Abstract
We analyze two parallel finite element implementations of the 2D time-dependent advection diffusion problem, one for multi-core clusters and one for CUDA-enabled GPUs, and compare their performances in terms of time and energy consumption. The parallel CUDA-enabled GPU implementation was derived from the multi-core cluster version. Our experimental results show that a desktop machine with a single CUDA-enabled GPU can achieve performance higher than a 24-machine (96 cores) cluster in this class of finite element problems. Also, the CUDA-enabled GPU implementation consumes less than one twentieth of the energy (Joules) consumed by the multi-core cluster implementation while solving a whole instance of the finite element problem. © 2013 Springer-Verlag.
Cite
CITATION STYLE
De Souza, A. F., Veronese, L., Lima, L. M., Badue, C., & Catabriga, L. (2013). Evaluation of two parallel finite element implementations of the time-dependent advection diffusion problem: GPU versus cluster considering time and energy consumption. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7851 LNCS, pp. 149–162). https://doi.org/10.1007/978-3-642-38718-0_17
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
