EXACT: Scalable Graph Neural Networks Training via Extreme Activation Compression
Training Graph Neural Networks (GNNs) on large graphs is a fundamental challenge due to the high memory usage, which is mainly occupied by activations (e.g., node embeddings). Previous works usually focus on reducing the number of nodes retained in memory. In parallel, unlike what has been developed for other types of neural networks, training with compressed activation maps is less explored for GNNs. This extension is notoriously difficult to implement due to the miss of necessary tools in common graph learning packages. To unleash the potential of this direction, we provide a optimized GPU implementation which supports training GNNs with compressed activations. Based on the implementation, we propose a memory-efficient framework called ``EXACT'', which for the first time demonstrate the potential and evaluate the feasibility of training GNNs with compressed activations. We systematically analyze the trade-off among the memory saving, time overhead, and accuracy drop. In practice, EXACT can reduce the memory footprint of activations by up to $32\times$ with $0.2$-$0.5\%$ accuracy drop and $10$-$25\%$ time overhead across different models and datasets. In addition, EXACT can trim down the hardware requirement of training a three-layer full-batch GraphSAGE on \textit{ogbn-products} from a 48GB GPU to a 12GB GPU.
PDF Abstract
OGB
Reddit
