4.5 Article

LETHOR: a thermal-aware proactive routing algorithm for 3D NoCs with less entrance to hot regions

Journal

JOURNAL OF SUPERCOMPUTING
Volume 78, Issue 6, Pages 8194-8218

Publisher

SPRINGER
DOI: 10.1007/s11227-021-04207-3

Keywords

3D NoC; Hot region; Proactive thermal-aware routing; Thermal management

Funding

  1. Shahid Rajaee Teacher Training Univesity

Ask authors/readers for more resources

This paper proposes an efficient proactive thermal-aware routing algorithm called LETHOR to reduce the temperature of 3D NoCs. By considering the temperature information of all nodes in each layer, LETHOR can decrease the standard deviation of chip temperature, statistical traffic load distribution, and global average delay compared to the state-of-the-art routing algorithm.
Although many Dynamic Thermal Management (DTM) techniques are employed to overcome thermal problems in 3D NoCs, none of them consider temperature information of all nodes of a layer at the same time, so that they cannot reduce the temperature of the network properly.To overcome this problem, this paper proposes an efficient proactive thermal-aware routing algorithm, called Less Entrance to Hot Regions (LETHOR), to keep the NoC temperature lower than a predefined thermal limit. LETHOR routes the network packets based on the temperature information of all nodes in the layers instead of considering only the neighbor nodes in each hop. To this aim, LETHOR introduces a Hot Region in each layer based on the gathered temperature information to determine the suitable routing path. Based on the gathered information, LETHOR examines available paths and selects the cooler and less crowded one. Experimental results show that the LETHOR routing algorithm decreases the standard deviation of chip temperature by 13.24-47.40% and the statistical traffic load distribution by 13.62-48.34%, under different traffic patterns. Moreover, the global average delay is reduced by 23.07-44.13% under different traffic patterns compared to the state-of-the-art routing algorithm.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.5
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available