Efficient SDS simulations on multi-GPU nodes of XSEDE high-end clusters

Samuel Schlachter, Stephen Herbein, Michela Taufer, Shuching Ou, Sandeep Patel, Jeremy S. Logan

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

1 Citation (Scopus)

Abstract

Efficiently studying Sodium Dodecyl Sulfate (SDS) molecules' formations in the presence of different molar concentrations on high-end GPU clusters whose nodes share accelerators exposes us to several challenges, including the need to dynamically adapt the job lengths. Neither virtualization nor lightweight OS solutions can easily support generality, portability, and maintainability in concert. Our solution complements rather than rewrites existing workflow and resource managers with a companion module that complements functions of the workflow manager and a wrapper module that extends functions of the resource managers. Results on the Keeneland cluster show how, by using our modules, accelerated SDS simulations more efficiently use the cluster's GPUs while leading to relevant scientific observations.

Original languageEnglish (US)
Title of host publicationProceedings - IEEE 9th International Conference on e-Science, e-Science 2013
PublisherIEEE Computer Society
Pages116-123
Number of pages8
ISBN (Print)9780768550831
DOIs
StatePublished - 2013
Externally publishedYes
Event9th IEEE International Conference on e-Science, e-Science 2013 - Beijing, China
Duration: Oct 22 2013Oct 25 2013

Other

Other9th IEEE International Conference on e-Science, e-Science 2013
CountryChina
CityBeijing
Period10/22/1310/25/13

Fingerprint

Sodium dodecyl sulfate
Managers
Maintainability
Particle accelerators
Molecules
Graphics processing unit

ASJC Scopus subject areas

  • Information Systems

Cite this

Schlachter, S., Herbein, S., Taufer, M., Ou, S., Patel, S., & Logan, J. S. (2013). Efficient SDS simulations on multi-GPU nodes of XSEDE high-end clusters. In Proceedings - IEEE 9th International Conference on e-Science, e-Science 2013 (pp. 116-123). [6683898] IEEE Computer Society. https://doi.org/10.1109/eScience.2013.22

Efficient SDS simulations on multi-GPU nodes of XSEDE high-end clusters. / Schlachter, Samuel; Herbein, Stephen; Taufer, Michela; Ou, Shuching; Patel, Sandeep; Logan, Jeremy S.

Proceedings - IEEE 9th International Conference on e-Science, e-Science 2013. IEEE Computer Society, 2013. p. 116-123 6683898.

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

Schlachter, S, Herbein, S, Taufer, M, Ou, S, Patel, S & Logan, JS 2013, Efficient SDS simulations on multi-GPU nodes of XSEDE high-end clusters. in Proceedings - IEEE 9th International Conference on e-Science, e-Science 2013., 6683898, IEEE Computer Society, pp. 116-123, 9th IEEE International Conference on e-Science, e-Science 2013, Beijing, China, 10/22/13. https://doi.org/10.1109/eScience.2013.22
Schlachter S, Herbein S, Taufer M, Ou S, Patel S, Logan JS. Efficient SDS simulations on multi-GPU nodes of XSEDE high-end clusters. In Proceedings - IEEE 9th International Conference on e-Science, e-Science 2013. IEEE Computer Society. 2013. p. 116-123. 6683898 https://doi.org/10.1109/eScience.2013.22
Schlachter, Samuel ; Herbein, Stephen ; Taufer, Michela ; Ou, Shuching ; Patel, Sandeep ; Logan, Jeremy S. / Efficient SDS simulations on multi-GPU nodes of XSEDE high-end clusters. Proceedings - IEEE 9th International Conference on e-Science, e-Science 2013. IEEE Computer Society, 2013. pp. 116-123
@inproceedings{b5c5ba38ce7144878eb20bd20f7c14d0,
title = "Efficient SDS simulations on multi-GPU nodes of XSEDE high-end clusters",
abstract = "Efficiently studying Sodium Dodecyl Sulfate (SDS) molecules' formations in the presence of different molar concentrations on high-end GPU clusters whose nodes share accelerators exposes us to several challenges, including the need to dynamically adapt the job lengths. Neither virtualization nor lightweight OS solutions can easily support generality, portability, and maintainability in concert. Our solution complements rather than rewrites existing workflow and resource managers with a companion module that complements functions of the workflow manager and a wrapper module that extends functions of the resource managers. Results on the Keeneland cluster show how, by using our modules, accelerated SDS simulations more efficiently use the cluster's GPUs while leading to relevant scientific observations.",
author = "Samuel Schlachter and Stephen Herbein and Michela Taufer and Shuching Ou and Sandeep Patel and Logan, {Jeremy S.}",
year = "2013",
doi = "10.1109/eScience.2013.22",
language = "English (US)",
isbn = "9780768550831",
pages = "116--123",
booktitle = "Proceedings - IEEE 9th International Conference on e-Science, e-Science 2013",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - Efficient SDS simulations on multi-GPU nodes of XSEDE high-end clusters

AU - Schlachter, Samuel

AU - Herbein, Stephen

AU - Taufer, Michela

AU - Ou, Shuching

AU - Patel, Sandeep

AU - Logan, Jeremy S.

PY - 2013

Y1 - 2013

N2 - Efficiently studying Sodium Dodecyl Sulfate (SDS) molecules' formations in the presence of different molar concentrations on high-end GPU clusters whose nodes share accelerators exposes us to several challenges, including the need to dynamically adapt the job lengths. Neither virtualization nor lightweight OS solutions can easily support generality, portability, and maintainability in concert. Our solution complements rather than rewrites existing workflow and resource managers with a companion module that complements functions of the workflow manager and a wrapper module that extends functions of the resource managers. Results on the Keeneland cluster show how, by using our modules, accelerated SDS simulations more efficiently use the cluster's GPUs while leading to relevant scientific observations.

AB - Efficiently studying Sodium Dodecyl Sulfate (SDS) molecules' formations in the presence of different molar concentrations on high-end GPU clusters whose nodes share accelerators exposes us to several challenges, including the need to dynamically adapt the job lengths. Neither virtualization nor lightweight OS solutions can easily support generality, portability, and maintainability in concert. Our solution complements rather than rewrites existing workflow and resource managers with a companion module that complements functions of the workflow manager and a wrapper module that extends functions of the resource managers. Results on the Keeneland cluster show how, by using our modules, accelerated SDS simulations more efficiently use the cluster's GPUs while leading to relevant scientific observations.

UR - http://www.scopus.com/inward/record.url?scp=84893504408&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84893504408&partnerID=8YFLogxK

U2 - 10.1109/eScience.2013.22

DO - 10.1109/eScience.2013.22

M3 - Conference contribution

SN - 9780768550831

SP - 116

EP - 123

BT - Proceedings - IEEE 9th International Conference on e-Science, e-Science 2013

PB - IEEE Computer Society

ER -