跳到正文
arXiv cs.AI· Yuanlin Wen, Elan Markowitz, Zubo Gu, Sami Abu-El-Haija, Bryan Perozzi, Mikhail Galkin, Ali Parviz, Aayush Gupta, Venkata Chivukula, Griffin Hadfield, Aravind Akella, Daniel J. Frame, Vahab Mirrokni, Amin Vahdat·· 10 小时前AI 评分52

HardwareDx:面向数据中心机器维修的图基础模型系统

HardwareDx: A Foundational Graph Model System for Datacenter Machine Repairs

AI 导读

Google 提出 AI 硬件诊断系统 HardwareDx,用图基础模型让单一模型诊断多种平台,并在生产规模部署。论文称其在 GPU 平台将生产停机时间降低 18.3%、在 TPU 上降低 16.1%,同时把 GPU 的 QA 时间缩短 28.3%、TPU 缩短 25.1%。该工作已被 NeurIPS 2026 接收。

正文 · 原文

Authors:Yuanlin Wen, Elan Markowitz, Zubo Gu, Sami Abu-El-Haija, Bryan Perozzi, Mikhail Galkin, Ali Parviz, Aayush Gupta, Venkata Chivukula, Griffin Hadfield, Aravind Akella, Daniel J. Frame, Vahab Mirrokni, Amin Vahdat

View PDF HTML (experimental)

Abstract:The rapid scale-up of AI infrastructure has made datacenter reliability a cornerstone of the cloud business economy. Since hardware failures in hyper-scale AI clusters can stall multi-million dollar training workloads, the downtime of individual machines can cause outsized economic losses. Minimizing machine downtime, therefore, is more important than ever.
To address this, we present Hardware Diagnoser (HardwareDx), an AI-driven hardware diagnosis system that applies a graph foundation model approach to automated datacenter maintenance. It allows a single model to diagnose many platforms and accelerate onboarding of new hardware products. Deployed at production scale at Google, HardwareDx yields substantial efficiency gains across diverse compute architectures. Real-world evaluations demonstrate that it reduces production downtime by 18.3% on GPU platforms and 16.1% on TPUs. Furthermore, it significantly accelerates hardware deployment, cutting quality assurance (QA) time by 28.3% for GPUs and 25.1% for TPUs. These results establish HardwareDx as a field-tested solution for maximizing the reliability and cost efficiency of next generation AI fleets.
Comments: Accepted to NeurIPS 2026
Subjects: Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR)
Cite as: arXiv:2611.23398 [cs.AI]
  (or arXiv:2611.23398v1 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2611.23398

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Elan Markowitz [view email]
[v1] Tue, 17 Nov 2026 19:01:11 UTC (125 KB)

来源:arXiv cs.AI · arxiv.org