A variable-pipeline on-chip router optimized to traffic pattern

Yuto Hirata, Hiroki Matsutani, Michihiro Koibuchi, Hideharu Amano

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

7 Citations (Scopus)

Abstract

Network-on-Chip (NoC) can be evaluated from various aspects, such as communication latency, throughput, and power consumption. The preference of these requirements depends on each application. An on-chip variable-pipeline (VP) router that can adapt to these requirements by dynamically reconfiguring its data path structure is proposed in this paper. In response to the communication pattern, it can change the pipeline structure, supply voltage, and operational frequency using the dynamic voltage and frequency scaling (DVFS). As the traffic load becomes high, the VP router uses a look-ahead two-cycle pipeline structure for exploiting the maximum frequency, while it behaves as a one-cycle router when a low latency is preferred. A three-cycle pipelined structure with an adaptive routing enables to dynamically avoid hotspots. Instead of a simple pipeline-stage unification which causes rapid decrease of the operating frequency, by speculatively executing multiple pipeline stages in parallel, the operating frequency gracefully decreases as the number of the pipeline stages increases. Simulation results show that the one-cycle mode offers the shortest communication latency, while the two-cycle mode achieves the highest throughput for SPLASH-2 benchmarks.

Original languageEnglish
Title of host publication3rd International Workshop on Network on Chip Architectures, NoCArc 2010, in Conjunction with the 43rd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-43
Pages57-62
Number of pages6
DOIs
Publication statusPublished - 2010
Event3rd International Workshop on Network on Chip Architectures, NoCArc 2010, in Conjunction with the 43rd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-43 - Atlanta, GA, United States
Duration: 2010 Dec 42010 Dec 4

Other

Other3rd International Workshop on Network on Chip Architectures, NoCArc 2010, in Conjunction with the 43rd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-43
CountryUnited States
CityAtlanta, GA
Period10/12/410/12/4

Fingerprint

Routers
Pipelines
Communication
Throughput
Electric power utilization
Electric potential

Keywords

  • interconnects
  • network-on-chip
  • router architecture

ASJC Scopus subject areas

  • Hardware and Architecture
  • Electrical and Electronic Engineering

Cite this

Hirata, Y., Matsutani, H., Koibuchi, M., & Amano, H. (2010). A variable-pipeline on-chip router optimized to traffic pattern. In 3rd International Workshop on Network on Chip Architectures, NoCArc 2010, in Conjunction with the 43rd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-43 (pp. 57-62) https://doi.org/10.1145/1921249.1921263

A variable-pipeline on-chip router optimized to traffic pattern. / Hirata, Yuto; Matsutani, Hiroki; Koibuchi, Michihiro; Amano, Hideharu.

3rd International Workshop on Network on Chip Architectures, NoCArc 2010, in Conjunction with the 43rd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-43. 2010. p. 57-62.

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

Hirata, Y, Matsutani, H, Koibuchi, M & Amano, H 2010, A variable-pipeline on-chip router optimized to traffic pattern. in 3rd International Workshop on Network on Chip Architectures, NoCArc 2010, in Conjunction with the 43rd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-43. pp. 57-62, 3rd International Workshop on Network on Chip Architectures, NoCArc 2010, in Conjunction with the 43rd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-43, Atlanta, GA, United States, 10/12/4. https://doi.org/10.1145/1921249.1921263
Hirata Y, Matsutani H, Koibuchi M, Amano H. A variable-pipeline on-chip router optimized to traffic pattern. In 3rd International Workshop on Network on Chip Architectures, NoCArc 2010, in Conjunction with the 43rd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-43. 2010. p. 57-62 https://doi.org/10.1145/1921249.1921263
Hirata, Yuto ; Matsutani, Hiroki ; Koibuchi, Michihiro ; Amano, Hideharu. / A variable-pipeline on-chip router optimized to traffic pattern. 3rd International Workshop on Network on Chip Architectures, NoCArc 2010, in Conjunction with the 43rd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-43. 2010. pp. 57-62
@inproceedings{56b70e14fd1347f39824ba7153bf98ca,
title = "A variable-pipeline on-chip router optimized to traffic pattern",
abstract = "Network-on-Chip (NoC) can be evaluated from various aspects, such as communication latency, throughput, and power consumption. The preference of these requirements depends on each application. An on-chip variable-pipeline (VP) router that can adapt to these requirements by dynamically reconfiguring its data path structure is proposed in this paper. In response to the communication pattern, it can change the pipeline structure, supply voltage, and operational frequency using the dynamic voltage and frequency scaling (DVFS). As the traffic load becomes high, the VP router uses a look-ahead two-cycle pipeline structure for exploiting the maximum frequency, while it behaves as a one-cycle router when a low latency is preferred. A three-cycle pipelined structure with an adaptive routing enables to dynamically avoid hotspots. Instead of a simple pipeline-stage unification which causes rapid decrease of the operating frequency, by speculatively executing multiple pipeline stages in parallel, the operating frequency gracefully decreases as the number of the pipeline stages increases. Simulation results show that the one-cycle mode offers the shortest communication latency, while the two-cycle mode achieves the highest throughput for SPLASH-2 benchmarks.",
keywords = "interconnects, network-on-chip, router architecture",
author = "Yuto Hirata and Hiroki Matsutani and Michihiro Koibuchi and Hideharu Amano",
year = "2010",
doi = "10.1145/1921249.1921263",
language = "English",
isbn = "9781450303972",
pages = "57--62",
booktitle = "3rd International Workshop on Network on Chip Architectures, NoCArc 2010, in Conjunction with the 43rd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-43",

}

TY - GEN

T1 - A variable-pipeline on-chip router optimized to traffic pattern

AU - Hirata, Yuto

AU - Matsutani, Hiroki

AU - Koibuchi, Michihiro

AU - Amano, Hideharu

PY - 2010

Y1 - 2010

N2 - Network-on-Chip (NoC) can be evaluated from various aspects, such as communication latency, throughput, and power consumption. The preference of these requirements depends on each application. An on-chip variable-pipeline (VP) router that can adapt to these requirements by dynamically reconfiguring its data path structure is proposed in this paper. In response to the communication pattern, it can change the pipeline structure, supply voltage, and operational frequency using the dynamic voltage and frequency scaling (DVFS). As the traffic load becomes high, the VP router uses a look-ahead two-cycle pipeline structure for exploiting the maximum frequency, while it behaves as a one-cycle router when a low latency is preferred. A three-cycle pipelined structure with an adaptive routing enables to dynamically avoid hotspots. Instead of a simple pipeline-stage unification which causes rapid decrease of the operating frequency, by speculatively executing multiple pipeline stages in parallel, the operating frequency gracefully decreases as the number of the pipeline stages increases. Simulation results show that the one-cycle mode offers the shortest communication latency, while the two-cycle mode achieves the highest throughput for SPLASH-2 benchmarks.

AB - Network-on-Chip (NoC) can be evaluated from various aspects, such as communication latency, throughput, and power consumption. The preference of these requirements depends on each application. An on-chip variable-pipeline (VP) router that can adapt to these requirements by dynamically reconfiguring its data path structure is proposed in this paper. In response to the communication pattern, it can change the pipeline structure, supply voltage, and operational frequency using the dynamic voltage and frequency scaling (DVFS). As the traffic load becomes high, the VP router uses a look-ahead two-cycle pipeline structure for exploiting the maximum frequency, while it behaves as a one-cycle router when a low latency is preferred. A three-cycle pipelined structure with an adaptive routing enables to dynamically avoid hotspots. Instead of a simple pipeline-stage unification which causes rapid decrease of the operating frequency, by speculatively executing multiple pipeline stages in parallel, the operating frequency gracefully decreases as the number of the pipeline stages increases. Simulation results show that the one-cycle mode offers the shortest communication latency, while the two-cycle mode achieves the highest throughput for SPLASH-2 benchmarks.

KW - interconnects

KW - network-on-chip

KW - router architecture

UR - http://www.scopus.com/inward/record.url?scp=79951639613&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79951639613&partnerID=8YFLogxK

U2 - 10.1145/1921249.1921263

DO - 10.1145/1921249.1921263

M3 - Conference contribution

AN - SCOPUS:79951639613

SN - 9781450303972

SP - 57

EP - 62

BT - 3rd International Workshop on Network on Chip Architectures, NoCArc 2010, in Conjunction with the 43rd Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-43

ER -