Self-Terminating Write of Multi-Level Cell ReRAM for Efficient Neuromorphic Computing
Zongwu Wang,Zhezhi He,Rui Yang,Shiquan Fan,Jie Lin,Fangxin Liu,Yueyang Jia,Chenxi Yuan,Qidong Tang,Li Jiang
DOI: https://doi.org/10.23919/date54114.2022.9774550
2022-01-01
Abstract:The Resistive Random-Access-Memory (ReRAM) in crossbar structure has shown great potential in accelerating the vector-matrix multiplication, owing to the fascinating computing complexity reduction (from O(n 2 ) to O(1) ). Nevertheless, the ReRAM cells still encounter device programming variation and resistance drifting during computation (known as read disturbance), which significantly hamper its analog computing precision. Inspired by prior precise memory programming works, we propose a Self-Terminating Write (STW) circuit for Multi-Level Cell (MLC) ReRAM. In order to minimize the area overhead, the design heavily reuses inherent computing peripherals (e.g., Analog-to-Digital Converter and Trans-Impedance Amplifier) in conventional dot-product engine. Thanks to the fast and precise programming capability of our design, the ReRAM cell can possess 4 linear distributed conductance levels, with minimum latency used for intermediate resistance refreshing. Our comprehensive cross-layer (device/circuit/architecture) simulation indicates that the proposed MLC STW scheme can effectively obtain 2-bit precision via a single programming pulse. Besides, our design outperforms the prior write&verify schemes by 4.7× and 2× in programming latency and energy, respectively.