ReadFastAFD: When Does Attention-FFN Disaggregation Pay on Rack-Scale NVLink?#Mixture of Experts (MoE)#vllmarXiv: Distributed ComputingarticleNov 6, 2026
ReadResource-Efficient Reinforcement Learning Post-Training through Elastic Expert Parallelism#post-training#Reinforcement learningarXiv: Artificial IntelligencearticleNov 6, 2026
ReadGC-MoE: Overlapping Expert-Parallel Token Dispatch with FP8/FP4 Grouped GEMM via CUDA Green Contexts#Mixture of Experts (MoE)#CUDAarXiv: Artificial IntelligencearticleNov 4, 2026
ReadRaMP: Runtime-Aware Megakernel Polymorphism for Mixture-of-Experts#Mixture of Experts (MoE)#LLM optimizationarXiv: Distributed ComputingarticleNov 3, 2026
ReadZeRO-Prefill: Zero Redundancy Overheads in MoE Prefill Serving#Mixture of Experts (MoE)#LLM optimizationarXiv: Machine LearningarticleOct 28, 2026
ReadNanoCP: Taming Load Imbalance in Data-Expert Parallel Decoding under Variable-Length Workloads#Distributed Systems & Parallel Computing#Load BalancingarXiv: Distributed ComputingarticleOct 22, 2026
Read3D-MoE: Complementarity-Guided Hardware-Software Co-Design for MoE Decode on 3D-Stacked NMP#Mixture of Experts (MoE)#High-Performance ComputingarXiv: Distributed ComputingnewsOct 21, 2026
ReadAchieving near-zero-downtime scaling in Elastic Expert Parallelism#Deployment & Scaling#vllmvLLM BlogarticleOct 19, 2026