Contention-Aware GPU Thread Block Scheduler for Efficient GPU-SSD
IEEE Computer Architecture Letters · 2025-07 · 24 (2) · 257-260
DOI: 10.1109/lca.2025.3586312 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Computer Architecture Letters, 2025-07. Verfasst von Xueyang Liu und 5 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 9 Literaturangaben. Gefördert von: NSF, research grant from Samsung Electronics.
Abstract (Originalsprache)
The growing dataset sizes in LLM have made low-cost SSDs a popular solution for extending GPU memory in mobile devices. In this paper, we introduce CA-Scheduler , a contention-aware scheduling scheme for GPU-initiated SSD access. The key insight behind CA-Scheduler is leveraging the BSP GPU programming model, which allows reordering work at the thread block level to optimize SSD throughput. By capitalizing on the predictable memory access patterns of GPU thread blocks, CA-Scheduler anticipates SSD locations to minimize contention and improve performance.
Förderung
- NSF
- research grant from Samsung Electronics
Zitieren
APA
Liu, X., Na, S., Chung, E., Cao, J., Yang, J., & Kim, H. (2025). Contention-Aware GPU Thread Block Scheduler for Efficient GPU-SSD. IEEE Computer Architecture Letters, 24(2), 257-260. https://doi.org/10.1109/lca.2025.3586312
Vancouver
Liu X, Na S, Chung E, Cao J, Yang J, Kim H. Contention-Aware GPU Thread Block Scheduler for Efficient GPU-SSD. IEEE Computer Architecture Letters. 2025;24(2):257-260. doi:10.1109/lca.2025.3586312
BibTeX
@article{Liu2025,
author = {Liu, Xueyang and Na, Seonjin and Chung, Euijun and Cao, Jiashen and Yang, Jing and Kim, Hyesoon},
title = {Contention-Aware GPU Thread Block Scheduler for Efficient GPU-SSD},
journal = {IEEE Computer Architecture Letters},
year = {2025},
volume = {24},
number = {2},
pages = {257--260},
doi = {10.1109/lca.2025.3586312}
}
EUVIMED