Scalable approaches for supporting MPI-IO atomicity

Peter M. Aarestad*, Ching Avery, George K. Thiruvathukal, Alok N. Choudhary

*Corresponding author for this work

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

5 Scopus citations

Abstract

Scalable atomic and parallel access to noncontiguous regions of a file is essential to exploit high performance I/O as required by large-scale applications. Parallel I/O frameworks such as MPI I/O conceptually allow I/O to be defined on regions of a file using derived datatypes. Access to regions of a file can be automatically computed on a perprocessor basis using the datatype, resulting in a list of (offset, length) pairs. We describe three approaches for implementing lock serving (whole file, region locking, and byterange locking) and compare the various approaches using three noncontiguous I/O benchmarks. We present the details of the lock server architecture and describe the implementation of a fully-functional prototype that makes use of a lightweight message passing library and red/black trees.

Original languageEnglish (US)
Title of host publicationSixth IEEE International Symposium on Cluster Computing and the Grid
Subtitle of host publicationSpanning the World and Beyond, 2006. CCGRID 06
Pages35-42
Number of pages8
DOIs
StatePublished - 2006
Event6th IEEE International Symposium on Cluster Computing and the Grid, 2006. CCGRID 06 - , Singapore
Duration: May 16 2006May 19 2006

Publication series

NameSixth IEEE International Symposium on Cluster Computing and the Grid, 2006. CCGRID 06

Other

Other6th IEEE International Symposium on Cluster Computing and the Grid, 2006. CCGRID 06
CountrySingapore
Period5/16/065/19/06

ASJC Scopus subject areas

  • Engineering(all)

Fingerprint Dive into the research topics of 'Scalable approaches for supporting MPI-IO atomicity'. Together they form a unique fingerprint.

Cite this