HiRy

An advanced theory on design of deadlock-free adaptive routing for arbitrary topologies

Ryuta Kawano, Ryota Yasudo, Hiroki Matsutani, Michihiro Koibuchi, Hideharu Amano

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Recently proposed irregular networks can reduce the latency for both on-chip and off-chip systems with a large number of computing nodes and thus can improve the performance of parallel application. However, these networks usually suffer from deadlocks in routing packets when using a naive minimal path routing algorithm. To solve this problem, we focus attention on a lately proposed theory that generalizes the turn model to maintain the network performance with deadlock-freedom. The theorems remain a challenge of applying themselves to arbitrary topologies including fully irregular networks. In this paper, we advance the theorems to completely general ones. To apply the idea of the turn model to arbitrary topologies, we introduce a concept of regions that define continuous directions of channels on an n-dimensional space. Moreover, we provide a feasible implementation of a deadlock-free routing method based on our advanced theorem. To reduce the latency and the number of required Virtual Channels (VCs) with this method, a heuristic approach is introduced to reduce the number of prohibited turns between channels. Experimental results show that the routing method based on our proposed theorem can improve the network throughput by up to 138 % compared to a conventional deterministic minimal routing method. Moreover, it can reduce the latency by up to 2.9 % compared to another fully adaptive routing method.

Original languageEnglish
Title of host publicationProceedings - 2017 IEEE 23rd International Conference on Parallel and Distributed Systems, ICPADS 2017
PublisherIEEE Computer Society
Pages664-673
Number of pages10
Volume2017-December
ISBN (Electronic)9781538621295
DOIs
Publication statusPublished - 2018 May 29
Event23rd IEEE International Conference on Parallel and Distributed Systems, ICPADS 2017 - Shenzhen, China
Duration: 2017 Dec 152017 Dec 17

Other

Other23rd IEEE International Conference on Parallel and Distributed Systems, ICPADS 2017
CountryChina
CityShenzhen
Period17/12/1517/12/17

Fingerprint

Topology
Routing algorithms
Network performance
Throughput

Keywords

  • Deadlock free Routing Algorithm
  • High Performance Computing
  • Interconnection Networks
  • Irregular Networks
  • Virtual Channels

ASJC Scopus subject areas

  • Hardware and Architecture

Cite this

Kawano, R., Yasudo, R., Matsutani, H., Koibuchi, M., & Amano, H. (2018). HiRy: An advanced theory on design of deadlock-free adaptive routing for arbitrary topologies. In Proceedings - 2017 IEEE 23rd International Conference on Parallel and Distributed Systems, ICPADS 2017 (Vol. 2017-December, pp. 664-673). IEEE Computer Society. https://doi.org/10.1109/ICPADS.2017.00091

HiRy : An advanced theory on design of deadlock-free adaptive routing for arbitrary topologies. / Kawano, Ryuta; Yasudo, Ryota; Matsutani, Hiroki; Koibuchi, Michihiro; Amano, Hideharu.

Proceedings - 2017 IEEE 23rd International Conference on Parallel and Distributed Systems, ICPADS 2017. Vol. 2017-December IEEE Computer Society, 2018. p. 664-673.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Kawano, R, Yasudo, R, Matsutani, H, Koibuchi, M & Amano, H 2018, HiRy: An advanced theory on design of deadlock-free adaptive routing for arbitrary topologies. in Proceedings - 2017 IEEE 23rd International Conference on Parallel and Distributed Systems, ICPADS 2017. vol. 2017-December, IEEE Computer Society, pp. 664-673, 23rd IEEE International Conference on Parallel and Distributed Systems, ICPADS 2017, Shenzhen, China, 17/12/15. https://doi.org/10.1109/ICPADS.2017.00091
Kawano R, Yasudo R, Matsutani H, Koibuchi M, Amano H. HiRy: An advanced theory on design of deadlock-free adaptive routing for arbitrary topologies. In Proceedings - 2017 IEEE 23rd International Conference on Parallel and Distributed Systems, ICPADS 2017. Vol. 2017-December. IEEE Computer Society. 2018. p. 664-673 https://doi.org/10.1109/ICPADS.2017.00091
Kawano, Ryuta ; Yasudo, Ryota ; Matsutani, Hiroki ; Koibuchi, Michihiro ; Amano, Hideharu. / HiRy : An advanced theory on design of deadlock-free adaptive routing for arbitrary topologies. Proceedings - 2017 IEEE 23rd International Conference on Parallel and Distributed Systems, ICPADS 2017. Vol. 2017-December IEEE Computer Society, 2018. pp. 664-673
@inproceedings{dc5d340164a74c868186c14df2c500e2,
title = "HiRy: An advanced theory on design of deadlock-free adaptive routing for arbitrary topologies",
abstract = "Recently proposed irregular networks can reduce the latency for both on-chip and off-chip systems with a large number of computing nodes and thus can improve the performance of parallel application. However, these networks usually suffer from deadlocks in routing packets when using a naive minimal path routing algorithm. To solve this problem, we focus attention on a lately proposed theory that generalizes the turn model to maintain the network performance with deadlock-freedom. The theorems remain a challenge of applying themselves to arbitrary topologies including fully irregular networks. In this paper, we advance the theorems to completely general ones. To apply the idea of the turn model to arbitrary topologies, we introduce a concept of regions that define continuous directions of channels on an n-dimensional space. Moreover, we provide a feasible implementation of a deadlock-free routing method based on our advanced theorem. To reduce the latency and the number of required Virtual Channels (VCs) with this method, a heuristic approach is introduced to reduce the number of prohibited turns between channels. Experimental results show that the routing method based on our proposed theorem can improve the network throughput by up to 138 {\%} compared to a conventional deterministic minimal routing method. Moreover, it can reduce the latency by up to 2.9 {\%} compared to another fully adaptive routing method.",
keywords = "Deadlock free Routing Algorithm, High Performance Computing, Interconnection Networks, Irregular Networks, Virtual Channels",
author = "Ryuta Kawano and Ryota Yasudo and Hiroki Matsutani and Michihiro Koibuchi and Hideharu Amano",
year = "2018",
month = "5",
day = "29",
doi = "10.1109/ICPADS.2017.00091",
language = "English",
volume = "2017-December",
pages = "664--673",
booktitle = "Proceedings - 2017 IEEE 23rd International Conference on Parallel and Distributed Systems, ICPADS 2017",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - HiRy

T2 - An advanced theory on design of deadlock-free adaptive routing for arbitrary topologies

AU - Kawano, Ryuta

AU - Yasudo, Ryota

AU - Matsutani, Hiroki

AU - Koibuchi, Michihiro

AU - Amano, Hideharu

PY - 2018/5/29

Y1 - 2018/5/29

N2 - Recently proposed irregular networks can reduce the latency for both on-chip and off-chip systems with a large number of computing nodes and thus can improve the performance of parallel application. However, these networks usually suffer from deadlocks in routing packets when using a naive minimal path routing algorithm. To solve this problem, we focus attention on a lately proposed theory that generalizes the turn model to maintain the network performance with deadlock-freedom. The theorems remain a challenge of applying themselves to arbitrary topologies including fully irregular networks. In this paper, we advance the theorems to completely general ones. To apply the idea of the turn model to arbitrary topologies, we introduce a concept of regions that define continuous directions of channels on an n-dimensional space. Moreover, we provide a feasible implementation of a deadlock-free routing method based on our advanced theorem. To reduce the latency and the number of required Virtual Channels (VCs) with this method, a heuristic approach is introduced to reduce the number of prohibited turns between channels. Experimental results show that the routing method based on our proposed theorem can improve the network throughput by up to 138 % compared to a conventional deterministic minimal routing method. Moreover, it can reduce the latency by up to 2.9 % compared to another fully adaptive routing method.

AB - Recently proposed irregular networks can reduce the latency for both on-chip and off-chip systems with a large number of computing nodes and thus can improve the performance of parallel application. However, these networks usually suffer from deadlocks in routing packets when using a naive minimal path routing algorithm. To solve this problem, we focus attention on a lately proposed theory that generalizes the turn model to maintain the network performance with deadlock-freedom. The theorems remain a challenge of applying themselves to arbitrary topologies including fully irregular networks. In this paper, we advance the theorems to completely general ones. To apply the idea of the turn model to arbitrary topologies, we introduce a concept of regions that define continuous directions of channels on an n-dimensional space. Moreover, we provide a feasible implementation of a deadlock-free routing method based on our advanced theorem. To reduce the latency and the number of required Virtual Channels (VCs) with this method, a heuristic approach is introduced to reduce the number of prohibited turns between channels. Experimental results show that the routing method based on our proposed theorem can improve the network throughput by up to 138 % compared to a conventional deterministic minimal routing method. Moreover, it can reduce the latency by up to 2.9 % compared to another fully adaptive routing method.

KW - Deadlock free Routing Algorithm

KW - High Performance Computing

KW - Interconnection Networks

KW - Irregular Networks

KW - Virtual Channels

UR - http://www.scopus.com/inward/record.url?scp=85048375136&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048375136&partnerID=8YFLogxK

U2 - 10.1109/ICPADS.2017.00091

DO - 10.1109/ICPADS.2017.00091

M3 - Conference contribution

VL - 2017-December

SP - 664

EP - 673

BT - Proceedings - 2017 IEEE 23rd International Conference on Parallel and Distributed Systems, ICPADS 2017

PB - IEEE Computer Society

ER -