Abstract

Computation in engineering and science can often benefit from acceleration due to lengthy calculation times for certain classes of numerical models. This paper, using a practical example drawn from...

Highlights

  • The boundary element method (BEM) is one of the established numerical methods for solving partial differential equations often of interest in the fields of engineering and science

  • The second author has investigated the possibility of use of graphics processing units (GPU) in solving for displacements and stresses at field points (Zsaki, 2011), at the time NVIDIA’s CUDA (NVIDIA, 2014) platform was used and no comparison was done regarding its performance with other hardware platforms

  • In the current climate of competing acceleration frameworks, such as NVIDIA’s CUDA (NVIDIA, 2014), the choice was made to use OpenCL since the code with minor modification can be complied on all platforms considered, which is not the case for CUDA, which currently only works on certain GPUs

Read more

Summary

PUBLIC INTEREST STATEMENT

Many problems in science and engineering require use of computers to create and analyse models to increase our understanding of the world around us. Most often the computation requires hours if not days to accomplish, any means to expedite the process is of interest. This paper presents a novel formulation of a numerical method used in engineering mechanics, developed such that it harnesses the power of various additional computer hardware, such as graphics cards, already found in a computer to achieve considerable reduction in time while maintaining the accuracy of computation. In addition to accelerated computing capabilities, the energy consumption was considered as well when ranking each computer hardware, catering to our energy-consciousness. The paper concludes with recommendations concerning the merits of each hardware accelerator. Power consumption were considered and recommendations are given concerning the merits of each hardware accelerator

Introduction
Dk υÞr
Write out solution to a file Clean up memory
Compute location of grid points
Obtain current index of execution thread
Cores Compute units Memory
Desktop GPU Tesla GPU FPGA
Findings
Conclusions
Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.