2011Unpublished venueOpen access

Development of a Chemically Reacting Flow Solver on the Graphic Processing Units

Hai Le

Open full text 0 citations

Abstract

The focus of the current research is to develop a numerical framework on the Graphic Processing Units (GPU) capable of modeling chemically reacting flow. The framework incorporates a high-order finite volume method coupled with an implicit solver for the chemical kinetics. Both the fluid solver and the kinetics solver are designed to take advantage of the GPU architecture to achieve high performance. The structure of the numerical framework is shown, detailing different aspects of the optimization implemented on the solver. The mathematical formulation of the core algorithms is presented along with a series of standard test cases, including both nonreactive and reactive flows, in order to validate the capability of the numerical solver. The performance results obtained with the current framework show the parallelization efficiency of the solver and emphasize the capability of the GPU in performing scientific calculations. Distribution A: Approved for public release; distribution unlimited. PA #11179

Open-access reader

About this research paper

What this paper is about

The focus of the current research is to develop a numerical framework on the Graphic Processing Units (GPU) capable of modeling chemically reacting flow. The framework incorporates a high-order finite volume method coupled with an implicit solver for the chemical kinetics. Both the fluid solver and the kinetics solver are designed to take advantage of the GPU architecture to achieve high performance. The structure of the numerical framework is shown, detailing different aspects of the optimization implemented on the solver. The mathematical formulation of the core algorithms is presented along with a series of standard test cases, including both nonreactive and reactive flows, in order to validate the capability of the numerical solver. The performance results obtained with the current framework show the parallelization efficiency of the solver and emphasize the capability of the GPU in performing scientific calculations. Distribution A: Approved for public release; distribution unlimited. PA #11179

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

The focus of the current research is to develop a numerical framework on the Graphic Processing Units (GPU) capable of modeling chemically reacting flow. The framework incorporates a high-order finite volume method coupled with an implicit solver for the chemical kinetics. Both the fluid solver and the kinetics solver are designed to take advantage of the GPU architecture to achieve high performance. The structure of the numerical framework is shown, detailing different aspects of the optimization implemented on the solver. The mathematical formulation of the core algorithms is presented along with a series of standard test cases, including both nonreactive and reactive flows, in order to validate the capability of the numerical solver. The performance results obtained with the current framework show the parallelization efficiency of the solver and emphasize the capability of the GPU in performing scientific calculations. Distribution A: Approved for public release; distribution unlimited. PA #11179

Key concepts: Solver, Computational science, Computer science, Focus (optics), Flow (mathematics), CUDA, Parallel computing, Mathematics

Related papers

Back to paper searchBrowse research topicsOriginal source
Development of a Chemically Reacting Flow Solver on the Graphic Processing Units — Research Paper | ScholarLens