TY - GEN
T1 - Multiple Factors Congestion Prediction Algorithm for Network-on-Chip
AU - Hu, Zhenyu
AU - Conrad Meyer, Michael
AU - Jiang, Xin
AU - Watanabe, Takahiro
N1 - Funding Information:
This work is partly supported by JSPS KAKENHI Grant Number 18K11226, Waseda University Tokutei-Kadai 2019C-289, and Int’l Sci. Tech. Cooperation Projects (No.2018031).
Publisher Copyright:
© 2020 IEICE.
PY - 2020/7
Y1 - 2020/7
N2 - Networks-on-Chip is a new paradigm to make the interconnections inside a System-on-Chip system, and the traffic in NoC is the most important issue that impacts the pereformance. Therefore congestion detection has become a hot issue in NoC. Congestion aware routing algorithms are created to aware and avoid the congestion area, lower the transmission latency and increase the throughput of the network. In this paper, a history-based congestion prediction algorithm named HCP is proposed, where broadcasted packets are used to carry the congestion information and the information is used to estimate the congestion condition inside the network. Besides the congestion, HCP also considers the historical routed packets and the path diversity to predict the congestion and balance the traffic in the network. Compared with conventional congestion aware routing algorithms, the proposed algorithm shows better results in both latency and throughput.
AB - Networks-on-Chip is a new paradigm to make the interconnections inside a System-on-Chip system, and the traffic in NoC is the most important issue that impacts the pereformance. Therefore congestion detection has become a hot issue in NoC. Congestion aware routing algorithms are created to aware and avoid the congestion area, lower the transmission latency and increase the throughput of the network. In this paper, a history-based congestion prediction algorithm named HCP is proposed, where broadcasted packets are used to carry the congestion information and the information is used to estimate the congestion condition inside the network. Besides the congestion, HCP also considers the historical routed packets and the path diversity to predict the congestion and balance the traffic in the network. Compared with conventional congestion aware routing algorithms, the proposed algorithm shows better results in both latency and throughput.
KW - Congestion Aware
KW - Network on Chips
KW - Routing Algorithm
KW - Selection Strategy
UR - http://www.scopus.com/inward/record.url?scp=85091447044&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85091447044&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85091447044
T3 - ITC-CSCC 2020 - 35th International Technical Conference on Circuits/Systems, Computers and Communications
SP - 211
EP - 216
BT - ITC-CSCC 2020 - 35th International Technical Conference on Circuits/Systems, Computers and Communications
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 35th International Technical Conference on Circuits/Systems, Computers and Communications, ITC-CSCC 2020
Y2 - 3 July 2020 through 6 July 2020
ER -