HFGCN: High-speed and Fully-optimized GCN Accelerator

graph convolutional network (GCN) is a type of neural network that inference new nodes based on the connectivity of the graphs. GCN requires high-calculation volume for processing, similar to other neural networks requiring significant calculation. In this paper, we propose a new hardware architectu...

Full description

Saved in:
Bibliographic Details
Main Authors: Han, MinSeok, Kim, Jiwan, Kim, Donggeon, Jeong, Hyunuk, Jung, Gilho, Oh, Myeongwon, Lee, Hyundong, Go, Yunjeong, Kim, HyunWoo, Kim, Jongbeom, Song, Taigon
Format: Conference Proceeding
Language:English
Subjects:
Convolution
Decoding
Delays
Hardware
Neural networks
Process control
Systolic arrays
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:graph convolutional network (GCN) is a type of neural network that inference new nodes based on the connectivity of the graphs. GCN requires high-calculation volume for processing, similar to other neural networks requiring significant calculation. In this paper, we propose a new hardware architecture for GCN that tackles the problem of wasted cycles during processing. We propose a new scheduler module that reduces memory access through aggregation and an optimized systolic array with improved delay. We compare our study with the state-of-the-art GCN accelerator and show outperforming results.
ISSN:1948-3295
DOI:10.1109/ISQED57927.2023.10129340