Graph neural networks (GNN) have achieved state-of-the-art performance on various industrial tasks. However, the poor efficiency of GNN inference and frequent Out-Of-Memory (OOM) problem limit the successful application of GNN on edge computing platforms. To tackle these problems, a feature decomposition approach is proposed for memory efficiency optimization of GNN inference. The proposed approach could achieve outstanding optimization on various GNN models, covering a wide range of datasets, which speeds up the inference by up to 3x. Furthermore, the proposed feature decomposition could significantly reduce the peak memory usage (up to 5x in memory efficiency improvement) and mitigate OOM problems during GNN inference.
@article{arxiv.2104.03058,
title = {Optimizing Memory Efficiency of Graph Neural Networks on Edge Computing Platforms},
author = {Ao Zhou and Jianlei Yang and Yeqi Gao and Tong Qiao and Yingjie Qi and Xiaoyi Wang and Yunli Chen and Pengcheng Dai and Weisheng Zhao and Chunming Hu},
journal= {arXiv preprint arXiv:2104.03058},
year = {2021}
}
Comments
This paper has been accepted by RTAS 2021(brief industry track), with link to publicly available code