Towards Performance Portability in a Compressible CFD Code

Howard, Micah; Bradley, Andrew M.; Bova, Steven W.; Overfelt, James R.; Wagnild, Ross; Dinzl, Derek; Hoemmen, Mark Frederick; Klinvex, Alicia

doi:10.2514/6.2017-4407

Cited by 43 publications

(11 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Next, in a nodeto-device comparison, HOMMEXX is faster on KNL than dual-socket HSW by up to ∼2.4 times; and HOMMEXX is faster on GPU than on dual-socket HSW by up to ∼3.2 times. These two sets of results establish that HOMMEXX has high performance on nonconventional architectures, at parity with or in some cases better than the end-to-end GPU performance reported for other codes for similar applications (Demeshko et al, 2018;Howard et al, 2017).…”

Section: Discussionsupporting

confidence: 58%

HOMMEXX 1.0: a performance-portable atmospheric dynamical core for the Energy Exascale Earth System Model

et al. 2019

View full text Add to dashboard Cite

Abstract. We present an architecture-portable and performant implementation of the atmospheric dynamical core (High-Order Methods Modeling Environment, HOMME) of the Energy Exascale Earth System Model (E3SM). The original Fortran implementation is highly performant and scalable on conventional architectures using the Message Passing Interface (MPI) and Open MultiProcessor (OpenMP) programming models. We rewrite the model in C++ and use the Kokkos library to express on-node parallelism in a largely architecture-independent implementation. Kokkos provides an abstraction of a compute node or device, layout-polymorphic multidimensional arrays, and parallel execution constructs. The new implementation achieves the same or better performance on conventional multicore computers and is portable to GPUs. We present performance data for the original and new implementations on multiple platforms, on up to 5400 compute nodes, and study several aspects of the single- and multi-node performance characteristics of the new implementation on conventional CPU (e.g., Intel Xeon), many core CPU (e.g., Intel Xeon Phi Knights Landing), and Nvidia V100 GPU.

show abstract

Section: Discussionsupporting

confidence: 58%

HOMMEXX 1.0: a performance-portable atmospheric dynamical core for the Energy Exascale Earth System Model

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Next, in a node-to-device comparison, HOMMEXX is faster on KNL than dual-socket HSW by up to ∼2.4×; and HOMMEXX is faster on GPU than on dual-socket HSW by up to ∼3.2×. These two sets of results establish that HOMMEXX has high performance on nonconventional architectures, at parity with or in some cases better than the end-to-end GPU performance reported for 10 other codes for similar applications (Demeshko et al, 2018;Howard et al, 2017).…”

mentioning

confidence: 59%

HOMMEXX 1.0: A Performance Portable Atmospheric Dynamical Core for the Energy Exascale Earth System Model

Bertagna¹,

Deakin²,

Guba³

et al. 2018

Preprint

View full text Add to dashboard Cite

Abstract. We present an architecture-portable and performant implementation of the atmospheric dynamical core (HOMME) of the Energy Exascale Earth System Model (E3SM). The original Fortran implementation is highly performant and scalable on conventional architectures using MPI and OpenMP. We rewrite the model in C++ and use the Kokkos library to express on-node parallelism in a largely architecture-independent implementation. Kokkos provides an abstraction of a compute node or device, layout-polymorphic multidimensional arrays, and parallel execution constructs. The new implementation achieves the same or better performance on conventional multicore computers and is portable to GPUs. We present performance data for the original and new implementations on multiple platforms, on up to 5400 compute nodes, and study several aspects of the single- and multi-node performance characteristics of the new implementation on conventional CPU, Intel Xeon Phi Knights Landing, and Nvidia V100 GPU.

show abstract

“…We demonstrate the aforementioned and forthcoming code-verification techniques on the Sandia Parallel Aerodynamics and Reentry Code (SPARC) [30,31,32,33] presently being developed at Sandia National Laboratories. SPARC is a compressible computational fluid dynamics code designed to model transonic and hypersonic reacting turbulent flows.…”

Section: Spatial-discretization Verification Resultsmentioning

confidence: 99%

Code-Verification Techniques for Hypersonic Reacting Flows in Thermochemical Nonequilibrium

Freno,

Carnes,

Weirs

2020

Preprint

View full text Add to dashboard Cite

The study of hypersonic flows and their underlying aerothermochemical reactions is particularly important in the design and analysis of vehicles exiting and reentering Earth's atmosphere. Computational physics codes can be employed to simulate these phenomena; however, code verification of these codes is necessary to certify their credibility. To date, few approaches have been presented for verifying codes that simulate hypersonic flows, especially flows reacting in thermochemical nonequilibrium. In this paper, we present our code-verification techniques for verifying the spatial accuracy and thermochemical source term in hypersonic reacting flows in thermochemical nonequilibrium. We demonstrate the effectiveness of these techniques on the Sandia Parallel Aerodynamics and Reentry Code (SPARC).

show abstract

Towards Performance Portability in a Compressible CFD Code

Cited by 43 publications

References 16 publications

HOMMEXX 1.0: a performance-portable atmospheric dynamical core for the Energy Exascale Earth System Model

HOMMEXX 1.0: a performance-portable atmospheric dynamical core for the Energy Exascale Earth System Model

HOMMEXX 1.0: A Performance Portable Atmospheric Dynamical Core for the Energy Exascale Earth System Model

Code-Verification Techniques for Hypersonic Reacting Flows in Thermochemical Nonequilibrium

Contact Info

Product

Resources

About