AI Post Transformers · Episode Companion
ASAP: Disaggregating Attention and Experts for Faster MoE Prefill
Weiwei Chen, Shuang Chen, Lele Li, Qiang Hu, Han Li, Xin Ye, Ming Yan, Zhibin Yu — Huawei · arXiv, June 21, 2026
arXiv:2606.22541
Mixture-of-Experts
Disaggregated Serving
Prefill / TTFT
CloudMatrix384
Overview
Straggler Diagnosis
Async Architecture
Results & Scrutiny
References