Scheduler-Driven Job Atomization

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Konopa, Michal, Fesl, Jan, Beránek, Ladislav
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866914052362469376
author Konopa, Michal
Fesl, Jan
Beránek, Ladislav
author_facet Konopa, Michal
Fesl, Jan
Beránek, Ladislav
contents Modern GPU clusters, particularly those built on NVIDIA's Multi-Instance GPU (MIG) architecture, often suffer from inefficiencies because jobs are treated as rigid, indivisible blocks that occupy a fixed slice until completion. The reliance on static peak memory estimates exacerbates fragmentation, underutilization, and job rejections. We propose Scheduler-Driven Job Atomization (SJA), a new paradigm that establishes a bidirectional interaction between scheduler and jobs. In SJA, the scheduler advertises available execution gaps, and jobs respond by signaling interest if they can potentially generate a subjob that fits the offered time-capacity window. The scheduler may collect multiple signals for the same slot and, based on its allocation policy (e.g., fairness, efficiency, or SLA priorities), selects which job is granted the slot. Only then does the chosen job materialize a safe, self-contained subjob tailored to that opportunity. Unlike migration or preemption, SJA proactively shapes workloads before execution, thereby avoiding costly state transfers and unpredictable interruptions. It aims to increase GPU utilization, reduce wait times, and minimize migration overhead by aligning jobs with opportunities in real time, ensuring that each admitted subjob is correct by construction. This paper is presented as a concept paper: it introduces the paradigm, defines its building blocks, and outlines future research directions, rather than offering a full experimental evaluation.
format Preprint
id arxiv_https___arxiv_org_abs_2509_19086
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Scheduler-Driven Job Atomization
Konopa, Michal
Fesl, Jan
Beránek, Ladislav
Distributed, Parallel, and Cluster Computing
D.4.1; C.4; C.1.4; D.1.3
Modern GPU clusters, particularly those built on NVIDIA's Multi-Instance GPU (MIG) architecture, often suffer from inefficiencies because jobs are treated as rigid, indivisible blocks that occupy a fixed slice until completion. The reliance on static peak memory estimates exacerbates fragmentation, underutilization, and job rejections. We propose Scheduler-Driven Job Atomization (SJA), a new paradigm that establishes a bidirectional interaction between scheduler and jobs. In SJA, the scheduler advertises available execution gaps, and jobs respond by signaling interest if they can potentially generate a subjob that fits the offered time-capacity window. The scheduler may collect multiple signals for the same slot and, based on its allocation policy (e.g., fairness, efficiency, or SLA priorities), selects which job is granted the slot. Only then does the chosen job materialize a safe, self-contained subjob tailored to that opportunity. Unlike migration or preemption, SJA proactively shapes workloads before execution, thereby avoiding costly state transfers and unpredictable interruptions. It aims to increase GPU utilization, reduce wait times, and minimize migration overhead by aligning jobs with opportunities in real time, ensuring that each admitted subjob is correct by construction. This paper is presented as a concept paper: it introduces the paradigm, defines its building blocks, and outlines future research directions, rather than offering a full experimental evaluation.
title Scheduler-Driven Job Atomization
topic Distributed, Parallel, and Cluster Computing
D.4.1; C.4; C.1.4; D.1.3
url https://arxiv.org/abs/2509.19086