Filters








1,905 Hits in 3.5 sec

Interconnect routing and scheduling---Adaptive routing in high-radix clos network

John Kim, William J. Dally, Dennis Abts
2006 Proceedings of the 2006 ACM/IEEE conference on Supercomputing - SC '06  
We show that adaptive routing, if done properly, outperforms oblivious routing by providing lower latency, lower latency variance, and higher throughput with limited buffering.  ...  The cost of adaptive routing, in terms of router latency and area, is increased in high-radix routers.  ...  This work has been supported in part by New Technology Endeavors, Inc. through DARPA subcontract CR03-C-0002 under U.S. government Prime Contract Number NBCH3039003.  ... 
doi:10.1145/1188455.1188552 dblp:conf/sc/KimDA06 fatcat:66eyzpt2nbgshd3dvk3ao4mlhq

Adaptive Routing in High-Radix Clos Network

John Kim, William Dally, J. Dally, Dennis Abts
2006 ACM/IEEE SC 2006 Conference (SC'06)  
We show that adaptive routing, if done properly, outperforms oblivious routing by providing lower latency, lower latency variance, and higher throughput with limited buffering.  ...  The cost of adaptive routing, in terms of router latency and area, is increased in high-radix routers.  ...  This work has been supported in part by New Technology Endeavors, Inc. through DARPA subcontract CR03-C-0002 under U.S. government Prime Contract Number NBCH3039003.  ... 
doi:10.1109/sc.2006.10 fatcat:n5udo3hnavbobkmlu6vrk455he

Throughput Unfairness in Dragonfly Networks under Realistic Traffic Patterns

Pablo Fuentes, Enrique Vallejo, Cristobal Camarero, Ramon Beivide, Mateo Valero
2015 2015 IEEE International Conference on Cluster Computing  
Non-minimal adaptive routing is employed to fully exploit the path diversity and increase the performance under adversarial traffic patterns.  ...  This traffic pattern is the most adversarial in terms of network fairness.  ...  Figure 2 shows average throughput and latency for the described oblivious, source adaptive and in-transit adaptive routing mechanisms under UN, ADV+1 and ADVc traffic patterns, using transit-over-injection  ... 
doi:10.1109/cluster.2015.136 dblp:conf/cluster/Fuentes0CBV15 fatcat:2fss2epk5ncqbm5uchbl6d435e

Network unfairness in dragonfly topologies

Pablo Fuentes, Enrique Vallejo, Cristóbal Camarero, Ramón Beivide, Mateo Valero
2016 Journal of Supercomputing  
Results show that age-based arbitration mitigates fairness issues especially when using in-transit adaptive routing.  ...  This makes source adaptive routing in dragonflies based on remote notifications prone to reduced performance, even when using age-based arbitration.  ...  Figure 6 shows average throughput and latency for the described oblivious, source adaptive and in-transit adaptive routing mechanisms under UN, ADV+1 and ADVc traffic patterns, using in-transit-over-injection  ... 
doi:10.1007/s11227-016-1758-z fatcat:qhevuzutrrgebif4eg5s5vlc6y

GOAL

Arjun Singh, William J. Dally, Amit K. Gupta, Brian Towles
2003 Proceedings of the 30th annual international symposium on Computer architecture - ISCA '03  
fully randomized routing and exceeding the worst-case performance of Chaos [2] , RLB [14], and minimal routing [8] by more than 40%.  ...  GOAL achieves global load balance by randomly choosing the direction to route in each dimension. Local load balance is then achieved by routing in the selected directions adaptively.  ...  Suppose the routing algorithm has obliviously chosen the minimal quadrant, (+1, +1). One possible route of the packet is shown in Figure 5 .  ... 
doi:10.1145/859618.859641 fatcat:fjso4ypbpjbfdpuwiznezphr3m

Randomized Partially-Minimal Routing: Near-Optimal Oblivious Routing for 3-D Mesh Networks

Rohit Sunkam Ramanujam, Bill Lin
2012 IEEE Transactions on Very Large Scale Integration (vlsi) Systems  
The main contribution of this paper is a new oblivious routing algorithm for 3-D mesh networks called randomized partially-minimal (RPM) routing.  ...  Finally, whereas VAL achieves optimal worst-case throughput at a penalty factor of 2 in average latency over DOR, RPM achieves (near) optimal worst-case throughput with a much smaller factor of 1.33.  ...  This results in the following expression for the average hop count of RPM The penalty factor in average latency for using a partially minimal routing algorithm like RPM instead of a minimal routing algorithm  ... 
doi:10.1109/tvlsi.2011.2167361 fatcat:64pom6acq5e2vih4nvns6picii

Near-optimal oblivious routing on three-dimensional mesh networks

Rohit Sunkam Ramanujam, Bill Lin
2008 2008 IEEE International Conference on Computer Design  
In practice, the average latency of RPM is expected to be closer to minimal routing because 3D mesh networks are not expected to be symmetric in 3D chip designs.  ...  The main contribution of this paper is the design of a new oblivious routing algorithm for 3D mesh networks called Randomized Partially-Minimal (RPM) routing.  ...  (c) On average latency, as measured in network hops, RPM does not achieve minimal-length routing because nonminimal routing is used in one of three dimensions.  ... 
doi:10.1109/iccd.2008.4751852 dblp:conf/iccd/RamanujamL08 fatcat:kob7wxywkjgnzpq4canm4wz3a4

A New Worst-Case Throughput Bound for Oblivious Routing in Odd Radix Mesh Network

Guang Sun, Chia-Wei Chang, Bill Lin
2013 IEEE computer architecture letters  
Moreover, U2TURN achieves good average-throughput at the expense of approximately 1.5x minimal average hop count.  ...  In addition, we propose a routing algorithm called U2TURN that can achieve this worst-case throughput bound.  ...  Latency Analysis For latency analysis, we use the average hop count Have(R) for a given routing algorithm R to measure the average latency.  ... 
doi:10.1109/l-ca.2012.5 fatcat:tgjrworn6zay3lhrewo7bodtwy

Near-Optimal Worst-Case Throughput Routing for Two-Dimensional Mesh Networks

Daeho Seo, Akif Ali, Won-Taek Lim, Nauman Rafique, Mithuna Thottethodi
2005 SIGARCH Computer Architecture News  
Minimizing latency and maximizing throughput are important goals in the design of routing algorithms for interconnection networks.  ...  The major contribution of this paper is the design of an oblivious routing algorithm-O1TURN-with provable nearoptimal worst-case throughput, good average-case throughput, low design complexity and minimal  ...  This work is supported in part by Purdue Research Foundation XR Grant No. 6904010 and Purdue University.  ... 
doi:10.1145/1080695.1070006 fatcat:aoewowwtsrbylft6zqz7fjqvue

Regional congestion awareness for load balance in networks-on-chip

Paul Gratz, Boris Grot, Stephen W. Keckler
2008 High-Performance Computer Architecture  
Our experiments show that RCA matches or exceeds the performance of conventional adaptive routing across all workloads examined, with a 16% average and 71% maximum latency reduction on SPLASH-2 benchmarks  ...  Despite somewhat higher implementation complexity, adaptive routing enjoys better fault tolerance characteristics, increases network throughput, and decreases latency compared to oblivious policies when  ...  Figure 6 . 6 Load-latency graphs comparing RCA, locally adaptive, and oblivious routing. Figure 7 . 7 Average latency for 100 permutations of random pair traffic at 30% injection bandwidth.  ... 
doi:10.1109/hpca.2008.4658640 dblp:conf/hpca/GratzGK08 fatcat:sko6ai7e2ban5bbxalynmfhrty

Globally Adaptive Load-Balanced Routing on Tori

A. Singh, W.J. Dally, B. Towles, A.K. Gupta
2004 IEEE computer architecture letters  
Using global information, GAL achieves the performance (latency and throughput) of minimal adaptive routing on benign traffic patterns and performs as well as the best obliviously load-balanced routing  ...  It further load balances the network by routing in the selected directions adaptively.  ...  GAL combines the best features of minimal algorithms (low latency at low load) and obliviously load balanced algorithms (high saturation throughput on adversarial patterns).  ... 
doi:10.1109/l-ca.2004.8 fatcat:wp7rjlroiba4zb6xhk3ad7znqy

Optimal and Heuristic Application-Aware Oblivious Routing

Michel A. Kinsy, Myong Hyon Cho, Keun Sup Shim, Mieszko Lis, G. Edward Suh, Srinivas Devadas
2013 IEEE transactions on computers  
Conventional oblivious routing algorithms do not take into account resource requirements (e.g., bandwidth, latency) of various flows in a given application.  ...  Our framework can be used to produce application-aware routes that target the minimization of latency, number of flows through a link, bandwidth, or any combination thereof.  ...  We would like to acknowledge the support of Intel Corporation for providing some of the workstations used in conducting this research. This research is partially funded by NSF grant CCF-0905208.  ... 
doi:10.1109/tc.2011.219 fatcat:o3r2d2uqurf5bgc5slkijxr5ly

Improving the Interaction between Overlay Routing and Traffic Engineering [chapter]

Gene Moo Lee, Taehwan Choi
2008 Lecture Notes in Computer Science  
In the meantime, traffic engineering (TE) are being used to reduce the whole network cost by adapting physical routing in response to varying traffic patterns.  ...  In this paper, we improve the interaction between overlay routing and TE by modifying the objectives of both parties.  ...  Lastly, we use average latency as an indicator in TE-aware overlay routing. But it may be difficult to get this value in reality.  ... 
doi:10.1007/978-3-540-79549-0_46 fatcat:f2scrp3atvcnlfi3iomzs2nkly

Locality-preserving randomized oblivious routing on torus networks

Arjun Singh, William J. Dally, Brian Towles, Amit K. Gupta
2002 Proceedings of the fourteenth annual ACM symposium on Parallel algorithms and architectures - SPAA '02  
Our results show that RLB outperforms deterministic algorithms (25% more bandwidth than Dimension Order Routing) and minimal oblivious algorithms (50% more bandwidth than 2 phase ROMM [9]) on worst-case  ...  RLBth (RLB threshold) improves the locality of RLB to match the throughput of minimal algorithms on very local traffic in exchange for a 4% reduction in worst-case throughput compared to RLB.  ...  RLBth to that of the three oblivious routing algorithms listed in Table 3 .  ... 
doi:10.1145/564870.564873 dblp:conf/spaa/SinghDTG02 fatcat:a4g3h7lpvfbhnn572ctt3chuii

Locality-preserving randomized oblivious routing on torus networks

Arjun Singh, William J. Dally, Brian Towles, Amit K. Gupta
2002 Proceedings of the fourteenth annual ACM symposium on Parallel algorithms and architectures - SPAA '02  
Our results show that RLB outperforms deterministic algorithms (25% more bandwidth than Dimension Order Routing) and minimal oblivious algorithms (50% more bandwidth than 2 phase ROMM [9]) on worst-case  ...  RLBth (RLB threshold) improves the locality of RLB to match the throughput of minimal algorithms on very local traffic in exchange for a 4% reduction in worst-case throughput compared to RLB.  ...  RLBth to that of the three oblivious routing algorithms listed in Table 3 .  ... 
doi:10.1145/564871.564873 fatcat:7ud6yd2kabhmxohx6titpg7d7a
« Previous Showing results 1 — 15 out of 1,905 results