In the past few years, GPUs have gained a lot of popularity as they offer an opportunity to accelerate many algorithms. In this paper we present a mono, a multi-GPU and GPU cluster implementation, based on CUDA, of a cellular automata. We derive a performance model and establish its accurateness.
Titre
Performance model for a cellular automata implementation on a GPU cluster
Date
2012-06
Publié dans
Applications, tools and techniques on the road to exascale computing
Editeur
Amsterdam, Washington D.C., IOS press
Pagination
pp. 191-198
ISBN
978-1-61499-040-6
Collection et n°
Advances in parallel computing, no. 22
Domaine
Ingénierie et Architecture
Ecole
HEPIA - Genève
Institut
inIT - Institut d'Ingénierie Informatique et des Télécommunications