Performance of FORTRAN and C GPU Extensions for a Benchmark Suite of Fourier Pseudospectral Algorithms

B. Cloutier; B. K. Muite; P. Rigge

Performance of FORTRAN and C GPU Extensions for a Benchmark Suite of Fourier Pseudospectral Algorithms

Computational Physics 2012-08-14 v2 Mathematical Software

Authors: B. Cloutier , B. K. Muite , P. Rigge

Abstract

A comparison of PGI OpenACC, FORTRAN CUDA, and Nvidia CUDA pseudospectral methods on a single GPU and GCC FORTRAN on single and multiple CPU cores is reported. The GPU implementations use CuFFT and the CPU implementations use FFTW. Porting pre-existing FORTRAN codes to utilize a GPUs is efficient and easy to implement with OpenACC and CUDA FORTRAN. Example programs are provided.

Keywords

GPU extensions Fourier pseudospectral algorithms OpenACC CUDA FORTRAN CuFFT FFTW performance benchmark parallel computing

Cite

@article{arxiv.1206.3215,
  title  = {Performance of FORTRAN and C GPU Extensions for a Benchmark Suite of Fourier Pseudospectral Algorithms},
  author = {B. Cloutier and B. K. Muite and P. Rigge},
  journal= {arXiv preprint arXiv:1206.3215},
  year   = {2012}
}

Comments

Revised version. New title. Forthcoming in proceedings of the Symposium on Application Accelerators in High-Performance computing (SAAHPC 2012). Related programs are available for download

Performance of FORTRAN and C GPU Extensions for a Benchmark Suite of Fourier Pseudospectral Algorithms

Abstract

Keywords

Cite

Comments

Related papers