Electronic Devices and Semiconductor Manufacturing Commons^™

Characterization Of Low Power Hfo2 Based Switching Devices For In-Memory Computing, Aseel Zeinati

Theses

Oxide based Resistive Random Access Memory (RRAM) devices are investigated as one of the promising non-volatile memories to be used for in-memory computing that will replace the classical von Neumann architecture and reduce the power consumption. These applications required multilevel cell (MLC) characteristics that can be achieved in RRAM devices. One of the methods to achieve this analog switching behavior is by performing an optimized electrical pulse. The RRAM device structure is basically an insulator between two metals as metal-insulator-metal (MIM) structure. Where one of the primary challenges is to assign an RRAM stack with both low power consumption and …

Iii-Nitride Triangular Microcantilevers For Multimodal Sensing Applications, Balaadithya Uppalapati

All Dissertations

Micro-electromechanical systems (MEMS)-based sensors have gained significant attention due to their ability to sense, measure, and process various physical, chemical, and biological parameters. The small size of MEMS sensors provides numerous advantages, including low power consumption, high sensitivity, and rapid response time, making them suitable for various applications in healthcare, automotive, aerospace, and consumer electronics.

In the past few years, AlGaN/GaN MEMS devices have been found to offer several advantages over silicon-based MEMS devices. One of the main advantages of AlGaN/GaN MEMS is their high sensitivity to surface stresses and forces due to their high piezoelectric coefficients. This sensitivity allows …

A Phase Change Memory And Dram Based Framework For Energy-Efficient And High-Speed In-Memory Stochastic Computing, Supreeth Mysore

Theses and Dissertations--Electrical and Computer Engineering

Convolutional Neural Networks (CNNs) have proven to be highly effective in various fields related to Artificial Intelligence (AI) and Machine Learning (ML). However, the significant computational and memory requirements of CNNs make their processing highly compute and memory-intensive. In particular, the multiply-accumulate (MAC) operation, which is a fundamental building block of CNNs, requires enormous arithmetic operations. As the input dataset size increases, the traditional processor-centric von-Neumann computing architecture becomes ill-suited for CNN-based applications. This results in exponentially higher latency and energy costs, making the processing of CNNs highly challenging.

To overcome these challenges, researchers have explored the Processing-In Memory (PIM) …