Accelerate all-to-all for deep learning with optical switching network architecture

Wenzhe Li; Xinhang Zhang; Guojun Yuan; Zhan Wang; Peiheng Zhang

doi:10.1117/12.3021853

18 March 2024 Accelerate all-to-all for deep learning with optical switching network architecture

Wenzhe Li, Xinhang Zhang, Guojun Yuan, Zhan Wang, Peiheng Zhang

Proceedings Volume 13104, Advanced Fiber Laser Conference (AFL2023); 1310417 (2024) https://doi.org/10.1117/12.3021853
Event: Advanced Fiber Laser Conference (AFL2023), 2023, Shenzhen, China

Abstract

We propose an optical switching network based on WDM technology for distributed training system to accelerate all-to-all communication. A corresponding all-to-all algorithm is proposed to reduce communication delay. Simulation results demonstrate the proposed network and algorithm can achieve up to 103× speedup compared with the traditional approach.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Wenzhe Li, Xinhang Zhang, Guojun Yuan, Zhan Wang, and Peiheng Zhang "Accelerate all-to-all for deep learning with optical switching network architecture", Proc. SPIE 13104, Advanced Fiber Laser Conference (AFL2023), 1310417 (18 March 2024); https://doi.org/10.1117/12.3021853

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available