A GPU implementation of the analog ensemble method

Crico, Ruben; Charles, Ines; Balsa, Carlos; Rufino, José

Publication

A GPU implementation of the analog ensemble method

2025Conference paper

datacite.subject.fos	Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática
datacite.subject.sdg	09:Indústria, Inovação e Infraestruturas
dc.contributor.author	Crico, Ruben
dc.contributor.author	Charles, Ines
dc.contributor.author	Balsa, Carlos
dc.contributor.author	Rufino, José
dc.date.accessioned	2025-06-18T14:50:23Z
dc.date.available	2025-06-18T14:50:23Z
dc.date.issued	2025
dc.description.abstract	The Analog Ensemble (AnEn) method can be used to reconstruct incomplete time series using correlated series. Since the AnEn method may use data including several variables through long periods of time, its storage and computational cost may be substantial, slowing down reconstructions. This paper presents a full GPU implementation of the AnEn method, based on PyCUDA, that leads to a significant a cceleration of its execution. The implementation resorts to several techniques that seek to minimize the consumption of GPU global memory in the various steps of the AnEn algorithm, thus making room for larger input datasets. This is further reinforced by the use of batch processing as a way to automatically adapt the datasets size to the GPU memory available. The GPU implementation was tested on a meteorological dataset spanning 10 years, exhibiting a 30-fold speedup in the reconstruction time against a comparable CPU-based multicore version executed with up to 48 cores. The impact on the reconstruction error of changes on several important parameters of the implementation was also assessed.	eng
dc.description.sponsorship	This work was supported by national funds through FCT/MCTES (PIDDAC): UID/05757 - Research Centre in Digitalization and Intelligent Robotics/ Centro de Investigação em Digitalização e Robótica Inteligente (CeDRI).
dc.identifier.citation	Crico, Ruben; Charles, Ines; Balsa, Carlos; Rufino, José (2025). A GPU implementation of the analog ensemble method. In 33rd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing. Toulouse. p. 1-8
dc.identifier.doi	10.1109/pdp66500.2025.00039
dc.identifier.isbn	979-833152493-7
dc.identifier.uri	http://hdl.handle.net/10198/34614
dc.language.iso	eng
dc.peerreviewed	yes
dc.publisher	Institute of Electrical and Electronics Engineers
dc.relation.ispartof	2025 33rd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing (PDP)
dc.rights.uri	http://creativecommons.org/licenses/by/4.0/
dc.subject	Hindcasting
dc.subject	Analog Ensemble
dc.subject	Scientific Computing
dc.subject	Parallel Computing: GPU Computing
dc.title	A GPU implementation of the analog ensemble method	eng
dc.type	conference paper
dspace.entity.type	Publication
oaire.citation.endPage	8
oaire.citation.startPage	1
oaire.citation.title	Proceedings - 33rd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2025
oaire.version	http://purl.org/coar/version/c_970fb48d4fbd8a85
person.familyName	Balsa
person.familyName	Rufino
person.givenName	Carlos
person.givenName	José
person.identifier	1721518
person.identifier.ciencia-id	DE1E-2F7A-AAB1
person.identifier.ciencia-id	C414-F47F-6323
person.identifier.orcid	0000-0003-2431-8665
person.identifier.orcid	0000-0002-1344-8264
person.identifier.rid	M-8735-2013
person.identifier.scopus-author-id	23391719100
person.identifier.scopus-author-id	55947199100
relation.isAuthorOfPublication	d0e5ccff-9696-4f4f-9567-8d698a6bf17d
relation.isAuthorOfPublication	1e24d2ce-a354-442a-bef8-eebadd94b385
relation.isAuthorOfPublication.latestForDiscovery	d0e5ccff-9696-4f4f-9567-8d698a6bf17d