An Enhanced Fault Identification Algorithm for PMC-Based Diagnosable Systems

Significant advancements in multi-core processors continue to unfold and boost contemporary high-performance computing (HPC) to gain attraction in computational technology. Fault identification plays a critical role in supporting an HPC system’s availability. Graph theory models a system’s interconnection, whose physical nodes and their communication links are abstractly represented by vertices and edges, respectively. In this paper, an enhanced algorithm is proposed to accomplish fault identification for PMC-based t-diagnosable systems, provided that every fault-free node cannot has no fault-free neighbor, and the total number of faulty nodes is bounded above by \(t+1\) .