Design and fabrication of a power-efficient AI-Processor for Edge computing

This project will allow us to develop a microelectronic chip prototype of an energy-efficient AI processor. The AI processor exhibits a 1000 times reduction in power profile compared to standard cloud-based GPUs. Tasks that can be transferred from server GPUs to this solution and deploying our AI-Processor solution to EDGE can estimate substantial energy savings when evaluated over ten years under realistic assumptions. We estimate a net reduction of 83K tonnes of CO2 equivalent over ten years, helping Quebec significantly achieve its clean environment objectives. Economically, this will help Aarish Technologies to launch its product with a solid foundation of intellectual property and to grasp the demands of the mar

Faculty Supervisor:

Zeljko Zilic

Student:

Kaustav Das Sharma;Garrett Kinman;Juan Morency Trudel;Mohammad Mustafa Baba

Partner:

Aarish Technologies

Discipline:

Engineering - computer / electrical

Sector:

Professional, scientific and technical services

University:

McGill University