Algorithmic Techniques for GPU Scheduling: A Comprehensive Survey

Збережено в:

Бібліографічні деталі
Опубліковано в::	Algorithms vol. 18, no. 7 (2025), p. 385-437
Автор:	Chab, Robert
Інші автори:	Li, Fei, Setia Sanjeev
Опубліковано:	MDPI AG
Предмети:	NVidia Corp Scheduling Machine learning Dynamic programming Task scheduling Deep learning Performance evaluation Mathematical analysis Artificial intelligence Formal method Graphics processing units Bandwidths Mathematical programming Greedy algorithms Architecture Energy efficiency Spectrum allocation Algorithms Resource management Cost control Workloads Decisions Queueing
Онлайн доступ:	Citation/Abstract Full Text + Graphics Full Text - PDF
Теги:	Додати тег Немає тегів, Будьте першим, хто поставить тег для цього запису!

Опис
Короткий огляд:	In this survey, we provide a comprehensive classification of GPU task scheduling approaches, categorized by their underlying algorithmic techniques and evaluation metrics. We examine traditional methods—including greedy algorithms, dynamic programming, and mathematical programming—alongside advanced machine learning techniques integrated into scheduling policies. We also evaluate the performance of these approaches across diverse applications. This work focuses on understanding the trade-offs among various algorithmic techniques, the architectural and job-level factors influencing scheduling decisions, and the balance between user-level and service-level objectives. The analysis shows that no one paradigm dominates; instead, the highest-performing schedulers blend the predictability of formal methods with the adaptability of learning, often moderated by queueing insights for fairness. We also discuss key challenges in optimizing GPU resource management and suggest potential solutions.
ISSN:	1999-4893
DOI:	10.3390/a18070385
Джерело:	Engineering Database