Network-on-Chip Congestion Prediction Using Spatiotemporal Graph Neural Networks and a Neo4j Graph Database

doi:10.5281/zenodo.22159229

ICTA 2026, Springer Lecture Notes in Networks and Systems (LNNS)

Network-on-Chip Congestion Prediction Using Spatiotemporal Graph Neural Networks and a Neo4j Graph Database

Tien-Hieu Le^1,2 , Duy-Hieu Bui¹, Xuan-Tu Tran¹

1VNU Information Technology Institute, Vietnam National University, Hanoi (VNU-ITI), Vietnam
2Microelectronics and Computer Science (MCS) Research Center, N.A Viet Nam Ltd., Nghe An, Vietnam

Abstract

Reproducibility artifact for an offline framework that predicts short-term per-router Network-on-Chip congestion, integrating a customised BookSim2 simulator, a Neo4j unified graph store for static topology and dynamic state, and a multi-seed comparison of topology-aware and topology-agnostic next-step buffer-occupancy predictors.

Record

Persistent URL: https://research.letienhieu.com/noc-congestion
DOI: 10.5281/zenodo.22159229
Venue: ICTA 2026, Springer Lecture Notes in Networks and Systems (LNNS)

This URL is permanent. The resources behind it may move; the address in the paper will not change.

Cite this work

Le, T.-H., Bui, D.-H., & Tran, X.-T. (2026). Network-on-Chip Congestion Prediction Using Spatiotemporal Graph Neural Networks and a Neo4j Graph Database [Software]. Zenodo. https://doi.org/10.5281/zenodo.22159229

BibTeX

@software{le2026noc_congestion,
  author    = {Le, Tien-Hieu and Bui, Duy-Hieu and Tran, Xuan-Tu},
  title     = {Network-on-Chip Congestion Prediction Using Spatiotemporal Graph Neural Networks and a Neo4j Graph Database},
  year      = {2026},
  publisher = {Zenodo},
  version   = {v1.0.0},
  doi       = {10.5281/zenodo.22159229},
  url       = {https://research.letienhieu.com/noc-congestion},
  license   = {MIT}
}