作者位次 #16/20OSDI 2026Oral· Operating Systems & Distributed Systems
Heterogeneity at Hyperscale: Characterization and Scheduling of Large Production AI Clusters at Alibaba (Operational Systems)
Suyi Li · Lingyun Yang · Haoxuan Yu · Sheng Yao · Tianyuan Wu · Xiaoxiao Jiang · Hanfeng Lu · Kangjin Wang · Chenhao Wang · Shenglin Xu · Lun Wang · Qingyang Duan · Shenghao Liang · Xiu Lin · Meng Zhang · Wenchao Wu · Yinghao Yu · Guodong Yang · Liping Zhang · Wei Wang
摘要
The rapid scaling of generative AI (GenAI), alongside the continued reliance on classical deep neural networks (DNNs), has pushed production AI infrastructure toward massive, heterogeneous GPU fleets. We present a comprehensive characterization of Alibaba Serverless Infrastructure (ASI), a hypers...