WireLightning: Harnessing Capacitances for In-Transit Massively Parallel Matrix Multiplication
Zoom Link: https://hku.zoom.us/j/3837289217?omn=95617077246 Abstract Analog computing-in-memory accelerators promise ultra-low-power, on-device AI by reducing data transfer and energy usage. Yet inherent device variations and high energy consumption for analog-digital conversion continue […]