Magnetic RAM based filter design for low power signal processing in IOT applications

Emerging memory technologies have attracted a lot of interest and has witnessed tremendous innovation in the past couple of years due to the continuous demand for greater performance by modern computing systems. This paper presents the design of Magnetic RAM (MRAM) and its implementation in memory based FIR filter. MRAM has obtained significant interest in the past decade due to its non-volatility property. MRAM stores data using spintronic device Magnetic Tunneling Junction (MTJ). In this paper a mathematical model is developed for MTJ using the LLG equation. A new configuration 3T2MTJ for MRAM is proposed and a dual port MRAM is designed using this configuration. A 9-tap FIR filter is constructed using the proposed dual port MRAM design. The performance of MRAM single cell, single and dual port multipliers and FIR filter is evaluated and compared with the conventional SRAM based system. It can be observed that dual port multiplier using MRAM reduces the power consumption by 77% as compared to SRAM based design and the power consumption in FIR filter is reduced by 37% by using MRAM in the place of SRAM.