MNSIM 2.0: A Behavior-Level Modeling Tool for Memristor-based Neuromorphic Computing Systems

• DOI: 10.1145/3386263.3407647
• Publication Date: 2020-09-04T17:34:23Z
• Authors: Zhenhua Zhu; Hanbo Sun; Kaizhong Qiu; Lixue Xia; Gokul Krishnan; Guohao Dai; Dimin Niu; Xiaoming Chen; X. Sharon Hu; Yu Cao; Yuan Xie; Yu Wang; Huazhong Yang
• Keywords: Neuromorphic engineering; Memristor; Design space exploration

Memristor based neuromorphic computing systems give alternative solutions to boost the energy efficiency of Neural Network (NN) algorithms. Because large-scale applications and large architecture design space, many factors will affect accuracy system's performance. In this work, we propose a behavior-level modeling tool for memristor-based systems, MNSIM 2.0, model performance help researchers realize an early-stage space exploration. Compared with former version other benchmarks, 2.0 has following new features: 1. algorithm level, supports inference simulation mixed-precision NNs considering non-ideal factors. 2. hierarchical structure PIM is proposed. Users can customize their designs from aspects devices, interfaces, processing units, buffer designs, interconnections. 3. Two hardware-aware optimization methods are integrated in software-hardware co-optimization.