The research results of Associate Professor Jiang Li's research group from Shanghai Jiaotong University Electric Power Institute won the DATE Best Paper Award

Recently, the European Conference on Design, Automation and Test in Europe (DATE) was successfully held in Antwerp, Belgium, under the guidance of Associate Professor Jiang Li and Assistant Professor He Zhezhi of the Advanced Computer Architecture Laboratory of the School of Electronic Information and Electrical Engineering.

The title of the award-winning paper is "Self-Terminated Write of Multi-Level Cell ReRAM for Efficient Neuromorphic Computing" , which proposes a number of solutions for the high-resistance instability common in existing memristor-based memory and computing systems.

Research Background

The memory-computing unit based on ReRAM memristive devices can significantly reduce the computational complexity of large-scale VMM operations in neural network inference from O(n2) to O(1), and greatly reduce the time and energy consumption of large-scale data transfer The overhead is expected to break the von Neumann bottleneck existing in the traditional architecture based on this technology

Research results

Since the Jingang Wu team of Fudan University proposed the Self-Adaptive Write Mode (SAWM) programming scheme on the VLSI Symposium in 2012 to realize single-bit precision programming, research teams from various countries have continued to improve the scheme

Left: Schematic diagram of adaptive circuit implementation; Right: circuit simulation waveform diagram

The above schematic diagrams show the schematic diagram of the designed circuit and the Spice simulation results respectively.

Monte Carlo Simulation Results of ReRAM Programming Resistance Distribution

European Conference on Design, Automation and Test in Europe (DATE)

DATE is a theme conference on electronic design automation held every year in Europe.

Paper link: https://github.