Filters








53 Hits in 4.2 sec

Strictly nonblocking conference networks using high-dimensional meshes

Yi Du, Gerald M. Masson
1999 Networks  
An r-dimensional conferencing mesh network is strictly nonblocking if, regardless of the existing conferences implemented, a new conference among any subset of the idle conferees can be implemented using  ...  For general r-dimensional meshes, M Å O(r ( r 0 1)/r N ( r / 1)/r ) nodes are sufficient to support strictly nonblocking capabilities.  ...  CONDITIONS FOR A STRICTLY NONBLOCKING r-DIMENSIONAL MESH 2.  ... 
doi:10.1002/(sici)1097-0037(199907)33:4<293::aid-net5>3.0.co;2-b fatcat:65ahjyeeh5hfzkbtejnblzyiuq

Page 2401 of Mathematical Reviews Vol. , Issue 2000c [page]

2000 Mathematical Reviews  
Masson, Strictly nonblocking conference networks using high- dimensional meshes (293-308).  ...  Hwang, Extended generalized shuf- fle networks: sufficient conditions for strictly nonblocking oper- ation [MR 99a:94073] (269-291); Yi Du [Yi Du] and Gerald 2000c:94004 M.  ... 

A new conference network for group communication

Yuanyuan Yang
2002 IEEE transactions on computers  
The conference network proposed is reammgeable nonblocking in general, and is strictly nonblocking under some conference service policy.  ...  However, using a single crossbar to build a large size conference network is very expensive because that an 71 x 71 crossbar has 0{7I2) hardware cost Thus, the crossbar implementation is only suitable  ...  -The research wock was suppcx1ed in part by the US National National Sci. ence Foondation under grant number CCR-00730SS and u.s. Army Research Office und~ grant number DAAH04-96-1-0234.  ... 
doi:10.1109/tc.2002.1032621 fatcat:bc3ahgb3ebezplqzzy6kyqvvcq

A new conference network for group communication

Y. Yang
2001 International Conference on Parallel Processing, 2001.  
The conference network proposed is reammgeable nonblocking in general, and is strictly nonblocking under some conference service policy.  ...  However, using a single crossbar to build a large size conference network is very expensive because that an 71 x 71 crossbar has 0{7I2) hardware cost Thus, the crossbar implementation is only suitable  ...  -The research wock was suppcx1ed in part by the US National National Sci. ence Foondation under grant number CCR-00730SS and u.s. Army Research Office und~ grant number DAAH04-96-1-0234.  ... 
doi:10.1109/icpp.2001.952057 dblp:conf/icpp/Yang01 fatcat:6gpyyiox3rbklf6ytnaza5rhbu

A class of multistage conference switching networks for group communication

Yuanyuan Yang, Jianchao Wang
2004 IEEE Transactions on Parallel and Distributed Systems  
The conference network proposed is rearrangeably nonblocking in general, and is strictly nonblocking under some conference service policy.  ...  conferences simultaneously present in the network.  ...  Du and Masson [4] also proposed a nonblocking conference network, in which the conference component adopted is an r-dimensional mesh with m ¼ Oðn rþ1 r Þ.  ... 
doi:10.1109/tpds.2004.1264808 fatcat:p4qvlgmnabfovmshsuce64gthm

Sparse collective operations for MPI

Torsten Hoefler, Jesper Larsson Traff
2009 2009 IEEE International Symposium on Parallel & Distributed Processing  
The second implementation works strictly for meshes or similar, regular topologies.  ...  First, it is not a natural abstraction for applications with a very small local neighborhood (e.g., 9-neighbor stencil in a 2-dimensional mesh) compared to the total number of processes.  ... 
doi:10.1109/ipdps.2009.5160935 dblp:conf/ipps/HoeflerT09 fatcat:5424axqvevdkla6m4rvgpj7nzu

A situated evaluation of the Object Management Group's (OMG) Object Management Architecture (OMA)

Evan Wallace, Kurt C. Wallnau
1996 Proceedings of the 11th ACM SIGPLAN conference on Object-oriented programming, systems, languages, and applications - OOPSLA '96  
We use CORBA to refer to only the message broker component of the OMA; we use OMA to refer to CORBA plus additional OMA services.  ...  In this paper we describe a case study in the use of the OMA to integrate legacy software components into a distributed object system.  ...  in a 1:many relationship with exodus3d objects to indicate that a single two-dimensional mesh can be extruded into several three-dimensional meshes.  ... 
doi:10.1145/236337.236353 dblp:conf/oopsla/WallaceW96 fatcat:i45dv6koorgu5ma4iuhrgxp6pm

On the Capacity of Optical Networks: A Framework for Comparing Different Transport Architectures

G. Weichenberg, V. W. S. Chan, M. Medard
2006 Proceedings IEEE INFOCOM 2006. 25TH IEEE International Conference on Computer Communications  
Our comparison is based on a notion of network capacity as the set of exogenous traffic rates that can be stably supported by a network under its operational constraints.  ...  In this work, we compare three optical transport network architectures: optical packet switching (OPS), optical flow switching (OFS), and optical burst switching (OBS).  ...  First, the network, when viewed as a switch, is nonblocking.  ... 
doi:10.1109/infocom.2006.107 dblp:conf/infocom/WeichenbergCM06 fatcat:ftm4k7wczngdlmy75x3ofnrwzm

LumiNOC

Cheng Li, Mark Browning, Paul V. Gratz, Samuel Palermo
2012 Proceedings of the 21st international conference on Parallel architectures and compilation techniques - PACT '12  
LumiNOC reduces latencies ∼40% versus an electrical 2D mesh NoCs on the PARSEC shared-memory, multithreaded benchmark suite.  ...  Silicon nanophotonics is a promising replacement for electronic on-chip interconnect due to its high bandwidth and low latency, however, prior techniques have required high static power for the laser and  ...  Although this straightforward structure provides strictly nonblocking connectivity, it requires a large number of transceivers O(r 2 ) and long waveguides crossing the chip, where r is the crossbar radix  ... 
doi:10.1145/2370816.2370876 dblp:conf/IEEEpact/LiBGP12 fatcat:d57qb5kik5hnzaoe5o263mwxvm

Scalable Algorithms for Molecular Dynamics Simulations on Commodity Clusters

Kevin Bowers, Edmond Chow, Huafeng Xu, Ron Dror, Michael Eastwood, Brent Gregersen, John Klepeis, Istvan Kolossvary, Mark Moraes, Federico Sacerdoti, John Salmon, Yibing Shan (+1 others)
2006 ACM/IEEE SC 2006 Conference (SC'06)  
We have also developed numerical techniques that maintain high accuracy while using single precision computation in order to exploit processor-level vector instructions.  ...  The use of SIMD gives Desmond very high singleprocessor performance.  ...  In many communication networks, the latency of communication calls may be somewhat overlapped, that is, it is possible to send k messages with nonblocking communications in time less than the product of  ... 
doi:10.1109/sc.2006.54 fatcat:yx2ivaz25zfppabjumlm7wvp7m

A universal approach to data center network design

Bruce Maggs
2014 Proceedings of the 26th ACM symposium on Parallelism in algorithms and architectures - SPAA '14  
In particular, we suggest that the principle goal of the network designer should be to build a single network that is provably competitive, for any application, with any network that can be built for the  ...  After describing our approach, we survey the technology choices available to network designers today, and examine several existing commercial data center networks.  ...  The nonblocking multi-Beneš network [3] , which is based on the multibutterfly can establish a set of circuits among N inputs and N outputs in any permutation using O(N log N ) crosspoints.  ... 
doi:10.1145/2612669.2612729 dblp:conf/spaa/Maggs14 fatcat:vpm3azwckbdxtp6thnitzcjqi4

High-Performance Energy-Efficient Multicore Embedded Computing

A. Munir, S. Ranka, A. Gordon-Ross
2012 IEEE Transactions on Parallel and Distributed Systems  
With Moore's law supplying billions of transistors on-chip, embedded systems are undergoing a transition from single-core to multicore to exploit this high-transistor density for high performance.  ...  We also discuss modern multicore processors that leverage these HPEEC techniques to deliver high performance per watt.  ...  ACKNOWLEDGMENTS This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) and the US National Science Foundation (NSF) (CNS-0953447 and CNS-0905308).  ... 
doi:10.1109/tpds.2011.214 fatcat:vagqmojdsjevvc2u2ewqrcjjpq

A Geometric Approach to Capacity Provisioning in WDM Networks with Dynamic Traffic

Li-wei Chen, Eytan Modiano
2006 2006 40th Annual Conference on Information Sciences and Systems  
In this paper, we use an asymptotic analysis similar to the sphere-packing argument in the proof of Shannon's channel capacity theorem to derive optimal provisioning requirements for networks with both  ...  on an individual basis, and show that by employing waveband switching, a link with only a few switches per user can achieve the same performance as a link provisioned with unlimited switches per user using  ...  INTRODUCTION O PTICAL networks are a common solution for high-speed communications.  ... 
doi:10.1109/ciss.2006.286404 fatcat:vbj2mnv2nzeq5k5urf3rvpalzm

Architectural approach to the role of optics in monoprocessor and multiprocessor machines

Jacques Henri Collet, Daniel Litaize, Jan Van Campenhout, Chris Jesshope, Marc Desmulliez, Hugo Thienpont, James Goodman, Ahmed Louri
2000 Applied Optics  
This relation leaves the choice of the best network open in terms of simplicity and latency reduction.  ...  Therefore the higher the connectivity ͑possibly with optics͒, the shorter the path to another node, but the more expensive the network and the more complex the structure of electronic nodes.  ...  of higher-order nonblocking caches.  ... 
doi:10.1364/ao.39.000671 pmid:18337941 fatcat:mowm54a6b5bidk3v5epnutx2a4

Spatial Channel Cross-Connect Architectures for Spatial Channel Networks

Masahiko Jinno
2020 IEEE Journal of Selected Topics in Quantum Electronics  
Based on the formulations, these architectures are compared with traditional SXC architectures based on a high-port count full-MS pair and a Clos network from the viewpoints of the freedom of connection  ...  A spatial channel network (SCN) is a novel optical network architecture targeting the space division multiplexing (SDM) era where the optical layer is explicitly evolved into the hierarchical SDM layer  ...  mesh networks, respectively.  ... 
doi:10.1109/jstqe.2020.2975660 fatcat:y5nivjzs7ngkdlintavfszv5tu
« Previous Showing results 1 — 15 out of 53 results