swDNN: A Library for Accelerating Deep Learning Applications on Sunway TaihuLight

Abstract

To explore the potential of training complex deep neural networks (DNNs) on other commercial chips rather than GPUs, we report our work on swDNN, which is a highly-efficient library for accelerating deep learning applications on the newly announced world-leading supercomputer, Sunway TaihuLight. Targeting SW26010 processor, we derive a performance model… (More)
DOI: 10.1109/IPDPS.2017.20

Topics

11 Figures and Tables

Cite this paper

@article{Fang2017swDNNAL, title={swDNN: A Library for Accelerating Deep Learning Applications on Sunway TaihuLight}, author={Jiarui Fang and Haohuan Fu and Wenlai Zhao and Bingwei Chen and Weijie Zheng and Guangwen Yang}, journal={2017 IEEE International Parallel and Distributed Processing Symposium (IPDPS)}, year={2017}, pages={615-624} }