dorsal/arxiv
View SchemaPeformance Isolation for Inference Processes in Edge GPU Systems
| Authors | Juan José Martín, José Flich, Carles Hernández |
|---|---|
| Categories | |
| ArXiv ID | 2601.07600vv2 |
| URL | https://arxiv.org/abs/2601.07600 |
| License | http://creativecommons.org/licenses/by/4.0/ |
Abstract
This work analyzes the main isolation mechanisms available in modern NVIDIA GPUs: MPS, MIG, and the recent Green Contexts, to ensure predictable inference time in safety-critical applications using deep learning models. The experimental methodology includes performance tests, evaluation of partitioning impact, and analysis of temporal isolation between processes, considering both the NVIDIA A100 and Jetson Orin platforms. It is observed that MIG provides a high level of isolation. At the same time, Green Contexts represent a promising alternative for edge devices by enabling fine-grained SM allocation with low overhead, albeit without memory isolation. The study also identifies current limitations and outlines potential research directions to improve temporal predictability in shared GPUs.
{
"annotation_id": "35afbbfb-e467-4cc7-828f-378cd08485aa",
"date_created": "2026-02-17T05:53:11.896000Z",
"date_modified": "2026-02-17T05:53:11.896000Z",
"file_hash": "9a0a1c327b16629720174b35072c15ad408f36fc1963952576aca693fa6a9554",
"private": false,
"record": {
"abstract": "This work analyzes the main isolation mechanisms available in modern NVIDIA GPUs: MPS, MIG, and the recent Green Contexts, to ensure predictable inference time in safety-critical applications using deep learning models. The experimental methodology includes performance tests, evaluation of partitioning impact, and analysis of temporal isolation between processes, considering both the NVIDIA A100 and Jetson Orin platforms. It is observed that MIG provides a high level of isolation. At the same time, Green Contexts represent a promising alternative for edge devices by enabling fine-grained SM allocation with low overhead, albeit without memory isolation. The study also identifies current limitations and outlines potential research directions to improve temporal predictability in shared GPUs.",
"arxiv_id": "2601.07600",
"authors": [
"Juan Jos\u00e9 Mart\u00edn",
"Jos\u00e9 Flich",
"Carles Hern\u00e1ndez"
],
"categories": [
"cs.OS",
"cs.DC"
],
"license": "http://creativecommons.org/licenses/by/4.0/",
"title": "Peformance Isolation for Inference Processes in Edge GPU Systems",
"url": "https://arxiv.org/abs/2601.07600",
"version": "v2"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "221f7744-9498-4863-ae95-30be9aec93aa",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}