An adaptive system-level diagnosis approach for hypercube multiprocessors

Abstract

This paper proposes a hierarchical adaptive system-level diagnosis approach f o r hypercube systems. Three measures for diagnosis cost (diagnosis time, number of tests and number of test links are analyzed f o r the proposed algorithm. It is provedthat the diagnosis cost required b y this algorithm is lower than in the previous diagnosis algorithms in most of the fault cases. It is shown that the diagnosis cost for the proposed algorithm depends on the number of faulty units in the system and the cost is extremely low when only a small number of faulty units exist. It is shown that this algorithm is even characterized b y lower costs than a pessimistic diagnosis algorithm which trades lower diagnosis cost f o r a lower degree of accuracy.