Metraipolis › archive › dc_f066cce76eb4

Where Is VLA Inference Time Spent? Bottlenecks and Optimization Interactions Across GPUs — arXiv:2611.06710

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.06710
License (as archived)
CC0-1.0
Archived
2026-11-05T18:30:47.585Z
SHA-256 of the archived text
052a3d803e3a3fea6d63c2ae89baeb99f7a0f2d0d3eac177fb60bedf1b275886
Stable URL of this page
https://metraipolis.com/documento/dc_f066cce76eb4

How to cite

Where Is VLA Inference Time Spent? Bottlenecks and Optimization Interactions Across GPUs — arXiv:2611.06710 — arXiv (en), https://arxiv.org/abs/2611.06710 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f066cce76eb4 (Metraipolis archive dc_f066cce76eb4, sha256 052a3d803e3a3fea6d63c2ae89baeb99f7a0f2d0d3eac177fb60bedf1b275886)
BibTeX
@misc{dcf066cce76eb4,
  title = {Where Is VLA Inference Time Spent? Bottlenecks and Optimization Interactions Across GPUs — arXiv:2611.06710},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.06710},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f066cce76eb4; sha256 052a3d803e3a3fea6d63c2ae89baeb99f7a0f2d0d3eac177fb60bedf1b275886},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.06710. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Vision-language-action (VLA) models map camera observations and language instructions to robot actions. However, their large backbones and repeated action generation can make inference take longer than the robot's real-time requirements for its next action. To reduce the load of VLA computations, researchers have proposed multiple optimization approaches, including quantization and pruning. These optimizations can be combined with other general optimizations such as better kernel compilation or using CUDA graphs instead of the eager mode. However, most of these optimization results are usually reported on a single GPU type, typically for a few VLA models, with no analysis of what the real bottlenecks are for a model on a particular hardware or how these optimizations would generalize beyond the reported results. In this paper, we provide a comprehensive measurement study of seven families of inference optimization methods, from CUDA graphs and compilation to token pruning and quantization, on seven state-of-the-art VLAs evaluated on four GPU families ranging from a datacenter H100 to an embedded Jetson Thor. We analyse how CPU launch overhead, memory bandwidth, and kernel efficiency jointly shape optimization gains, and how applying one method changes the benefit of another. From this analysis, we derive VLA-Triage, guidelines for stacking optimizations under latency budgets. The guidelines identify which stages can provide the required savings and which combinations need testing in the deployed execution mode.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.RO, nuovi lavori del giorno).
Categorie arXiv: Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2611.06710 — 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 2026-11-05T18:20:03.993Z dalla pagina https://arxiv.org/list/cs.RO/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).