发表机构
Faculty of Computer Science and Artificial Intelligence, Shenzhen University of Advanced Technology（深圳先进技术大学计算机科学与人工智能学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
RingStitch提出一种基于光电路交换的碎片整理调度器，通过本地优先策略和低成本逻辑环拼接，将碎片化机架容量转化为可调度资源，提升TPU集群的可调度性。
AI 中文摘要
大规模AI训练集群日益采用光电路交换（OCS）来重新配置机架级互连并创建弹性加速器切片。然而，在多租户TPU风格集群中，中小型作业常常在机架内留下部分空闲容量。尽管总空闲容量可能足以容纳新作业，但本地放置或粗粒度的整机架拼接无法利用这些容量。本文提出RingStitch，一种基于OCS的碎片整理调度器，将碎片化的机架容量转化为可调度资源。RingStitch遵循本地优先策略，仅在需要时拼接紧凑的跨机架碎片，并将选定的机架排序为低成本逻辑环。在TPU 8t类SuperPod模型上的模拟表明，RingStitch相较于本地和整机架基线提高了可调度性。
英文摘要
Large-scale AI training clusters increasingly use optical circuit switching (OCS) to reconfigure rack-level interconnects and create elastic accelerator slices. In multi-tenant TPU-style clusters, however, small and medium jobs often leave partial free capacity stranded inside racks. Although the aggregate free capacity may be sufficient for a new job, it cannot be used by local placement or coarse full-rack stitching. This paper presents RingStitch, an OCS-based defragmentation scheduler that turns fragmented rack capacity into schedulable resources. RingStitch follows a local-first policy, stitches compact cross-rack fragments only when needed, and orders the selected racks into a low-cost logical ring. Simulations on a TPU 8t-like SuperPod model show that RingStitch improves schedulability over local and full-rack baselines.
CommentsAccepted by ACP 2026, Top-Score Papers
