Graph Neural Networks (GNNs) is a promising approach for applications with nonEuclidean data. However, training GNNs on large scale graphs with hundreds of millions nodes is both resource and time consuming. Different from DNNs, GNNs usually have larger memory footprints, and thus the GPU memory capacity and PCIe bandwidth are the main resource bottlenecks in GNN training. To address this problem, we present BiFeat: a graph feature quantization methodology to accelerate GNN training by significantly reducing the memory footprint and PCIe bandwidth requirement so that GNNs can take full advantage of GPU computing capabilities. Our key insight is that unlike DNN, GNN is less prone to the information loss of input features caused by quantization. We identify the main accuracy impact factors in graph feature quantization and theoretically prove that BiFeat training converges to a network where the loss is within $\epsilon$ of the optimal loss of uncompressed network. We perform extensive evaluation of BiFeat using several popular GNN models and datasets, including GraphSAGE on MAG240M, the largest public graph dataset. The results demonstrate that BiFeat achieves a compression ratio of more than 30 and improves GNN training speed by 200%-320% with marginal accuracy loss. In particular, BiFeat achieves a record by training GraphSAGE on MAG240M within one hour using only four GPUs.
翻译:神经网图( GNN) 是一个很有希望的方法, 用于使用非 Euclidean 数据的应用。 但是, 以数亿节点的大型图形培训 GNNN 既消耗资源,也消耗时间。 不同于 DNNN, GNN通常有较大的内存足迹, 因此GPU内存能力和 PCIe 带宽是GNN 培训中的主要资源瓶颈。 为了解决这一问题, 我们介绍了 BiFeat: 一个图形特征量化方法, 以通过大量减少记忆足迹和 PCIe 带宽要求来加速 GNNNN 培训, 从而让 GPU 充分利用 GPU 计算能力。 我们的关键洞察力是, 不同于 DNNNN, GNN, GNNN, GNF, 与 DNG 相比, GNF, GM 最大的数字率比 GNF 高。 我们从图表中找出图形中找到主要精确度影响因素因素。 我们从理论上证明BF 损失低于 GNF 的四度 GV 数据, 中, 显示具体的GNF 。