Parallel Primitive Optimization for GPU and Multicore
Leopold-Franzens-University Innsbruck
Leopold-Franzens-University Innsbruck, 2013
@article{hormandinger2013parallel,
title={Parallel Primitive Optimization for GPU and Multicore},
author={H{"o}rmandinger, Andreas},
year={2013}
}
This thesis focuses on the use of automatic code generation to combine different classes of optimizations to find the best optimization for parallel reduction in OpenCL on various devices. It also introduces the optimizations used. In the end the results of the combinations will be evaluated and discussed.
September 20, 2014 by hgpu
