SAILLARD, Emmanuelle; CARRIBAULT, Patrick; BARTHOU, Denis

doi:10.1177/1094342014552204

Métadonnées

Afficher la notice complète

Licence d’utilisation du document

SAILLARD, Emmanuelle
DAM Île-de-France [DAM/DIF]

CARRIBAULT, Patrick
DAM Île-de-France [DAM/DIF]

BARTHOU, Denis
Efficient runtime systems for parallel architectures [RUNTIME]
Laboratoire Bordelais de Recherche en Informatique [LaBRI]

Langue

Article de revue

Ce document a été publié dans

International Journal of High Performance Computing Applications. 2014p. 10.1177/1094342014552204

SAGE Publications

Résumé en anglais

Nowadays most scientific applications are parallelized based on MPI communications. Collective MPI communications have to be executed in the same order by all processes in their communicator and the same number of times, otherwise it is not conforming to the standard and a deadlock or other undefined behavior can occur. As soon as the control-flow involving these collective operations becomes more complex, in particular including conditionals on process ranks, ensuring the correction of such code is error-prone. We propose in this paper a static analysis to detect when such situation occurs, combined with a code transformation that prevents from deadlocking. We focus on blocking MPI collective operations in SPMD applications, assuming MPI calls are not nested in multithreaded regions. We show on several benchmarks the small impact on performance and the ease of integration of our techniques in the development process.< Réduire

Mots clés en anglais

MPI

debugging

collective

static analysis

correctness

Métadonnées

Partager cette publication !

Licence d’utilisation du document

PARCOACH: Combining static and dynamic validation of MPI collective communications

Langue

Ce document a été publié dans

Résumé en anglais

Mots clés en anglais

URI

DOI

Origine

Unités de recherche