Bienvenue aux Mines Paristech
Bienvenue à MINES ParisTech
Newsletter International
Website
Théorie & Pratique
Vous êtes

webTV

Lecture

Séminaire PSL - Écosystèmes de médias I Session 4 Partie 4

Lecture

Séminaire PSL - Écosystèmes de médias I Session 4 Partie 2

Lecture

Séminaire PSL - Écosystèmes de médias I Session 4 Partie 3

Lecture

Séminaire PSL - Écosystèmes de médias I Session 4 Partie 1

Lecture

Lancement de la Chaire DEEP unissant HUTCHINSON, l'ESPCI et MINES ParisTech

+ Toutes les vidéos

Partager

Synthèse de code hybride à partir d'un programme à mémoire purement distribuée

Synthèse de code hybride à partir d'un programme à mémoire purement distribuée

Hybrid code synthesis from a pure distributed memory program

Proposition de thèse

Spécialité

Informatique temps réel, robotique et automatique - Fontainebleau

Ecole doctorale

SMI - Sciences des Métiers de l'Ingénieur

Directeur de thèse

TADONKI Claude

Unité de recherche

Mathématiques et Systèmes

ContactClaude TADONKI
Date de validité

01/10/2018

Site Web
Mots-clés

parallélisme, mémoire

parallelism, memory

Résumé

With the advent of multicore architectures we need us to reconsider the way we organize the tasks of our parallel code and plan their cooperation. Current (and future) supercomputers are built up with multicore processors probably coupled with accelerators such as GPUs. Regarding the connectivity, latency and network bandwidth have certainly evolved significantly, but the gap between the virtual topology and physical topology greatly increases the effective cost of interprocessor communications, especially for applications involving a high connectivity like stencil computation. Therefore, it becomes crucial to really consider hybrid implementations, which consider distributed memory model on top and shared memory model on the nodes. However, several factors are likely to discourage this effort: the required skill to design and implement hybrid codes is not common; in addition, number of quite complex codes have already been written based on the distributed memory model exclusively, and it is hard to consider modifying them; moreover, the belief of a clear effective reward from a hybrid code is moderate, because getting a benefit as close as expected is indeed not trivial. Therefore, designing a quasi-systematic methodology and a corresponding framework, that allow to easily switching from a pure MPI code to an efficient corresponding MPI + OpenMP version would be a valuable contribution. In this thesis, we propose to tackle the problem quite rigorously, on the basis of a tasks graph model and a clearly quantified accounting of major hardware and network aspects. The first goal will be to understand the criteria for an efficient and scalable hybrid code, followed by a methodology for a quick and pragmatic implementation. Applications envisaged for conducting this study are: Lattice QCD; pattern matching on a large sequence of high resolution images; partial differential equations; shortest paths computation in a graph.

With the advent of multicore architectures we need us to reconsider the way we organize the tasks of our parallel code and plan their cooperation. Current (and future) supercomputers are built up with multicore processors probably coupled with accelerators such as GPUs. Regarding the connectivity, latency and network bandwidth have certainly evolved significantly, but the gap between the virtual topology and physical topology greatly increases the effective cost of interprocessor communications, especially for applications involving a high connectivity like stencil computation. Therefore, it becomes crucial to really consider hybrid implementations, which consider distributed memory model on top and shared memory model on the nodes. However, several factors are likely to discourage this effort: the required skill to design and implement hybrid codes is not common; in addition, number of quite complex codes have already been written based on the distributed memory model exclusively, and it is hard to consider modifying them; moreover, the belief of a clear effective reward from a hybrid code is moderate, because getting a benefit as close as expected is indeed not trivial. Therefore, designing a quasi-systematic methodology and a corresponding framework, that allow to easily switching from a pure MPI code to an efficient corresponding MPI + OpenMP version would be a valuable contribution. In this thesis, we propose to tackle the problem quite rigorously, on the basis of a tasks graph model and a clearly quantified accounting of major hardware and network aspects. The first goal will be to understand the criteria for an efficient and scalable hybrid code, followed by a methodology for a quick and pragmatic implementation. Applications envisaged for conducting this study are: Lattice QCD; pattern matching on a large sequence of high resolution images; partial differential equations; shortest paths computation in a graph.

Contexte

Parallélisme

Encadrement

Claude Tadonki, CRI, Centre de recherche en informatique

Profil candidat

Please send a CV, a motivation letter, the list of grades for your last academic year and 1 or 2 recommendation letters.

Please see description in French

Objectif

Innovation

Références

-

Type financement

Concours pour un contrat doctoral

Retour à la liste des propositions

actualité

La finance au service du climat

Formation La finance au service du climat Et si l’Europe décidait de mettre la finance au service…
> En savoir +

Employabilité des diplômés

Formation Employabilité des diplômés Dans l'édition 2018 du Global Employability University Ranking , publiée le 14 novembre…
> En savoir +

#PACTE : un nouveau statut pour l'entreprise

Formation #PACTE : un nouveau statut pour l'entreprise Votée le 9 octobre 2018, en première lecture à…
> En savoir +

Les nouveaux visages de MINES ParisTech

Formation Les nouveaux visages de MINES ParisTech Conformément aux nouvelles orientations de son Plan…
> En savoir +

Peut-on apprendre à sauver la planète ?

Formation Peut-on apprendre à sauver la planète ? Faut-il être un « super héros » pour changer le monde ? Il faut…
> En savoir +

+ Toutes les actualités

agenda

Du 18 septembre au 14 décembre 2018 Musique aux Mines 2018

Du 1 octobre 2018 au 31 janvier 2019 Mai 68 aux Mines

Le 19 décembre 2018 Réduire le gaspillage & manger à petit prix

+ Tous les événements

contact

Régine MOLINS
Direction de l'Enseignement
Service du Doctorat
> envoyer un mail

Plan du site
MINES
ParisTech

60, Boulevard Saint-Michel
75272 PARIS Cedex 06
Tél. : +33 1 40 51 90 00

Implantations
Formation
Mentions légales | efil.fr | ©2012 MINES ParisTech | +33 1 40 51 90 00 |