An efficient implementation for broadcasting data in parallel applications over ethernet clusters
Material type:
ArticleSubject(s): Online resources: Summary: This paper introduces a natural implementation for broadcasting data on Ethernet based clusters used for parallel computing. Initially, it will be shown that libraries for Message Passing processes such as PVM implementations of MPI do not implement efficiently this operation there is no reliability in terms of its performance. An implementation for broadcast messages is presented, taking into account the Ethernetbased hardware layer found in most of the clusters used for parallel computing. The proposed implementation for broadcasting data is compared with the broadcast message available in PVM LAM/MPI. Also, some comments are made about the proposed broadcast messages when the hardware layer is not Ethernet based. Finally, it will be shown by experimentation how the proposed broadcast message is used in the context of parallel linear algebra operations, specifically for parallel matrix multiplication.
| Item type | Home library | Collection | Call number | URL | Status | Date due | Barcode | |
|---|---|---|---|---|---|---|---|---|
Capítulo de libro
|
Biblioteca de la Facultad de Informática | Biblioteca digital | A0243 (Browse shelf(Opens below)) | Link to resource | No corresponde |
Browsing Biblioteca de la Facultad de Informática shelves, Collection: Biblioteca digital Close shelf browser (Hides shelf browser)
Formato de archivo: PDF. -- Este documento es producción intelectual de la Facultad de Informática-UNLP (Colección BIPA / Biblioteca.) -- Disponible también en línea (Cons. 09/05/2011)
This paper introduces a natural implementation for broadcasting data on Ethernet based clusters used for parallel computing. Initially, it will be shown that libraries for Message Passing processes such as PVM implementations of MPI do not implement efficiently this operation there is no reliability in terms of its performance. An implementation for broadcast messages is presented, taking into account the Ethernetbased hardware layer found in most of the clusters used for parallel computing. The proposed implementation for broadcasting data is compared with the broadcast message available in PVM LAM/MPI. Also, some comments are made about the proposed broadcast messages when the hardware layer is not Ethernet based. Finally, it will be shown by experimentation how the proposed broadcast message is used in the context of parallel linear algebra operations, specifically for parallel matrix multiplication.
AINA ༿ Proceedings of the 17th International Conference on Advanced Information Networking Applications (17th : 2003 : Xi'an)