arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2507.13601cs.DCcs.ETcs.PF
收藏

通过可塑任务调度利用多实例 GPU

Leveraging Multi-Instance GPUs through moldable task scheduling

Jorge Villarrubia, Luis Costero, Francisco D. Igual, Katzalin Olcoz

AI总结:

本文针对 NVIDIA MIG 多实例 GPU 环境提出三阶段可塑任务调度算法 FAR,通过动态重配置和重分区树启发式等方法最小化完工时间,在 A30 和 A100/H100 上分别取得 7/4 和 2 的近似比,实验表现显著优于现有方案。

摘要:

NVIDIA MIG(多实例 GPU)允许将物理 GPU 划分为多个具有完全隔离资源的逻辑实例,并且这些实例可以动态重新配置。本工作通过带有动态重新配置的可塑任务调度,揭示了 MIG 尚未被开发的潜力。具体而言,我们提出了一个在 MIG 约束下多任务执行的完工时间最小化问题。我们的性能分析表明,在多核调度中通常假设任务工作量随资源单调变化,但在 MIG 环境下这一假设并不成立。基于一个不需要此类假设的最新方案,我们提出了 FAR,一个用于解决该问题的三阶段算法。FAR 的第一阶段建立在经典的任务可塑性方法之上;第二阶段将最长处理时间优先和列表调度与一种针对 MIG 约束量身定制的新型重分区树启发式方法相结合;第三阶段通过任务移动和交换进行局部搜索。FAR 以离线方式分批调度任务,并以一种有利于资源复用的改进方式动态拼接各批次的调度方案。在不考虑重新配置成本的情况下,列表调度的证明表明在 NVIDIA A30 模型上近似比为 7/4。我们将该技术适配到 NVIDIA A100/H100 的特定约束上,获得了近似比为 2 的结果。在考虑重新配置成本的情况下,我们的实际实验表明,对于一组知名的基准测试套件,完工时间相对于最优值不劣于 1.22 倍;对于受真实内核启发的合成输入,该比值为 1.10 倍。我们不仅在每批任务上取得了良好的实验结果,在各批次拼接方面也表现优异,相比现有最先进方法和不进行 GPU 重新配置的方案有大幅改进。除算法本身外,本文还展示了 MIG 技术的研究潜力,并为该领域未来的工作提出了有用的度量指标、工作负载特征描述和评估技术。

英文摘要:

NVIDIA MIG (Multi-Instance GPU) allows partitioning a physical GPU into multiple logical instances with fully-isolated resources, which can be dynamically reconfigured. This work highlights the untapped potential of MIG through moldable task scheduling with dynamic reconfigurations. Specifically, we propose a makespan minimization problem for multi-task execution under MIG constraints. Our profiling shows that assuming monotonicity in task work with respect to resources is not viable, as is usual in multicore scheduling. Relying on a state-of-the-art proposal that does not require such an assumption, we present FAR, a 3-phase algorithm to solve the problem. Phase 1 of FAR builds on a classical task moldability method, phase 2 combines Longest Processing Time First and List Scheduling with a novel repartitioning tree heuristic tailored to MIG constraints, and phase 3 employs local search via task moves and swaps. FAR schedules tasks in batches offline, concatenating their schedules on the fly in an improved way that favors resource reuse. Excluding reconfiguration costs, the List Scheduling proof shows an approximation factor of 7/4 on the NVIDIA A30 model. We adapt the technique to the particular constraints of an NVIDIA A100/H100 to obtain an approximation factor of 2. Including the reconfiguration cost, our real-world experiments reveal a makespan with respect to the optimum no worse than 1.22x for a well-known suite of benchmarks, and 1.10x for synthetic inputs inspired by real kernels. We obtain good experimental results for each batch of tasks, but also in the concatenation of batches, with large improvements over the state-of-the-art and proposals without GPU reconfiguration. Beyond the algorithm, the paper demonstrates the research potential of the MIG technology and suggests useful metrics, workload characterizations and evaluation techniques for future work in this field.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑