Distributed NVRAM Cache – Optimization and Evaluation with Power of Adjacency Matrix

Abstract : In this paper we build on our previously proposed MPI I/O NVRAM distributed cache for high performance computing. In each cluster node it incorporates NVRAMs which are used as an intermediate cache layer between an application and a file for fast read/write operations supported through wrappers of MPI I/O functions. In this paper we propose optimizations of the solution including handling of write requests with a synchronous mode, additional modes preventing data preloading from a file and synchronization on file close if the solution is used as temporary cache only. Furthermore, we have evaluated the solution for a real application that computes powers of an adjacency matrix of a graph in parallel. We demonstrated superiority of our solution compared to a regular MPI I/O implementation for various powers and numbers of graph nodes. Finally, we presented good scalability of the solution for more than 600 processes running on a large HPC cluster.
Type de document :
Communication dans un congrès
Khalid Saeed; Władysław Homenda; Rituparna Chaki. 16th IFIP International Conference on Computer Information Systems and Industrial Management (CISIM), Jun 2017, Bialystok, Poland. Springer International Publishing, Lecture Notes in Computer Science, LNCS-10244, pp.15-26, 2017, Computer Information Systems and Industrial Management. 〈10.1007/978-3-319-59105-6_2〉
Liste complète des métadonnées

Littérature citée [19 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01656233
Contributeur : Hal Ifip <>
Soumis le : mardi 5 décembre 2017 - 14:58:06
Dernière modification le : mercredi 6 décembre 2017 - 01:20:58

Fichier

 Accès restreint
Fichier visible le : 2020-01-01

Connectez-vous pour demander l'accès au fichier

Licence


Distributed under a Creative Commons Paternité 4.0 International License

Identifiants

Citation

Artur Malinowski, Paweł Czarnul. Distributed NVRAM Cache – Optimization and Evaluation with Power of Adjacency Matrix. Khalid Saeed; Władysław Homenda; Rituparna Chaki. 16th IFIP International Conference on Computer Information Systems and Industrial Management (CISIM), Jun 2017, Bialystok, Poland. Springer International Publishing, Lecture Notes in Computer Science, LNCS-10244, pp.15-26, 2017, Computer Information Systems and Industrial Management. 〈10.1007/978-3-319-59105-6_2〉. 〈hal-01656233〉

Partager

Métriques

Consultations de la notice

151