FlashMoE: Fast Distributed MoE in a Single Kernel

arXiv:2506.04667 · cs.DC, cs.AR, cs.LG · Submitted 2025-06-05 · Read on arXiv

cs.DC, cs.AR, cs.LG

Submitted: 2025-06-05

Updated: 2025-11-08

Code: https://github.com/osayamenja/FlashMoE

Terminology

• Mixture-of-Experts (MoE) — 1× in this paper · explained in 7 episodes
• Mixture-of-Experts (MoE) Models — 1× in this paper · explained in 3 episodes
• Sparsity — 1× in this paper · explained in 4 episodes

Sources

• Efficient AllReduce with Stragglers
• Comet: Fine-grained Computation-communication Overlapping for Mixture-of-Experts
• Triton-distributed: Programming Overlapping Kernels on Distributed AI Systems with the Triton Compiler
• BlackMamba: Mixture of Experts for State-Space Models
• Muon is Scalable for LLM Training
• Qwen3 Technical Report
• Mixtral of Experts
• DeepSeek-V3 Technical Report

Related papers

• iScheduler: Reinforcement Learning-Driven Continual Optimization for Large-Scale Resource Investment Problems
• SAMM: Sharded Automated Market Maker
• InferScale: GPU-Native KV Injection for Personalized LLM Serving
• Vigil: Accountable Liveness against Selective Silence
• Steelhead: Interleaving Partially Synchronous and Asynchronous Commit Rules on a Shared DAG
• Democratizing MoE inference on commodity GPUs with CoMoE