In a microprogrammed digital computer, the instructions are carried out in two stages#x2014;an instruction set is fetched from the main memory; is decoded and then a sequence of microoperations is performed by returning the control to a particular location of the control memory containing the microprogramme. The instruction sets are specified in terms of the required microoperations and the definite precedence relations. In this paper, an algorithm has been described to find out an optimal specification of the microprogramme control memory of a digital computer, given the total set of distinct instructions to be executed by the system. The objective of the algorithm has been to minimize the cost of implementation in terms of the total number of bits in the control memory. A loop-free structure of the microprogramme has been assumed, and multiple occurrence of a single micro-event in a single instruction set has been allowed.
展开▼
