On optimization techniques for the matrix multiplication on hybrid CPU+GPU platforms
Technical University of Cartagena
Annals of Multicore and GPU Programming, Vol 1, No 1, 2014
@article{canovas2014optimization,
title={On optimization techniques for the matrix multiplication on hybrid CPU+ GPU platforms},
author={C{‘a}novas, Domingo Gim{‘e}nez and Garc{‘i}a, Luis-Pedro and Cuenca, Javier},
journal={Annals of Multicore and GPU Programming},
volume={1},
number={1},
year={2014}
}
The use of auto-tuning techniques in a matrix multiplication routine for hybrid CPU+GPU platforms is analyzed. Basic models of the execution time of the hybrid routine and information obtained during its installation are used to optimize the execution time with a balanced assignation of the computation to the computing components in the heterogeneous system. Satisfactory results are obtained, with experimental execution times close to the lowest achievable.
April 16, 2014 by hgpu
