Jul 2026
ExpertPlex: A High-Goodput Disaggregated Serving System for MoE LLMs with Adaptive Persistent Kernels
ExpertPlex is presented, which shares massive MoE experts across phases while disaggregating lightweight attention modules to eliminate over 95% of duplicate model weights and multiplexes dynamically sparse computation, while attention disaggregation reduces attention communication cost.
Bing-Ya Wu, Chao Jin, Zili Zhang et al.
· arXiv.org · 1 citation