A New Book on Computer Arithmetic
Amos Omondi
uunet!Comp.VUW.AC.NZ!Amos.Omondi
Mon Feb 21 18:15:11 PST 1994
ANNOUNCING A NEW BOOK ON COMPUTER ARITHMETIC
============================================
Title: Computer Arithmetic Systems -- Algorithms, Architecture,
Implementation. 522 pages.
Author: Amos R. Omondi
Publisher: Prentice-Hall International, UK
(CAR Hoare series in Computer Science)
ISBN: 0-13-334301-4
Date: 1994
Short description of the book
============================
This book attempts to do for the 1990s what Flores's "Logic of Computer
Arithmetic" did for the 1960s. It covers much more than any other book
currently available on the subject.
Detailed description of the book
===============================
This book is primarily intended to be a text for postgraduate and advanced
undergraduate students in computer science, computer engineering, and
electrical engineering, and the chief concern has been to cater to the needs
of these groups. Nevertheless, professionals engaged in the design and
and implementation of digital computers will also find the book useful
as a reference manual, as will prospective researchers in the field of
computer arithmetic.
Small selected parts of the book can be used as supplementary
material in courses on computer architecture and implementation. Larger
parts, comprising about half the book,
can be used as the basis of an undergraduate course devoted entirely to
computer arithmetic; a similar course at the postgraduate level can
reasonably cover
most, if not all, of the book.
Whatever the use of the book,
it is assumed that the reader has a reasonable background
knowledge of digital logic design techniques,
such as can be obtained from a good first course.
For a few sections of the book,
an acquaintance with basic numerical analysis would be helpful but
is not altogether indispensable.
The book is divided into three major
parts. The first part consists of four chapters and is devoted to
algorithms and hardware for basic arithmetic
operations in conventional fixed-point number systems.
The second part consists of three chapters,
the subject of which is various aspects of
floating-point number systems and
algorithms and hardware for operations in these systems.
The third part is a single chapter that
deals with unconventional number representations and arithmetic.
(The third part, because of limitations
of space,
is intended to be no more than a brief introduction to the subjects
discussed.)
A number of case studies are included where appropriate. In reviewing
these case studies, especially the older ones,
the reader should distinguish between some logical
design technique and an instance of the same; that is, architecture,
implementation, and realization should be separated and the effects of
one on another recognized.
Part I: The first chapter is a brief introduction to the commonly used
fixed-point number representations.
The second, third, and fourth chapters deal with algorithms
and hardware for addition and subtraction, multiplication, and division.
Part II:
The fifth chapter is mainly a discussion of various
abstract features of floating-point number
systems and their concrete representation within a computer.
Basic error analysis is introduced in this chapter, and related topics,
such as rounding, are also discussed. The sixth
chapter deals with hardware structures for the implementation of the basic
floating-point arithmetic operations, and
the seventh covers the elementary functions:
square root,
sine, cosine, logarithms, exponential, and so forth.
Part III:
The eighth chapter of the book
deals with unconventional
number systems and corresponding algorithms and hardware for the basic
arithmetic
operations. Three number systems are discussed there:
redundant signed-digit, residue, and decimal.
There are three appendices. The first is a brief
introduction to pipelining, an important technique in the design of
high-performance
computers. Pipelining is normally covered in depth in
a course on computer architecture but is nevertheless
of some importance
here. The second appendix is a discussion on the design of shifters, a
subject that gets short shrift in most textbooks and courses on digital logic
design.
The last appendix is a design library. It
consists of detailed descriptions of some commercial
integrated-circuit modules and is
to be used as the basis of various design exercises. This appendix
has been bound separately and can be obtained by writing to the publisher;
alternatively, a suitable databook may be substituted.
Table of contents
=================
(A detailed table of contents is available, in the form of a LaTeX "toc" file,
by e-mail.)
1. Fixed-point number systems
1.1 Introduction
1.2 Representing signed numbers
1.3 Other fixed-point number systems
1.4 Summary
1.5 Bibliographic notes
1.6 Exercises
2. Fixed-point addition and subtraction
2.1 Algorithms
2.2 Implementation
2.3 Subtraction
2.4 Case studies
2.5 Bibliographic notes
2.6 Exercises
3. Fixed-point multiplication
3.1 Basic algorithms
3.2 Fast multiplication
3.3 Evaluation of powers
3.4 Implementation
3.5 Case studies
3.6 Bibliographic notes
3.7 Exercises
4. Fixed-point division
4.1 Basic algorithms
4.2 Fast division
4.3 Implementation
4.4 Case studies
4.5 Bibliographic notes
4.6 Exercises
5. Floating-point arithmetic
5.1 Introduction
5.2 Normalization
5.3 Basic arithmetic operations
5.4 Overflow and underflow
5.5 Representation format
5.6 Error analysis
5.7 Case studies
5.8 Bibliographic notes
5.9 Exercises
6. Basic floating-point operations: implementation
6.1 Rounding
6.2 Addition and subtraction
6.3 Multiplication and division
6.4 Summary
6.5 Bibliographic notes
6.6 Exercises
7. Elementary functions
7.1 Square root and related functions
7.2 Trigonometric functions
7.3 Exponential and logarithm functions
7.4 Polynomial approximations
7.5 Summary
7.6 Case study
7.7 Bibliographic notes
7.8 Exercises
8. Unconventional number systems and arithmetic
8.1 Residue number systems and arithmetic
8.2 Decimal number systems and arithmetic
8.3 Redundant signed-digit number systems and arithmetic
8.4 Bibliographic notes
8.5 Exercises
Bibliography
Appendix A: Pipelining
Appendix B: Design of shifters
More information about the Numeric-interest
mailing list