On the autotuning of task-based numerical libraries for heterogeneous architectures

Agullo, Emmanuel; Cámara, Jesús; Cuenca Muñoz, Antonio Javier; Giménez, Domingo

Publication:
On the autotuning of task-based numerical libraries for heterogeneous architectures

dc.contributor.author	Agullo, Emmanuel
dc.contributor.author	Cámara, Jesús
dc.contributor.author	Cuenca Muñoz, Antonio Javier
dc.contributor.author	Giménez, Domingo
dc.contributor.department	Ingeniería y Tecnología de Computadores
dc.contributor.other	Facultad de Informática
dc.date.accessioned	2025-11-12T07:45:46Z
dc.date.available	2025-11-12T07:45:46Z
dc.date.copyright	© 2020 The authors and IOS Press
dc.date.issued	2020
dc.description.abstract	A roadmap for autotuning task-based numerical libraries is presented. Carefully chosen experiments are carried out when the numerical library is being installed to assess its performance. Real and simulated executions are considered to optimize the routine. The discussion is illustrated with a task-based tile Cholesky factorization, and the aim is to find the optimum tile size for any problem size, using the Chameleon numerical linear algebra package on top of the StarPU runtime system and also with the SimGrid simulator. The study shows that combining a smart exploration strategy of the search space with both real and simulated executions results in a fast, reliable autotuning process.
dc.format	application/pdf
dc.format.extent	10
dc.identifier.citation	Agullo, E., Cámara, J., Cuenca, J., Giménez, D.: On the Autotuning of Task- Based Numerical Libraries fo Heterogeneous Architectures. In: Proceedings of the 2019 International Conference on Parallel Computing. Advances in Parallel Computing, vol. 36, pp. 157–166 (2020)
dc.identifier.issn	0927-5452
dc.identifier.uri	http://hdl.handle.net/10201/172349
dc.language	eng
dc.publisher	IOS Press
dc.relation	Este trabajo forma parte del proyecto de investigación RTI2018-098156-B-C53 financiado por la Agencia Estatal de Investigación (AEI) y el Ministerio de Ciencia, Innovación y Universidades (MCIU)
dc.relation.ispartof	Proceedings of the 2019 International Conference on Parallel Computing. Advances in Parallel Computing, vol. 36, pp. 157–166 (2020)
dc.relation.publisherversion	https://ebooks.iospress.nl/volumearticle/53915
dc.rights.accessRights	info:eu-repo/semantics/openAccess
dc.subject	Autotuning
dc.subject	Linear Algebra
dc.subject	Task-based Programming
dc.subject	Heterogeneous Computing
dc.subject	Simulation
dc.subject.ods	Objetivo 7: Energía
dc.subject.ods	Objetivo 9: Infraestructura
dc.title	On the autotuning of task-based numerical libraries for heterogeneous architectures
dc.type	info:eu-repo/semantics/article
dspace.entity.type	Publication	es
relation.isAuthorOfPublication	7f8123b6-544c-4956-8229-538e4d177c31
relation.isAuthorOfPublication.latestForDiscovery	7f8123b6-544c-4956-8229-538e4d177c31