Exploiting system level heterogeneity to improve the performance of a GeoStatistics multi-phase task-based application - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2021

Exploiting system level heterogeneity to improve the performance of a GeoStatistics multi-phase task-based application

Abstract

Heterogeneity is part of HPC infrastructures, not only at the intranode but at the system level. Applications with multiple phases with distinct resource necessities can take advantage of this internode heterogeneity to improve performance and reduce resource idleness. Such an application is ExaGeoStat, a task-based machine learning framework specifically designed for geostatistics data. This work presents strategies to efficiently distribute multi-phase applications in system-level heterogeneous resources. We both (1) improve application phase overlap by optimizing runtime and scheduling decisions and (2) compute the optimal distribution for all the phases using a linear program leveraging node heterogeneity while limiting communication overhead. The performance gains of our phase overlap improvements are between 36% and 50% compared to the original base synchronous and homogeneous execution. We show that by adding some slow nodes to a homogeneous set of fast nodes, we can improve the performance by another 25% compared to a standard block-cyclic distribution, thereby harnessing any machine.
Fichier principal
Vignette du fichier
ICPP.pdf (2.34 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-03280459 , version 1 (16-07-2021)

Identifiers

Cite

Lucas Leandro Nesi, Arnaud Legrand, Lucas Mello Schnorr. Exploiting system level heterogeneity to improve the performance of a GeoStatistics multi-phase task-based application. ICPP 2021 - 50th International Conference on Parallel Processing, Aug 2021, Lemont, United States. pp.1-10, ⟨10.1145/3472456.3472516⟩. ⟨hal-03280459⟩
157 View
202 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More