Accéder directement au contenu Accéder directement à la navigation
Communication dans un congrès

MHPM : Multi-Scale Hybrid Programming Model A flexible parallelization Medthodology

Nader Khammassi 1 Jean-Christophe Le Lann 2 Jean-Philippe Diguet 3 Alexandre Skrzyniarz 4
1 Lab-STICC_ENSTAB_CACS_MOCS ; IDM
STIC - Pôle STIC [Brest], Lab-STICC - Laboratoire des sciences et techniques de l'information, de la communication et de la connaissance
2 Lab-STICC_ENSTAB_CACS_MOCS ; IDM
STIC - Pôle STIC [Brest]
3 Lab-STICC_UBS_CACS_MOCS
Lab-STICC - Laboratoire des sciences et techniques de l'information, de la communication et de la connaissance
Abstract : The continuous proliferation of multicore architectures has placed a great pressure on developers to parallelize their applications accordingly with what such platforms can offer . Unfortunately, traditional low-level programming model exacerbate the difficulties of building large and complex parallel applications. Highlevel parallel programming models are in high-demand as they reduce the burdens of programmers significantly and provide enough abstraction to accommodate hardware heterogeneity. In this paper, we propose a flexible parallelization methodology, and we introduce a new task-based hybrid programming model (MHPM) designed to provide high productivity and expressiveness without sacrificing performance. We show that MHPM allows easy expression of both sequential execution and several types of parallelism including task, data and temporal parallelism at all levels of granularity inside a single structured homogeneous programming model. In order to demonstrate the potential of our approach, we present a pure C++ implementation of MHPM, and we show that, despite its high abstraction, it provides comparable performances to lower-level programming models.
Liste complète des métadonnées

https://hal-ensta-bretagne.archives-ouvertes.fr/hal-00728364
Contributeur : Annick Billon-Coat <>
Soumis le : mercredi 5 septembre 2012 - 16:40:12
Dernière modification le : mercredi 24 juin 2020 - 16:19:21

Identifiants

  • HAL Id : hal-00728364, version 1

Citation

Nader Khammassi, Jean-Christophe Le Lann, Jean-Philippe Diguet, Alexandre Skrzyniarz. MHPM : Multi-Scale Hybrid Programming Model A flexible parallelization Medthodology. HPCC 2012, Jun 2012, Liverpool, United Kingdom. ⟨hal-00728364⟩

Partager

Métriques

Consultations de la notice

2322