Variation-tolerant, low-power, and high endurance read scheme for memristor memories
- PDF / 2,300,349 Bytes
- 16 Pages / 595.276 x 790.866 pts Page_size
- 49 Downloads / 157 Views
Variation-tolerant, low-power, and high endurance read scheme for memristor memories V. Ravi1
•
K. Chitra1 • S. R. S. Prabaharan2
Received: 16 January 2020 / Revised: 16 January 2020 / Accepted: 31 July 2020 Ó Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Reading the memristor memory cell without changing its resistance state is one of the potential problems to be addressed in the memristor-based memory design. This paper presents a novel read scheme that achieves a non-destructive read operation, consumes less power, provides high endurance and adapts itself based on the process variations. The proposed scheme uses built-in self-tuning circuitry to obtain the optimum amplitude and width of the refresh pulse required to completely retrieve the state of the memristor after the read cycle. As the scheme uses refresh pulse only when needed, the scheme saves nearly 50% of average power when compared with a conventional fixed pulse read method. The self-tuning circuits are validated by a generic, accurate, and efficient ‘‘voltage threshold adaptive memristor’’ model. The validation results prove that the proposed tuning circuitry achieves optimum refresh pulse size under various read disturbance faults. Graphic abstract
Keywords Memristor Endurance Variation-tolerant Low-power Reliability
1 Introduction & V. Ravi [email protected] 1
School of Electronics Engineering, Vellore Institute of Technology, Chennai, India
2
SRM Research Institute, SRM Institute of Science and Technology, Kattankulathur, India
Memristor is a two-terminal, non-linear, passive, and fundamental circuit component that relates flux (U) and charge (q). It was theoretically postulated by Chua [27] and experimentally fabricated by Hewlett Packard Enterprise [42]. Memristor stores data in the form of resistance and
123
Analog Integrated Circuits and Signal Processing
the resistance offered by the device depends on the amplitude, duration, and polarity of the applied electrical bias. Memristors are presently investigated for various applications such as non-volatile memory [13, 20], neural network [36], signal processing [31, 33], analog [2, 28], digital [14], in-memory [19], reconfigurable and neuromorphic computations [3, 21]. Based on the material used, the memristor can be classified as metal oxide memristor [42], ionic (polymeric) memristor [4], spintronic memristor [43], ferroelectric memristor [5], and Carbon nanotube memristor [1]. Among all the materials, metal oxide-based devices demonstrate good data retention, endurance (number of successful write/read operations that can be performed on a single cell without a major change in the state of the cell) and reliability [23, 41]. Knowm Inc. is the first industry to bring the commercial version of the memristor into the market [35]. Other key industries namely HP, Crossbar Inc., AMD Inc., Hynix, Samsung, Rambus, IBM, Micron Technology, and Panasonic Corporation are working on to bring the memristor-based memory into the market [29, 40]. Though the m
Data Loading...
