previous | contents | next

thus to be charged to capital budgets rather than current expenses, and to be reviewed by a Facilities Committee. The success of minicomputers is gradually wearing down some of these myths, though it leaves others untouched. The sort of digital systems designed in this book give the lie to all of them. But even realities dispel myths only slowly. Thus, we feel it is legitimate to state the case for register-transfer systems designs in specific and unequivocal terms. This book is our attempt to do so.

The book is organized as follows: Chapter 1 provides an introduction, describing both what digital systems are and their applications. It includes a basic description of register-transfer level modules in the functional PMS language used throughout the book It also introduces ways to specify the behavior of desired systems. It provides a set of diverse examples of digital systems to define concretely the scope of the book. Finally, it discusses the processes of design. Chapter 2 is a systematic description of the specific set of PDP- 16 modules.

Chapter 3 gives the solutions of 8 small systems built from RTM's: These should give the reader a good feeling about the design process and what RTM systems look like, since the problems being solved are small and straightforward. These problems are presented in a relatively standard format which we carry throughout.

Chapter 4 is the main reference on the process of RTM design. It too uses a problem, that of a multiplier, to illustrate both the design process and the many trade-offs that are possible even with a limited set of modules. These tradeoffs include: hardware-microprogram-program; parallelism (time-size) in the form of concurrency, array processing, and pipeline processing; the algorithm; and memory-compute.

Chapters 5 and 6 proceed by cases to design a wide array of systems. Some are introduced at length with intensive discussion, some are given simply as problems. Always there are suggestions for further work, even where our own treatment is most complete. Generally, designs are separated from the statement of the design problem, so that a student can attempt the design first, before seeing how we approached the problem. Any of the readers mentioned earlier-in the preface should find some designs to fit his interest or tickle his fancy. We include the design of several real computers so that the mystery of the operation of computers can be dispelled once and for all, for all readers. In Chapter 6, we urge the reader who has never designed, built, and programmed his very own stored program computer to take the few days to do so. All in all, about one hundred design problems are posed and treated, and a similar number of tasks are posed.

Chapter 7 disc the design of register transfer modules themselves. This is necessary, not only to keep them from appearing to be black boxes, but because no one set of modules is appropriate for all tasks. As a by-product, we hope that some integrated circuit logical designer will read this and derive designs for control schemes that people can actually easily use with integrated circuit logical design. We also hope to encourage the logic design textbook writer to think about control schemes. Chapter 8 treats register transfer design generally, freed from the dependence on the PDP-16 modules. This chapter


viii

previous | contents | next