OpenMP application experiences: Porting to accelerated nodes

被引：21

作者：

Bak, Seonmyeong ^{[3
]}

Bertoni, Colleen ^{[1
]}

Boehm, Swen ^{[6
]}

Budiardja, Reuben ^{[6
]}

Chapman, Barbara M. ^{[2
]}

Doerfert, Johannes ^{[1
]}

Eisenbach, Markus ^{[6
]}

Finkel, Hal ^{[1
]}

Hernandez, Oscar ^{[6
]}

Huber, Joseph ^{[6
]}

Iwasaki, Shintaro ^{[1
]}

Kale, Vivek ^{[2
]}

Kent, Paul R. C. ^{[6
]}

Kwack, JaeHyuk ^{[1
]}

Lin, Meifeng ^{[2
]}

Luszczek, Piotr ^{[8
]}

Luo, Ye ^{[1
]}

Pham, Buu ^{[4
]}

Pophale, Swaroop ^{[6
]}

Ravikumar, Kiran ^{[3
]}

Sarkar, Vivek ^{[3
]}

Scogland, Thomas ^{[5
]}

Tian, Shilei ^{[7
]}

Yeung, P. K. ^{[3
]}

机构：

[1] Argonne Natl Lab, Lemont, IL 60439 USA

[2] Brookhaven Natl Lab, Upton, NY 11973 USA

[3] Georgia Inst Technol, Atlanta, GA 30332 USA

[4] Iowa State Univ, Ames, IA 50011 USA

[5] Lawrence Livermore Natl Lab, Livermore, CA 94550 USA

[6] Oak Ridge Natl Lab, One Bethel Valley Rd, Oak Ridge, TN 37830 USA

[7] SUNY Stony Brook, Stony Brook, NY 11794 USA

[8] Univ Tennessee, Knoxville, TN 37996 USA

来源：

PARALLEL COMPUTING | 2022年 / 109卷

基金：

美国国家科学基金会;

关键词：

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

D O I：

10.1016/j.parco.2021.102856

中图分类号：

TP301 [理论、方法];

学科分类号：

081202 ;

摘要：

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 leadershipclass 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.

引用

页数：12

共 34 条

[1]

[Anonymous], 2020, LLVM COMPILER INFRAS

[2]

[Anonymous], 2021, OPENMP 5 0 REFERENCE

[3] RAJA: Portable Performance for Large-Scale Scientific Applications [J].