MPI Runtime Error Detection with MUST: Advanced Error Reports

Joachim Protze (), Tobias Hilbrich, Bronis R. de Supinski, Martin Schulz, Matthias S. Müller and Wolfgang E. Nagel
Additional contact information
Joachim Protze: Technische Universität Dresden, Center for Information Services and High Performance Computing (ZIH)
Tobias Hilbrich: Technische Universität Dresden, Center for Information Services and High Performance Computing (ZIH)
Bronis R. de Supinski: Lawrence Livermore National Laboratory
Martin Schulz: Lawrence Livermore National Laboratory
Matthias S. Müller: Technische Universität Dresden, Center for Information Services and High Performance Computing (ZIH)
Wolfgang E. Nagel: Technische Universität Dresden, Center for Information Services and High Performance Computing (ZIH)

A chapter in Tools for High Performance Computing 2012, 2013, pp 25-38 from Springer

Abstract: Abstract The Message Passing Interface (MPI) is a widely used paradigm for distributed memory programming. Its API is primarily designed for good performance and less for usability; it provides only very limited abstractions that help enforce its correct use. As a result, application developers need tools that aid in the detection and removal of MPI usage errors. Our runtime error detection tool MUST addresses this issue and provides a wide range of automatic correctness checks. MUST uses state-of-the-art approaches to cope with complex MPI semantics like derived datatypes, collective operations, and wildcard receive operations. However, equally important to detecting correctness violations, is that such correctness tools present all details of the violating MPI call(s) required to pinpoint the problem in the source code and to remove the error. In this paper we focus on the error reports presented by MUST and propose a new set of error reports that present complex errors with fine-grained details of the error situation. This includes a deadlock view and a view for usage errors in complex MPI datatypes.

Keywords: Runtime Error Detection; Message Passing Interface (MPI); Error Reports; Datatype; Correctness Checking (search for similar items in EconPapers)
Date: 2013
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:sprchp:978-3-642-37349-7_3

Ordering information: This item can be ordered from
http://www.springer.com/9783642373497

DOI: 10.1007/978-3-642-37349-7_3

Access Statistics for this chapter

More chapters in Springer Books from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().