Adaptive routing technique on hypercrossbar network and its evaluation

Boku, Taisuke; Nakazawa, Koyomi; Nakamura, Hiroshi; Sone, Takeshi; Mishima, Taketoshi; Itakura, Keiichi

doi:10.1002/scj.4690270405

Cited by 1 publication

(3 citation statements)

References 4 publications

(2 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This results in a particular case of our proposed family of topologies with very poor expected performance due to using a simple tree for connecting the nodes of a dimension. The Bcube [20] and Hypercrossbar network [10] also resemble our proposal. Each processing node is connected to several dimensions by using several NICs.…”

Section: Related Worksupporting

confidence: 73%

“…The resulting topologies could seem similar to BCube [20] or Hypercrossbar network [10], but the great difference between our proposal and BCube is that the processing nodes do not eject messages from the network; the new topology family forwards the messages in the network interface which highly reduces latency. However, these cases could be considered as particular configurations of the KNS family of topologies.…”

Section: Description Of the Familymentioning

confidence: 95%

“…Other hybrid topologies have been proposed previously in the literature (see Section 6). However, most of them are hierarchical and therefore introduce long latency for the communication of non-local processing nodes and others are intended for particular purposes [20,10]. In this paper we propose a topology family that can be configured to meet different scenario requirements.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

The k-ary n-direct s-indirect family of topologies for large-scale interconnection networks

Penaranda

Gómez

et al. 2016

J Supercomput

View full text Add to dashboard Cite

In large-scale supercomputers, the interconnection network plays a key role in system performance. Network topology highly defines the performance and cost of the interconnection network. Direct topologies are sometimes used due to its reduced hardware cost, but the number of network dimensions is limited by the physical 3-D space, which leads to an increase of the communication latency and a reduction of network throughput for large machines. Indirect topologies can provide better performance for large machines, but at higher hardware cost. In this paper, we propose a new family of hybrid topologies, the k-ary n-direct s-indirect, that combines the best features from both direct and indirect topologies to efficiently connect an extremely high number of processing nodes. The proposed network is an n-dimensional topology where the k nodes of each dimension are connected through a small indirect topology of s stages. This combination results in a family of topologies that provides high performance, with latency and throughput figures of merit close to indirect topologies, but at a lower hardware cost. In particular, it doubles the throughput obtained per cost unit compared with indirect topologies in most of the cases. Moreover, their fault-tolerance degree is similar to the one achieved by direct topologies built with switches with the same number of ports.