ABSTRACT
This chapter illustrates the very large scale integration design methodology by discussing the top-down design flows of two projects: a simple microprocessor and a field programmable gate array. The essence of register transfer is that each operation in the data path involves source and destination registers. A mask-programmed gate-array, or simply gate-array, is a chip that consists of an array of prefabricated transistors. The design of a high performance microprocessor is unarguably a very complicated task that will consume hundreds of man-months. The concept of dividing a system into a datapath unit and a control unit is essential in the design of computing structures. An alternative approach called microprogramming implements the finite-state machine with a memory unit. A memory system has its own timing and handshaking requirements. These requirements are quite independent of the microprocessor. The programmable routing switches provide the programmable connections between the horizontal and vertical routing interconnects.
