COIN: Communication-Aware In-Memory Acceleration for Graph Convolutional Networks

TL;DR

This paper introduces COIN, a communication-aware in-memory computing architecture designed to accelerate Graph Convolutional Networks by reducing communication overheads, significantly improving energy efficiency and performance.

Contribution

The paper proposes a novel GCN-specific hardware architecture that minimizes communication overheads through in-memory computing, enhancing energy efficiency and computational speed.

Findings

Up to 105x energy consumption reduction compared to existing accelerators

Effective minimization of intra- and inter-CE communication in GCN processing

Significant performance improvements demonstrated on standard datasets

Abstract

Graph convolutional networks (GCNs) have shown remarkable learning capabilities when processing graph-structured data found inherently in many application areas. GCNs distribute the outputs of neural networks embedded in each vertex over multiple iterations to take advantage of the relations captured by the underlying graphs. Consequently, they incur a significant amount of computation and irregular communication overheads, which call for GCN-specific hardware accelerators. To this end, this paper presents a communication-aware in-memory computing architecture (COIN) for GCN hardware acceleration. Besides accelerating the computation using custom compute elements (CE) and in-memory computing, COIN aims at minimizing the intra- and inter-CE communication in GCN operations to optimize the performance and energy efficiency. Experimental evaluations with widely used datasets show up to 105x improvement in energy consumption compared to state-of-the-art GCN accelerator.