Adapting a Finite-Element Type Solver for Bioelectromagnetics to the DEEP-ER Platform

Abstract : In this paper, we report on our recent efforts towards adapting a Discon-tinuous Galerkin Time-Domain solver for computational bioelectromagnetics to the novel, heterogeneous architecture proposed in the DEEP-ER european project on exascale computing. This architecture is based on the Cluster/Booster division concept which will be recalled. As a first step, we summarize the key features of the application and present the outcomes of a profiling of the code using the tools developed by DEEP-ER partners. We then go through the subsequent general improvements of the application as well as specific developments aimed at exploiting efficiently the DEEP-ER platform. This particularly includes porting the application to the Intel (R) Many Integrated Core Architecture. We conclude with an outlook on next steps, including the different Cluster/Booster division strategies.
Complete list of metadatas

Cited literature [12 references]  Display  Hide  Download

https://hal.inria.fr/hal-01243708
Contributor : Raphaël Léger <>
Submitted on : Thursday, December 17, 2015 - 2:12:08 PM
Last modification on : Thursday, May 3, 2018 - 1:32:55 PM
Long-term archiving on : Saturday, April 29, 2017 - 3:16:27 PM

File

paper.pdf
Files produced by the author(s)

Licence


Copyright

Identifiers

Citation

Raphaël Léger, Damian Alvarez Mallon, Alejandro Duran, Stéphane Lanteri. Adapting a Finite-Element Type Solver for Bioelectromagnetics to the DEEP-ER Platform. ParCo 2015, Edinburgh, United Kingdom. Parallel Computing: On the Road to Exascale (27), IOS Press, pp.850, 2016, Advances in parallel computing, ⟨10.3233/978-1-61499-621-7-349⟩. ⟨hal-01243708⟩

Share

Metrics

Record views

516

Files downloads

440