arXiv cs.AI· Francesco Lettich, Emanuele Carlini, Franco Maria Nardini, Raffaele Perego, Salvatore Trani·· 12 小时前AI 评分23
PWR：面向 GPU 数据中心的功耗与碎片感知在线调度策略
Power- and Fragmentation-aware Online Scheduling for GPU Datacenters
AI 导读
研究提出在线调度策略 PWR，通过选择高能效的 GPU 与 CPU 组合来降低 GPU 数据中心功耗，其简化功耗模型已集成进 Kubernetes score 插件。在模拟集群实验中，PWR 与碎片梯度下降（FGD）策略结合后，在降低功耗与减少 GPU 碎片之间取得平衡。该成果发表于 IEEE CCGrid 2025。
正文 · 原文
Abstract:The rise of Artificial Intelligence and Large Language Models is driving increased GPU usage in data centers for complex training and inference tasks, impacting operational costs, energy demands, and the environmental footprint of large-scale computing infrastructures. This work addresses the online scheduling problem in GPU datacenters, which involves scheduling tasks without knowledge of their future arrivals. We focus on two objectives: minimizing GPU fragmentation and reducing power consumption. GPU fragmentation occurs when partial GPU allocations hinder the efficient use of remaining resources, especially as the datacenter nears full capacity. A recent scheduling policy, Fragmentation Gradient Descent (FGD), leverages a fragmentation metric to address this issue. Reducing power consumption is also crucial due to the significant power demands of GPUs. To this end, we propose PWR, a novel scheduling policy to minimize power usage by selecting power-efficient GPU and CPU combinations. This involves a simplified model for measuring power consumption integrated into a Kubernetes score plugin. Through an extensive experimental evaluation in a simulated cluster, we show how PWR, when combined with FGD, achieves a balanced trade-off between reducing power consumption and minimizing GPU fragmentation.
| Comments: | Published in the proceedings of IEEE CCGrid 2025 |
| Subjects: | Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2412.17484 [cs.DC] |
| (or arXiv:2412.17484v2 [cs.DC] for this version) | |
| https://doi.org/10.48550/arXiv.2412.17484 arXiv-issued DOI via DataCite |
|
| Journal reference: | 2025 IEEE 25th International Symposium on Cluster, Cloud and Internet Computing (CCGrid), pp. 43-52 (2025) |
| Related DOI: | https://doi.org/10.1109/CCGRID64434.2025.00015
DOI(s) linking to related resources |
Submission history
From: Francesco Lettich [view email]
[v1]
Mon, 23 Dec 2024 11:27:17 UTC (310 KB)
[v2]
Fri, 23 Oct 2026 15:37:26 UTC (186 KB)
来源：arXiv cs.AI · arxiv.org