Gemini: Bridging Tomorrow’s Deep Neural Network Frontiers with Unrivaled Chiplet Accelerator Mastery

A research team introduces Gemini, an innovative framework, focusing on both architecture and mapping co-exploration, aims to propel large-scale DNN chiplet accelerators to new heights, achieving an impressive average performance improvement of 1.98× and an energy efficiency boost of 1.41× compared to the state-of-the-art Simba architecture.

accelerator accelerators ai architecture art artificial intelligence boost chip design deep neural network deep-neural-networks dnn efficiency energy energy efficiency exploration framework frontiers gemini improvement machine learning machine learning & data science mapping ml network neural network performance research research team scale state team technology