| Issue |
EPJ Web Conf.
Volume 337, 2025
27th International Conference on Computing in High Energy and Nuclear Physics (CHEP 2024)
|
|
|---|---|---|
| Article Number | 01079 | |
| Number of page(s) | 9 | |
| DOI | https://doi.org/10.1051/epjconf/202533701079 | |
| Published online | 07 October 2025 | |
https://doi.org/10.1051/epjconf/202533701079
Improving overall GPU sharing and usage efficiency with Kubernetes
CERN
* e-mail: diana.gaponcic@cern.ch
** e-mail: ricardo.rocha@cern.ch
Published online: 7 October 2025
GPUs and accelerators are enabling High Energy Physics (HEP) to keep pace with the growing data volume and computational complexity. The challenge remains to improve overall efficiency and sharing opportunities of what are currently expensive and scarce resources.
In this paper, we describe the common patterns of GPU usage in HEP, including spiky requirements with low overall usage for interactive access, as well as more predictable but potentially bursty workloads. We then explore the multiple mechanisms to share and partition GPUs, covering time-slicing, and physical partitioning (MIG) for NVIDIA devices.
We conclude with the results of an extensive set of benchmarks for representative HEP use cases. We highlight the limitations of each option and the use cases where they fit best. Finally, we cover the deployment aspects and the different options available targeting a centralized GPU pool that can significantly push the overall GPU usage efficiency.
© The Authors, published by EDP Sciences, 2025
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.