A Class of Non-optimum-time FSSP Algorithms

Abstract

Synchronization of large-scale networks is an important and fundamental computing primitive in parallel and distributed systems. The synchronization in cellular automata, known as the firing squad synchronization problem (FSSP), has been studied extensively for more than fifty years, and a rich variety of synchronization algorithms has been proposed. In the present chapter, we give a survey on a class of non-optimum-time 3n-step FSSP algorithms for synchronizing one-dimensional (1D) cellular automata of length n in \(3n \pm O(\log n)\) steps and present a comparative study of a relatively large-number of their implementations. We also propose two smallest-state, known at present, implementations of the 3n-step algorithm. This chapter gives the first complete transition rule sets implemented on finite state automata for the class of non-optimum-time 3n-step FSSP algorithms developed so far.