Skip to Main Content
Instead of using a very expensive and powerful central processing unit, the Hypercube F8 uses independent parallel microprocessors that are slow and inexpensive. This architecture is modeled in order to determine the proper number of microprocessors and to validate the system. The model is hierarchical. Different levels of the model are considered. Each level model corresponds to a subsystem of the Hypercube F8. Most of the common analytical methods are used and direct calculations are made. For each level several methods are compared and the approximations validated. The level 1 direct method is a general solution of polling models. The global model gives the throughput of the system the utilizations of the processors and the service times. We discuss the values of the parameters of the system.