Towards Accelerated Computation of Atmospheric Equations using CUDA

Towards Accelerated Computation of Atmospheric Equations using CUDA

Paper in proceedings (conference paper)

Main objective ofthis paper is to outline possible ways how to achieve a substantialacceleration in case of advection-diffusion equation (A-DE) calculation, whichis commonly used for a description of the pollutant behavior in atmosphere.A-DE is a kind of partial differential equation (PDE) and in general case it isusually solved by numerical integration due to its high complexity. These typesof calculations are time consuming thus the main idea of our work is to adoptCUDA platform and commodity GPU card to do the calculations in a faster way.The solution is based on method of lines with 4th order Runge-Kuttascheme to handle the integration. As a matter of fact, the selected approachinvolves number of auxiliary variables and thus the memory management iscritical in order to achieve desired performance. We have implemented severalpossible solutions that use different memory access schemes. Detailedevaluation is provided in this paper where the obtained results show a tremendousprocessing speed up in comparison to CPU.

Main objective ofthis paper is to outline possible ways how to achieve a substantialacceleration in case of advection-diffusion equation (A-DE) calculation, whichis commonly used for a description of the pollutant behavior in atmosphere.A-DE is a kind of partial differential equation (PDE) and in general case it isusually solved by numerical integration due to its high complexity. These typesof calculations are time consuming thus the main idea of our work is to adoptCUDA platform and commodity GPU card to do the calculations in a faster way.The solution is based on method of lines with 4th order Runge-Kuttascheme to handle the integration. As a matter of fact, the selected approachinvolves number of auxiliary variables and thus the memory management iscritical in order to achieve desired performance. We have implemented severalpossible solutions that use different memory access schemes. Detailedevaluation is provided in this paper where the obtained results show a tremendousprocessing speed up in comparison to CPU.

CUDA, GPU, advection-diffusion equation, partial differential equation, Runge-Kutta, CFD

CUDA, GPU, advection-diffusion equation, partial differential equation, Runge-Kutta, CFD

ŠIMEK, V.; DVOŘÁK, R.; ZBOŘIL, F.; KUNOVSKÝ, J.

2010

24.03.2009

IEEE Computer Society

Cambridge

978-0-7695-3593-7

Proceedings of Eleventh International Conference on Computer Modelling and Simulation

449

454

6