Impact of virtual channels and adaptive routing on application performance

Research output: Contribution to journalArticle

25 Citations (Scopus)

Abstract

Research on multiprocessor interconnection networks has primarily focused on wormhole switching, virtual channel flow control, and routing algorithms to enhance their performance. The rationale behind this research is that by alleviating the network latency for high network loads, the overall system performance would improve. Many studies have used synthetic workloads to support this claim. However, such workloads may not necessarily capture the behavior of real applications. In this paper, we have used parallel applications for a closer examination of the network behavior. In particular, the performance benefit from enhancing a 2D mesh with virtual channels (VCs) and a fully adaptive routing algorithm is examined with a set of shared-memory and message passing applications. Execution time and average message latency of shared memory applications are measured using execution-driven simulation and by varying many architectural attributes that affect the network workload. The communication traces of message passing applications, collected on an IBM-SP2, are used to run a trace-driven simulation of the mesh architecture to obtain message latency. Simulation results show that VCs and adaptive routing can reduce the network latency to varying degrees depending on the application. However, these modest benefits do not translate to significant improvements in the overall execution time because the load on the network is not high enough to exploit the advantages of the network enhancements. Moreover, this benefit may be negated if the architectural enhancements increase the network cycle time. Rather, emphasis should be placed on improving the raw network bandwidth and faster network interfaces.

Original languageEnglish (US)
Pages (from-to)223-237
Number of pages15
JournalIEEE Transactions on Parallel and Distributed Systems
Volume12
Issue number2
DOIs
StatePublished - Feb 1 2001

Fingerprint

Message passing
Routing algorithms
Data storage equipment
Interconnection networks (circuit switching)
Channel flow
Adaptive algorithms
Flow control
Interfaces (computer)
Bandwidth
Communication

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Hardware and Architecture
  • Computational Theory and Mathematics

Cite this

@article{a25e255942b6460daa7aa6e32a55be41,
title = "Impact of virtual channels and adaptive routing on application performance",
abstract = "Research on multiprocessor interconnection networks has primarily focused on wormhole switching, virtual channel flow control, and routing algorithms to enhance their performance. The rationale behind this research is that by alleviating the network latency for high network loads, the overall system performance would improve. Many studies have used synthetic workloads to support this claim. However, such workloads may not necessarily capture the behavior of real applications. In this paper, we have used parallel applications for a closer examination of the network behavior. In particular, the performance benefit from enhancing a 2D mesh with virtual channels (VCs) and a fully adaptive routing algorithm is examined with a set of shared-memory and message passing applications. Execution time and average message latency of shared memory applications are measured using execution-driven simulation and by varying many architectural attributes that affect the network workload. The communication traces of message passing applications, collected on an IBM-SP2, are used to run a trace-driven simulation of the mesh architecture to obtain message latency. Simulation results show that VCs and adaptive routing can reduce the network latency to varying degrees depending on the application. However, these modest benefits do not translate to significant improvements in the overall execution time because the load on the network is not high enough to exploit the advantages of the network enhancements. Moreover, this benefit may be negated if the architectural enhancements increase the network cycle time. Rather, emphasis should be placed on improving the raw network bandwidth and faster network interfaces.",
author = "Vaidya, {Aniruddha S.} and Anand Sivasubramaniam and Chitaranjan Das",
year = "2001",
month = "2",
day = "1",
doi = "10.1109/71.910875",
language = "English (US)",
volume = "12",
pages = "223--237",
journal = "IEEE Transactions on Parallel and Distributed Systems",
issn = "1045-9219",
publisher = "IEEE Computer Society",
number = "2",

}

Impact of virtual channels and adaptive routing on application performance. / Vaidya, Aniruddha S.; Sivasubramaniam, Anand; Das, Chitaranjan.

In: IEEE Transactions on Parallel and Distributed Systems, Vol. 12, No. 2, 01.02.2001, p. 223-237.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Impact of virtual channels and adaptive routing on application performance

AU - Vaidya, Aniruddha S.

AU - Sivasubramaniam, Anand

AU - Das, Chitaranjan

PY - 2001/2/1

Y1 - 2001/2/1

N2 - Research on multiprocessor interconnection networks has primarily focused on wormhole switching, virtual channel flow control, and routing algorithms to enhance their performance. The rationale behind this research is that by alleviating the network latency for high network loads, the overall system performance would improve. Many studies have used synthetic workloads to support this claim. However, such workloads may not necessarily capture the behavior of real applications. In this paper, we have used parallel applications for a closer examination of the network behavior. In particular, the performance benefit from enhancing a 2D mesh with virtual channels (VCs) and a fully adaptive routing algorithm is examined with a set of shared-memory and message passing applications. Execution time and average message latency of shared memory applications are measured using execution-driven simulation and by varying many architectural attributes that affect the network workload. The communication traces of message passing applications, collected on an IBM-SP2, are used to run a trace-driven simulation of the mesh architecture to obtain message latency. Simulation results show that VCs and adaptive routing can reduce the network latency to varying degrees depending on the application. However, these modest benefits do not translate to significant improvements in the overall execution time because the load on the network is not high enough to exploit the advantages of the network enhancements. Moreover, this benefit may be negated if the architectural enhancements increase the network cycle time. Rather, emphasis should be placed on improving the raw network bandwidth and faster network interfaces.

AB - Research on multiprocessor interconnection networks has primarily focused on wormhole switching, virtual channel flow control, and routing algorithms to enhance their performance. The rationale behind this research is that by alleviating the network latency for high network loads, the overall system performance would improve. Many studies have used synthetic workloads to support this claim. However, such workloads may not necessarily capture the behavior of real applications. In this paper, we have used parallel applications for a closer examination of the network behavior. In particular, the performance benefit from enhancing a 2D mesh with virtual channels (VCs) and a fully adaptive routing algorithm is examined with a set of shared-memory and message passing applications. Execution time and average message latency of shared memory applications are measured using execution-driven simulation and by varying many architectural attributes that affect the network workload. The communication traces of message passing applications, collected on an IBM-SP2, are used to run a trace-driven simulation of the mesh architecture to obtain message latency. Simulation results show that VCs and adaptive routing can reduce the network latency to varying degrees depending on the application. However, these modest benefits do not translate to significant improvements in the overall execution time because the load on the network is not high enough to exploit the advantages of the network enhancements. Moreover, this benefit may be negated if the architectural enhancements increase the network cycle time. Rather, emphasis should be placed on improving the raw network bandwidth and faster network interfaces.

UR - http://www.scopus.com/inward/record.url?scp=0035248113&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0035248113&partnerID=8YFLogxK

U2 - 10.1109/71.910875

DO - 10.1109/71.910875

M3 - Article

VL - 12

SP - 223

EP - 237

JO - IEEE Transactions on Parallel and Distributed Systems

JF - IEEE Transactions on Parallel and Distributed Systems

SN - 1045-9219

IS - 2

ER -