MEGA.mini: A NPU with Novel Heterogeneous AI Processing Architecture Balancing Efficiency, Performance, and Intelligence for the Era of Generative AI

Download PDF
Open Webpage

Donghyeon Han, Anantha P. Chandrakasan

Published: 2025, Last Modified: 23 Jan 2026HCS 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: NPU with a Novel big.LITTLECore Architecture to Balance 3 Key Aspects of AI Acceleration–Efficiency: > 95% computations w/ Low-precision FXP–Performance: 3 hierarchical solutions @ MEGA+mini–Intelligence: Hybrid IA (FP for < 5% outlier data)