Metraipolis › archive › dc_0359eb84be77

Characterizing the Scalability and Performance of Large-Scale AI Training Under Multi-Tenancy — arXiv:2609.00817

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.00817
License (as archived)
CC0-1.0
Archived
2026-10-17T06:41:58.447Z
SHA-256 of the archived text
5ade62f58f3a6e9960cdc80eddc4fe39d2c1be43903a5fd680124f6888981b5b
Stable URL of this page
https://metraipolis.com/documento/dc_0359eb84be77

How to cite

Characterizing the Scalability and Performance of Large-Scale AI Training Under Multi-Tenancy — arXiv:2609.00817 — arXiv (en), https://arxiv.org/abs/2609.00817 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0359eb84be77 (Metraipolis archive dc_0359eb84be77, sha256 5ade62f58f3a6e9960cdc80eddc4fe39d2c1be43903a5fd680124f6888981b5b)
BibTeX
@misc{dc0359eb84be77,
  title = {Characterizing the Scalability and Performance of Large-Scale AI Training Under Multi-Tenancy — arXiv:2609.00817},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.00817},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0359eb84be77; sha256 5ade62f58f3a6e9960cdc80eddc4fe39d2c1be43903a5fd680124f6888981b5b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.00817. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Characterising AI workload performance on modern HPC systems requires understanding both their scalability in isolation and their behaviour under concurrent execution. However, the interplay among parallelisation strategies, network congestion, compute capability, and interconnect technologies remains poorly understood. This work investigates the performance and scalability of AI models up to 2400 GPUs. We quantify the communication overheads and their impact across different interconnects by evaluating scale-up, scale-out, and rack-scale configurations under multiple allocation schemes. Finally, we study how multiple concurrent training jobs interfere with each other by designing a realistic noise model. We design a benchmark suite of AI models to evaluate the performance of five distinct parallelisation strategies across different supercomputing clusters, including Alps, Leonardo, LUMI, JUPITER, NVL72 GB300, and DGX A100. Our work provides a systematic characterization of the scalability and execution efficiency of distributed AI training, while offering key insights into performance behavior under realistic multi-tenant scenarios.

Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-09-02 in cs.DC).
Categorie arXiv: Distributed, Parallel, and Cluster Computing (cs.DC).
Pagina del lavoro: https://arxiv.org/abs/2609.00817 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-17T06:34:26.606Z dalla pagina https://arxiv.org/catchup/cs.DC/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).