Metraipolis › archive › dc_64e43e6d5be2

Learning to Formulate: Agentic Hardware-Software Co-Design of AI Infrastructure through Progressive Optimization and Verification — arXiv:2610.39380

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.39380
License (as archived)
CC0-1.0
Archived
2026-10-31T06:13:13.514Z
SHA-256 of the archived text
b8160615aacc2bc57bbe51e509e29c7ed6ae750e0a8893f96a6134f9ee027465
Stable URL of this page
https://metraipolis.com/documento/dc_64e43e6d5be2

How to cite

Learning to Formulate: Agentic Hardware-Software Co-Design of AI Infrastructure through Progressive Optimization and Verification — arXiv:2610.39380 — arXiv (en), https://arxiv.org/abs/2610.39380 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_64e43e6d5be2 (Metraipolis archive dc_64e43e6d5be2, sha256 b8160615aacc2bc57bbe51e509e29c7ed6ae750e0a8893f96a6134f9ee027465)
BibTeX
@misc{dc64e43e6d5be2,
  title = {Learning to Formulate: Agentic Hardware-Software Co-Design of AI Infrastructure through Progressive Optimization and Verification — arXiv:2610.39380},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.39380},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_64e43e6d5be2; sha256 b8160615aacc2bc57bbe51e509e29c7ed6ae750e0a8893f96a6134f9ee027465},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.39380. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

AI infrastructure, the chips and software that train and serve AI models, is designed in layers. Layers make the system manageable, but they also fix which choices an optimizer can consider. Gains that need the chip and its software to change together are excluded before optimization begins: the formulation leaves them out. Language-model agents make it possible to revisit these boundaries, but an agent's proposal cannot show by itself that a design is correct or optimal. We present Formulaic, a flow that divides the work. The agent writes the formulation, a mathematical solver optimizes within it, and verification and measurement check its assumptions and results. The agent refines the problem step by step, from the computation graph toward the chip. Each step is a template: a family of programs with the conditions under which they are correct. Wall relaxation searches the chip by pricing the hardware limits that bind the software. This paper takes a first step, through separate studies rather than one end-to-end run. On NVIDIA GB10, templates the agent wrote run complete Qwen3-8B layers $1.075\times$ faster than SGLang. In a simulated Ventus GPU study, joint design cuts the worst-scenario cycles of Qwen2.5-0.5B to 0.428 of the reference, against 0.934 for software alone. On a five-step GEMM refinement chain, reusing proofs makes checking $2.51\times$ faster.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Hardware Architecture (cs.AR) ; Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC).
Pagina del lavoro: https://arxiv.org/abs/2610.39380 — 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 2026-10-31T06:07:08.712Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).