A NUMA Aware Scheduler for a Parallel Sparse Direct Solver
| hal.structure.identifier | Laboratoire Bordelais de Recherche en Informatique [LaBRI] | |
| hal.structure.identifier | Efficient runtime systems for parallel architectures [RUNTIME] | |
| hal.structure.identifier | Algorithms and high performance computing for grand challenge applications [SCALAPPLIX] | |
| dc.contributor.author | FAVERGE, Mathieu | |
| hal.structure.identifier | Algorithms and high performance computing for grand challenge applications [SCALAPPLIX] | |
| dc.contributor.author | LACOSTE, Xavier | |
| hal.structure.identifier | Laboratoire Bordelais de Recherche en Informatique [LaBRI] | |
| hal.structure.identifier | Algorithms and high performance computing for grand challenge applications [SCALAPPLIX] | |
| dc.contributor.author | RAMET, Pierre | |
| dc.date.accessioned | 2024-04-15T09:53:12Z | |
| dc.date.available | 2024-04-15T09:53:12Z | |
| dc.date.created | 2008 | |
| dc.date.issued | 2008 | |
| dc.date.conference | 2008 | |
| dc.identifier.uri | https://oskar-bordeaux.fr/handle/20.500.12278/198590 | |
| dc.description.abstractEn | Over the past few years, parallel sparse direct solvers made significant progress and are now able to solve efficiently industrial three-dimensional problems with several millions of unknowns. An hybrid MPI-thread implementation of our direct solver PaStiX is already well suited for SMP nodes or new multi-core architectures and drastically reduced the memory overhead and improved scalability. In the context of distributed NUMA architectures, a dynamic scheduler based on a work-stealing algorithm has been developed to fill in communication idle times. On these architectures, it is important to take care of NUMA effects and to preserve memory affinity during the work-stealing. The scheduling of communications also needs to be adapted, especially to ensure the overlap by computations. Experiments on numerical test cases will be presented to prove the efficiency of the approach on NUMA architectures. If memory is not large enough to treat a given problem, disks must be used to store data that cannot fit in memory (out-of-core storage). The idle-times due to disk access have to be managed by our dynamic scheduler to prefetch and save datasets. Thus, we design and study specific scheduling algorithms in this particular context. | |
| dc.description.sponsorship | Adaptation et Optimisation des Performances Applicatives sur architectures NUMA. Etude et Mise en Œuvre sur des Applications en SISmologie. - ANR-05-CIGC-0002 | |
| dc.language.iso | en | |
| dc.subject.en | sparse direct solver | |
| dc.subject.en | NUMA architecture | |
| dc.subject.en | multi-cores | |
| dc.subject.en | dynamic scheduling | |
| dc.title.en | A NUMA Aware Scheduler for a Parallel Sparse Direct Solver | |
| dc.type | Communication dans un congrès | |
| dc.subject.hal | Informatique [cs]/Calcul parallèle, distribué et partagé [cs.DC] | |
| bordeaux.hal.laboratories | Laboratoire Bordelais de Recherche en Informatique (LaBRI) - UMR 5800 | * |
| bordeaux.institution | Université de Bordeaux | |
| bordeaux.institution | Bordeaux INP | |
| bordeaux.institution | CNRS | |
| bordeaux.conference.title | PMAA'08 | |
| bordeaux.country | CH | |
| bordeaux.conference.city | Neuchâtel | |
| bordeaux.peerReviewed | oui | |
| hal.identifier | inria-00344709 | |
| hal.version | 1 | |
| hal.invited | non | |
| hal.proceedings | non | |
| hal.popular | non | |
| hal.audience | Internationale | |
| hal.origin.link | https://hal.archives-ouvertes.fr//inria-00344709v1 | |
| bordeaux.COinS | ctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.date=2008&rft.au=FAVERGE,%20Mathieu&LACOSTE,%20Xavier&RAMET,%20Pierre&rft.genre=unknown |
Files in this item
| Files | Size | Format | View |
|---|---|---|---|
|
There are no files associated with this item. |
|||