Deterministic versus Adaptive Routing in Fat-Trees

Gómez, C.; Gilabert, Francisco Jurado; Gómez, María E.; López, Pedro; Duato, J.

doi:10.1109/ipdps.2007.370482

Cited by 123 publications

(70 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Existing deterministic single-path routing algorithms for ( + , ) are either the Source-mod-k routing (S-m-k) [6], [8], [9] or the Destination-mod-k routing (D-m-k) [3], [9], [12]. A good summary of routing in generalized fat-trees can be found in [9].…”

Section: A Worst-case Permutation Load Of Existing Routing Algorithmsmentioning

confidence: 99%

“…In this routing scheme, one path is used to carry all traffics from a source to a destination. Existing single-path deterministic routing for generalized fat trees include Sourcemod-k routing [6], [8], [9] and Destination-mod-k routing [3], [9], [12]. As will be shown later, all of these existing routing schemes are not ideal for many common communication patterns.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Improving Performance of Deterministic Single-Path Routing on 2-Level Generalized Fat-Trees

Nienaber

Mahapatra

Yuan

2011

2011 IEEE International Symposium on Parallel and Distributed Processing Workshops and PHD Forum

View full text Add to dashboard Cite

Abstract-This paper focuses on deterministic single-path routing schemes on 2-level generalized fat-trees. We develop a routing algorithm that is optimal in terms of worst-case permutation performance. In comparison to existing routing schemes for such topologies, our algorithm also improves the average performance of common communication patterns including bisect patterns, full permutation patterns, and dissemination (Bruck) patterns on various 2-level generalized fat-trees as demonstrated in our evaluation results.

show abstract

Section: A Worst-case Permutation Load Of Existing Routing Algorithmsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Improving Performance of Deterministic Single-Path Routing on 2-Level Generalized Fat-Trees

Nienaber

Mahapatra

Yuan

2011

2011 IEEE International Symposium on Parallel and Distributed Processing Workshops and PHD Forum

View full text Add to dashboard Cite

show abstract

“…Select a node, a, in ND; (4) for each S i ∈ S do (5) partition paths in S i that splits at node a into multiple sets…”

Section: B Split-merge Heuristicsmentioning

confidence: 99%

“…(2) workinggraph = split graph;color=1; (3) repeat (4) repeat (5) Select a node in workinggraph with the largest degree (6) Assign color to the node (7) Remove the node and all adjacent nodes from workinggraph (8) until no more nodes in workinggraph (9) workinggraph = all nodes without colors plus the edges between them (10) color ++; (11) until workinggraph is empty (all nodes are colored) Consider the example in Fig. 8.…”

Section: Graph Coloring Heuristicsmentioning

confidence: 99%

On LID assignment in infiniBand networks

Nienaber

Yuan

Duan

2007

Proceedings of the 3rd ACM/IEEE Symposium on Architecture for Networking and Communications Systems

View full text Add to dashboard Cite

Abstract-To realize a path in an InfiniBand network, an address, known as Local IDentifier (LID) in the InfiniBand specification, must be assigned to the destination of the path and used in the forwarding tables of intermediate switches to direct the traffic following the path. Hence, routing in InfiniBand has two components: (1) computing all paths, and (2) assigning LIDs to destinations and using them in intermediate switches to realize the paths. We refer to the task of computing paths as path computation and the task of assigning LIDs as LID assignment. This paper focuses on the LID assignment component, whose major issue is to minimize the number of LIDs required to support a given set of paths. We prove that the problem of realizing a given set of paths with a minimum number of LIDs is NPcomplete, develop an integer linear programming formulation for this problem, design a number of heuristics that are effective and efficient in practical cases, and evaluate the performance of the heuristics through simulation. The experimental results indicate that the performance of our best performing heuristic is very close to optimal. We further demonstrate that by separating path computation from LID assignment and using the schemes that are known to achieve good performance for path computation and LID assignment separately, more effective routing schemes than existing ones can be developed.

show abstract

“…To provide a deterministic routing algorithm, as required by the reference architecture, a unique path must be selected for each origin-destination pair among all the possible ones in the ascending phase. To address this issue, the deterministic routing algorithm presented in [66] is used. In this algorithm, during the ascending phase, consecutive destinations are shuffled among the different ascending links of the switches, as in Figure 3.36: each ascending port is labeled in italics with the destination cores that are reachable through it.…”

Section: Topologies Under Testmentioning

confidence: 99%

Design Space Exploration for Networks On-chip

Villamón¹

View full text Add to dashboard Cite

Deterministic versus Adaptive Routing in Fat-Trees

Abstract: Abstract

Cited by 123 publications

References 6 publications

Improving Performance of Deterministic Single-Path Routing on 2-Level Generalized Fat-Trees

Improving Performance of Deterministic Single-Path Routing on 2-Level Generalized Fat-Trees

On LID assignment in infiniBand networks

Design Space Exploration for Networks On-chip

Contact Info

Product

Resources

About