swdnn: A library for accelerating deep learning applications on sunway taihulight

J Fang, H Fu, W Zhao, B Chen… - 2017 IEEE …, 2017 - ieeexplore.ieee.org
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 that guides us in the process of identifying the most suitable approach
for mapping the convolutional neural networks (CNNs) onto the 260 cores within the chip. By …

[CITATION][C] swDNN: Alibraryfor acceleratingdeeplearningapplicationsonsunwaytaihulight

FH FangJ - … DistributedProcessingSymposium. Orlando, USA, 2017
Showing the best results for this search. See all results