Skip to main navigation Skip to search Skip to main content

Parallel Auto-Scheduling of Counting Queries in Machine Learning Applications on HPC Systems

  • Częstochowa University of Technology

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

We introduce a parallel mechanism for auto-scheduling data access queries in machine learning applications. Our solution combines the advantages of three individual strategies to reduce the time of query stream execution. Using bayesian network learning as a use case, we achieve several times speedup compared to the best possible strategy on two different computing servers.

Original languageEnglish
Title of host publicationEuro-Par 2023
Subtitle of host publicationParallel Processing Workshops - Euro-Par 2023 International Workshops, 2023, Revised Selected Papers
EditorsDemetris Zeinalipour, Dora Blanco Heras, George Pallis, Herodotos Herodotou, Demetris Trihinas, Daniel Balouek, Patrick Diehl, Terry Cojean, Karl Fürlinger, Maja Hanne Kirkeby, Matteo Nardelli, Pierangelo Di Sanzo
PublisherSpringer Science and Business Media Deutschland GmbH
Pages327-333
Number of pages7
ISBN (Print)9783031488023
DOIs
StatePublished - 2024
EventInternational workshops held at the 29th International Conference on Parallel and Distributed Computing, Euro-Par 2023 - Limassol, Cyprus
Duration: Aug 28 2023Sep 1 2023

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume14352 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceInternational workshops held at the 29th International Conference on Parallel and Distributed Computing, Euro-Par 2023
Country/TerritoryCyprus
CityLimassol
Period08/28/2309/1/23

Fingerprint

Dive into the research topics of 'Parallel Auto-Scheduling of Counting Queries in Machine Learning Applications on HPC Systems'. Together they form a unique fingerprint.

Cite this