Name
Title | ||
|---|---|---|
FusedGCN: A Systolic Three-Matrix Multiplication Architecture for Graph Convolutional Networks |
Abstract | ||
|---|---|---|
Machine-learning applications have garnered widespread adoption over the last several years. Graph Neural Networks have been proposed as an extension of machine-learning models to graph-structured data. The training and inference tasks on graph neural networks involve graph convolution operations that can be equivalently expressed as three-matrix multiplications. In this work, we propose FusedGCN, a custom systolic architecture that computes in a fused, i.e., combined, manner the product of three matrices. FusedGCN supports compressed sparse representations and tiled computations, which allow the design to adapt to the available input/output bandwidth without losing the regularity of a systolic architecture. The experimental results show that FusedGCN achieves lower execution times than the current best-performing state-of-the-art architecture for computing representative GCN applications. Most importantly, this result is achieved by consuming only marginally more area/power than a traditional systolic array used for two-matrix multiplications. |
Year | DOI | Venue |
|---|---|---|
2022 | 10.1109/ASAP54787.2022.00024 | 2022 IEEE 33rd International Conference on Application-specific Systems, Architectures and Processors (ASAP) |
Keywords | DocType | ISSN |
Graph Neural Networks,Systolic arrays,Graph Convolution,Machine Learning Accelerators | Conference | 2160-0511 |
ISBN | Citations | PageRank |
978-1-6654-8309-4 | 0 | 0.34 |
References | Authors | |
11 | 4 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Christodoulos Peltekis | 1 | 0 | 0.34 |
| Dionysios Filippas | 2 | 0 | 0.34 |
| Chrysostomos Nicopoulos | 3 | 0 | 0.34 |
| Giorgos Dimitrakopoulos | 4 | 215 | 27.31 |