Prefix computation on faulty hypercubes

In a parallel computer system, fault-tolerance is more obtainable since there is more than one processor cooperating in the system. For solving parallel prefix computation, we need not use all processors all the time. Thus, we propose a fault-tolerant mechanism by solving the prefix computation problem on faulty hypercube computers. By this mechanism, prefix computation can be done in O(log n) time on a faulty hypercube computer if there exists a fault-free subcube with size of at least a half of the hypercube, and in O(log/sup 2/n) time in the worst case.<<ETX>>