# MoESys: A Distributed and Efficient Mixture-of-Experts Training and Inference System for Internet Services

arXiv: https://arxiv.org/abs/2205.10034 · HTML: https://mlsanity.com/paper/2205.10034/

## Key facts

- Cited by 37 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 34 of its 60 references that are in the map.
- First posted on arXiv 20 May 2022.

## 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 34)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [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)
- [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://mlsanity.com/paper/1909.08053/) (2019)
- [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://mlsanity.com/paper/1901.02860/) (2019)
- [mT5: A massively multilingual pre-trained text-to-text transformer](https://mlsanity.com/paper/2010.11934/) (2020)
- [GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding](https://mlsanity.com/paper/2006.16668/) (2020)
- [fairseq: A Fast, Extensible Toolkit for Sequence Modeling](https://mlsanity.com/paper/1904.01038/) (2019)
- [GLaM: Efficient Scaling of Language Models with Mixture-of-Experts](https://mlsanity.com/paper/2112.06905/) (2021)
- [Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LM](https://mlsanity.com/paper/2104.04473/) (2021)
- [Using DeepSpeed and Megatron to Train Megatron-Turing NLG 530B, A Large-Scale Generative Language Model](https://mlsanity.com/paper/2201.11990/) (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)
- [ZeRO-Infinity: Breaking the GPU Memory Wall for Extreme Scale Deep Learning](https://mlsanity.com/paper/2104.07857/) (2021)
- [Massively Multilingual Neural Machine Translation](https://mlsanity.com/paper/1903.00089/) (2019)
- [Efficient Large Scale Language Modeling with Mixtures of Experts](https://mlsanity.com/paper/2112.10684/) (2021)
- [MLPerf Training Benchmark](https://mlsanity.com/paper/1910.01500/) (2019)
- [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)
- [Software-Hardware Co-design for Fast and Scalable Training of Deep Learning Recommendation Models](https://mlsanity.com/paper/2104.05158/) (2021)
- [Scalable and Efficient MoE Training for Multitask Multilingual Models](https://mlsanity.com/paper/2109.10465/) (2021)
- [Pre-trained Summarization Distillation](https://mlsanity.com/paper/2010.13002/) (2020)
- [Hard Mixtures of Experts for Large Scale Weakly Supervised Vision](https://mlsanity.com/paper/1704.06363/) (2017)
- [Distributed Hierarchical GPU Parameter Server for Massive Scale Deep Learning Ads Systems](https://mlsanity.com/paper/2003.05622/) (2020)
- [CPM-2: Large-scale Cost-effective Pre-trained Language Models](https://mlsanity.com/paper/2106.10715/) (2021)
- [Multi-task Learning for Multilingual Neural Machine Translation](https://mlsanity.com/paper/2010.02523/) (2020)
- [ERNIE 3.0 Titan: Exploring Larger-scale Knowledge Enhanced Pre-training for Language Understanding and Generation](https://mlsanity.com/paper/2112.12731/) (2021)

## Built on by (most cited first, 30 of 37)

- [A Survey on Mixture of Experts in Large Language Models](https://mlsanity.com/paper/2407.06204/) (2024)
- [Accelerating Distributed MoE Training and Inference with Lina](https://mlsanity.com/paper/2210.17223/) (2022)
- [Pre-gated MoE: An Algorithm-System Co-Design for Fast and Scalable Mixture-of-Expert Inference](https://mlsanity.com/paper/2308.12066/) (2023)
- [A Hybrid Tensor-Expert-Data Parallelism Approach to Optimize Mixture-of-Experts Training](https://mlsanity.com/paper/2303.06318/) (2023)
- [A Survey on Inference Optimization Techniques for Mixture of Experts Models](https://mlsanity.com/paper/2412.14219/) (2024)
- [SiDA-MoE: Sparsity-Inspired Data-Aware Serving for Efficient and Scalable Large Mixture-of-Experts Models](https://mlsanity.com/paper/2310.18859/) (2023)
- [Towards MoE Deployment: Mitigating Inefficiencies in Mixture-of-Expert (MoE) Inference](https://mlsanity.com/paper/2303.06182/) (2023)
- [Exploiting Inter-Layer Expert Affinity for Accelerating Mixture-of-Experts Model Inference](https://mlsanity.com/paper/2401.08383/) (2024)
- [SwapMoE: Serving Off-the-shelf MoE-based Large Language Models with Tunable Memory Budget](https://mlsanity.com/paper/2308.15030/) (2023)
- [Towards Efficient Mixture of Experts: A Holistic Study of Compression Techniques](https://mlsanity.com/paper/2406.02500/) (2024)
- [ExpertFlow: Efficient Mixture-of-Experts Inference via Predictive Expert Caching and Token Scheduling](https://mlsanity.com/paper/2410.17954/) (2024)
- [From Google Gemini to OpenAI Q* (Q-Star): A Survey of Reshaping the Generative Artificial Intelligence (AI) Research Landscape](https://mlsanity.com/paper/2312.10868/) (2023)
- [Mixture of Experts (MoE): A Big Data Perspective](https://mlsanity.com/paper/2501.16352/) (2025)
- [MoE Parallel Folding: Heterogeneous Parallelism Mappings for Efficient Large-Scale MoE Model Training with Megatron Core](https://mlsanity.com/paper/2504.14960/) (2025)
- [Mixture of Cache-Conditional Experts for Efficient Mobile Device Inference](https://mlsanity.com/paper/2412.00099/) (2024)
- [GRACE-MoE: Grouping and Replication with Locality-Aware Routing for Efficient Distributed MoE Inference](https://mlsanity.com/paper/2509.25041/) (2025)
- [Optimizing Mixture-of-Experts Inference Time Combining Model Deployment and Communication Scheduling](https://mlsanity.com/paper/2410.17043/) (2024)
- [Advancing MoE Efficiency: A Collaboration-Constrained Routing (C2R) Strategy for Better Expert Parallelism Design](https://mlsanity.com/paper/2504.01337/) (2025)
- [LAER-MoE: Load-Adaptive Expert Re-layout for Efficient Mixture-of-Experts Training](https://mlsanity.com/paper/2602.11686/) (2026)
- [Theory of Mixture-of-Experts for Mobile Edge Computing](https://mlsanity.com/paper/2412.15690/) (2024)
- [SYMI: Efficient Mixture-of-Experts Training via Model and Optimizer State Decoupling](https://mlsanity.com/paper/2504.19925/) (2025)
- [DuoServe-MoE: Dual-Phase Expert Prefetch and Caching for LLM Inference QoS Assurance](https://mlsanity.com/paper/2509.07379/) (2025)
- [Efficient MoE Serving in the Memory-Bound Regime: Balance Activated Experts, Not Tokens](https://mlsanity.com/paper/2512.09277/) (2025)
- [A Study on Transformer Configuration and Training Objective](https://mlsanity.com/paper/2205.10505/) (2022)
- [SpecMoE: A Fast and Efficient Mixture-of-Experts Inference via Self-Assisted Speculative Decoding](https://mlsanity.com/paper/2604.10152/) (2026)
- [HAP: Hybrid Adaptive Parallelism for Efficient Mixture-of-Experts Inference](https://mlsanity.com/paper/2508.19373/) (2025)
- [When MoE Meets Blockchain: A Trustworthy Distributed Framework of Large Models](https://mlsanity.com/paper/2509.12141/) (2025)
- [Scaling Multi-Node Mixture-of-Experts Inference Using Expert Activation Patterns](https://mlsanity.com/paper/2604.23150/) (2026)
- [Disaggregating Embedding Recommendation Systems with FlexEMR](https://mlsanity.com/paper/2410.12794/) (2024)
- [Fine-grained MoE Load Balancing with Linear Programming](https://mlsanity.com/paper/2511.16947/) (2025)

## Abstract (arXiv)

While modern internet services, such as chatbots, search engines, and online advertising, demand the use of large-scale deep neural networks (DNNs), distributed training and inference over heterogeneous computing systems are desired to facilitate these DNN models. Mixture-of-Experts (MoE) is one the most common strategies to lower the cost of training subject to the overall size of models/data through gating and parallelism in a divide-and-conquer fashion. While DeepSpeed has made efforts in carrying out large-scale MoE training over heterogeneous infrastructures, the efficiency of training and inference could be further improved from several system aspects, including load balancing, communication/computation efficiency, and memory footprint limits. In this work, we present a novel MoESys that boosts efficiency in both large-scale training and inference. Specifically, in the training procedure, the proposed MoESys adopts an Elastic MoE training strategy with 2D prefetch and Fusion communication over Hierarchical storage, so as to enjoy efficient parallelisms. For scalable inference in a single node, especially when the model size is larger than GPU memory, MoESys builds the CPU-GPU memory jointly into a ring of sections to load the model, and executes the computation tasks across the memory sections in a round-robin manner for efficient inference. We carried out extensive experiments to evaluate MoESys, where MoESys successfully trains a Unified Feature Optimization (UFO) model with a Sparsely-Gated Mixture-of-Experts model of 12B parameters in 8 days on 48 A100 GPU cards. The comparison against the state-of-the-art shows that MoESys outperformed DeepSpeed with 33% higher throughput (tokens per second) in training and 13% higher throughput in inference in general. Particularly, under unbalanced MoE Tasks, e.g., UFO, MoESys achieved 64% higher throughput with 18% lower memory footprints.