ReGraphX: NoC-enabled 3D Heterogeneous ReRAM Architecture for Training Graph Neural Networks

TL;DR

ReGraphX is a novel 3D heterogeneous ReRAM-based architecture designed for efficient GNN training, combining ReRAM crossbars and multicast-enabled NoC to outperform GPUs in speed and energy efficiency.

Contribution

This work introduces ReGraphX, the first architecture to integrate ReRAM crossbars with a 3D NoC specifically optimized for GNN training.

Findings

ReGraphX achieves up to 3.5X faster training than GPUs.

ReGraphX reduces energy consumption by up to 11X.

ReGraphX effectively handles the dual requirements of DNN and graph computations.

Abstract

Graph Neural Network (GNN) is a variant of Deep Neural Networks (DNNs) operating on graphs. However, GNNs are more complex compared to traditional DNNs as they simultaneously exhibit features of both DNN and graph applications. As a result, architectures specifically optimized for either DNNs or graph applications are not suited for GNN training. In this work, we propose a 3D heterogeneous manycore architecture for on-chip GNN training to address this problem. The proposed architecture, ReGraphX, involves heterogeneous ReRAM crossbars to fulfill the disparate requirements of both DNN and graph computations simultaneously. The ReRAM-based architecture is complemented with a multicast-enabled 3D NoC to improve the overall achievable performance. We demonstrate that ReGraphX outperforms conventional GPUs by up to 3.5X (on an average 3X) in terms of execution time, while reducing energy consumption by as much as 11X.