A fast diagnosis algorithm for interconnection network in high performance computers. Issue 1 (2nd January 2016)

• Record Type: Journal Article
• Authors: He, Li; Li, Furong
• Is Part Of: International journal of parallel, emergent and distributed systems. Volume 31:Issue 1(2016)
• Page Start: 34
• Page End: 46
• Publication Date: 2016-01-02
• Subjects: high performance computers -- folded hypercube -- diagnosability -- diagnosis algorithm; Parallel computers -- Periodicals; Electronic data processing -- Distributed processing -- Periodicals; Computer algorithms -- Periodicals; 004.35
• DOI: 10.1080/17445760.2014.958823
• Languages: English
• ISSNs: 1744-5760

Abstract

Abstract : In high performance computers, a popular interconnection network, the folded hypercube (FHC), possesses smaller diameter, larger connectivity, better reliability and fault tolerance capability as compared with a hypercube counterpart. This paper addresses the fault identification of FHC multiprocessor interconncection systems under the MM * model. The pessimistic one-step diagnosability of FHC networks is first determined. On the basis, a pessimistic one-step diagnosis algorithm tailored for FHC multiprocessor systems is proposed. The presented algorithm can isolate all faulty nodes to within a set which has at most one fault-free node, and can run in linear time.