Execution time (and memory) optimization in microprograms

Availability of dynamically writable control memories for microprogramming motivates researchers in the design of appropriate languages. Certain processors allow simultaneous activation of several microoperations. Exploitation of parallelism existing in a source program therefore becomes a necessity in the implementation of some compilers. In this paper, the improvements on the execution time of programs by a near optimal compiler are given. Then an algorithm is suggested for obtaining the optimal object microcode. It is then applied to an example.