Skip to main content
NASA Advanced Supercomputing (NAS) Division World-class modeling, simulation, and supercomputing in support of the NASA mission

PyFR: High-Order Accurate Cross-Platform Petascale Computational Fluid Dynamics with Python

AMS Seminar Series

Speakers: Freddie Witherden, Stanford University and Peter Vincent, Imperial College London

Presentation

Seminar Slide Deck (PDF-5.8MB)

Abstract

High-order numerical methods for unstructured grids combine the superior accuracy of high-order spectral or finite difference methods with the geometrical flexibility of low-order finite volume or finite element schemes. The Flux Reconstruction (FR) approach unifies various high-order schemes for unstructured grids within a single framework. Additionally, the FR approach exhibits a significant degree of element locality, and is thus able to run efficiently on modern many-core hardware platforms, such as graphics processing units (GPUs). The aforementioned properties of FR mean it offers a promising route to performing affordable, and hence industrially relevant, scale-resolving simulations of hitherto intractable unsteady flows within the vicinity of real-world engineering geometries. In this talk we will present PyFR, an open-source, Python-based framework for solving advection-diffusion type problems using the FR approach. The framework is designed to solve a range of governing systems on mixed unstructured grids containing various element types. It is also designed to target a range of hardware platforms via use of a custom Mako-derived domain specific language. The latest release of PyFR is able to solve the compressible Euler and Navier-Stokes equations on grids of quadrilateral and triangular elements in two dimensions, and hexahedral, tetrahedral, prismatic, and pyramidal elements in three dimensions, targeting clusters of multi-core CPUs, NVIDIA GPUs, AMD GPUs, Intel Xeon Phis, and heterogeneous mixtures thereof. Results will be presented for various benchmark and "real-world" flow problems, and scalability/performance of PyFR will be demonstrated on clusters with thousands of NVIDIA GPUs. Throughout the talk the importance of algorithm-software-hardware co-design, in the context of next-generation computational fluid dynamics, will be highlighted.

Biography

Freddie Witherden studied Physics with Theoretical Physics at Imperial College London between 2008–2012 earning an M.Sci. degree with first class honours. In September of 2012, Witherden started a Ph.D. in computational fluid dynamics in the department of Aeronautics at Imperial College London under the supervision of Dr. Peter Vincent. His research focused on the development and implementation of high-order numerical schemes for solving the compressible Navier-Stokes equations with a particular focus on the co-design of software that is well suited to the massively parallel hardware of today and tomorrow. Graduating in December of 2015, Witherden is currently a postdoctoral scholar in the department of Aeronautics and Astronautics at Stanford University under the supervision of Prof. Antony Jameson.

Peter Vincent is a Senior Lecturer and EPSRC Early Career Fellow working in the department of Aeronautics at Imperial College London. He specializes in development of next-generation numerical methods for CFD simulations of unsteady flows, as well as the efficient implementation of these methods for massively parallel "many-core" systems, and their application to solve industrial flow problems. Prior to his current role at Imperial College London, Vincent was a postdoctoral scholar in the department of Aeronautics and Astronautics at Stanford University. He holds a Ph.D. in high-order numerical methods from the department of Aeronautics at Imperial College London, and a first class undergraduate degree (top-of-year) from the department of Physics at Imperial College London.