An adaptive, scalable, and portable technique for speeding up MPI-based applications

Rosa Filgueira*, Malcolm Atkinson, Alberto Nuñez, Javier Fernández

*Corresponding author for this work

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

6 Citations (Scopus)


This paper presents a portable optimization for MPI communications, called PRAcTICaL-MPI (Portable Adaptive Compression Library- MPI). PRAcTICaL-MPI reduces the data volume exchanged among processes by using lossless compression and offers two main advantages. Firstly, it is independent of the MPI implementation and the application used. Secondly, it allows for turning the compression on and off and selecting the most appropriate compression algorithm at run-time, depending on the characteristics of each message and on network performance. We have validated PRAcTICaL-MPI in different MPI implementations and HPC clusters. The evaluation shows that compressing MPI messages with the best algorithm and only when it is worthwhile, we obtain a great reduction in the overall execution time for many of the scenarios considered.

Original languageEnglish
Title of host publicationEuro-Par 2012 Parallel Processing. Euro-Par 2012
EditorsC. Kaklamanis, T. Papatheodorou, P. G. Spirakis
Number of pages12
ISBN (Electronic)9783642328206
ISBN (Print)9783642328190
Publication statusPublished - 2012
Event18th International Conference on Parallel Processing 2012 - Rhodes Island, Greece
Duration: 27 Aug 201231 Aug 2012

Publication series

NameLecture Notes in Computer Science
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference18th International Conference on Parallel Processing 2012
Abbreviated titleEuro-Par 2012
CityRhodes Island


  • Adaptive systems
  • Compression algorithms
  • High-Performance Computing
  • MPI Library
  • Parallel techniques
  • Portable optimizations

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science


Dive into the research topics of 'An adaptive, scalable, and portable technique for speeding up MPI-based applications'. Together they form a unique fingerprint.

Cite this