A Hybrid Tensor-Expert-Data Parallelism Approach to Optimize Mixture-of-Experts Training

Thumbnail Image


Singh et al.pdf (731.24 KB)
No. of downloads: 20

Publication or External Link





Siddharth Singh, Olatunji Ruwase, Ammar Ahmad Awan, Samyam Rajbhandari, Yuxiong He, and Abhinav Bhatele. 2023. A Hybrid Tensor-Expert-Data Parallelism Approach to Optimize Mixture-of-Experts Training. In 2023 International Conference on Supercomputing (ICS ’23), June 21–23, 2023, Orlando, FL, USA. ACM, New York, NY, USA, 12 pages.


Mixture-of-Experts (MoE) is a neural network architecture that adds sparsely activated expert blocks to a base model, increasing the number of parameters without impacting computational costs. However, current distributed deep learning frameworks are limited in their ability to train high-quality MoE models with large base models. In this work, we present DeepSpeed-TED, a novel, threedimensional, hybrid parallel algorithm that combines data, tensor, and expert parallelism to enable the training of MoE models with 4–8× larger base models than the current state-of-the-art. We also describe memory optimizations in the optimizer step, and communication optimizations that eliminate unnecessary data movement. We implement our approach in DeepSpeed and achieve speedups of 26% over a baseline (i.e. without our communication optimizations) when training a 40 billion parameter MoE model (6.7 billion base model with 16 experts) on 128 V100 GPUs.