Designing a parallel graph reduction machine with dynamic granularity control
Type doc. :
Thèses / mémoires
Langue :
Anglais
Année de soutenance:
1991
Afficher le Résumé
A new partitioning technique is proposed, in which several function calls are coalesced into a single task, increasing the granularity of the computation. Ideal- ly, a task is only created if its execution time exceeds the scheduling overhead, and if there will be a processor free to execute it. Unlike other approaches, the decision can be postponed till run-time, thus taking the advantage of testing the value of some important parameters, whose in?uence has been determined using static analysis techniques.
| N° Bulletin | Date / Année de parution | Titre N° Spécial | Sommaire |
|---|
| Cote | Localisation | Type de Support | Type de Prêt | Statut | Date de Restitution Prévue | Réservation |
|---|---|---|---|---|---|---|
| 004 RAB TH 1 | BIB-Centrale / Thèses | interne | disponible |
Rabhi, F. et al. (1991). Designing a parallel graph reduction machine with dynamic granularity control (P.H.D.) . Sheffield (angleterre).