OpenMP application experiences: Porting to accelerated nodes

Seonmyeong Bak; Colleen Bertoni; Swen Boehm; Reuben Budiardja; Barbara M. Chapman; Johannes Doerfert; Markus Eisenbach; Hal Finkel; Oscar Hernandez; Joseph Huber; Shintaro Iwasaki; Vivek Kale; Paul R.C. Kent; Jae Hyuk Kwack; Meifeng Lin; Piotr Luszczek; Ye Luo; Buu Pham; Swaroop Pophale; Kiran Ravikumar; Vivek Sarkar; Thomas Scogland; Shilei Tian; P. K. Yeung

doi:10.1016/j.parco.2021.102856

OpenMP application experiences: Porting to accelerated nodes

Seonmyeong Bak, Colleen Bertoni, Swen Boehm, Reuben Budiardja, Barbara M. Chapman, Johannes Doerfert, Markus Eisenbach, Hal Finkel, Oscar Hernandez, Joseph Huber, Shintaro Iwasaki, Vivek Kale, Paul R.C. Kent, Jae Hyuk Kwack, Meifeng Lin, Piotr Luszczek, Ye Luo, Buu Pham, Swaroop Pophale, Kiran RavikumarVivek Sarkar, Thomas Scogland, Shilei Tian, P. K. Yeung

Research output: Contribution to journal › Article › peer-review

26 Scopus citations

Abstract

As recent enhancements to the OpenMP specification become available in its implementations, there is a need to share the results of experimentation in order to better understand the OpenMP implementation's behavior in practice, to identify pitfalls, and to learn how the implementations can be effectively deployed in scientific codes. We report on experiences gained and practices adopted when using OpenMP to port a variety of ECP applications, mini-apps and libraries based on different computational motifs to accelerator-based leadership-class high-performance supercomputer systems at the United States Department of Energy. Additionally, we identify important challenges and open problems related to the deployment of OpenMP. Through our report of experiences, we find that OpenMP implementations are successful on current supercomputing platforms and that OpenMP is a promising programming model to use for applications to be run on emerging and future platforms with accelerated nodes.

Original language	English
Article number	102856
Journal	Parallel Computing
Volume	109
DOIs	https://doi.org/10.1016/j.parco.2021.102856
State	Published - Mar 2022

Funding

This work was funded in part by the Exascale Computing Project (17-SC-20-SC), a collaborative effort of the U.S. Department of Energy Office of Science and the National Nuclear Security Administration, in particular its subproject on Scaling OpenMP with LLVM for Exascale performance and portability (SOLLVE). This manuscript has been authored by UT-Battelle, LLC under Contract No. DE-AC05-00OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan). The development of some of the numerical software libraries tested in this work was supported by the National Science Foundation under OAC grant No. 2004541. This work was supported in part by the Argonne Leadership Computing Facility, which is a DOE Office of Science User Facility supported under Contract DE-AC02-06CH11357. This work was funded in part by the Exascale Computing Project ( 17-SC-20-SC ), a collaborative effort of the U.S. Department of Energy Office of Science and the National Nuclear Security Administration, in particular its subproject on Scaling OpenMP with LLVM for Exascale performance and portability (SOLLVE). This manuscript has been authored by UT-Battelle, LLC under Contract No. DE-AC05-00OR22725 with the U.S. Department of Energy . The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan ( http://energy.gov/downloads/doe-public-access-plan ). The development of some of the numerical software libraries tested in this work was supported by the National Science Foundation under OAC grant No. 2004541 . This work was supported in part by the Argonne Leadership Computing Facility, which is a DOE Office of Science User Facility supported under Contract DE-AC02-06CH11357.

Funders	Funder number
DOE Public Access Plan
United States Government
National Science Foundation
U.S. Department of Energy
Office of Advanced Cyberinfrastructure	2004541
Office of Science	DE-AC02-06CH11357
National Nuclear Security Administration	DE-AC05-00OR22725

Keywords

Accelerators
Application porting experiences
GAMESS
GESTS
GenASiS
GridQCD
High performance computing
LSMS
OpenMP implementations
QMCPACK
RAJA
SLATE

Access to Document

10.1016/j.parco.2021.102856

Cite this

Bak, S., Bertoni, C., Boehm, S., Budiardja, R., Chapman, B. M., Doerfert, J., Eisenbach, M., Finkel, H., Hernandez, O., Huber, J., Iwasaki, S., Kale, V., Kent, P. R. C., Kwack, J. H., Lin, M., Luszczek, P., Luo, Y., Pham, B., Pophale, S., ... Yeung, P. K. (2022). OpenMP application experiences: Porting to accelerated nodes. Parallel Computing, 109, Article 102856. https://doi.org/10.1016/j.parco.2021.102856

@article{daae01661e1f44fa981a0ac0868a6e82,

title = "OpenMP application experiences: Porting to accelerated nodes",

abstract = "As recent enhancements to the OpenMP specification become available in its implementations, there is a need to share the results of experimentation in order to better understand the OpenMP implementation's behavior in practice, to identify pitfalls, and to learn how the implementations can be effectively deployed in scientific codes. We report on experiences gained and practices adopted when using OpenMP to port a variety of ECP applications, mini-apps and libraries based on different computational motifs to accelerator-based leadership-class high-performance supercomputer systems at the United States Department of Energy. Additionally, we identify important challenges and open problems related to the deployment of OpenMP. Through our report of experiences, we find that OpenMP implementations are successful on current supercomputing platforms and that OpenMP is a promising programming model to use for applications to be run on emerging and future platforms with accelerated nodes.",

keywords = "Accelerators, Application porting experiences, GAMESS, GESTS, GenASiS, GridQCD, High performance computing, LSMS, OpenMP implementations, QMCPACK, RAJA, SLATE",

author = "Seonmyeong Bak and Colleen Bertoni and Swen Boehm and Reuben Budiardja and Chapman, {Barbara M.} and Johannes Doerfert and Markus Eisenbach and Hal Finkel and Oscar Hernandez and Joseph Huber and Shintaro Iwasaki and Vivek Kale and Kent, {Paul R.C.} and Kwack, {Jae Hyuk} and Meifeng Lin and Piotr Luszczek and Ye Luo and Buu Pham and Swaroop Pophale and Kiran Ravikumar and Vivek Sarkar and Thomas Scogland and Shilei Tian and Yeung, {P. K.}",

note = "Publisher Copyright: {\textcopyright} 2021 Elsevier B.V.",

year = "2022",

month = mar,

doi = "10.1016/j.parco.2021.102856",

language = "English",

volume = "109",

journal = "Parallel Computing",

issn = "0167-8191",

publisher = "Elsevier",

}

Bak, S, Bertoni, C, Boehm, S , Budiardja, R, Chapman, BM, Doerfert, J, Eisenbach, M, Finkel, H, Hernandez, O, Huber, J, Iwasaki, S, Kale, V, Kent, PRC, Kwack, JH, Lin, M, Luszczek, P, Luo, Y, Pham, B, Pophale, S, Ravikumar, K, Sarkar, V, Scogland, T, Tian, S & Yeung, PK 2022, 'OpenMP application experiences: Porting to accelerated nodes', Parallel Computing, vol. 109, 102856. https://doi.org/10.1016/j.parco.2021.102856

TY - JOUR

T1 - OpenMP application experiences

T2 - Porting to accelerated nodes

AU - Bak, Seonmyeong

AU - Bertoni, Colleen

AU - Boehm, Swen

AU - Budiardja, Reuben

AU - Chapman, Barbara M.

AU - Doerfert, Johannes

AU - Eisenbach, Markus

AU - Finkel, Hal

AU - Hernandez, Oscar

AU - Huber, Joseph

AU - Iwasaki, Shintaro

AU - Kale, Vivek

AU - Kent, Paul R.C.

AU - Kwack, Jae Hyuk

AU - Lin, Meifeng

AU - Luszczek, Piotr

AU - Luo, Ye

AU - Pham, Buu

AU - Pophale, Swaroop

AU - Ravikumar, Kiran

AU - Sarkar, Vivek

AU - Scogland, Thomas

AU - Tian, Shilei

AU - Yeung, P. K.

PY - 2022/3

Y1 - 2022/3

N2 - As recent enhancements to the OpenMP specification become available in its implementations, there is a need to share the results of experimentation in order to better understand the OpenMP implementation's behavior in practice, to identify pitfalls, and to learn how the implementations can be effectively deployed in scientific codes. We report on experiences gained and practices adopted when using OpenMP to port a variety of ECP applications, mini-apps and libraries based on different computational motifs to accelerator-based leadership-class high-performance supercomputer systems at the United States Department of Energy. Additionally, we identify important challenges and open problems related to the deployment of OpenMP. Through our report of experiences, we find that OpenMP implementations are successful on current supercomputing platforms and that OpenMP is a promising programming model to use for applications to be run on emerging and future platforms with accelerated nodes.

AB - As recent enhancements to the OpenMP specification become available in its implementations, there is a need to share the results of experimentation in order to better understand the OpenMP implementation's behavior in practice, to identify pitfalls, and to learn how the implementations can be effectively deployed in scientific codes. We report on experiences gained and practices adopted when using OpenMP to port a variety of ECP applications, mini-apps and libraries based on different computational motifs to accelerator-based leadership-class high-performance supercomputer systems at the United States Department of Energy. Additionally, we identify important challenges and open problems related to the deployment of OpenMP. Through our report of experiences, we find that OpenMP implementations are successful on current supercomputing platforms and that OpenMP is a promising programming model to use for applications to be run on emerging and future platforms with accelerated nodes.

KW - Accelerators

KW - Application porting experiences

KW - GAMESS

KW - GESTS

KW - GenASiS

KW - GridQCD

KW - High performance computing

KW - LSMS

KW - OpenMP implementations

KW - QMCPACK

KW - RAJA

KW - SLATE

UR - http://www.scopus.com/inward/record.url?scp=85118705309&partnerID=8YFLogxK

U2 - 10.1016/j.parco.2021.102856

DO - 10.1016/j.parco.2021.102856

M3 - Article

AN - SCOPUS:85118705309

SN - 0167-8191

VL - 109

JO - Parallel Computing

JF - Parallel Computing

M1 - 102856

ER -

OpenMP application experiences: Porting to accelerated nodes

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this