English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Graph neural networks (GNNs) emerge as a powerful approach to processnon-euclidean data structures and have been proved powerful in variousapplication domains such as social networks and e-commerce. While such graphdata maintained in real-world systems can be extremely large and sparse, thusemploying GNNs to deal with them requires substantial computational and memoryoverhead, which induces considerable energy and resource cost on CPUs and GPUs.In this work, we present a specialized accelerator architecture, EnGN, toenable high-throughput and energy-efficient processing of large-scale GNNs. Theproposed EnGN is designed to accelerate the three key stages of GNNpropagation, which is abstracted as common computing patterns shared by typicalGNNs. To support the key stages simultaneously, we propose thering-edge-reduce(RER) dataflow that tames the poor locality ofsparsely-and-randomly connected vertices, and the RER PE-array to practice RERdataflow. In addition, we utilize a graph tiling strategy to fit large graphsinto EnGN and make good use of the hierarchical on-chip buffers throughadaptive computation reordering and tile scheduling. Overall, EnGN achievesperformance speedup by 1802.9X, 19.75X, and 2.97X and energy efficiency by1326.35X, 304.43X, and 6.2X on average compared to CPU, GPU, and astate-of-the-art GCN accelerator HyGCN, respectively.

EnGN: A High-Throughput and Energy-Efficient Accelerator for Large Graph Neural Networks