Book Reviews
A Top-Down Constraint-Driven Design
Methodology for Analog Integrated
Circuits
H. Chang, E. Charbon, U. Choudhury, A. Demir,
E. Felt, E. Liu, E. Malavasi, A. SangiovanniVincentelli and I. Vassiliou, Kluwer Academic
Publishers, NorweU, MA,
1997, 366pp.,
US$115.00, UK£81.65, Dfl.215.00
The first comment that may be made on this
book is the remarkable number of authors
who are listed as having written its contents,
even though there is no indication that individual chapters are the work of individual
authors. Indeed, the style throughout appears
to be the same, which may indicate only one
principal writer of this text. It is left for the
reader to ponder which one, if indeed this is
the case.
The material of this text is based on on-going
research activity at the University of California at
Berkeley, which is considering new computeraided design methodologies for analogue and
mixed analogue/digital IC design, working from
a top-down hierarchical level. Chapter 1 is a
good introduction to the past work on analogue
CAD, and notes silicon compilation, knowledge-based techniques, hybrid systems and
human-driven systems, leading to the specific
methodology of constraint-driven design, which
is the subject of the following chapters.
Chapter 2 introduces this methodology, which
briefly consists of two phases, namely: (i) the
mapping of performance specifications onto
bounds on all physical parasitics relevant to the
implementation; and (ii) enforcing each bound
during the physical assembly of the design. A
technique of 'constraint graphs' is employed in
this procedure. Simulation and behaviour
modelling is central to the procedures, together
with architectural mapping and optimization.
Bottom-up verification is done after layout
synthesis.
364
Chapters 3-8 cover this material, following
which three further chapters giving examples of
practical design exercises are given. Overall, this
is an interesting text for all who are concerned
with analogue CAD and research in software
developments in this subject area. The claim that
this design methodology will result in a high
probability of right-first-time silicon must be
judged against other existing methodologies, and
is not necessarily proven by this publication.
This is, however, a book which should be available to students and others who are actively
engaged in analogue CAD research, if only for
the very extensive list of over 300 references
with which the book concludes.
M.S. Harris
Synthesis of Finite State Machines:
Functional Optimization
T. Kam, T. Villa, R. Brayton and A. SangiovanniVincentelli, Kluwer Academic Publishers, Norwell,
MA, 1997, 282pp., US$110.00, UK£78.10,
Dr.205.00
This is yet a further book detailing activities at
the University of California at Berkeley,
although the affiliation of one of the authors
(T.K.) is now listed as the Intel Corporation.
The book is stated to be the first of two monographs addressing the synthesis of finite state
machines, this one addressing functional optimization, with the subsequent one addressing
logic optimization. It is possibly a pity that these
two associated subject areas could not have been
combined into one text, at the expense of losing
a score of one in the tally of university department publications.
Functional optimization is defined as the
computation of all permissible sequential functions for a given topology of interconnected
finite state machines and the selection of a 'best'
function from the set of permissible ones, the