In this section, the performance of the sequential and parallel implementations of the FSM-based Hoshen-Kopelman algorithm is presented. Actual timing results obtained on a Sun SPARCstation 10 and Thinking Machines CM-5 are provided. The FSM sequential implementation is compared to the Hoshen-Kopelman cluster identification textbook implementation from Berry et al. (1994), which was coded verbatim from the algorithm presented in Hoshen and Kopelman (1976). The FSM sequential and parallel performance results will be compared for accuracy and speedup - the reduction in execution time due to the code parallelization.