A note on two simulation benchmarks

Abstract

This note presents the results of a comparative study done on the popular simulation languages SLAM and GPSS. The models that we tested were executed on a CDC Cyber 750, a multiprogramming system with 3770008 words of main memory and no virtual memory. A single server queueing model and a multiqueuing model for simulating a token ring local area network were used for the comparison. In addition to SLAM and GPSS, we also develop FORTRAN versions of the models to obtain an idea of how a low level model performs in comparison to a model constructed in a well developed simulation language. The note includes a discussion and some graphical results that may be useful to a modelling practitioner in planning a simulation study.