ANAS: Asynchronous Neuromorphic Hardware Architecture Search Based on System-Level Simulator
DescriptionAsynchronous neuromorphic hardware is emerging for edge computing with high energy efficiency, while it is challenging to find an optimal hardware architecture from the non-numerical design space. To address this problem, we propose an asynchronous neuromorphic hardware architecture search (ANAS) method which uses an evolutionary algorithm to optimize both the numerical and non-numerical design space. Meanwhile, we introduce a configurable asynchronous neuromorphic hardware simulator (CanMore) to offer system-level modeling and performance estimation. Experimental results show that ANAS rivals the best human-designed architectures by 7× EDP reduction, and offers 2.3× EDP reduction than methods that only optimize numerical design space.
Event Type
Research Manuscript
TimeWednesday, July 12th10:40am - 10:55am PDT
Location3010, 3rd Floor
Emerging Models of Computation