A continuous benchmarking infrastructure for high-performance computing applications

被引：0

作者：

Alt, Christoph ^{[1
,2
]}

Lanser, Martin ^{[3
,4
]}

Plewinski, Jonas ^{[1
]}

Janki, Atin ^{[6
]}

Klawonn, Axel ^{[3
,4
]}

Koestler, Harald ^{[1
,5
]}

Selzer, Michael ^{[7
]}

Ruede, Ulrich ^{[1
,8
]}

机构：

[1] Friedrich Alexander Univ Erlangen Nurnberg, Chair Comp Sci Syst Simulat 10, Cauer Str 11, D-91058 Erlangen, Germany

[2] Paderborn Univ, Paderborn Ctr Parallel Comp, Warburger Str 100, D-33098 Paderborn, Germany

[3] Univ Cologne, Dept Math & Comp Sci, Cologne, Germany

[4] Univ Cologne, Ctr Data & Simulat Sci, Cologne, Germany

[5] Erlangen Natl High Performance Comp Ctr NHR FAU, Erlangen, Germany

[6] Karlsruhe Inst Technol KIT, Inst Appl Mat IAM, Karlsruhe, Germany

[7] Karlsruhe Inst Technol KIT, Inst Nanotechnol INT, Eggenstein Leopoldshafen, Germany

[8] CERFACS, Toulouse, France

来源：

INTERNATIONAL JOURNAL OF PARALLEL EMERGENT AND DISTRIBUTED SYSTEMS | 2024年 / 39卷 / 04期

关键词：

Continuous integration; continuous Benchmarking; finite elements method; computational homogenization; lattice Boltzmann method; FETI-DP;

D O I：

10.1080/17445760.2024.2360190

中图分类号：

TP301 [理论、方法];

学科分类号：

081202 ;

摘要：

For scientific software, especially those used for large-scale simulations, achieving good performance and efficiently using the available hardware resources is essential. It is important to regularly perform benchmarks to ensure the efficient use of hardware and software when systems are changing and the software evolves. However, this can become quickly very tedious when many options for parameters, solvers, and hardware architectures are available. We present a continuous benchmarking strategy that automates benchmarking new code changes on high-performance computing clusters. This makes it possible to track how each code change affects the performance and how it evolves.

引用

页码：501 / 523

页数：23

共 55 条

[1]

[Anonymous], 2024, GIT ACCESSED

[2]

[Anonymous], DE RSEORG RES SOFTWA

[3]

[Anonymous], MUMPS : a parallel sparse direct solver

[4]

[Anonymous], googlecloudplatform

[5]

Anzt H., 2021, F1000RESEARCH, V9, P295, DOI [10.12688/f1000research, DOI 10.12688/F1000RESEARCH]

[6] GINKGO: A Modern Linear Operator Algebra Framework for High Performance Computing [J].

Anzt, Hartwig ;

Cojean, Terry ;

Flegar, Goran ;

Gobel, Fritz ;

Grutzmacher, Thomas ;

Nayak, Pratik ;

Ribizel, Tobias ;

Tsai, Yuhsiang Mike ;

Quintana-Orti, Enrique S. .

ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE, 2022, 48 (01)

[7] Towards Continuous Benchmarking: An Automated Performance Evaluation Framework for High Performance Software [J].

Anzt, Hartwig ;

Chen, Yen-Chen ;

Cojean, Terry ;

Dongarra, Jack ;

Flegar, Goran ;

Nayak, Pratik ;

Quintana-Orti, Enrique S. ;

Tsai, Yuhsiang M. ;

Wang, Weichung .

PROCEEDINGS OF THE PLATFORM FOR ADVANCED SCIENTIFIC COMPUTING CONFERENCE (PASC '19), 2019,

[8]

Bader A., 2017, Datenbanksysteme fur Business, Technologie und Web (BTW 2017)-Workshopband

[9] xSDK foundations: Toward an extreme-scale scientific software development kit [J].

Bartlett R. ;

Demeshko I. ;

Gamblin T. ;

Hammond G. ;

Heroux M. ;

Johnson J. ;

Klinvex A. ;

Li X. ;

McInnes L.C. ;

Moulton J.D. ;

Osei-Kuffuor D. ;

Sarich J. ;

Smith B. ;

Willenbring J. ;

Yang U.M. .

Supercomputing Frontiers and Innovations, 2017, 4 (01) :69-82

[10] lbmpy: Automatic code generation for efficient parallel lattice Boltzmann methods [J].

Bauer, Martin ;

Koestler, Harald ;

Ruede, Ulrich .

JOURNAL OF COMPUTATIONAL SCIENCE, 2021, 49

← 1 2 3 4 5 6 →