Coarse-grain Load Distribution in Heterogeneous Computing

Alonso, Fernando; González Escribano, Arturo; Torres de la Sierra, Yuri; Llanos Ferraris, Diego Rafael

Por favor, use este identificador para citar o enlazar este ítem:http://uvadoc.uva.es/handle/10324/39043

Título

Coarse-grain Load Distribution in Heterogeneous Computing

Autor

Alonso, Fernando

González Escribano, Arturo

Torres de la Sierra, Yuri

Llanos Ferraris, Diego Rafael

Editor

CMMSE 2019

Año del Documento

2019

Documento Fuente

Coarse-grain Load Distribution in Heterogeneous Computing. Fernando Alonso, Arturo González-Escribano, Yuri Torres, Diego R. Llanos. Tipo de Participación: poster. 19th International Computational and Mathematical Methods in Science and Engineering 2019 – CMMSE 2019, Costa Ballena (Rota), Cádiz (España). Fecha: 1-5 julio 2019.

Resumo

HPC heterogeneous clusters are composed by different type of machines (various types of component manufacturers, varying computational capacities), and different hardware accelerators. TThe most common type of data distributions is the equal division of the data across all the nodes. A more sophisticated policy of data distribution is needed to explode the computational capacity of the entire system.