PERFORMANCE PREDICTION OF CONSERVATIVE PARALLEL DISCRETE EVENT SIMULATION

In a previous paper, a quantitative criterion has been proposed for efficient execution of the Null Message Protocol, the best- known conservative parallel discrete event simulation (PDES) protocol. The criterion is based on a novel concept of the coupling factor, and allows one to use intuitive and easy-to-measure input parameters. The criterion can be used to assess simulation models 'potential for parallel execution as well as the maximum partitioning that may still potentially yield good performance. This paper deals with experimental verification of the proposed rule that a large coupling factor is a necessary precondition for getting a good speed-up with conservative parallel simulation. A closed queueing network is used as simulation model, and it is run on up to 24 CPU cores. It is shown that the criterion provides a quick and convenient way to determine whether a simulation model has a potential for speed-up before actually investing work in the parallelization.