AGULLO, Emmanuel; BEAUMONT, Olivier; EYRAUD-DUBOIS, Lionel; HERRMANN, Julien; KUMAR, Suraj; MARCHAL, Loris; THIBAULT, Samuel

doi:10.1109/IPDPSW.2015.35

Métadonnées

Licence d’utilisation du document

AGULLO, Emmanuel
High-End Parallel Algorithms for Challenging Numerical Simulations [HiePACS]

BEAUMONT, Olivier
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Reformulations based algorithms for Combinatorial Optimization [Realopt]

EYRAUD-DUBOIS, Lionel
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Reformulations based algorithms for Combinatorial Optimization [Realopt]

Langue

Communication dans un congrès

Ce document a été publié dans

Heterogeneity in Computing Workshop 2015, 2015-05-25, Hyderabad. 2015-05-25

Résumé en anglais

We consider the problem of allocating and scheduling dense linear application on fully heterogeneous platforms made of CPUs and GPUs. More specifically, we focus on the Cholesky factorization since it exhibits the main features of such problems. Indeed, the relative performance of CPU and GPU highly depends on the sub-routine: GPUs are for instance much more efficient to process regular kernels such as matrix-matrix multiplications rather than more irregular kernels such as matrix factorization. In this context, one solution consists in relying on dynamic scheduling and resource allocation mechanisms such as the ones provided by PaRSEC or StarPU. In this paper we analyze the performance of dynamic schedulers based on both actual executions and simulations, and we investigate how adding static rules based on an offline analysis of the problem to their decision process can indeed improve their performance, up to reaching some improved theoretical performance bounds which we introduce.< Réduire

Mots clés en anglais

starpu-simgrid

Dense Linear Algebra

Cholesky Factorization

Resource Allocation

Scheduling

Dynamic Schedulers

Heterogeneous Resources

Simulation

URI

https://oskar-bordeaux.fr/handle/20.500.12278/194454

DOI

http://dx.doi.org/10.1109/IPDPSW.2015.35

Project ANR

Simulation de systèmes de prochaine génération - ANR-11-INFR-0013
Solveurs pour architectures hétérogènes utilisant des supports d'exécution - ANR-13-MONU-0007

Origine

Importé de hal

Unités de recherche

Institut de Mathématiques de Bordeaux (IMB) - UMR 5251