BOA Constrictor: Squeezing Performance out of GPUs in the Cloud via Budget-Optimal Allocation

Zhouzi Li, Cindy Zhu, Arpan Mukhopadhyay, Mor Harchol-Balter, Benjamin Berg · 2026

Read BOA Constrictor: Squeezing Performance out of GPUs in the Cloud via Budget-Optimal Allocation by Zhouzi Li, Cindy Zhu, Arpan Mukhopadhyay, Mor Harchol-Balter, Benjamin Berg (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-01404.txt.

Description

Zhouzi Li and co-authors propose BOA Constrictor, a scheduler squeezing top performance from rented GPUs under a budget. Most organizations rent GPUs rather than build clusters, and must split them among active jobs, though extra GPUs yield ever smaller gains while billing by the hour. The team casts this as budget-constrained scheduling, deriving the Budget-Optimal Allocation policy for minimum mean completion time under budget. Against heuristic schedulers, BOA Constrictor cuts mean completion time by a factor of 1.6 in small experiments and 2 in large simulations.

Catalog record

CREATORZhouzi Li, Cindy Zhu, Arpan Mukhopadhyay, Mor Harchol-Balter, Benjamin Berg
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE91761
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602.01404

Browse the full library · About the Tower of Babel