Autotuning under Tight Budget Constraints: A Transparent Design of Experiments Approach

Pedro Bruel 1, 2 Steven Quinito Masnada 3 Brice Videau 4 Arnaud Legrand 2 Jean-Marc Vincent 2 Alfredo Goldman 1
2 POLARIS - Performance analysis and optimization of LARge Infrastructures and Systems
Inria Grenoble - Rhône-Alpes, LIG - Laboratoire d'Informatique de Grenoble
3 MISTIS - Modelling and Inference of Complex and Structured Stochastic Systems
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann, INPG - Institut National Polytechnique de Grenoble
Abstract : A large amount of resources is spent writing, port-ing, and optimizing scientific and industrial High Performance Computing applications, which makes autotuning techniques fundamental to lower the cost of leveraging the improvements on execution time and power consumption provided by the latest software and hardware platforms. Despite the need for economy, most autotuning techniques still require a large budget of costly experimental measurements to provide good results, while rarely providing exploitable knowledge after optimization. The contribution of this paper is a user-transparent autotuning technique based on Design of Experiments that operates under tight budget constraints by significantly reducing the measurements needed to find good optimizations. Our approach enables users to make informed decisions on which optimizations to pursue and when to stop. We present an experimental evaluation of our approach and show it is capable of leveraging user decisions to find the best global configuration of a GPU Laplacian kernel using half of the measurement budget used by other common autotuning techniques. We show that our approach is also capable of finding speedups of up to 50×, compared to gcc's-O3, for some kernels from the SPAPT benchmark suite, using up to 10× less measurements than random sampling.
Type de document :
Pré-publication, Document de travail
2018
Liste complète des métadonnées

https://hal.inria.fr/hal-01953287
Contributeur : Arnaud Legrand <>
Soumis le : mercredi 12 décembre 2018 - 17:43:03
Dernière modification le : mercredi 26 décembre 2018 - 10:23:06

Fichier

ipdps19.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01953287, version 1

Citation

Pedro Bruel, Steven Quinito Masnada, Brice Videau, Arnaud Legrand, Jean-Marc Vincent, et al.. Autotuning under Tight Budget Constraints: A Transparent Design of Experiments Approach. 2018. 〈hal-01953287〉

Partager

Métriques

Consultations de la notice

48

Téléchargements de fichiers

42