Compute Unified Device Architecture Application Suitability
University of Illinois
Computing in Science and Engg., Vol. 11, No. 3. (2009), pp. 16-26
@article{hwu2009compute,
title={Compute unified device architecture application suitability},
author={Hwu, W.M. and Rodrigues, C. and Ryoo, S. and Stratton, J.},
journal={Computing in Science & Engineering},
volume={11},
number={3},
pages={16–26},
issn={1521-9615},
year={2009},
publisher={IEEE}
}
Graphics processing units (GPUs) can provide excellent speedups on some, but not all, general-purpose workloads. Using a set of computational GPU kernels as examples, the authors show how to adapt kernels to utilize the architectural features of a GeForce 8800 GPU and what finally limits the achievable performance.
November 25, 2010 by hgpu
