Towards a Unified CPU-GPU code hybridization: A GPU Based Optimization Strategy Efficient on Other Modern Architectures
ONERA, 91123 Palaiseau, France
hal-01742774, (5 April 2018)
@misc{oteski2018towards,
title={Towards a Unified CPU–GPU code hybridization: A GPU Based Optimization Strategy Efficient on Other Modern Architectures},
author={Oteski, Ludomir and De Verdi{`e}re, Guillaume and Contassot-Vivier, Sylvain and Vialle, Stephane and Ryan, Juliet},
year={2018}
}
In this paper, we suggest a different methodology to shorten the code optimization development time while getting a unified code with good performance on different targeted devices. In the scope of this study, experiments are illustrated on a Discontinuous Galerkin code applied to Computational Fluid Dynamics. Tests are performed on CPUs, KNL Xeon-Phi and GPUs where performance comparison confirms that the GPU optimization guideline leads to efficient versions on CPU and Xeon-Phi for this kind of scientific applications. Based on these results, we finally suggest a methodology to end-up with an efficient hybridized CPU-GPU implementation.
April 15, 2018 by hgpu
