Process arrival pattern aware algorithms for acceleration of scatter and gather operations

Author:

Proficz JerzyORCID

Abstract

AbstractImbalanced process arrival patterns (PAPs) are ubiquitous in many parallel and distributed systems, especially in HPC ones. The collective operations, e.g. in MPI, are designed for equal process arrival times, and are not optimized for deviations in their appearance. We propose eight new PAP-aware algorithms for the scatter and gather operations. They are binomial or linear tree adaptations introducing additional process ordering and (in some cases) additional activities in a special background thread. The solution was implemented using one of the most popular open source MPI compliant library (OpenMPI), and evaluated in a typical HPC environment using a specially developed benchmark as well as a real application: FFT. The experimental results show a significant advantage of the proposed approach over the default OpenMPI implementation, showing good scalability and high performance with the FFT acceleration for the communication run time: 16.7% and for the total application execution time: 3.3%.

Publisher

Springer Science and Business Media LLC

Subject

Computer Networks and Communications,Software

Reference29 articles.

1. Arap, O., Swany, M., Brown, G., Himebaugh, B.: Adaptive recursive doubling algorithm for collective communication. In: 2015 IEEE International Parallel and Distributed Processing Symposium Workshop, pp. 121–128. IEEE (2015)

2. Bailey, D.H.: NAS parallel benchmarks. In: Padua, D. (ed.) Encyclopedia of Parallel Computing, pp. 1254–1259. Springer, Boston (2011)

3. Balducci, M., Choudary, A., Hamaker, J.: Comparative analysis of FFT algorithms in sequential and parallel form. In: Mississippi State University Conference on Digital Signal Processing, pp. 5–16 (1996)

4. Butenhof, D.R.: Programming with POSIX Threads. Addison-Wesley Professional, Boston (1997)

5. Czarnul, P., Kuchta, J., Matuszek, M., Proficz, J., Rościszewski, P., Wójcik, M., Szymański, J.: MERPSYS: an environment for simulation of parallel application execution on large scale HPC systems. Simul. Model. Pract. Theory 77, 124–140 (2017)

Cited by 2 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. All-gather Algorithms Resilient to Imbalanced Process Arrival Patterns;ACM Transactions on Architecture and Code Optimization;2021-12-31

2. Improving Clairvoyant: reduction algorithm resilient to imbalanced process arrival patterns;The Journal of Supercomputing;2020-11-20

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3