# FloE: On-the-Fly MoE Inference on Memory-constrained GPU

arXiv: https://arxiv.org/abs/2505.05950 · HTML: https://mlsanity.com/paper/2505.05950/

## Key facts

- Cited by 9 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 28 of its 63 references that are in the map.
- First posted on arXiv 9 May 2025.

## Placed in

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

## Builds on (most cited first, 28 of 28)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions](https://mlsanity.com/paper/1905.10044/) (2019)
- [Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone](https://mlsanity.com/paper/2404.14219/) (2024)
- [Mixtral of Experts](https://mlsanity.com/paper/2401.04088/) (2024)
- [GLU Variants Improve Transformer](https://mlsanity.com/paper/2002.05202/) (2020)
- [A Simple and Effective Pruning Approach for Large Language Models](https://mlsanity.com/paper/2306.11695/) (2023)
- [Phi-4 Technical Report](https://mlsanity.com/paper/2412.08905/) (2024)
- [Few-Shot Learning via Learning the Representation, Provably](https://mlsanity.com/paper/2002.09434/) (2020)
- [WinoGrande: An Adversarial Winograd Schema Challenge at Scale](https://mlsanity.com/paper/1907.10641/) (2019)
- [LLM in a flash: Efficient Large Language Model Inference with Limited Memory](https://mlsanity.com/paper/2312.11514/) (2023)
- [PowerInfer-2: Fast Large Language Model Inference on a Smartphone](https://mlsanity.com/paper/2406.06282/) (2024)
- [ReLU Strikes Back: Exploiting Activation Sparsity in Large Language Models](https://mlsanity.com/paper/2310.04564/) (2023)
- [Fast Inference of Mixture-of-Experts Language Models with Offloading](https://mlsanity.com/paper/2312.17238/) (2023)
- [MoE-Lightning: High-Throughput MoE Inference on Memory-constrained GPUs](https://mlsanity.com/paper/2411.11217/) (2024)
- [Fiddler: CPU-GPU Orchestration for Fast Inference of Mixture-of-Experts Models](https://mlsanity.com/paper/2402.07033/) (2024)
- [ReLU2 Wins: Discovering Efficient Activation Functions for Sparse LLMs](https://mlsanity.com/paper/2402.03804/) (2024)
- [Training-Free Activation Sparsity in Large Language Models](https://mlsanity.com/paper/2408.14690/) (2024)
- [HOBBIT: A Mixed Precision Expert Offloading System for Fast MoE Inference](https://mlsanity.com/paper/2411.01433/) (2024)
- [MoE-Infinity: Efficient MoE Inference on Personal Machines with Sparsity-Aware Expert Cache](https://mlsanity.com/paper/2401.14361/) (2024)
- [ProSparse: Introducing and Enhancing Intrinsic Activation Sparsity within Large Language Models](https://mlsanity.com/paper/2402.13516/) (2024)
- [ProMoE: Fast MoE-based LLM Serving using Proactive Caching](https://mlsanity.com/paper/2410.22134/) (2024)
- [Turbo Sparse: Achieving LLM SOTA Performance with Minimal Activated Parameters](https://mlsanity.com/paper/2406.05955/) (2024)
- [Sample Efficient Linear Meta-Learning by Alternating Minimization](https://mlsanity.com/paper/2105.08306/) (2021)
- [CHESS: Optimizing LLM Inference via Channel-Wise Thresholding and Selective Sparsification](https://mlsanity.com/paper/2409.01366/) (2024)

## Built on by (most cited first, 9 of 9)

- [SpecVLM: Enhancing Speculative Decoding of Video LLMs via Verifier-Guided Token Pruning](https://mlsanity.com/paper/2508.16201/) (2025)
- [MoE-SpeQ: Speculative Quantized Decoding with Proactive Expert Prefetching and Offloading for Mixture-of-Experts](https://mlsanity.com/paper/2511.14102/) (2025)
- [Efficient Inference for Large Vision-Language Models: Bottlenecks, Techniques, and Prospects](https://mlsanity.com/paper/2604.05546/) (2026)
- [ZipMoE: Efficient On-Device MoE Serving via Lossless Compression and Cache-Affinity Scheduling](https://mlsanity.com/paper/2601.21198/) (2026)
- [DyMoE: Dynamic Expert Orchestration with Mixed-Precision Quantization for Efficient MoE Inference on Edge](https://mlsanity.com/paper/2603.19172/) (2026)
- [VersatileFFN: Achieving Parameter Efficiency in LLMs via Adaptive Wide-and-Deep Reuse](https://mlsanity.com/paper/2512.14531/) (2025)
- [MM-ShiftKV: Decode-Aware Prefill-Stage KV Selection for Multimodal Large Language Models](https://mlsanity.com/paper/2607.22586/) (2026)
- [Mira: Memory-Efficient MoE Inference Using Adaptive Caching and Predictive Expert Staging](https://mlsanity.com/paper/2609.38090/) (2026)
- [ITC-MoE: Importance-guided Token-aware Compression for MoE Diffusion Language Models](https://mlsanity.com/paper/2610.01296/) (2026)

## Abstract (arXiv)

With the widespread adoption of Mixture-of-Experts (MoE) models, there is a growing demand for efficient inference on memory-constrained devices. While offloading expert parameters to CPU memory and loading activated experts on demand has emerged as a potential solution, the large size of activated experts overburdens the limited PCIe bandwidth, hindering the effectiveness in latency-sensitive scenarios. To mitigate this, we propose FloE, an on-the-fly MoE inference system on memory-constrained GPUs. FloE is built on the insight that there exists substantial untapped redundancy within sparsely activated experts. It employs various compression techniques on the expert's internal parameter matrices to reduce the data movement load, combined with low-cost sparse prediction, achieving perceptible inference acceleration in wall-clock time on resource-constrained devices. Empirically, FloE achieves a 9.3x compression of parameters per expert in Mixtral-8x7B; enables deployment on a GPU with only 11GB VRAM, reducing the memory footprint by up to 8.5x; and delivers a 48.7x inference speedup compared to DeepSpeed-MII on a single GeForce RTX 3090 - all with only a 4.4% - 7.6% average performance degradation.