Performance Tradeoff Spectrum of Integer and Floating Point Applications Kernels on Various GPUs
Computer Science, Massey University, North Shore 102-904, Auckland, New Zealand
The 2013 International Conference on Computer Design (CDES’13), 2013
@inproceedings{johnson2013performance,
title={Performance tradeoff spectrum of integer and floating point applications kernels on various gpus},
author={Johnson, MGB and Playne, DP and Hawick, KA},
booktitle={Proc. 13th International Conference on Computer Design (CDES’13), number CSTN-180, WorldComp, p. CDE4086},
year={2013}
}
Floating point precision and performance and the ratio of floating point units to integer processing elements on a graphics processing unit accelerator all continue to present complex tradeoffs for optimising core utilisation on modern devices. We investigate various hybrid CPU and GPU combinations using a range of different GPU models occupying different points in this tradeoff space. We analyse some performance data for a range of numerical simulation kernels and discuss their use as benchmark problems for characterising such devices.
March 6, 2014 by hgpu
