A holistic dataflow-inspired system design

Stephane Zuckerman, Haitao Wei, Guang R. Gao, Howard Wong, Jean Luc Gaudiot, Ahmed Louri

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

1 Citation (Scopus)

Abstract

Computer systems have undergone a fundamental transformation recently, from single-core processors to devices with increasingly higher core counts within a single chip. The semi-conductor industry now faces the infamous power and utilization walls. To meet these challenges, heterogeneity in design, both at the architecture and technology levels, will be the prevailing approach for energy efficient computing as specialized cores, accelerators, etc., can eliminate the energy overheads of general-purpose homogeneous cores. However, with future technological challenges pointing in the direction of on-chip heterogeneity, and because of the traditional difficulty of parallel programming, it becomes imperative to produce new system software stacks that can take advantage of the heterogeneous hardware. As a case in point, the core count per chip continues to increase dramatically while the available on-chip memory per core is only getting marginally bigger. Thus, data locality, already a must-have in high-performance computing, will become even more critical as memory technology progresses. In turn, this makes it crucial that new execution models be developed to better exploit the trends of future heterogeneous computing in many-core chips. To solve these issues, we propose a cross-cutting cross-layer approach to address the challenges posed by future heterogeneous many-core chips.

Original languageEnglish (US)
Title of host publicationProceedings - 2014 4th Workshop on Data-Flow Execution Models for Extreme Scale Computing, DFM 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages46-49
Number of pages4
ISBN (Print)9781479980956
DOIs
StatePublished - Apr 17 2014
Event2014 4th Workshop on Data-Flow Execution Models for Extreme Scale Computing, DFM 2014 - Edmonton, Canada
Duration: Aug 24 2014 → …

Other

Other2014 4th Workshop on Data-Flow Execution Models for Extreme Scale Computing, DFM 2014
CountryCanada
CityEdmonton
Period8/24/14 → …

Fingerprint

Systems analysis
Data storage equipment
Parallel programming
Particle accelerators
Computer systems
Hardware
Industry

Keywords

  • Codelets
  • Dataflow
  • Heterogeneous architecture
  • Streaming

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Computer Science Applications

Cite this

Zuckerman, S., Wei, H., Gao, G. R., Wong, H., Gaudiot, J. L., & Louri, A. (2014). A holistic dataflow-inspired system design. In Proceedings - 2014 4th Workshop on Data-Flow Execution Models for Extreme Scale Computing, DFM 2014 (pp. 46-49). [7089029] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/DFM.2014.16

A holistic dataflow-inspired system design. / Zuckerman, Stephane; Wei, Haitao; Gao, Guang R.; Wong, Howard; Gaudiot, Jean Luc; Louri, Ahmed.

Proceedings - 2014 4th Workshop on Data-Flow Execution Models for Extreme Scale Computing, DFM 2014. Institute of Electrical and Electronics Engineers Inc., 2014. p. 46-49 7089029.

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

Zuckerman, S, Wei, H, Gao, GR, Wong, H, Gaudiot, JL & Louri, A 2014, A holistic dataflow-inspired system design. in Proceedings - 2014 4th Workshop on Data-Flow Execution Models for Extreme Scale Computing, DFM 2014., 7089029, Institute of Electrical and Electronics Engineers Inc., pp. 46-49, 2014 4th Workshop on Data-Flow Execution Models for Extreme Scale Computing, DFM 2014, Edmonton, Canada, 8/24/14. https://doi.org/10.1109/DFM.2014.16
Zuckerman S, Wei H, Gao GR, Wong H, Gaudiot JL, Louri A. A holistic dataflow-inspired system design. In Proceedings - 2014 4th Workshop on Data-Flow Execution Models for Extreme Scale Computing, DFM 2014. Institute of Electrical and Electronics Engineers Inc. 2014. p. 46-49. 7089029 https://doi.org/10.1109/DFM.2014.16
Zuckerman, Stephane ; Wei, Haitao ; Gao, Guang R. ; Wong, Howard ; Gaudiot, Jean Luc ; Louri, Ahmed. / A holistic dataflow-inspired system design. Proceedings - 2014 4th Workshop on Data-Flow Execution Models for Extreme Scale Computing, DFM 2014. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 46-49
@inproceedings{77b271b88804467b8328248a70310cdc,
title = "A holistic dataflow-inspired system design",
abstract = "Computer systems have undergone a fundamental transformation recently, from single-core processors to devices with increasingly higher core counts within a single chip. The semi-conductor industry now faces the infamous power and utilization walls. To meet these challenges, heterogeneity in design, both at the architecture and technology levels, will be the prevailing approach for energy efficient computing as specialized cores, accelerators, etc., can eliminate the energy overheads of general-purpose homogeneous cores. However, with future technological challenges pointing in the direction of on-chip heterogeneity, and because of the traditional difficulty of parallel programming, it becomes imperative to produce new system software stacks that can take advantage of the heterogeneous hardware. As a case in point, the core count per chip continues to increase dramatically while the available on-chip memory per core is only getting marginally bigger. Thus, data locality, already a must-have in high-performance computing, will become even more critical as memory technology progresses. In turn, this makes it crucial that new execution models be developed to better exploit the trends of future heterogeneous computing in many-core chips. To solve these issues, we propose a cross-cutting cross-layer approach to address the challenges posed by future heterogeneous many-core chips.",
keywords = "Codelets, Dataflow, Heterogeneous architecture, Streaming",
author = "Stephane Zuckerman and Haitao Wei and Gao, {Guang R.} and Howard Wong and Gaudiot, {Jean Luc} and Ahmed Louri",
year = "2014",
month = "4",
day = "17",
doi = "10.1109/DFM.2014.16",
language = "English (US)",
isbn = "9781479980956",
pages = "46--49",
booktitle = "Proceedings - 2014 4th Workshop on Data-Flow Execution Models for Extreme Scale Computing, DFM 2014",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - A holistic dataflow-inspired system design

AU - Zuckerman, Stephane

AU - Wei, Haitao

AU - Gao, Guang R.

AU - Wong, Howard

AU - Gaudiot, Jean Luc

AU - Louri, Ahmed

PY - 2014/4/17

Y1 - 2014/4/17

N2 - Computer systems have undergone a fundamental transformation recently, from single-core processors to devices with increasingly higher core counts within a single chip. The semi-conductor industry now faces the infamous power and utilization walls. To meet these challenges, heterogeneity in design, both at the architecture and technology levels, will be the prevailing approach for energy efficient computing as specialized cores, accelerators, etc., can eliminate the energy overheads of general-purpose homogeneous cores. However, with future technological challenges pointing in the direction of on-chip heterogeneity, and because of the traditional difficulty of parallel programming, it becomes imperative to produce new system software stacks that can take advantage of the heterogeneous hardware. As a case in point, the core count per chip continues to increase dramatically while the available on-chip memory per core is only getting marginally bigger. Thus, data locality, already a must-have in high-performance computing, will become even more critical as memory technology progresses. In turn, this makes it crucial that new execution models be developed to better exploit the trends of future heterogeneous computing in many-core chips. To solve these issues, we propose a cross-cutting cross-layer approach to address the challenges posed by future heterogeneous many-core chips.

AB - Computer systems have undergone a fundamental transformation recently, from single-core processors to devices with increasingly higher core counts within a single chip. The semi-conductor industry now faces the infamous power and utilization walls. To meet these challenges, heterogeneity in design, both at the architecture and technology levels, will be the prevailing approach for energy efficient computing as specialized cores, accelerators, etc., can eliminate the energy overheads of general-purpose homogeneous cores. However, with future technological challenges pointing in the direction of on-chip heterogeneity, and because of the traditional difficulty of parallel programming, it becomes imperative to produce new system software stacks that can take advantage of the heterogeneous hardware. As a case in point, the core count per chip continues to increase dramatically while the available on-chip memory per core is only getting marginally bigger. Thus, data locality, already a must-have in high-performance computing, will become even more critical as memory technology progresses. In turn, this makes it crucial that new execution models be developed to better exploit the trends of future heterogeneous computing in many-core chips. To solve these issues, we propose a cross-cutting cross-layer approach to address the challenges posed by future heterogeneous many-core chips.

KW - Codelets

KW - Dataflow

KW - Heterogeneous architecture

KW - Streaming

UR - http://www.scopus.com/inward/record.url?scp=84949924625&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84949924625&partnerID=8YFLogxK

U2 - 10.1109/DFM.2014.16

DO - 10.1109/DFM.2014.16

M3 - Conference contribution

AN - SCOPUS:84949924625

SN - 9781479980956

SP - 46

EP - 49

BT - Proceedings - 2014 4th Workshop on Data-Flow Execution Models for Extreme Scale Computing, DFM 2014

PB - Institute of Electrical and Electronics Engineers Inc.

ER -