Visualizing, Measuring, and Tuning Adaptive MPI Parameters

Matthias Diener, Sam White, Laxmikant V. Kale

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Adaptive MPI (AMPI) is an advanced MPI runtime environment that offers several features over traditional MPI runtimes, which can lead to a better utilization of the underlying hardware platform and therefore higher performance. These features are overdecomposition through virtualization, and load balancing via rank migration. Choosing which of these features to use, and finding the optimal parameters for them is a challenging task however, since different applications and systems may require different options. Furthermore, there is a lack of information about the impact of each option. In this paper, we present a new visualization of AMPI in its companion Projections tool, which depicts the operation of an MPI application and details the impact of the different AMPI features on its resource usage. We show how these visualizations can help to improve the efficiency and execution time of an MPI application. Applying optimizations indicated by the performance analysis to two MPI-based applications results in performance improvements of up 18% from overdecomposition and load balancing.

Original languageEnglish (US)
Title of host publicationProgramming and Performance Visualization Tools - International Workshops, ESPT 2017 and VPA 2017, Revised Selected Papers
EditorsAbhinav Bhatele, David Boehme, Joshua A. Levine, Allen D. Malony, Martin Schulz
PublisherSpringer
Pages219-230
Number of pages12
ISBN (Print)9783030178710
DOIs
StatePublished - 2019
Event6th Workshop on Extreme-Scale Programming Tools, ESPT 2017 and 4th International Workshop on Visual Performance Analysis, VPA 2017 and Workshop on Extreme-Scale Programming Tools, ESPT 2018 and 5th International Workshop on Visual Performance Analysis, VPA 2018 - Dallas, United States
Duration: Nov 11 2018Nov 16 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11027 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference6th Workshop on Extreme-Scale Programming Tools, ESPT 2017 and 4th International Workshop on Visual Performance Analysis, VPA 2017 and Workshop on Extreme-Scale Programming Tools, ESPT 2018 and 5th International Workshop on Visual Performance Analysis, VPA 2018
Country/TerritoryUnited States
CityDallas
Period11/11/1811/16/18

Keywords

  • AMPI
  • Load balancing
  • MPI
  • Migration
  • Overdecomposition

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Visualizing, Measuring, and Tuning Adaptive MPI Parameters'. Together they form a unique fingerprint.

Cite this