A Study of the Clarity of Functionally and Structurally Composed High-level Simulation Models

Simulation is a widely used tool in evaluating and guiding high-level hardware design decisions. As a result, the accuracy of the simulation model is critical; a poor model can result in poor design choices. Since most simulation models are complex, some form of validation is needed to ensure this accuracy. The most common validation technique is that of checking the simulation model by hand. Clearly, the reliability and efficiency of this technique is closely tied to how easy it is for someone to understand (i.e. the clarity of) the simulation model. This is especially true for validation by a third party. The clarity is also critical if third parties are to correctly modify the simulation model for their own purposes. This technical report presents the results of an experiment that compares the clarity of models specified with a structural specification system (in particular, LSE [3]) versus a simulation model written in a sequential programming language (i.e. a sequential simulator). The organization of the paper is as follows. Section 2 gives a brief description of how sequential simulators differ from structural models. Section 3 describes the overall experimental setup. Section 4 presents the results of the experiment.