TY - GEN
T1 - Performance modeling of a fully adaptive and fault-tolerant wormhole switching strategy in 2-D mesh
AU - Safaei, Farshad
AU - Fathy, Mahmood
AU - Khonsari, Ahmad
AU - Ould-Khaoua, Mohamed
PY - 2006
Y1 - 2006
N2 - In recent years, many researchers have devoted much efforts to construct high performance interconnect networks resilient to faults. Their studies motivated by the fact that a network can be a major performance bottleneck in parallel processors; such as multiprocessors system-on-chip (Mp-SoCs), multicomputers and cluster computers. This paper proposes a new analytical model to predict message latency in 2-dimensional wormhole-switched mesh with a routing scheme suggested by Linder and Harden [1], as an instance of a fault-tolerant routing widely used in literature to achieve high adaptivity. Furthermore, the validity of the proposed model is demonstrated by comparing analytical results to those conducted through simulation experiments of the actual system and show a good degree of accuracy with as many as 10% nodes faulty.
AB - In recent years, many researchers have devoted much efforts to construct high performance interconnect networks resilient to faults. Their studies motivated by the fact that a network can be a major performance bottleneck in parallel processors; such as multiprocessors system-on-chip (Mp-SoCs), multicomputers and cluster computers. This paper proposes a new analytical model to predict message latency in 2-dimensional wormhole-switched mesh with a routing scheme suggested by Linder and Harden [1], as an instance of a fault-tolerant routing widely used in literature to achieve high adaptivity. Furthermore, the validity of the proposed model is demonstrated by comparing analytical results to those conducted through simulation experiments of the actual system and show a good degree of accuracy with as many as 10% nodes faulty.
UR - http://www.scopus.com/inward/record.url?scp=33745937496&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=33745937496&partnerID=8YFLogxK
U2 - 10.1007/11751649_13
DO - 10.1007/11751649_13
M3 - Conference contribution
AN - SCOPUS:33745937496
SN - 3540340793
SN - 9783540340799
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 118
EP - 128
BT - Computational Science and Its Applications - ICCSA 2006
PB - Springer Verlag
T2 - ICCSA 2006: International Conference on Computational Science and Its Applications
Y2 - 8 May 2006 through 11 May 2006
ER -