TTN: a high performance hierarchical interconnection network for massively parallel computers

Interconnection networks play a crucial role in the performance of massively parallel computers. Hierarchical interconnection networks provide high performance at low cost by exploring the locality that exists in the communication patterns of massively parallel computers. A Tori connected Torus Netw...

Full description

Bibliographic Details
Main Authors: Rahman, M.M. Hafizur, Inoguchi, Yasushi, Sato, Yukinori, Horiguchi, Susumu
Format: Article
Language:English
Published: The Institute of Electronics, Information and Communication Engineers, Japan 2009
Subjects:
Online Access:http://irep.iium.edu.my/7669/1/IEICE_2009.pdf
_version_ 1825644935349534720
author Rahman, M.M. Hafizur
Inoguchi, Yasushi
Sato, Yukinori
Horiguchi, Susumu
author_facet Rahman, M.M. Hafizur
Inoguchi, Yasushi
Sato, Yukinori
Horiguchi, Susumu
author_sort Rahman, M.M. Hafizur
collection IIUM
description Interconnection networks play a crucial role in the performance of massively parallel computers. Hierarchical interconnection networks provide high performance at low cost by exploring the locality that exists in the communication patterns of massively parallel computers. A Tori connected Torus Network (TTN) is a 2D-torus network of multiple basic modules, in which the basic modules are 2D-torus networks that are hierarchically interconnected for higher-level networks. This paper addresses the architectural details of the TTN and explores aspects such as node degree, network diameter, cost, average distance, arc connectivity, bisection width, and wiring complexity. We also present a deadlock-free routing algorithm for the TTN using four virtual channels and evaluate the network's dynamic communication performance using the proposed routing algorithm under uniform and various non-uniform traffic patterns. We evaluate the dynamic communication performance of TTN, TESH, MH3DT, mesh, and torus networks by computer simulation. It is shown that the TTN possesses several attractive features, including constant node degree, small diameter, low cost, small average distance, moderate (neither too low, nor too high) bisection width, and high throughput and very low zero load latency, which provide better dynamic communication performance than that of other conventional and hierarchical networks.
first_indexed 2024-03-05T22:40:11Z
format Article
id oai:generic.eprints.org:7669
institution International Islamic University Malaysia
language English
last_indexed 2024-03-05T22:40:11Z
publishDate 2009
publisher The Institute of Electronics, Information and Communication Engineers, Japan
record_format dspace
spelling oai:generic.eprints.org:76692012-01-16T05:28:13Z http://irep.iium.edu.my/7669/ TTN: a high performance hierarchical interconnection network for massively parallel computers Rahman, M.M. Hafizur Inoguchi, Yasushi Sato, Yukinori Horiguchi, Susumu T Technology (General) TK7885 Computer engineering Interconnection networks play a crucial role in the performance of massively parallel computers. Hierarchical interconnection networks provide high performance at low cost by exploring the locality that exists in the communication patterns of massively parallel computers. A Tori connected Torus Network (TTN) is a 2D-torus network of multiple basic modules, in which the basic modules are 2D-torus networks that are hierarchically interconnected for higher-level networks. This paper addresses the architectural details of the TTN and explores aspects such as node degree, network diameter, cost, average distance, arc connectivity, bisection width, and wiring complexity. We also present a deadlock-free routing algorithm for the TTN using four virtual channels and evaluate the network's dynamic communication performance using the proposed routing algorithm under uniform and various non-uniform traffic patterns. We evaluate the dynamic communication performance of TTN, TESH, MH3DT, mesh, and torus networks by computer simulation. It is shown that the TTN possesses several attractive features, including constant node degree, small diameter, low cost, small average distance, moderate (neither too low, nor too high) bisection width, and high throughput and very low zero load latency, which provide better dynamic communication performance than that of other conventional and hierarchical networks. The Institute of Electronics, Information and Communication Engineers, Japan 2009-05 Article PeerReviewed application/pdf en http://irep.iium.edu.my/7669/1/IEICE_2009.pdf Rahman, M.M. Hafizur and Inoguchi, Yasushi and Sato, Yukinori and Horiguchi, Susumu (2009) TTN: a high performance hierarchical interconnection network for massively parallel computers. IEICE Transactions on Information and Systems, E92D (5). pp. 1062-1078. ISSN 1745-1361 (O), 0916-8532 (P) http://search.ieice.org/bin/index.php?category=D&year=2009&vol=E92-D&num=5&lang=&abst=
spellingShingle T Technology (General)
TK7885 Computer engineering
Rahman, M.M. Hafizur
Inoguchi, Yasushi
Sato, Yukinori
Horiguchi, Susumu
TTN: a high performance hierarchical interconnection network for massively parallel computers
title TTN: a high performance hierarchical interconnection network for massively parallel computers
title_full TTN: a high performance hierarchical interconnection network for massively parallel computers
title_fullStr TTN: a high performance hierarchical interconnection network for massively parallel computers
title_full_unstemmed TTN: a high performance hierarchical interconnection network for massively parallel computers
title_short TTN: a high performance hierarchical interconnection network for massively parallel computers
title_sort ttn a high performance hierarchical interconnection network for massively parallel computers
topic T Technology (General)
TK7885 Computer engineering
url http://irep.iium.edu.my/7669/1/IEICE_2009.pdf
work_keys_str_mv AT rahmanmmhafizur ttnahighperformancehierarchicalinterconnectionnetworkformassivelyparallelcomputers
AT inoguchiyasushi ttnahighperformancehierarchicalinterconnectionnetworkformassivelyparallelcomputers
AT satoyukinori ttnahighperformancehierarchicalinterconnectionnetworkformassivelyparallelcomputers
AT horiguchisusumu ttnahighperformancehierarchicalinterconnectionnetworkformassivelyparallelcomputers