An optimal scheme of parallel processing for test generation in a distributed system

This paper presents an approach to parallel processing of test generation for logic circuits in a loosely-coupled distributed network of general purpose computers. We first analyze the relation between the number of processors and the total processing time for a client-server model (CS model). The result of the analysis shows that for the CS model there exists an optimal number of processors which minimizes the total processing time. In order to get a more efficient scheme than the CS model, we propose another model called a client-agent-server model (CAS model), which is derived by adding agent processors to the CS model. We formulate the problem of test generation on the CAS model, and analyze the relation between the number of processors and the total processing time. Our analysis leads to an optimal scheme for the CAS model which minimizes the total processing time for a given number of processors. We present experimental results for the ISCAS benchmark circuits by implementing the CAS model on a network of workstations.<>