. 10: 822. D. 2, ‘05. 160w.

 * “Told with charm and well-rendered Oriental atmosphere.”
+ Outlook. 81: 683. N. 18, ‘05. 50w.

* Long, John Luther. Seffy; a little comedy of country

manners. †$1.50. Bobbs.
Old Baumgarten, a Pennsylvania-German and Maryland farmer,
has set his heart upon marrying his great slow going son,
Seffy, to a red-headed, tempestuous girl, named Sally, who
owns the lands lying between his farm and the railroad. He
almost brings this about, but Seffy’s reticence allows another
lover to come between him and his sweetheart. Sally marries
out of spite and comes to bitterly repent of it, while old
Baumgarten curses his son, knocks him down and sends him
out into the world where he learns to fight for things and to
get them. In the end he comes back to claim all that he lost in
his youth.
* + Dial. 39: 447. D. 16, ‘05. 140w.

Long, William Joseph. Northern trails: stories of animal life in

the far north. *$1.50. Ginn.
“Mr. Long takes the reader with him ... to the barren shores of
Labrador and Newfoundland. Wolves, we meet, that guide lost
children home, and then disappear into the wilderness; a wild
goose that caresses his mate goodbye at the approach of the
hunter, before going out to fight for his home and young; and
Pequam, of the weasel family, that tempts an Indian to
abandon his trail, by killing a deer and leaving it across the
track. These animals and many more—whales, polar bears and
salmon—are all introduced to us in the midst of their wild,
unfrequented haunts. All are endowed with almost human
intelligence and reason, after the manner of interpreting their
actions which Mr. Long has made so popular.”—Ind.
* “There is a charm about Mr. Long’s book that few writers for
children attain.”
+ Acad. 68: 1287. D. 9, ‘05. 150w.
+
* “His stories have a charm and an excellence of their own.”
May Estelle Cook.
+ Dial. 39: 373. D. 1, ‘05. 190w.
“We are willing to let the disputed question of instinct or
intelligence go, however, and on the strength of the splendid
descriptions of nature and the always evident love of the wild,
accord this volume a high place among ‘books of the trail.’”
+ Ind. 59: 873. O. 12, ‘05. 250w.
+
“Mr. Long assures us of the accuracy of his data, and maintains
the reasonableness of his inferences.”
+ Nation. 81: 340. O. 26, ‘05. 290w.
+
+
* “Aside from the controversial side as to whether these eight
stories are to be classified as natural history or fiction, these
tales of the Northern trails are dull and lifeless.” Mabel Osgood
Wright.
— N. Y. Times. 10: 872. D. 9, ‘05. 560w.
+

* + Outlook. 81: 718. N. 25, ‘05. 100w.

“His this year’s story is vigorous, delightful, and refreshing.”
+ Pub. Opin. 39: 601. N. 4, ‘05. 100w.
+
 * + R. of Rs. 32: 754. D. ‘05. 130w.

Long day. See Richardson, Dorothy.

Loomis, Charles Battell. Minerva’s manoeuvres: the cheerful

facts of a return to nature. †$1.50. Barnes.
A novel which “recounts the experiences of Mr. and Mrs. Philip
Vernon and their city-reared colored cook, Minerva, during a
summer sojourn in the country.... A fine silk thread of a plot
runs through the book, stringing together the many humorous
situations.”—Pub. Opin.
* + Critic. 47: 478. N. ‘05. 80w.
“It is a good book to read aloud, but only a chapter or two at a
time.”
+ Ind. 59: 696. S. 21, ‘05. 130w.
* “The unexpected endings of the many humorous situations
will keep the reader in a gale of mirth, and when he lays the
book down after the last chapter, he will feel that he has found
a new friend in Minerva.”
+ Lit. D. 31: 797. N. 25, ‘05. 520w.
* “Is more in the nature of a vaudeville show than anything
else, and it is not possible to describe all of the attractions
which Mr. Loomis offers. They are surely worth a reading.”
+ N. Y. Times. 10: 800. N. 25, ‘05. 460w.

+ Outlook. 81: 88. S. 9, ‘05. 50w.

“It is not an uproarious story; its humor is quiet; it possesses
the subtle turn which is symptomatic of its author.”
 + Pub. Opin. 39: 384. S. 16, ‘05. 230w.

Lord, Eliot; Trenor, John J. D.; and Barrows, Samuel

June. Italian in America. $1.50. Buck.
“The pretty evident purpose of this volume is to reverse the
prevailing American prejudice against the Italian as an
immigrant and material for United States citizenship.... [It]
uses ... the argument ... of statistics, and its authors ...
attempt to show first of all that the Italian settler is
economically a good thing for the country.... Secondly, they
produce evidence that in the particulars of disease and crime
he does not supply more than his quota ... and, thirdly, they
argue from data which they present that he ... adapts himself
very completely ... to American ways of doing and thinking.”—
N. Y. Times.
“The book as a whole is general in its treatment, somewhat
objectionable because of frequent quotations, and partakes too
much of the loose character of magazine articles. The spirit of
the book is much to be commended.” Emily Fogg Meade.
+ Ann. Am. Acad. 26: 609. S. ‘05. 460w.
—
“The book is optimistic, discriminating and instructive.”
+ Engin. N. 53: 532. My. 18, ‘05. 110w.
+
“Is of normal simplicity and clearness.”
+ Ind. 59: 579. S. 7, ‘05. 120w.
“There is room for believing that ‘The Italian in America’ will be
a potent instrument in molding a saner public opinion.”
+ Lit. D. 31: 666. N. 4, ‘05. 800w.
+
 “But the labors of others are here presented in logical
sequence and in a sympathetic spirit, resulting in an interesting
and readable book. The book is not free from dubious
assertions.”
+ Nation. 80: 361. My. 4, ‘05. 1040w.
+
—
“It will strike many perhaps that Messrs. Lord, Trenor, and
Barrows have omitted some essential facts, but both the facts
presented and the inferences drawn are interesting in
substance—even when the manner of presentation is dry.
Taken all together the cumulative evidence for the Italian
collected by the authors is impressive.”
+ N. Y. Times. 10: 259. Ap. 22, ‘05. 1050w.
—
“Welcome as a wholesome corrective of fallacy and prejudice.”
+ Outlook. 80: 140. My. 13, ‘05. 460w.
+

Lorenz, Daniel Edward. Mediterranean traveller, *$2.50.

Revell.
A compact practical guide-book which covers southern Spain,
Morocco, Algiers, the chief cities of northern Africa, Greece,
Asia Minor, Palestine, and Egypt. Much historical and general
information is given, a bibliography precedes each chapter, and
there are many maps and pictures.
“This compact work ‘fills a long felt want.’ The proof reading
has not been done by a classical expert.”
+ Critic. 47: 190. Ag. ‘05. 160w.
+
—
 “The text is in some portions accurate and business-like, but in
others it reveals amateurishness, and some inaccuracies and
misprints.”
+ Nation. 80: 211. Mr. 16, ‘05. 370w.
+
—
“The excellence of its method and treatment of the many
countries bordering on the great interior sea of Europe are
unquestionable.”
+ N. Y. Times. 10: 140. Mr. 4, ‘05. 260w.
+
+
“‘The Mediterranean traveler’ will find here in one volume what
elsewhere must be picked out of several.”
+ Outlook. 79: 451. F. 18. ‘05. 60w.

Loring, Andrew. Rhymer’s lexicon; with an introd. by George

Saintsbury. *$2.50. Dutton.
“The lexicon is divided into three parts—Finals, Penults, and
Antepenults. The words have been grouped according to the
accented vowel sound and placed in columns in the
alphabetical sequence of the letters which follow this sound....
Each part of the lexicon has fourteen vowel divisions, adopted
for reference purposes; and the divisions are enumerated in a
table of contents, which also includes key words illustrating the
vowel sounds.”—N. Y. Times.
“In size and arrangement it is admirable; it might have been
larger still without being any better.”
+ Acad. 68: 678. Jl. 1, ‘05. 280w.
—
 + Ath. 1905, 1: 623. My. 20. 450w.

+ N. Y. Times. 10: 409. Je. 17, ‘05. 290w.

+
“Altogether an able book, full of aid to those who make
rhymes.”
+ N. Y. Times. 10: 531. Ag. 12, ‘05. 320w.
“The arrangement is novel, at first sight a little intricate, but
truly scientific.”
+ Outlook. 80: 694. Jl. 15, ‘05. 100w.
“This may be a very useful book.”
+ Spec. 94: 791. My. 27, ‘05. 240w.

Loring, J. Alden. Art of preserving animal tracks, $1. J. A.

Loring, Owega, Tioga co., N. Y.
“Mr. Loring describes in this pamphlet a very ingenious and
apparently effective method of making molds, and from the
mold, casts of the tracks of mammals and birds, large and
small.... The operation itself is clearly and minutely described,
and seemingly could be easily managed by any intelligent
boy.”—Outlook.
Outlook. 79: 757 Mr. 25, ‘05. 130w.

* Lothrop, Harriet Mulford Stone (Margaret Sidney,

pseud.) Ben Pepper. †$1.50. Lothrop.
This is the tenth volume in the popular “Little Pepper” series.
“The hero is Ben, Mother Pepper’s first-born, her ‘steady-as-a-
rock’ boy. Christmas shopping in which the Little Peppers take
 a lively hand, Christmas philanthropies, the usual quota of
accidents and pranks, and, finally, Ben’s decision as to whether
he will go to college or enter a business office, ‘beginning at
the very bottom,’ are the features of the story.”—Outlook.
* “Mrs. Sidney has made him as interesting as others of the
Pepper family.”
+ N. Y. Times. 10: 708. O. 21, ‘05. 100w.

* + Outlook. 81: 278. S. 30, ‘05. 70w.

* + R. of Rs. 32: 765. D. ‘05. 70w.

Lott, Noah, pseud. See Hobart, George Vere.

* Lottridge, Silas A. Animal snap-shots and how made. **$2.

Holt.
A simple narrative concerning the birds and mammals which
the author has come to know in the course of various
vacations spent in studying and photographing them. The
pictures illustrate the facts and some of the series represent
the work of years. The object of the book is to arouse,
especially in young people, a living interest in the animals
about them. There are chapters on the woodchuck, skunk,
muskrat, fox, mouse, squirrel, blue bird, robin, bobolink, crow,
owl, hawk and others.

Lovett, W. J. Complete class book of naval architecture;

practical, laying off, theoretical, with numerous il. and nearly
200 full, worked-out answers to recent education department
examination questions. *$2.50. Longmans.
 “This work is intended primarily for British students attending
technical classes.” It “covers the whole field of naval
architecture, theoretical and practical.”—Engin. N.
“In this country its field as a text-book will necessarily be
limited, and as a reference book its treatment of the various
subjects, except elementary ship calculations, is inadequate.”
D. W. Taylor.
+ Engin. N. 53: 529. My. 18, ‘05. 1030w.
—

Low, Berthe Julienne. French home cooking. **$1.20.

McClure.
The author, tho a Frenchwoman by birth, has lived in this
country twenty-five years. She says: “This is not a book for
restaurants, hotels, or people who can afford a chef. Most
Americans have formed their ideas of French cooking from
hotels, restaurants, or formal dinners, and have never known
the home cooking, which is more simple and more wholesome.
It is also less complicated.... The recipes which I shall give are
used in well-to-do families and constitute what is called in
French the ‘bonne cuisine bourgeoise.’” She starts with the
very arrangement of the kitchen and instructs in those little
tricks by which the French are able to obtain distinction and
flavor in their cookery.
+ Critic. 46: 565. Je. ‘05. 60w.
“Mrs. Low’s formulas are in the main so excellent that it would
be invidious to discriminate. Her success is unequivocal and
decisive.”
+ Nation. 80: 78. Ja. 25, ‘05. 490w.
+
 + N. Y. Times. 10: 34. Ja. 21, ‘05. 330w.
+

Low, Sidney. The governance of England. *$2.25. Putnam.

“A well-informed, well-written, and interesting description of
the government of Great Britain, beginning with a definition of
the British constitution, so difficult of characterization, but
explained by Mr. Low in a thoroughly rational and
comprehensive way.... Mr. Low gives a very interesting account
of the place and function of the prime minister, of the cabinet,
of the privy council, of both houses of parliament, and of every
other form and function of government in Great Britain.”
(Outlook). “The main view of Mr. Low is that of Lord Salisbury
and Mr. Balfour, that the power of the house of commons is
declining and must continue to decline, while that of the
cabinet, and especially of the inner cabinet, is increasing.”
(Ath.)
“Is a most able and valuable production, marked, too, by
unusual excellence of style. If we name points on which we
have doubts as to whether Mr. Low is right, it is with the
profound feeling that he has given great attention to a subject
in which he evidently takes much interest, and the facts of
which, so far as they are generally available, he has mastered.”
+ Ath. 1905, 1: 79. Ja. 21. 1980w.
+
—
“There can be little but praise for the author’s literary style. It
is easy, strong and clear, and with a light touch and aptness of
allusion that never detract from the weighty theme.” John
William Russell.
+ Bookm. 22: 57. S. ‘05. 1400w.
+
 +
“There are many clever and some acute observations in the
book; but, in our judgment, the view given of the English
constitution is superficial, and in some cases erroneous.”
— Nation. 80: 400. My. 18, ‘05. 2010w.
+
“Very admirable book. The plan of the work is so excellently
conceived and executed that only one or two objections are
suggested by a first reading. One is to the title. The other
objection is to an occasional drop into triviality and the college
graduate habit of quoting mere hackneyed phrases and tags
from other languages. There are occasional slips in the printing
and in the statements. The particular excellence of this work
has already been indicated as being an interpretation of the
English constitution as it operates to-day. The value of this
book is very greatly increased for American readers by the
frequent comparisons instituted between the English and
American political systems.”
+ N. Y. Times. 10: 132. Mr. 4, ‘05. 1650w.
+
—
“Probably no book has yet appeared which, in so untechnical
and comprehensive a way, places before the reader the
elaborate, highly complex, and thoroughly democratic
governmental system.”
+ Outlook. 79: 448. F. 18, ‘05. 200w.
+
+

+ R. of Rs. 31: 382. Mr. ‘05. 70w.

+
* Lowell, Mrs. Carrie Thompson, comp. Art lovers’ treasury;
famous pictures described in poems; forty-eight
reproductions of famous pictures accompanied by poems of
noted writers, with text by Carrie Thompson Lowell. **$1.20.
Estes.
“Forty-eight pictures are reproduced in half-tone, and the
editor writes a running comment, treating a group of paintings
and sculpture under some general heading, such as ‘Mythology
in poetry and sculpture,’ ‘Legends of the saints,’ or ‘Pictures
translated into verse.’” (Dial.) “An attempt has been made to
assemble famous pictures and pieces of sculpture, and to bring
into association with these certain poems that have been
inspired by the various works of art that are pictured, or which
have been written descriptive of them.... Many of the best
artists are represented, as well as poets such as Dante, Keats,
Browning, Longfellow, Whittier, Markham, and some others.”
(Ind.)
* “An excellent companion volume to Miss Singleton’s ‘Great
portraits’ is this compilation of Mrs. Lowell.”
+ Critic. 47: 572. D. ‘05. 30w.
* “Pictures and poetry are thoroughly representative, and the
arrangement, though necessarily loose holds the reader’s
interest.”
+ Dial. 39: 446. D. 16, ‘05. 130w.

* + Ind. 59: 1376. D. 14, ‘05. 100w.

* + N. Y. Times. 10: 874. D. 9, ‘05. 150w.

* Lowery, Woodbury. Spanish settlements within the present

limits of the United States: Florida. 1562-1574. **$2.50.
Putnam.
 “This, the second of Mr. Lowery’s monographs on the history of
Spanish colonization within the present limits of the United
States, deals with the Florida settlements of the period 1562-
1574, and like its predecessors, is based on a careful study of
original sources.... An interesting feature is—comprehensive
exposition of the tribal organization, characteristics and
customs of the Florida Indians. The work contains several
maps, more than thirty bibliographical and critical appendices,
and a good index.”—Outlook.
* “Scholarly work.”
+ Ind. 59: 1156. N. 16, ‘05. 20w.
* “A treatise not only of prime interest but of solid value, as
embodying a broader and more than usually judicial statement
of the vexed themes involved.”
+ Lit. D. 31: 797. N. 25, ‘05. 600w.
+
—

* + N. Y. Times. 10: 786. N. 18, ‘05. 290w.

+
* “So cautious is he, and so frequent are his references to and
citations from authorities, that from the narrative standpoint
his book is at times arid and tedious. But it is unquestionably
of distinct value to the historical student.”
+ Outlook. 81: 431. O. 21, ‘05. 340w.
+
—

Lowrie, Rev. Walter. Gaudium crucis: a meditation for Good

Friday upon the seven words from the cross. *90c.
Longmans.
 Meditations upon mercy, judgment, love, joy and sacrifice,
confirmation, accomplishment and duty, and filial trust. The
book is designed for those who are unable to attend the Good
Friday services, and to assist the clergy in preparing their
sermons.
Outlook. 79: 705. Mr. 18, ‘05. 70w.

* Loyson, Mme. Emilie Jane (Butterfield) Meriman (Mme.

Hyacinthe Loyson). To Jerusalem through the lands of
Islam, among Jews, Christians, and Moslems. $2.50. Open
ct.
“‘A tour of Christian exploration.’ Pere Hyacinthe and his wife
(who is an American) travelled from Algeria to Jerusalem, by
way of Arabia and Egypt, and the travels are described in a
lively and vigorous style.... The idea of the book is not the
travel, so much as the relativity of religions of the peoples
studied ... and Madame Hyacinthe Loyson’s point is the
universal brotherhood of ... the religions of Allah and Jehovah
and the Christian religion. In the co-operation of the three—
and in the honouring by modern Christianity of some of the
grander and simpler elements of the other two faiths, she sees
the regeneration of the world.”—Acad.
* “There is a breadth of view in the book, enthusiasm and
some little of that spirit which sees good in ‘every country but
its own.’ It will not please theologians, but it may stimulate the
thoughts of the ordinary religiously-minded man or woman.”
+ Acad. 68: 1237. N. 25, ‘05. 260w.
* “Everywhere in the book there is the intense spiritual
earnestness of a good woman holding conferences with the
leading representatives of Islam.” H. E. Coblentz.
+ Dial. 39: 379. D. 1, ‘05. 340w.
 * + Outlook. 81: 384. O. 14, ‘05. 230w.

* + R. of Rs. 32: 636. N. ‘05. 60w.

Lucas, Abner H. Call of to-day. *50c. Meth. bk.

Sermons preached in the First Methodist Episcopal church,
Montclair, N. J. They include: The religion for to-day; Work for
to-day; The commanded strength; Joy for the morning; The
mighty appeal of usefulness; Re-enlisted strength; and The
complete life.

Lucas, Edward Verrall, comp. Book of verses for children. $1.

Holt.
Some 200 verses which Stevenson, Browning, Shakespeare,
Goldsmith, Lewis Carroll, Riley, Longfellow, Scott, Rossetti, and
many others have written for little folks are gathered into this
delightful volume, with old ballads, rhymes and songs of
Christmas.
“Altogether, the little volume is one of the most desirable of
such collections (in small space) now to be got at. There
seems to be something in it for all good juvenile tastes.”
+ N. Y. Times. 10: 575. S. 2. ‘05. 230w.
+

* Lucas, Edward Verrall. Life of Charles Lamb. 2v. *$6.

Putnam.
Mr. Lucas “has tried as far as possible to keep the story of the
life to the words of the original performers and their
contemporaries.... At a wave of his hand witness after witness
gets up to testify in his own words and tell the reader what he
knew of Lamb during the period in question.... We are able to
see the actual environment of Lamb between 1815 and 1825,
surrounded ... by the normal frequenters of these ‘noctes’ such
as George Dyer, Fenwick, Robert Fell, Martin Burney, G.
Burnett, Randal Norris, George Dawe, Ayrton, Phillips, Alsager,
and Barren Field. The portraits of most of these intimates of
the Mitrecourt and Inner Temple-lane are limned with a
delicate and artistic curiosity. Lamb is depicted in this circle as
he lived.... For all the very happiest things that have ever been
said about Lamb the enthusiast will find a happy-hunting-
ground in these two volumes.”—Lond. Times.
* “Only once, so far as we have noticed, is he betrayed into
something like over-confidence in his minute research.”
+ Acad. 95: 999. S. 30, ‘05. 1810w.
+
—
* “Of the man Charles Lamb—the ‘human mortal,’ as
distinguished from the thinker and writer—Mr. Lucas’s pages
reflect a true and lively image. He is less successful in
reproducing the intellectual features of his subject; while his
portraits of certain of Lamb’s contemporaries—notably that of
Coleridge—are not far removed from travesty.”
+ Ath. 1905, 2: 756. D. 2. 880w.
+
—
* “Is likely to prove of more importance than the recent edition
of ‘The works and letters of Charles and Mary Lamb,’ of which
he was the editor. It will not supersede the ‘Life and final
memorials’ of Talfourd, but it contains, mainly in the form of
letter and anecdote, much of supplementary value, and some
matter which is absolutely fresh.” H. W. Boynton.
+ Atlan. 96: 844. D. ‘05. 1080w.
+
 —
* “Taking Mr. Lucas’s biography as a whole there is a wealth of
entertainment in its pages which it would be difficult to
overestimate. The part that we are least sanguine of
recommending is the appendix, which seems to us a heavy
incubus upon a book which ought to carry not an ounce of
superfluous material.”
+ Lond. Times. 4: 297. S. 22, ‘05. 2440w.
+
—
* “Mr. Lucas has drawn upon a large fund of fresh material,
and has so generously told the story of both lives in the
language of his subjects that this biography is really an
autobiography.” Hamilton W. Mabie.
+ N. Y. Times. 10: 827. D. 2, ‘05. 380w.
+
+

* + N. Y. Times. 10: 836. D. 2, ‘05. 200w.

+
* “Above all other things Mr. Lucas’s work is one which
abounds in the essential characteristic of biographical work—
sympathy. The vast compilation of tiny details of personality
and character are not thrown together haphazard but are
arranged chronologically, and indexed in a masterly manner.”
+ Pub. Opin. 39: 820. D. 23, ‘05. 690w.
+
* “The first really complete and adequate Life of that singularly
delightful writer and admirable man.”
+ Spec. 95: 653. O. 28, ‘05. 1550w.
+
 +

Lucas, Edward Verrall. Wanderer in Holland. *$1.75.

Macmillan.
“The combination of Mr. Lucas as narrator with Mr. Herbert
Marshall as illustrator has given us a charming volume.... It
was a happy idea to intersperse photographs of some of the
more famous Dutch pictures. Mr. Lucas is an admirable guide
and visitors to Holland could not have a more agreeable
commentator on their travels past or future.... He not only
abounds in wise and quaint comments himself, but is the cause
of our remembering the wisdom of others.”—Sat. R.
“‘A wanderer in Holland’ is, of course, no substitute for Murray
or Baedeker, rather is it their essential complement.”
+ Acad. 68: 920. S. 9, ‘05. 1050w.
+
“If the success of a book of travels is to be measured by the
travel-fever it excites in the veins of its readers, this volume
should have a warm welcome.”
+ Ath. 1905, 2:571. O. 28. 2300w.
+
“And now we have found all the fault we care to find with this
charming guide. To say that it ranks a long way after ‘The
inland voyage’ is only to say that Stevenson is dead. We
welcome in it a like sense of intimacy—it wears the face of a
friend—it talks.”
+ Lond. Times. 4: 292. S. 15, ‘05. 1150w.
+
—
* “Mr. Lucas makes no pretension to connoisseurship, but his
untechnical remarks on pictures are nearly always interesting,
and, to one reader at least, prove the most attractive part of
his writing.”
+ Nation. 81: 449. N. 30, ‘05. 150w.
“The fact is Mr. Lucas comes near being in his book exactly
what one would like a well-informed and companionable
fellow-traveler to be if one were seeing Holland with one’s own
eyes.”
+ N. Y. Times. 10: 725. O. 28, ‘05. 920w.
+
“In short the book is notable among books of travel and
description for its readable qualities and discriminating and
individual taste.”
+ Outlook. 81: 580. N. 4, ‘05. 120w.
+
+
* “A book of more than ordinary merit—a book with genuinely
original qualities.”
+ Pub. Opin. 39: 826. D. 23, ‘05. 150w.
+

* + R. of Rs. 32: 755. D. ‘05. 60w.

“As we might have expected from his record, he neither bores
nor dogmatises but his book is full of information and not a
little wise reflection.”
+ Sat. R. 100: 346. S. 9, ‘05. 370w.
+
—
“It is as a critic of character and manners and a chronicler of
art that Mr. Lucas interprets his function as a guide.”
 + Spec. 95: 468. S. 30, ‘05. 1330w.
+

Luccock, Rev. Naphtali. Royalty of Jesus and other sermons.

*50c. Meth. bk.
Beauty of thought and simplicity of language mark these
sermons which apply the teachings of Christ to the conditions
of to-day under the titles: The royalty of Jesus; The fullness of
Christ; The power of a surrendered life; The face of Jesus
Christ; The brook in the way; The gospel for an opulent
civilization; The cry of the disinherited; The song of Moses and
of the Lamb.

Lucian (Lucianus Samosatensis). Work of Lucian of

Samosata; trans, by H. W. Fowler, and F. G. Fowler. 4v. *$4.
Oxford.
Four handy volumes in which the translators have happily
rendered idiom by idiom and “literary allusions, quotations,
and technicalities of law, philosophy, or art are neatly turned to
apt analogues. They sound every note in Lucian’s compass,
from the mock-heroic serio-satiric eloquence of the Nigrinus,
the angry contempt of the False prophet and the Death of
Peregrine ... to the solemn trifling of the Fly ... and the
demonstration by Socratic induction in the ‘Parasite’ that dining
out is better than dining.” (Dial.) The fourth volume contains a
list of notes which explain all allusions to classical biography
and mythology.
* “The renderings of Messrs. Fowler have all the ease and
‘élan’ of a work originally written in English.” R. Y. Tyrrell.
+ Acad. 68: 846. Ag. 19, ‘05. 1680w.
+
 “Their translation is decidedly good; they have ventured on
some daring modernisms, but these we can tolerate if only
lightness is secured.”
+ Ath. 1905, 2: 294. S. 2. 2420w.
+
“The translation is admirably executed in the freer manner of
Jowett’s ‘Plato.’” Paul Shorey.
+ Dial. 39: 233. O. 16, ‘05. 1850w.
+
“The editors ... deserve high praise for the clearness and
vigour of their translations.”
+ Lond. Times. 4: 265. Ag. 25, ‘05. 1520w.
+
“This Fowler translation is a work of high art, for which its
authors are to be thanked.”
+ N. Y. Times. 10: 717. O. 21, ‘05. 610w.
+
+
“The translators have with admirable fidelity, vigour, and
vivacity reproduced the writings of one whom such a critic as
Erasmus reckoned not only among the most entertaining, but
also the most instructive, of ancient authors.”
+ Spec. 95: 713. N. 4, ‘05. 1870w.
+
+

Lucke, Charles Edward. Gas-engine design. **$3. Van

Nostrand.
 “The book is divided into three parts: 1, Power and efficiency,
with rules for deciding on the necessary piston displacement;
2, Stresses on the various parts of the engine and also with
the various cylinder arrangements as affecting the turning
effort and balance; 3, The necessary dimensions of the various
parts to resist the stresses with both empirical and theoretical
formulæ for the computation.”—Engin. N.
“It is a very notable addition to the literature on the gas
engine.” Storm Bull.
+ Engin. N. 53: 526. My. 18, ‘05. 1170w.
+
—

Lützow, Francis, count. Lectures on the historians of

Bohemia. Oxford.
The Ilchester lectures for the year 1904 “have their origin in a
wish to do something for the Bohemian cause by illustrating
before a foreign audience the wealth of Czechish traditions....
While Count Lützow alludes briefly to his contemporaries, the
chief of his attention is devoted to the chronicles of the Middle
ages and the era of the Reformation. Here the conditions fixed
by a popular course of lectures compel him to be brief in his
notice of all save the now famous authorities like Cosmas,
Benes of Weitmil, Lawrence of Brezof, Sixt of Ottersdorf, and
Paul Skála.... He gives us simply and tersely the results of the
most recent research on technical points in conjunction with
Palacky’s views on the larger issues.” (Nation.)
“The count, who is a master of our language, goes through the
list of Bohemian historians, estimating their merits and
furnishing characteristic extracts. These are translated into
very clear and succinct English. Excellent book.”
 + Ath. 1905, 1: 710. Je. 10. 1860w.
+
+
“To many of our readers we can best convey an impression
regarding the style and quality of his work by stating that it
resembles a compressed Wattenbach with an element of
current political interest added.”
+ Nation. 81: 85. Jl. 27, ‘05. 1300w.
+
+

Lyall, Sir Alfred Comyn. Lord Dufferin, the life of the Marquis
of Dufferin and Ava. *$7.50. Scribner.
The life of a man to whom fate gave great opportunities, and
who was big enough to handle and hold them. He was a
central figure in many of the political events of the last half of
the nineteenth century, he served as Governor-general of
Canada, Viceroy of India, Ambassador to St. Petersburg, to
Constantinople, Rome and France. This biography is compiled
from his journal, his letters, and the recollections of his friends.
“But the real value of the book lies in the information it
supplies in regard to the great movements in foreign and
colonial politics that have been going on during the last thirty
years.”
+ Acad. 68: 144. F. 18, ‘05. 1580w.
+

+ Ath. 1905, 1: 201. F. 18. 1560w.

+
“Making every deduction for the imperfections inseparable
from even the best biographies, one reaches the conclusions
 that here a really great subject has been treated both
adequately and effectively.” Lawrence J. Burpee.
+ Dial. 39: 58. Ag. 1, ‘05. 1520w.
+

Ind. 58: 1358. Je. 15, ‘05. 1040w.

* “His book is frank yet discreet, and marked in all its parts by
delicacy of perception.”
+ Nation. 81: 342. O. 26, ‘05. 3370w.
+
+
Reviewed by Joseph O’Connor.
N. Y. Times. 10: 193. Ap. 1, ‘05. 3250w.
“What is to be regarded as the official biography. It is official
also in its discreetness—a discreetness at times carried to
extremes, dimming perception—and in the highly eulogistic
tone maintained throughout. It may safely be said that Sir
Alfred, while presenting a work obviously open to criticism, has
also presented one of direct value to the historical student, and
of interest to the general reader.”
+ Outlook. 79: 760. Mr. 25, ‘05. 280w.
+
—
“Sir Alfred Lyall seems to us to have chosen the best way in
which to tell the story of Lord Dufferin’s life.”
+ Spec. 94: 253. F. 18, ‘05. 930w.
+

Lydston, G. Frank. Diseases of society. **$3. Lippincott.

“A study of social conditions in this country. The police
criminal, the anarchist, and the large number of moral and
physical law-breakers are here discussed. The author also
deals with such questions as the oppression of wealth, the
rights and wrongs of organized capital and labour, the negro
question, and the offences of society at large. The book is well
illustrated.”—Bookm.
“The style, although brilliant at times, is open to much
criticism. It is verbose, often disconnected and rambling. In
spite of many blemishes the book is of great value. With the
general thesis of the book and a large percentage of the
conclusions, the reviewer is in hearty sympathy and heartily
commends it to students of social problems.” C. Kelsey.
+ Ann. Am. Acad. 25: 350. Mr. ‘05. 1040w.
+
—
“His examination of the question of crime seems exhaustive,
his inferences inevitable.” Albert Warren Ferris.
+ Bookm. 21: 528. Jl. ‘05. 660w.
+
“While here and there is much that is interesting, although at
times crudely presented, the author like many others who
write upon the subject errs in trying to prove too much from
insufficient premises and newspaper gossip, and this is
especially true when he treats of craniometry and
physiognomy.” Allan McLane Hamilton.
+ Critic. 47: 183. Ag. ‘05. 1140w.
—
“It has not the air of a serious book of science, and indeed
contains here and there a misplaced facetiousness.”
 + Ind. 59: 213. Jl. 27, ‘05. 360w.
—
“As monographs the parts are incomplete, and the whole is
neither sufficiently unified for the ordinary reader, nor clearly
cut for the student. Nor is the style attractive.”
— Nation. 81: 63. Jl. 20, ‘05. 680w.
+

Pub. Opin. 38: 57. Ja. 12, ‘05. 600w.

“This is really a study of the vice and crime problem from a
medical standpoint.”
+ R. of Rs. 31: 255. F. ‘05. 170w.
+

Lyle, Eugene P. Missourian. †$1.50. Doubleday.

Mr. Lyle finds material for his first story within the tottering
Empire of Maximilian. The hero is one of Jo. Shelby’s band
who, refusing to surrender after the fall of the Confederacy,
offered their services to Maximilian in Mexico. Din Driscoll,
Missourian, Confederate officer, the “storm center” in every
fight, and the exquisite, capricious Jacquelin d’Aumerle, secret
emissary of Napoleon on business of state, figure almost
grotesquely in a series of thrilling adventures which result from
defending each other from intrigue and death. In the end this
airy coquette of two imperial courts chooses to find her
happiness within the confines of a shut-away Missouri farm.
* “The fact is Mr. Lyle has been absorbed by his material,
instead of absorbing it.”
+ Ath. 1905, 2: 718. N. 25. 200w.
—
 “Here, for instance, is an example of literary over-seasoning,
which, far from being exceptional, is fairly characteristic of the
book’s style.” Frederic Taber Cooper.
+ Bookm. 22: 135. O. ‘05. 550w.
—
“We should be grateful to Mr. Lyle for having given in this novel
a new and adequate setting for the American hero of love and
war.”
+ Ind. 59: 930. O. 19, ‘05. 860w.
+
“Yet, dramatic, picturesque, brilliant in attack and technique as
the book undoubtedly is, the interest in it is largely
spectacular.”
+ Lit. D. 31: 586. O. 21, ‘05. 510w.
“It is crude enough in certain details, but its reading leaves no
doubt as to the fact that Mr. Lyle possesses extraordinary
vision and power to communicate what his imagination sees.”
L. L.
+ N. Y. Times. 10: 496. Jl. 29, ‘05. 880w.
—
* “Admirably fresh and lively tale.”
+ N. Y. Times. 10: 823. D. 2, ‘05. 160w.
“The history is accurate, but unimportant; the romance is of
vast importance and fairly accurate.”
+ Pub. Opin. 39: 220. Ag. 12, ‘05. 130w.

Lyman, Olin Linus. Oliver Hazard Perry and the war on the
Lakes. $1.25. Amsterdam.
 The brief career of Commodore Perry (1785-1819),
midshipman, lieutenant, commander of Lake Erie, and the
American squadron in the Mediterranean is given in this
volume which is “a eulogy rather than a biography.” (N. Y.
Times.)
“As an elementary history the book is good. It should make
rather a good ‘reader.’ Mr. Lyman has padded his book
tremendously, and has indulged in ‘fine writing’ of the worst
sort. The author is very chary of dates.”
+ N. Y. Times. 10: 275. Ap. 29, ‘05. 370w.
—
—

Lynch, Frederick. Is life worth living? **30c. Crowell.

A new volume in the “What is worth while series.” A message
of comfort showing that life in God’s world is truly worth living,
that there is but one answer to the question for those who
believe in the life eternal.

Lytton, Lord Bulwer-. Last days of Pompeii. $1.25. Crowell.

Uniform with the “Thin paper classics,” this pocket volume is
printed on opaque “Bible” paper in large clear type, is bound in
limp leather, and contains a frontispiece of the author.
 M

Maartens, Maarten (J. M. W. van der Poorten-Schwartz).

My poor relations. †$1.50. Appleton.
Fourteen unpleasant stories of life in a little Dutch village,
where the people are degraded and low in mind and morals. In
“The mother” Mary Quint vainly struggles to help her son
conquer his inherited love of drink. “Jan Hunkum’s money,”
“Fair lover,” “The summer Christmas,” “The notary’s love story,”
“The banquet,” and the rest, are all horridly true, and are told
in a vivid style that makes them almost too real.
“The book is as oppressive as a nightmare.”
— Critic. 47: 189. Ag. ‘05. 90w.
“Most of the fourteen stories herein told are pathetic almost to
tragedy.”
+ N. Y. Times. 10: 213. Ap. 8, ‘05. 340w.
—
“All the stories, while not calculated to make one laugh, will
undoubtedly keep one’s interest alive.”
+ N. Y. Times. 10: 390. Je. 17, ‘05. 200w.
“It would be hard to name a book in which the characters are
so uniformly disagreeable as in this collection of short stories.”
+ Outlook. 79: 858. Ap. 1, ‘05. 170w.
—
 “One may go so far as to compare them to De Maupassant’s
though hardly to that master’s best.”
+ R. of Rs. 31: 764. Je. ‘05. 120w.
+

Mabie, Hamilton Wright. Fairy tales every child should know.

**90c. Doubleday.
Twenty four “once upon a time” fairy tales collected from
various countries to amuse and stimulate the imagination of
the child of today. They include such familiar stories as, Hans
and Gretel, Ali Baba, The golden goose, One eye, two eyes,
three eyes, Blue beard, Red riding hood, The ugly duckling,
Tom Thumb, Jack the giant killer, Jack and the bean stalk, and
Puss in boots.
Dial. 39: 20. Jl. 1, ‘05. 60w.

+ Nation. 81: 12. Jl. 6, ‘05. 160w.

+
—

+ N. Y. Times. 10: 371. Je. 10, ‘05. 260w.

+
—
“In one respect the book appears to us defective, in that it
does not state by whom the particular version of each of these
child classics was written.”
+ Outlook. 80: 443. Je. 17, ‘05. 260w.
+
—
 * + R. of Rs. 32: 768. D. ‘05. 80w.
—

* Mabie, Hamilton Wright, ed. Myths every child should

know: a selection of the classic myths of all times for young
people. **90c. Doubleday.
“This volume is uniform with ‘Fairy tales every child should
know.’ It collects for children’s reading and for school use
sixteen myth-stories which belong to the world’s literature and
appeal to the young imagination. Hawthorne’s ‘Wonder-book’
and ‘Tanglewood tales’ furnish half the material.... Charles
Kingsley’s ‘Greek heroes,’ Mr. Brown’s ‘In the days of the
giants,’ Mr. A. J. Church’s ‘Stories from Homer,’ Mr. Mabie’s
‘Norse stories,’ and Miss Emerson’s ‘Indian myths’ are the other
sources. Mr. Mabie furnishes an introduction.”—Outlook.
* + Lond. Times. 4: 432. D. 8, ‘05. 40w.
—
* “We could wish that Mr. Mabie had put his interesting preface
before a more consecutive and less heterogeneous collection.”
* + Nation. 81: 490. D. 14, ‘05. 230w.
—

* + Outlook. 81: 629. N. 11, ‘05. 120w.

* “They are rather stiffly told and frequently the style is too
difficult and elaborate to be easily understood by children.”
— Sat. R. 100: sup. 10. D. 9, ‘05. 110w.
+

McAlilly, Alice. Hilda Lane’s adoptions. $1.50. Meth. bk.

 Hilda Lane, kept from the man she loves for twenty years by a
war time misunderstanding, adopts a sturdy waif named
Robert, and a negro girl, Liberty, and educates them. Liberty
grows up to offer her life to white fever sufferers, and Robert,
on the eve of a successful career and engaged to marry a
lovely southern girl, discovers that there is negro blood in his
veins and nobly consecrates his life to the uplifting of the black
race. The book is chiefly occupied with the negro question.

McAlilly, Alice. Larkins wedding. $1. Moffat.

“An apotheosis of good nature and neighborly kindness. A
worthy washerwoman related grammatically to Mrs. Partington
arranges the wedding of her daughter. The respect both have
won in their town inspires the interested villagers of higher
social position to make the pathetic efforts of Mrs. Larkins turn
out a happy success. A change in bridegrooms adds to the
general jollity, and the two Larkins, mother and daughter,
disappear in a haze of prosperity and sentiment.”—Outlook.
+ N. Y. Times. 10:650. O. 7, ‘05. 340w.
“The story is told with many touches of humor.”
+ Outlook. 81: 383. O. 14, ‘05. 80w.

Macaulay, Thomas Babington Macaulay, 1st baron.

Essays; ed. by Lady Trevelyan. $6. Putnam.
These six compact little volumes contain nothing but the text
of the essays and preface as edited by Lady Trevelyan, the
author’s sister. There are several illustrations in each volume—
mainly engravings and portraits.
“Edition is as satisfactory for the purposes of the reader of
Macaulay as a modest man can desire, handy enough to
permit you on occasion to put a volume in your coat pocket
 and take it with you upon a journey, yet entirely fit for the
library shelves. For it sacrifices to compactness not size of type
(and the eyes of the reader) but an easily dispensable surplus
of margin.”
+ N. Y. Times. 10:92. F. 11, ‘05. 250w.
+
“Admirably planned for thoroughly comfortable reading, and to
take up small space in a library. For a good edition which
meets all the requirements of the average reader, and of a size
which makes it possible to carry the volumes about when one
travels, we do not recall a better edition than this.”
+ Outlook. 79: 348. F. 4, ‘05. 120w.
+

Macbean, L., and Brown, John. Marjorie Fleming. **$1.40.

Putnam.
The famous account of “Pet Marjorie” by Dr. John Brown is
here reprinted, with much later information and her journals
and letters hitherto unpublished. There are fourteen
illustrations, including pictures of the little girl taken alone and
with Scott.
* “We commend this book, sure that it will become a precious
possession.”
+ Acad. 68: 1098. O. 21, ‘05. 1000w.
+
“Should be welcomed by all admirers of Dr. Brown’s earlier
story of her.”
+ Critic. 46: 283. Mr. ‘05. 50w.
+
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

textbookfull.com