Metraipolis › archive › dc_bbf985d4e508

Look Before You Launch: Learned Throughput Prediction for Hardware Selection in a Production Training Fleet — arXiv:2611.04057

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.04057
License (as archived): CC0-1.0
Archived: 2026-11-04T18:34:18.369Z
SHA-256 of the archived text: 1e1a93ef76513e3cade417780b013b1198c6e4ef131407b0c24ab49aa9ea6d66
Stable URL of this page: https://metraipolis.com/documento/dc_bbf985d4e508

How to cite

Look Before You Launch: Learned Throughput Prediction for Hardware Selection in a Production Training Fleet — arXiv:2611.04057 — arXiv (en), https://arxiv.org/abs/2611.04057 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bbf985d4e508 (Metraipolis archive dc_bbf985d4e508, sha256 1e1a93ef76513e3cade417780b013b1198c6e4ef131407b0c24ab49aa9ea6d66)

BibTeX

@misc{dcbbf985d4e508,
  title = {Look Before You Launch: Learned Throughput Prediction for Hardware Selection in a Production Training Fleet — arXiv:2611.04057},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.04057},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bbf985d4e508; sha256 1e1a93ef76513e3cade417780b013b1198c6e4ef131407b0c24ab49aa9ea6d66},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04057. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

In mixed-vendor accelerator fleets, many training jobs can run on several hardware types, and the best choice depends on each job's throughput on each type. Routing a job needs a throughput estimate for every candidate type before the job runs. We present LEAP, a machine learning performance prediction service that predicts each training job's throughput on every candidate hardware type at submission time and ranks the candidates by predicted throughput per unit cost. It routes tens of thousands of training job attempts a day across hundreds of different model types and more than 20 hardware types in a production fleet. LEAP predicts throughput from submission-time training configuration, including each job's prior throughput. LEAP reaches an accuracy of 77% over 30 nightly retrains, higher than the lookup table it replaced and at least as high as every alternative model we tested, including fine-tuned language models. Over six months, the share of eligible training GPU-hours that received a ranking by LEAP grew from 58% to 95%, and we estimate that it saves about 3% of training computational cost. In this paper, we identify the drivers that actually move accuracy and those that do not. We offer these as lessons for teams building performance predictors at fleet scale.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.04057 — 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 2026-11-04T18:32:37.269Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).