A method to evaluate the performance of a multiprocessor machine based on data flow principles

A method is presented to model a static dataflow-oriented multiprocessor system. This methodology for modeling can be used to examine the machine behavior for executing a program according to three scheduling strategies: static, dynamic and quasi-dynamic policies. The processing elements (PEs) of the machine go through different states in order to complete tasks allotted to them. Hence, the time taken by the machine to execute a program is directly dependent on the time spent by the PEs in various states during the execution of tasks. The authors adopt a state diagram approach to model the machine. This modeling scheme can be used for a class of machines, having similar execution paradigms. By introducing wait states in the state diagram of a PE at appropriate places, the delays that are incurred by the PE waiting on events are captured. The communication media are modeled as queuing networks and the delay introduced by the wait state (of a PE for accessing a medium) is specified by the queuing delay of the corresponding network model.<>