학술논문
Lightspeed Binary Neural Networks using Optical Phase-Change Materials
Document Type
Conference
Author
Source
2023 Design, Automation & Test in Europe Conference & Exhibition (DATE) Design, Automation & Test in Europe Conference & Exhibition (DATE), 2023. :1-2 Apr, 2023
Subject
Language
ISSN
1558-1101
Abstract
This paper investigates the potential of a compute-in-memory core based on optical Phase Change Materials (oPCMs) to speed up and reduce the energy consumption of the Matrix-Matrix-Multiplication operation. The paper also proposes a new data mapping for Binary Neural Networks (BNNs) tailored for our oPCM core. The preliminary results show a significant latency improvement irrespective of the evaluated network structure and size. The improvement varies from network to network and goes up to ~1053x.