Plugging anti and output dependence removal techniques into loop parallelization algorithms

Pierre-Yves Calland 1 Alain Darte 1 Yves Robert 1 Frédéric Vivien 1
1 REMAP - Regularity and massive parallel computing
Inria Grenoble - Rhône-Alpes, LIP - Laboratoire de l'Informatique du Parallélisme
Abstract : In this paper we shortly survey some loop transformation techniques which break anti or output dependences, or artificial cycles involving such 'false' dependences. These false dependences are removed through the introduction of temporary buffer arrays. Next we show how to plug these techniques into loop parallelization algorithms (such as Allen and Kennedy's algorithm). The goal is to extract as many parallel loops as the intrinsic degree of parallelism of the nest authorizes, while avoiding a full memory expansion. We try to reduce the number of temporary arrays that we introduce, as well as their dimension.
Type de document :
Article dans une revue
Parallel Computing, Elsevier, 1997, 23 (1-2), pp.251-266. 〈10.1016/S0167-8191(96)00108-1〉
Liste complète des métadonnées

https://hal.inria.fr/hal-00856884
Contributeur : Equipe Roma <>
Soumis le : lundi 2 septembre 2013 - 16:19:06
Dernière modification le : vendredi 20 avril 2018 - 15:44:24

Lien texte intégral

Identifiants

Collections

Citation

Pierre-Yves Calland, Alain Darte, Yves Robert, Frédéric Vivien. Plugging anti and output dependence removal techniques into loop parallelization algorithms. Parallel Computing, Elsevier, 1997, 23 (1-2), pp.251-266. 〈10.1016/S0167-8191(96)00108-1〉. 〈hal-00856884〉

Partager

Métriques

Consultations de la notice

104