Energy aware approach for HPC systems - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Chapitre D'ouvrage Année : 2014

Energy aware approach for HPC systems

Résumé

High‐performance computing (HPC) systems require energy during their full life cycle from design and production to transportation to usage and recycling/dismanteling. Because of increase of ecological and cost awareness, energy performance is now a primary focus. This chapter focuses on the usage aspect of HPC and how adapted and optimized software solutions could improve energy efficiency. It provides a detailed explanation of server power consumption, and discusses the application of HPC, phase detection, and phase identification. The chapter also suggests that having the load and memory access profiles is insufficient for an effective evaluation of the power consumed by an application. The available leverages in HPC systems are also shown in detail. The chapter proposes some solutions for modeling the power consumption of servers, which allows designing power prediction models for better decision making.These approaches allow the deployment and usage of a set of available green leverages, permitting energy reduction.
Fichier principal
Vignette du fichier
complexhpc_ch_book.pdf (568.49 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-00925310 , version 1 (07-01-2014)

Identifiants

Citer

Basmadjian Robert, Georges da Costa, Ghislain Landry Tsafack Chetsa, Laurent Lefevre, Olieksiak Ariel, et al.. Energy aware approach for HPC systems. Emmanuel Jeannot; Julius Zilinskas. High-Performance Computing on Complex Environments, Chapter 1, Wiley-IEEE Press, pp.341-363, 2014, 978-1118712054. ⟨10.1002/9781118711897.ch18⟩. ⟨hal-00925310⟩
403 Consultations
429 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More