Extending Energy-Efficient and Scalable DNN Training and Inference With 3-D Photonic Accelerator
- 23 July 2025
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Emerging and Selected Topics in Circuits and Systems
- Vol. 15 (4) , 560-576
Abstract
As deep neural network (DNN) models continue to grow in complexity, analog computing architectures have emerged as a promising solution to meet increasing computational demands. Among these, silicon photonic computing excels at efficiently executing dot product operations while leveraging inherent parallelism. Photonic phase change memory (photonic-PCM) further enhances photonic computing by enabling scalable, non-volatile storage. In this work, we introduce the 3D Large-Scale Photonic Accelerator (LSPA), a novel photonic computing architecture designed for large-scale DNN models. LSPA employs multi-layered 3D stacking of non-volatile photonic-PCM cells, creating a high-density computational fabric that optimizes energy efficiency, flexibility, and scalability. LSPAs custom 3D photonic network enables simultaneous data multicast in two dimensions and accumulation in three dimensions, optimizing communication patterns essential for efficient DNN training. A distinctive feature of LSPA is its ability to execute multiple forward and backward passes in parallel within each mini-batch, reducing latency associated with data movement and photonic-PCM programming. This unique capability combined with high-bandwidth photonic interconnects allows LSPA to sustain efficient training across a wide range of DNN workloads. When evaluated against a range of neural network models including VGG-16, ResNet-50, GoogLeNet, Transformer, GNMT, LLaMA 7B, and LLaMA 30B, LSPA reduces execution time by up to 92% and energy consumption by up to 90%. These results highlight LSPA as a transformative advancement in scalable, high-performance photonic computing for deep learning.Keywords
Funding Information
- NSF (CCF-2311543)
- NSF (23224644)
- NSF (CCF-213946)
- NSF (CCF-1936794)
- NSF (CCF-2324645)
- NSF (CCF-2311544)
This publication has 68 references indexed in Scilit:
- Phase change material thermal energy storage systems for cooling applications in buildings: A reviewRenewable and Sustainable Energy Reviews, 2020
- A phase‐field method for modeling cracks with frictional contactInternational Journal for Numerical Methods in Engineering, 2019
- Increasing the density of passive photonic-integrated circuits via nanophotonic cloakingNature Communications, 2016
- Training and operation of an integrated neuromorphic network based on metal-oxide memristorsNature, 2015
- Pattern classification by memristive crossbar circuits using ex situ and in situ trainingNature Communications, 2013
- Memristive devices for computingNature Nanotechnology, 2012
- Silicon microring resonatorsLaser & Photonics Reviews, 2011
- On the genesis and evolution of Integrated Quantum OpticsLaser & Photonics Reviews, 2011
- Phase-change materials for rewriteable data storageNature Materials, 2007
- Micrometre-scale silicon electro-optic modulatorNature, 2005