Md Maatar の履歴(No.6)

Members-Internal

Thermal-power Aware Mapping Method and Architecture/Design for (Fault-tolerant) 3D-NoC-based Neuromorphic Systems

Background†

(This is still a draft) Artificial intelligence, neural networks, deep learning, ANNs, SNNs, neuromorphic systems and computing, reduce energy cost and higher bandwith (speed).

Applications: edge computing (IoT), real time application, devices with limited energy supply.

↑

Motivation†

In the last decade, neural networks and deep learning had improved and gained a lot of interest.

Thermal-aware mapping in neuromorphic computing.

In conventional computers NoC (architecture), thermal awarness researches is advanced but not for neuromorphic HW specially 3D-NoC neuromorphic systems.

Take into concideration the heat generated from neurons' activities in the neuuromorphic hardware and how to make it (spread) equally on the chip.

↑

Goal†

↑

Research Plan†

Step 1: 1- Read and gain knowledge about neuromorphic systems 2- Survey on Routing algorithms for neuromorphic systems

Step2 (doing): 1- Document on mapping and thermal-awarness in NoC 2- Survey thermal-aware mapping and routing for 3D-NoC neuromorphic systems 3- Propose a thermal-aware mapping method for NASH

↑

Research Schedule†

Date	Task
☑️Oct 1st - Dec 31, 2021	Tutorials and reading about neuromorphic systems
Jan 1st - May 31, 2022	Survey on Routing methods and algorithms for neuromorphic systems
☑️Feb 1st - May 16, 2022	Understand SNPC Source code and do the Tutorials

↑

My Shared GDrive†

My-Shared-GoogleDrive

Ebooks

↑

Key references†

"Thermal-aware application mapping strategy for network-on-chip based system design." IEEE Transactions on Computers 2017
"Thermal-aware IP virtualization and placement for networks-on-chip architecture." IEEE International Conference on Computer Design 2004
"Thermal-aware mapping and placement for 3-D NoC designs." Proceedings 2005 IEEE International SOC Conference