Addressing Materials Science Challenges Using GPU-accelerated POWER8 Nodes

Baumeister, Paul F.; Bühler, Markus; Zeller, Rudolf; Hater, Thorsten; Krill, Benjamin; Bornemann, Marcel; Pleiter, Dirk

doi:10.1007/978-3-319-43659-3_6

Contribution to a conference proceedings/Contribution to a book

FZJ-2016-05342

Addressing Materials Science Challenges Using GPU-accelerated POWER8 Nodes

Baumeister, P. F. (Corresponding author)FZJ* ; Bornemann, M.FZJ* ; Bühler, M. ; Hater, T.FZJ* ; Krill, B. ; Pleiter, D.FZJ* ; Zeller, R.FZJ*

2016
Springer International Publishing Cham
ISBN: 978-3-319-43658-6 (print), 978-3-319-43659-3 (electronic)

Euro-Par 2016: Parallel Processing / Dutot, Pierre-François (Editor) ; Cham : Springer International Publishing, 2016, Chapter 6 ; ISSN: 0302-9743=1611-3349 ; ISBN: 978-3-319-43658-6=978-3-319-43659-3
Euro-Par 2016, Grenoble, France, 24 Aug 2016 - 26 Aug 2016 Cham : Springer International Publishing, Lecture Notes in Computer Science 9833, 77 - 89 (2016) [10.1007/978-3-319-43659-3_6]

This record in other databases:

Please use a persistent id in citations: doi:10.1007/978-3-319-43659-3_6

Abstract: Materials research is an area that is expected to strongly benefit from the growing performance capabilities of future supercomputers towards exascale. Density functional theory (DFT) has become one of the most important methods for numerical materials science. In this paper we present results of a performance model based analysis of a particular, scalable DFT-based application on GPU-accelerated compute nodes with POWER8 processors. These technologies are part of a future roadmap for pre-exascale architectures. With power consumption becoming a major design constraint, we also determine the energy required for executing the most performance critical kernel.

Classification: