Automatic generation of benchmarks for I/O-intensive parallel applications

Meng Hao, Weizhe Zhang, You Zhang, Marc Snir, Laurence T. Yang

Research output: Contribution to journalArticle

Abstract

The benchmarks of I/O-intensive parallel applications are important for evaluating and optimizing HPC softwares and hardwares. However, extracting high-fidelity benchmarks which can fully reflect computation, communication, and I/O behaviors of original I/O-intensive parallel applications is very difficult. This work contributes a framework which can automatically generate benchmarks for I/O-intensive parallel applications. We demonstrate our framework on Taub and TianHe-2 supercomputers with five NAS Parallel Benchmarks (NPB) and four I/O-intensive parallel applications. The results show that our trace merging algorithm and trace compressing algorithm are better than others, and the generated benchmarks can accurately mimic the computation, communication, and I/O behaviors of original I/O-intensive parallel applications. Also, these generated benchmarks can be used to predict the performance of original applications, while reducing the prediction overhead by scaling down the execution time of benchmark proportionally.

LanguageEnglish (US)
Pages1-13
Number of pages13
JournalJournal of Parallel and Distributed Computing
Volume124
DOIs
StatePublished - Feb 1 2019

Fingerprint

Parallel Applications
Benchmark
Trace
Supercomputers
Communication
Merging
Supercomputer
Fidelity
Execution Time
Hardware
Scaling
Predict
Software
Prediction
Demonstrate

Keywords

  • Benchmark generation
  • I/O-intensive applications
  • Performance prediction
  • Trace compressing
  • Trace merging

ASJC Scopus subject areas

  • Software
  • Theoretical Computer Science
  • Hardware and Architecture
  • Computer Networks and Communications
  • Artificial Intelligence

Cite this

Automatic generation of benchmarks for I/O-intensive parallel applications. / Hao, Meng; Zhang, Weizhe; Zhang, You; Snir, Marc; Yang, Laurence T.

In: Journal of Parallel and Distributed Computing, Vol. 124, 01.02.2019, p. 1-13.

Research output: Contribution to journalArticle

Hao, Meng ; Zhang, Weizhe ; Zhang, You ; Snir, Marc ; Yang, Laurence T. / Automatic generation of benchmarks for I/O-intensive parallel applications. In: Journal of Parallel and Distributed Computing. 2019 ; Vol. 124. pp. 1-13.
@article{850ea608f1084bb38ca70127241c3296,
title = "Automatic generation of benchmarks for I/O-intensive parallel applications",
abstract = "The benchmarks of I/O-intensive parallel applications are important for evaluating and optimizing HPC softwares and hardwares. However, extracting high-fidelity benchmarks which can fully reflect computation, communication, and I/O behaviors of original I/O-intensive parallel applications is very difficult. This work contributes a framework which can automatically generate benchmarks for I/O-intensive parallel applications. We demonstrate our framework on Taub and TianHe-2 supercomputers with five NAS Parallel Benchmarks (NPB) and four I/O-intensive parallel applications. The results show that our trace merging algorithm and trace compressing algorithm are better than others, and the generated benchmarks can accurately mimic the computation, communication, and I/O behaviors of original I/O-intensive parallel applications. Also, these generated benchmarks can be used to predict the performance of original applications, while reducing the prediction overhead by scaling down the execution time of benchmark proportionally.",
keywords = "Benchmark generation, I/O-intensive applications, Performance prediction, Trace compressing, Trace merging",
author = "Meng Hao and Weizhe Zhang and You Zhang and Marc Snir and Yang, {Laurence T.}",
year = "2019",
month = "2",
day = "1",
doi = "10.1016/j.jpdc.2018.10.004",
language = "English (US)",
volume = "124",
pages = "1--13",
journal = "Journal of Parallel and Distributed Computing",
issn = "0743-7315",
publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - Automatic generation of benchmarks for I/O-intensive parallel applications

AU - Hao, Meng

AU - Zhang, Weizhe

AU - Zhang, You

AU - Snir, Marc

AU - Yang, Laurence T.

PY - 2019/2/1

Y1 - 2019/2/1

N2 - The benchmarks of I/O-intensive parallel applications are important for evaluating and optimizing HPC softwares and hardwares. However, extracting high-fidelity benchmarks which can fully reflect computation, communication, and I/O behaviors of original I/O-intensive parallel applications is very difficult. This work contributes a framework which can automatically generate benchmarks for I/O-intensive parallel applications. We demonstrate our framework on Taub and TianHe-2 supercomputers with five NAS Parallel Benchmarks (NPB) and four I/O-intensive parallel applications. The results show that our trace merging algorithm and trace compressing algorithm are better than others, and the generated benchmarks can accurately mimic the computation, communication, and I/O behaviors of original I/O-intensive parallel applications. Also, these generated benchmarks can be used to predict the performance of original applications, while reducing the prediction overhead by scaling down the execution time of benchmark proportionally.

AB - The benchmarks of I/O-intensive parallel applications are important for evaluating and optimizing HPC softwares and hardwares. However, extracting high-fidelity benchmarks which can fully reflect computation, communication, and I/O behaviors of original I/O-intensive parallel applications is very difficult. This work contributes a framework which can automatically generate benchmarks for I/O-intensive parallel applications. We demonstrate our framework on Taub and TianHe-2 supercomputers with five NAS Parallel Benchmarks (NPB) and four I/O-intensive parallel applications. The results show that our trace merging algorithm and trace compressing algorithm are better than others, and the generated benchmarks can accurately mimic the computation, communication, and I/O behaviors of original I/O-intensive parallel applications. Also, these generated benchmarks can be used to predict the performance of original applications, while reducing the prediction overhead by scaling down the execution time of benchmark proportionally.

KW - Benchmark generation

KW - I/O-intensive applications

KW - Performance prediction

KW - Trace compressing

KW - Trace merging

UR - http://www.scopus.com/inward/record.url?scp=85055754292&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055754292&partnerID=8YFLogxK

U2 - 10.1016/j.jpdc.2018.10.004

DO - 10.1016/j.jpdc.2018.10.004

M3 - Article

VL - 124

SP - 1

EP - 13

JO - Journal of Parallel and Distributed Computing

T2 - Journal of Parallel and Distributed Computing

JF - Journal of Parallel and Distributed Computing

SN - 0743-7315

ER -