A fault tolerant protocol for massively parallel systems

On the Probability of Facing Fault Patterns: A Performance and Comparison Measure of Network Fault-Tolerance

“…In order to calculate the parameter ( ; ) X P X S , we should enumerate the number of all existing paths facing the − X shape fault pattern and divide them by the number of all existing paths in the connected × R C torus network. This probability is expressed formally as hit The number of minimal paths crossing the fault region P The number of all minimal paths existing in the network (1) The following theorem provides the total number of paths with minimal length in the network. ( , )…”

“…To be able to adapt with faults without serious degradation of the service, networks and routing protocols have to be set up so that they are fault-tolerant. Several recent studies address faulttolerance in a diverse range of systems and applications [1][2][3][4][5][6][7][8][9][10][11][12]. Almost all of the performance evaluation studies for functionality of these systems, however, have made use solely of simulation experiments.…”

On Quantifying Fault Patterns of the Mesh Interconnect Networks

“…In recent years, many researchers have addressed to several issues in the field of fault-tolerance and reliability analysis of large scale parallel and distributed systems [4][5][6][7][8][9][10][11][12][13][14][15][16]. These researches span a diverse range of systems and applications such as massively parallel processors [8], cluster-based systems [9], mobile systems [10], sensor networks [11], and more recently network on chip [1].…”

“…These researches span a diverse range of systems and applications such as massively parallel processors [8], cluster-based systems [9], mobile systems [10], sensor networks [11], and more recently network on chip [1].…”

ParFUM: a parallel framework for unstructured meshes for scalable dynamic physics applications

“…Charm++ consists of a variety of broadly applicable high-performance tools integrated in a single run-time system. Virtualization techniques are employed for hiding latency via message-driven execution [2], automatic applicationindependent load balancing [3], automatic communication optimization [4], check-pointing [5], fault tolerance [6,7], and performance visualization and analysis [8]. All of these tools help make a parallel code run better, but even with Charm++, developing a new parallel program still requires many hours of effort.…”