Towards exascale co-design in a runtime system

Thomas Sterling; Matthew Anderson; P. Kevin Bohan; Maciej Brodowicz; Abhishek Kulkarni; Bo Zhang

doi:10.1007/978-3-319-15976-8_6

Towards exascale co-design in a runtime system

Thomas Sterling
, Matthew Anderson
, P. Kevin Bohan
, Maciej Brodowicz
, Abhishek Kulkarni
, Bo Zhang

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

3 Scopus citations

Abstract

Achieving the performance potential of an Exascale machine depends on realizing both operational efficiency and scalability in high performance computing applications. This requirement has motivated the emergence of several new programming models which emphasize fine and medium grain task parallelism in order to address the aggravating effects of asynchrony at scale. The performance modeling of Exascale systems for these programming models requires the development of fundamentally new approaches due to the demands of both scale and complexity. This work presents a performance modeling case study of the Livermore Unstructured Lagrangian Explicit Shock Hydrodynamics (LULESH) proxy application where the performance modeling approach has been incorporated directly into a runtime system with two modalities of operation: computation and performance modeling simulation. The runtime system exposes performance sensitivies and projects operation to larger scales while also realizing the benefits of removing global barriers and extracting more parallelism from LULESH. Comparisons between the computation and performance modeling simulation results are presented.

Original language	English
Title of host publication	Solving Software Challenges for Exascale - International Conference on Exascale Applications and Software, EASC 2014, Revised Selected Papers
Editors	Stefano Markidis, Erwin Laure
Publisher	Springer Verlag
Pages	85-99
Number of pages	15
ISBN (Electronic)	9783319159751
DOIs	https://doi.org/10.1007/978-3-319-15976-8_6
State	Published - 2015
Externally published	Yes
Event	International Conference on Exascale Applications and Software, EASC 2014 - Stockholm, Sweden Duration: Apr 2 2014 → Apr 3 2014

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	8759
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	International Conference on Exascale Applications and Software, EASC 2014
Country/Territory	Sweden
City	Stockholm
Period	04/2/14 → 04/3/14

Access to Document

10.1007/978-3-319-15976-8_6

Cite this

Sterling, T., Anderson, M., Bohan, P. K., Brodowicz, M., Kulkarni, A., & Zhang, B. (2015). Towards exascale co-design in a runtime system. In S. Markidis, & E. Laure (Eds.), Solving Software Challenges for Exascale - International Conference on Exascale Applications and Software, EASC 2014, Revised Selected Papers (pp. 85-99). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8759). Springer Verlag. https://doi.org/10.1007/978-3-319-15976-8_6

Sterling, Thomas ; Anderson, Matthew ; Bohan, P. Kevin et al. / Towards exascale co-design in a runtime system. Solving Software Challenges for Exascale - International Conference on Exascale Applications and Software, EASC 2014, Revised Selected Papers. editor / Stefano Markidis ; Erwin Laure. Springer Verlag, 2015. pp. 85-99 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{ba7b6b353f5c41b4b2e3b1a331091ba4,

title = "Towards exascale co-design in a runtime system",

abstract = "Achieving the performance potential of an Exascale machine depends on realizing both operational efficiency and scalability in high performance computing applications. This requirement has motivated the emergence of several new programming models which emphasize fine and medium grain task parallelism in order to address the aggravating effects of asynchrony at scale. The performance modeling of Exascale systems for these programming models requires the development of fundamentally new approaches due to the demands of both scale and complexity. This work presents a performance modeling case study of the Livermore Unstructured Lagrangian Explicit Shock Hydrodynamics (LULESH) proxy application where the performance modeling approach has been incorporated directly into a runtime system with two modalities of operation: computation and performance modeling simulation. The runtime system exposes performance sensitivies and projects operation to larger scales while also realizing the benefits of removing global barriers and extracting more parallelism from LULESH. Comparisons between the computation and performance modeling simulation results are presented.",

author = "Thomas Sterling and Matthew Anderson and Bohan, \{P. Kevin\} and Maciej Brodowicz and Abhishek Kulkarni and Bo Zhang",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2015.; International Conference on Exascale Applications and Software, EASC 2014 ; Conference date: 02-04-2014 Through 03-04-2014",

year = "2015",

doi = "10.1007/978-3-319-15976-8\_6",

language = "English",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "85--99",

editor = "Stefano Markidis and Erwin Laure",

booktitle = "Solving Software Challenges for Exascale - International Conference on Exascale Applications and Software, EASC 2014, Revised Selected Papers",

}

Sterling, T, Anderson, M, Bohan, PK, Brodowicz, M, Kulkarni, A & Zhang, B 2015, Towards exascale co-design in a runtime system. in S Markidis & E Laure (eds), Solving Software Challenges for Exascale - International Conference on Exascale Applications and Software, EASC 2014, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8759, Springer Verlag, pp. 85-99, International Conference on Exascale Applications and Software, EASC 2014, Stockholm, Sweden, 04/2/14. https://doi.org/10.1007/978-3-319-15976-8_6

Towards exascale co-design in a runtime system. / Sterling, Thomas; Anderson, Matthew; Bohan, P. Kevin et al.
Solving Software Challenges for Exascale - International Conference on Exascale Applications and Software, EASC 2014, Revised Selected Papers. ed. / Stefano Markidis; Erwin Laure. Springer Verlag, 2015. p. 85-99 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8759).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Towards exascale co-design in a runtime system

AU - Sterling, Thomas

AU - Anderson, Matthew

AU - Bohan, P. Kevin

AU - Brodowicz, Maciej

AU - Kulkarni, Abhishek

AU - Zhang, Bo

PY - 2015

Y1 - 2015

N2 - Achieving the performance potential of an Exascale machine depends on realizing both operational efficiency and scalability in high performance computing applications. This requirement has motivated the emergence of several new programming models which emphasize fine and medium grain task parallelism in order to address the aggravating effects of asynchrony at scale. The performance modeling of Exascale systems for these programming models requires the development of fundamentally new approaches due to the demands of both scale and complexity. This work presents a performance modeling case study of the Livermore Unstructured Lagrangian Explicit Shock Hydrodynamics (LULESH) proxy application where the performance modeling approach has been incorporated directly into a runtime system with two modalities of operation: computation and performance modeling simulation. The runtime system exposes performance sensitivies and projects operation to larger scales while also realizing the benefits of removing global barriers and extracting more parallelism from LULESH. Comparisons between the computation and performance modeling simulation results are presented.

AB - Achieving the performance potential of an Exascale machine depends on realizing both operational efficiency and scalability in high performance computing applications. This requirement has motivated the emergence of several new programming models which emphasize fine and medium grain task parallelism in order to address the aggravating effects of asynchrony at scale. The performance modeling of Exascale systems for these programming models requires the development of fundamentally new approaches due to the demands of both scale and complexity. This work presents a performance modeling case study of the Livermore Unstructured Lagrangian Explicit Shock Hydrodynamics (LULESH) proxy application where the performance modeling approach has been incorporated directly into a runtime system with two modalities of operation: computation and performance modeling simulation. The runtime system exposes performance sensitivies and projects operation to larger scales while also realizing the benefits of removing global barriers and extracting more parallelism from LULESH. Comparisons between the computation and performance modeling simulation results are presented.

UR - https://www.scopus.com/pages/publications/84928885906

U2 - 10.1007/978-3-319-15976-8_6

DO - 10.1007/978-3-319-15976-8_6

M3 - Conference contribution

AN - SCOPUS:84928885906

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 85

EP - 99

BT - Solving Software Challenges for Exascale - International Conference on Exascale Applications and Software, EASC 2014, Revised Selected Papers

A2 - Markidis, Stefano

A2 - Laure, Erwin

PB - Springer Verlag

T2 - International Conference on Exascale Applications and Software, EASC 2014

Y2 - 2 April 2014 through 3 April 2014

ER -

Sterling T, Anderson M, Bohan PK, Brodowicz M, Kulkarni A, Zhang B. Towards exascale co-design in a runtime system. In Markidis S, Laure E, editors, Solving Software Challenges for Exascale - International Conference on Exascale Applications and Software, EASC 2014, Revised Selected Papers. Springer Verlag. 2015. p. 85-99. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-15976-8_6

Towards exascale co-design in a runtime system

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this