A Distributed and Parallel Asynchronous Unite and Conquer Method to Solve Large Scale Non-Hermitian Linear Systems

Wu, Xinzhe; Petiton, Serge

Document type :

Autre communication scientifique (congrès sans actes - poster - séminaire...): Communication dans un congrès avec actes

DOI :

10.1145/3149457.3154481

Title :

A Distributed and Parallel Asynchronous Unite and Conquer Method to Solve Large Scale Non-Hermitian Linear Systems

Author(s) :

Wu, Xinzhe [Auteur]
Centre National de la Recherche Scientifique [CNRS]
Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 [CRIStAL]
Maison de la Simulation [MDLS]
Petiton, Serge [Auteur]

Centre National de la Recherche Scientifique [CNRS]
Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 [CRIStAL]
Maison de la Simulation [MDLS]

Conference title :

HPC Asia 2018 - International Conference on High Performance Computing in Asia-Pacific Region

City :

Tokyo

Country :

Japon

Start date of the conference :

2018-01-28

Publication date :

2018-01

English keyword(s) :

iterative methods
asynchronous communication
linear algebra

HAL domain(s) :

Informatique [cs]/Calcul parallèle, distribué et partagé [cs.DC]
Informatique [cs]/Analyse numérique [cs.NA]

English abstract : [en]

Parallel Krylov Subspace Methods are commonly used for solving large-scale sparse linear systems. Facing the development of extreme scale platforms, the minimization of synchronous global communication becomes critical to ...
Show more >Parallel Krylov Subspace Methods are commonly used for solving large-scale sparse linear systems. Facing the development of extreme scale platforms, the minimization of synchronous global communication becomes critical to obtain good efficiency and scal-ability. This paper highlights a recent development of a hybrid (unite and conquer) method, which combines three computation algorithms together with asynchronous communication to accelerate the resolution of non-Hermitian linear systems and to improve its fault tolerance and reusability. Experimentation shows that our method has an up to 5× speedup and better scalability than the conventional methods for the resolution on hierarchical clusters with hundreds of nodes.Show less >

Language :

Anglais

Peer reviewed article :

Oui

Audience :

Internationale

Popular science :

Non

ANR Project :

MYX: MUST correctness checking for YML and XMP programs

Collections :

Centre de Recherche en Informatique, Signal et Automatique de Lille (CRIStAL) - UMR 9189

Source :

Harvested from HAL

Files

https://hal.archives-ouvertes.fr/hal-01677110/document
Open access
Access the document

https://hal.archives-ouvertes.fr/hal-01677110/document
Open access
Access the document

https://hal.archives-ouvertes.fr/hal-01677110/document
Open access
Access the document

https://hal.archives-ouvertes.fr/hal-01677110/document
Open access
Access the document

document
Open access
Access the document

HPCAsia2018_pap114_camera_ready_version.pdf
Open access
Access the document

HPCAsia2018_pap114_camera_ready_version.pdf
Open access
Access the document

document
Open access
Access the document

HPCAsia2018_pap114_camera_ready_version.pdf
Open access
Access the document

A Distributed and Parallel Asynchronous ... BibTeX CSV Excel RIS

Files

A Distributed and Parallel Asynchronous ...

BibTeX

CSV

Excel

RIS