Using Fermi architecture knowledge to speed up CUDA and OpenCL programs
Dpto. Informatica, Univ. Valladolid, Spain
International Workshop on Heterogeneus Architectures and Computing (ISPA 2012), 2012
@inproceedings{torres2012using,
title={Using Fermi architecture knowledge to speed up CUDA and OpenCL programs},
author={Torres, Y. and Gonzalez-Escribano, A. and Llanos, D.R.},
booktitle={Proc. ISPA},
year={2012}
}
The NVIDIA graphics processing units (GPUs) are playing an important role as general purpose programming devices. The implementation of parallel codes to exploit the GPU hardware architecture is a task for experienced programmers. The threadblock size and shape choice is one of the most important user decisions when a parallel problem is coded. The threadblock configuration has a significant impact on the global performance of the program. While in CUDA parallel programming model it is always necessary to specify the threadblock size and shape, the OpenCL standard also offers an automatic mechanism to take this delicate decision. In this paper we present a study of these criteria for Fermi architecture, introducing a general approach for threadblock choice, and showing that there is considerable room for improvement in OpenCL automatic strategy.
June 13, 2012 by hgpu
