Checkpoint and Restart: An Energy Consumption Characterization in Clusters

The fault tolerance method currently used in High Perfor- mance Computing (HPC) is the rollback-recovery method by using check- points. This, like any other fault tolerance method, adds an additional energy consumption to that of the execution of the application. The objective of this work is to...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Morán, Marina, Balladini, Javier, Rexachs, Dolores, Luque, Emilio
Formato: Articulo article acceptedVersion
Lenguaje:Inglés
Publicado: arXiv 2024
Materias:
Checkpoint
Restart
Energy consumption
Power
Fault tol- erance methods
Ciencias de la Computación e Información
Artículos
Acceso en línea:https://rdi.uncoma.edu.ar/handle/uncomaid/19173
Aporte de:
Repositorio Institucional UNCo de Universidad Nacional del Comahue
id I22-R178-uncomaid-19173
record_format dspace
spelling I22-R178-uncomaid-191732025-12-23T14:11:07Z Checkpoint and Restart: An Energy Consumption Characterization in Clusters Morán, Marina Balladini, Javier Rexachs, Dolores Luque, Emilio Checkpoint Restart Energy consumption Power Fault tol- erance methods Ciencias de la Computación e Información Artículos The fault tolerance method currently used in High Perfor- mance Computing (HPC) is the rollback-recovery method by using check- points. This, like any other fault tolerance method, adds an additional energy consumption to that of the execution of the application. The objective of this work is to determine the factors that affect the energy consumption of the computing nodes on homogeneous cluster, when per- forming checkpoint and restart operations, on SPMD (Single Program Multiple Data) applications. We have focused on the energetic study of compute nodes, contemplating different configurations of hardware and software parameters. We studied the effect of performance states (states P) and power states (states C) of processors, application problem size, checkpoint software (DMTCP) and distributed file system (NFS) config- uration. The results analysis allowed to identify opportunities to reduce the energy consumption of checkpoint and restart operations. Fil: Morán, Marina. Universidad Nacional del Comahue. Facultad de Informática; Argentina. Fil: Balladini, Javier. Universidad Nacional del Comahue. Facultad de Informática; Argentina. Fil: Rexachs, Dolores. Universitat Autónoma de Barcelona. Departamento de Arquitectura de Computadores y Sistemas Operativos; España. Fil: Luque, Emilio. Universitat Autónoma de Barcelona. Departamento de Arquitectura de Computadores y Sistemas Operativos; España. 2024 2025-12-17T14:54:56Z 2025-12-17T14:54:56Z Articulo article acceptedVersion https://rdi.uncoma.edu.ar/handle/uncomaid/19173 eng https://arxiv.org/abs/2409.02214 Atribución-NoComercial-CompartirIgual 4.0 https://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf 15 p. application/pdf arXiv
institution Universidad Nacional del Comahue
institution_str I-22
repository_str R-178
collection Repositorio Institucional UNCo
language Inglés
topic Checkpoint
Restart
Energy consumption
Power
Fault tol- erance methods
Ciencias de la Computación e Información
Artículos
spellingShingle Checkpoint
Restart
Energy consumption
Power
Fault tol- erance methods
Ciencias de la Computación e Información
Artículos
Morán, Marina
Balladini, Javier
Rexachs, Dolores
Luque, Emilio
Checkpoint and Restart: An Energy Consumption Characterization in Clusters
topic_facet Checkpoint
Restart
Energy consumption
Power
Fault tol- erance methods
Ciencias de la Computación e Información
Artículos
description The fault tolerance method currently used in High Perfor- mance Computing (HPC) is the rollback-recovery method by using check- points. This, like any other fault tolerance method, adds an additional energy consumption to that of the execution of the application. The objective of this work is to determine the factors that affect the energy consumption of the computing nodes on homogeneous cluster, when per- forming checkpoint and restart operations, on SPMD (Single Program Multiple Data) applications. We have focused on the energetic study of compute nodes, contemplating different configurations of hardware and software parameters. We studied the effect of performance states (states P) and power states (states C) of processors, application problem size, checkpoint software (DMTCP) and distributed file system (NFS) config- uration. The results analysis allowed to identify opportunities to reduce the energy consumption of checkpoint and restart operations.
format Articulo
article
acceptedVersion
author Morán, Marina
Balladini, Javier
Rexachs, Dolores
Luque, Emilio
author_facet Morán, Marina
Balladini, Javier
Rexachs, Dolores
Luque, Emilio
author_sort Morán, Marina
title Checkpoint and Restart: An Energy Consumption Characterization in Clusters
title_short Checkpoint and Restart: An Energy Consumption Characterization in Clusters
title_full Checkpoint and Restart: An Energy Consumption Characterization in Clusters
title_fullStr Checkpoint and Restart: An Energy Consumption Characterization in Clusters
title_full_unstemmed Checkpoint and Restart: An Energy Consumption Characterization in Clusters
title_sort checkpoint and restart: an energy consumption characterization in clusters
publisher arXiv
publishDate 2024
url https://rdi.uncoma.edu.ar/handle/uncomaid/19173
work_keys_str_mv AT moranmarina checkpointandrestartanenergyconsumptioncharacterizationinclusters
AT balladinijavier checkpointandrestartanenergyconsumptioncharacterizationinclusters
AT rexachsdolores checkpointandrestartanenergyconsumptioncharacterizationinclusters
AT luqueemilio checkpointandrestartanenergyconsumptioncharacterizationinclusters
_version_ 1854720037399560192

Ejemplares similares