Contact Us

Colin Britcher, Ph.D.
Director of Graduate Education

Mary Catherine Bunde, M.Ed.
Senior Education Administrator

6.23.16 Parasani

Title: Efficiency of high order spectral element methods on three of the largest petascale supercomputers in the World

Speaker: Dr. Matteo Parasani

Date: Thursday, June 23, 2016

Time: 10:30-11:30

Location: NIA, Room 137

Abstract: High order methods for the solution of PDEs expose a trade- off between computational cost and accuracy on a per degree of freedom basis. In many cases, the cost increases due to higher arithmetic intensity while affecting data movement minimally. As architectures tend towards wider vector instructions and expect higher arithmetic intensities, the best order for a particular simulation may change. This study highlights preferred orders by identifying the high order efficiency frontier of the spectral element method implemented in NekBox: the set of orders and meshes that minimize computational cost at fixed accuracy. From a performance point of view, we demonstrate up to 60% full application bandwidth utilization at scale and achieve ≈ 1.1 PFlop/s of compute performance in most flop-intense methods.

Bio: Matteo Parsani is an Assistant Professor in Applied Mathematics & Computational Science and a core faculty in the Extreme Computing Research Center at KAUST. Dr. Parsani holds a master degree in Aerospace Engineering and Computational Aerodynamics from Politecnico di Milano. He received his doctorate in Mechanical Engineering in November 2010 from the Vreije Universities Brussel, Belgium. Before joining KAUST in September 2015, he was a NASA NPP Fellow at NASA Langley Research Center, for almost three years. Matteo Parsani’s research interests are related to the design and implementation of robust and scalable numerical methods for hyperbolic and mixed hyperbolic/parabolic partial differential equations, as well as their application to solve realistic flow problems in computational aerodynamics and dense gas flows simulations.



100 Exploration Way
Hampton, VA 23666