Non-data-communication overheads in MPI: Analysis on blue Gene/P

Pavan Balaji, Anthony Chan, William Gropp, Rajeev Thakur, Ewing Lusk

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


Modern HEC systems, such as Blue Gene/P, rely on achieving high-performance by using the parallelism of a massive number of low-frequency/low-power processing cores. This means that the local pre- and post-communication processing required by the MPI stack might not be very fast, owing to the slow processing cores. Similarly, small amounts of serialization within the MPI stack that were acceptable on small/medium systems can be brutal on massively parallel systems. In this paper, we study different non-data-communication overheads within the MPI implementation on the IBM Blue Gene/P system.

Original languageEnglish (US)
Title of host publicationRecent Advances in Parallel Virtual Machine and Message Passing Interface - 15th European PVM/MPI Users' Group Meeting, Proceedings
Number of pages10
StatePublished - 2008
Event15th European PVM/MPI Users' Group Meeting, EuroPVM/MPI 2008 - Dublin, Ireland
Duration: Sep 7 2008Sep 10 2008

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5205 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Other15th European PVM/MPI Users' Group Meeting, EuroPVM/MPI 2008

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science


Dive into the research topics of 'Non-data-communication overheads in MPI: Analysis on blue Gene/P'. Together they form a unique fingerprint.

Cite this