# FSMoE: A Flexible and Scalable Training System for Sparse Mixture-of-Experts Models

arXiv: https://arxiv.org/abs/2501.10714 · HTML: https://mlsanity.com/paper/2501.10714/

## Key facts

- Cited by 17 papers in the map (data through 14 Oct 2026).
- Placed in Efficient MoE training and inference systems, one of 12 directions in Mixture-of-experts.
- Builds on 33 of its 55 references that are in the map.
- First posted on arXiv 18 Jan 2025.

## Placed in

- Mixture-of-experts / [Efficient MoE training and inference systems](https://mlsanity.com/moe/efficient-moe-training-and-inference-systems/)

## Builds on (most cited first, 30 of 33)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [PaLM: Scaling Language Modeling with Pathways](https://mlsanity.com/paper/2204.02311/) (2022)
- [Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer](https://mlsanity.com/paper/1701.06538/) (2017)
- [Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity](https://mlsanity.com/paper/2101.03961/) (2021)
- [PaLM-E: An Embodied Multimodal Language Model](https://mlsanity.com/paper/2303.03378/) (2023)
- [Mixtral of Experts](https://mlsanity.com/paper/2401.04088/) (2024)
- [GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding](https://mlsanity.com/paper/2006.16668/) (2020)
- [DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model](https://mlsanity.com/paper/2405.04434/) (2024)
- [Large Batch Optimization for Deep Learning: Training BERT in 76 minutes](https://mlsanity.com/paper/1904.00962/) (2019)
- [Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LM](https://mlsanity.com/paper/2104.04473/) (2021)
- [Mixture-of-Experts with Expert Choice Routing](https://mlsanity.com/paper/2202.09368/) (2022)
- [DeepSpeed Inference: Enabling Efficient Inference of Transformer Models at Unprecedented Scale](https://mlsanity.com/paper/2207.00032/) (2022)
- [DeepSpeed-MoE: Advancing Mixture-of-Experts Inference and Training to Power Next-Generation AI Scale](https://mlsanity.com/paper/2201.05596/) (2022)
- [BASE Layers: Simplifying Training of Large, Sparse Models](https://mlsanity.com/paper/2103.16716/) (2021)
- [Tutel: Adaptive Mixture-of-Experts at Scale](https://mlsanity.com/paper/2206.03382/) (2022)
- [On the Representation Collapse of Sparse Mixture of Experts](https://mlsanity.com/paper/2204.09179/) (2022)
- [Taming Sparsely Activated Transformer with Stochastic Experts](https://mlsanity.com/paper/2110.04260/) (2021)
- [FastMoE: A Fast Mixture-of-Expert Training System](https://mlsanity.com/paper/2103.13262/) (2021)
- [Accelerating Distributed MoE Training and Inference with Lina](https://mlsanity.com/paper/2210.17223/) (2022)
- [StableMoE: Stable Routing Strategy for Mixture of Experts](https://mlsanity.com/paper/2204.08396/) (2022)
- [Highly Scalable Deep Learning Training System with Mixed-Precision: Training ImageNet in Four Minutes](https://mlsanity.com/paper/1807.11205/) (2018)
- [FLUX: Fast Software-based Communication Overlap On GPUs Through Kernel Fusion](https://mlsanity.com/paper/2406.06858/) (2024)
- [FlexMoE: Scaling Large-scale Sparse Pre-trained Model Training via Dynamic Device Placement](https://mlsanity.com/paper/2304.03946/) (2023)
- [SpeechMoE: Scaling to Large Acoustic Models with Dynamic Routing Mixture of Experts](https://mlsanity.com/paper/2105.03036/) (2021)
- [A Hybrid Tensor-Expert-Data Parallelism Approach to Optimize Mixture-of-Experts Training](https://mlsanity.com/paper/2303.06318/) (2023)
- [Breaking the Computation and Communication Abstraction Barrier in Distributed Machine Learning Workloads](https://mlsanity.com/paper/2105.05720/) (2021)
- [HetuMoE: An Efficient Trillion-scale Mixture-of-Expert Distributed Training System](https://mlsanity.com/paper/2203.14685/) (2022)
- [PIT: Optimization of Dynamic Sparse Deep Learning Models via Permutation Invariant Transformation](https://mlsanity.com/paper/2301.10936/) (2023)
- [Gating Dropout: Communication-efficient Regularization for Sparsely Activated Transformers](https://mlsanity.com/paper/2205.14336/) (2022)

## Built on by (most cited first, 17 of 17)

- [UltraEP: Unleash MoE Training and Inference on Rack-Scale Nodes with Near-Optimal Load Balancing](https://mlsanity.com/paper/2606.04101/) (2026)
- [SYMI: Efficient Mixture-of-Experts Training via Model and Optimizer State Decoupling](https://mlsanity.com/paper/2504.19925/) (2025)
- [DynaMo: Runtime Switchable Quantization for MoE with Cross-Dataset Adaptation](https://mlsanity.com/paper/2503.21135/) (2025)
- [HierMoE: Accelerating MoE Training with Hierarchical Token Deduplication and Expert Swap](https://mlsanity.com/paper/2508.09591/) (2025)
- [FlowMoE: A Scalable Pipeline Scheduling Framework for Distributed Mixture-of-Experts Training](https://mlsanity.com/paper/2510.00207/) (2025)
- [HybridEP: Scaling Expert Parallelism to Cross-Datacenter Scenario via Hybrid Expert/Data Transmission](https://mlsanity.com/paper/2510.19470/) (2025)
- [10Cache: Heterogeneous Resource-Aware Tensor Caching and Migration for LLM Training](https://mlsanity.com/paper/2511.14124/) (2025)
- [PC-MoE: Memory-Efficient and Privacy-Preserving Collaborative Training for Mixture-of-Experts LLMs](https://mlsanity.com/paper/2506.02965/) (2025)
- [Efficient MoE Inference with Fine-Grained Scheduling of Disaggregated Expert Parallelism](https://mlsanity.com/paper/2512.21487/) (2025)
- [ZipCCL: Efficient Lossless Data Compression of Communication Collectives for Accelerating LLM Training](https://mlsanity.com/paper/2604.27844/) (2026)
- [Multi-level context Modeling for consistent expert selection in Mixture-of-Experts](https://mlsanity.com/paper/2607.16427/) (2026)
- [Expert Streaming: Accelerating Low-Batch MoE Inference via Multi-chiplet Architecture and Dynamic Expert Trajectory Scheduling](https://mlsanity.com/paper/2603.27624/) (2026)
- [CogniRoute: Learning to Route Social Evidence in Omni-Modal Models](https://mlsanity.com/paper/2606.20970/) (2026)
- [Director: Accelerating Distributed MoE Serving via Online Proactive Expert Placement](https://mlsanity.com/paper/2607.08782/) (2026)
- [X-Stage: Modeling Post-Issue Backpressure in GPU Communication--Computation Fusion](https://mlsanity.com/paper/2607.23264/) (2026)
- [The Life of a Token: from Words to Bits on the Wire](https://mlsanity.com/paper/2609.19924/) (2026)
- [Mixture-of-Kittens: MoE Megakernel for NVL72s](https://mlsanity.com/paper/2609.36070/) (2026)

## Abstract (arXiv)

Recent large language models (LLMs) have tended to leverage sparsity to reduce computations, employing the sparsely activated mixture-of-experts (MoE) technique. MoE introduces four modules, including token routing, token communication, expert computation, and expert parallelism, that impact model quality and training efficiency. To enable versatile usage of MoE models, we introduce FSMoE, a flexible training system optimizing task scheduling with three novel techniques: 1) Unified abstraction and online profiling of MoE modules for task scheduling across various MoE implementations. 2) Co-scheduling intra-node and inter-node communications with computations to minimize communication overheads. 3) To support near-optimal task scheduling, we design an adaptive gradient partitioning method for gradient aggregation and a schedule to adaptively pipeline communications and computations. We conduct extensive experiments with configured MoE layers and real-world MoE models on two GPU clusters. Experimental results show that 1) our FSMoE supports four popular types of MoE routing functions and is more efficient than existing implementations (with up to a 1.42× speedup), and 2) FSMoE outperforms the state-of-the-art MoE training systems (DeepSpeed-MoE and Tutel) by 1.18×-1.22× on 1458 MoE layers and 1.19×-3.01× on real-world MoE models based on GPT-2 and Mixtral using a popular routing function.