Reproducible Parallel Simulations in HPC - LIRMM - Laboratoire d’Informatique, de Robotique et de Microélectronique de Montpellier Accéder directement au contenu
Communication Dans Un Congrès Année : 2017

Reproducible Parallel Simulations in HPC

Résumé

Post Moore's era supercomputing will certainly require more hierarchical parallelism and variable precision floating-point arithmetic to satisfy the computing need of exascale numerical simulations. Nevertheless floating-point addition will remain non asso-ciative and so parallel computations will still be prone to return results being different from one run to another one. These failures of the numerical reproducibility reduce the simulation reliability and complicate the debugging and the validating steps of large scale software. We present two case studies to illustrate how to recover this numerical reproducibility without jeopardizing the computing efficiency. Hydrodynamics parallel simulations with the openTelemac code rely on finite element modelization, subdomain decomposition and iterative solvers. Two openTelemac modules have been modified to provide reproducible results for any number of computing units thanks to targeted compensation techniques. We also describe and analyze generic solutions that are also provided by reproducible and accurately rounded BLAS.
Fichier principal
Vignette du fichier
cse2017.pdf (38.69 Ko) Télécharger le fichier
cse2017-slides.pdf (13.6 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

lirmm-02091186 , version 1 (05-04-2019)

Identifiants

  • HAL Id : lirmm-02091186 , version 1

Citer

Chemseddine Chohra, Philippe Langlois, Rafife Nheilli, David Parello. Reproducible Parallel Simulations in HPC. CSE: Computational Science and Engineering, Feb 2017, Altanta, Georgia, United States. ⟨lirmm-02091186⟩
105 Consultations
65 Téléchargements

Partager

Gmail Facebook X LinkedIn More