Design tools for digital macro systems

Abstract

The early stages of system design and evaluation are rarely treated in a systematic way. Furthermore, system design tools have mostly been applied to the later phases of the design cycle. The paper presents a quantitative basis for the high-level design of digital systems in the form of probabilistic and simulation models and several performance measures. These tools form a major part of an overall system design methodology which has been valuable in the design, optimisation and evaluation of large-scale digital systems. The application of the modelling tools to the design of a distributed processing system is demonstrated. Validation of the results by simulation reveal an average error of less than 3% for this example.