Mind The Gap! A study of some pitfalls preventing peak performance in polyhedral compilation using a polyhedral antidote

Philippe Clauss 1, 2, 3
1 CAMUS - Compilation pour les Architectures MUlti-coeurS
Inria Nancy - Grand Est, ICube - Laboratoire des sciences de l'ingénieur, de l'informatique et de l'imagerie
Abstract : The polyhedral model is a wonderful but imperfect world. While many advanced and fully automatic program analysis and optimization techniques have been developed thanks to its accuracy and expressiveness, these techniques may fail in producing efficient codes in some circumstances. Recently, this has been identified more clearly through the proposition of a new programming structure called xfor (multifor). This structure eases the manual application of optimizing transformations on loop nests for expert programmers and allows to generate executable codes that may be significantly faster than those generated automatically using well-established polyhedral strategies. In this paper, we highlight five main gaps regarding these strategies and discuss some ideas on how to bridge them.
Type de document :
Communication dans un congrès
IMPACT 2015, Fifth International Workshop on Polyhedral Compilation Techniques, In conjunction with HiPEAC 2015, Jan 2015, Amsterdam, Netherlands. 2015, 〈http://impact.gforge.inria.fr/impact2015〉
Liste complète des métadonnées

https://hal.inria.fr/hal-01099583
Contributeur : Philippe Clauss <>
Soumis le : dimanche 4 janvier 2015 - 19:24:35
Dernière modification le : mercredi 29 novembre 2017 - 15:07:10

Identifiants

  • HAL Id : hal-01099583, version 1

Collections

Citation

Philippe Clauss. Mind The Gap! A study of some pitfalls preventing peak performance in polyhedral compilation using a polyhedral antidote. IMPACT 2015, Fifth International Workshop on Polyhedral Compilation Techniques, In conjunction with HiPEAC 2015, Jan 2015, Amsterdam, Netherlands. 2015, 〈http://impact.gforge.inria.fr/impact2015〉. 〈hal-01099583〉

Partager

Métriques

Consultations de la notice

207