--- 格式版本: 2 标题: "An HPC Approach to Accelerate Tensor Decompositions" 原文链接: "https://arxiv.org/abs/2608.24307" 发布日期: "2026-08-25" 发布时间校准状态: "found" 发布时间需复核: "否" 发布时间来源: "rule:local:strict_original_body" 发布时间证据: "**\\[v1\\]** Tue, 25 Aug 2026 09:36:54 UTC (47" 发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body" 发布时间校准置信度: "high" 发布时间候选数量: 20 发布时间严格候选数量: 7 发布时间原页读取状态: "source template page reused from URL open" 发布时间未找到原因: "" 发布时间校准时间: "2026-08-26T21:54:09+08:00" 发布时间仲裁状态: "skipped" 发布时间仲裁尝试次数: 0 发布时间仲裁耗时毫秒: 0 发现时间: "2026-08-26T21:50:39+08:00" 入库时间: "2026-08-26T13:54:09.660Z" 来源平台: "arXiv 学术论文搜索" 搜索渠道: "source_template" 搜索词: "https://arxiv.org/search/?query=Scale-up&searchtype=all" 匹配关键词: - "Scale-up" - "GPU" - "AI" 相关厂家: - "NVIDIA" 相关专家: [] 内容类型: "网页" 抓取工具: "Free Fetch + Defuddle" 清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取" 原始附件: [] AI优质: "否" AI打分: 29 AI分档: "非优质" AI质检状态: "不通过" AI打分理由: "正文主线是单张NVIDIA H100上的CUDA Jacobi张量分解算法，用于量子系统模拟，并非超节点或机架级AI基础设施。arXiv为原始预印本来源，但当前页面仅提供摘要。相较固定知识库，新增事实是支持任意阶张量、验证至九维且较MATLAB参考实现快两个数量级以上；未能据此确认历史首次。没有机架拓扑、互连、供电、散热、RAS或生产部署信息，也无客户、量产和交付信号，命中单一工作负载优化强否决项。" AI质检模型: "gpt-5.6-sol" AI质检时间: "2026-08-26T21:54:25+08:00" AI主题相关性: 1 AI来源权威性: 11 AI新颖性: 6 AI技术细节: 4 AI商业部署信号: 0 AI完整性: 7 AI评分提示词版本: "v17-精简生产版" AI评分提示词SHA256: "48fb9777f386026761b4873eaff30807694fb11e9b352d7c69bf2dfde750cc7d" AI评分知识库版本: "knowledge_base_v1-20260819+runtime.3" AI评分知识库SHA256: "dbc02c7552b478ae5aae514533e58a5de9ce72d2911e4aaac0f744c08178e4a6" AI评分知识库检索词: "[\"Scale-up\",\"GPU\",\"NVIDIA\",\"HPC\",\"PDF\",\"arxiv.org/pdf/2608.24307\",\"HTML\",\"arxiv.org/html/2608.24307v1\",\"v1\",\"CUDA-based\",\"H100\",\"GPUs\"]" AI评分知识库命中: "[{\"id\":\"july-correct-0081\",\"title\":\"AAI 2026: 6th Gen AMD EPYC Server CPUs Power the Agentic Data Center\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"GPU\",\"NVIDIA\",\"HPC\",\"PDF\",\"HTML\",\"GPUs\"],\"rank\":-12.257843607351044},{\"id\":\"july-correct-0078\",\"title\":\"AAI 2026: AMD Launches AMD Instinct MI400 Series GPUs for Frontier AI, HPC\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"GPU\",\"HPC\",\"HTML\",\"GPUs\"],\"rank\":-11.326805051101733},{\"id\":\"july-correct-0016\",\"title\":\"AMD launches Instinct MI400 Series GPUs for AI workloads\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"Scale-up\",\"GPU\",\"HPC\",\"GPUs\"],\"rank\":-10.76259043490252},{\"id\":\"july-correct-0111\",\"title\":\"StrataCL: Fabric-Native Communication Library for Production Supernodes\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"PDF\",\"HTML\",\"v1\"],\"rank\":-10.496705137689483},{\"id\":\"july-correct-0089\",\"title\":\"Setting a World Record for MoE Pre-Training on NVIDIA GB300 NVL72\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"Scale-up\",\"GPU\",\"NVIDIA\",\"HTML\",\"GPUs\"],\"rank\":-9.360241728473463}]" AI摘要: "研究人员提出一种Jacobi型张量分解算法，并开发了支持任意阶张量的CUDA实现，在NVIDIA H100 GPU上成功处理至九维可对角化张量，运行时间随阶数增长呈可预测扩展。该算法比原始MATLAB参考实现快两个数量级以上。" AI摘要模型: "ali-deepseek-v4-flash" AI摘要时间: "2026-08-27T08:12:42.071Z" 采集批次: "2026年8月26日20点37分07秒" 采集批次ID: "20260826-203707-934" 去重键: "https://arxiv.org/abs/2608.24307" --- ## Computer Science > Distributed, Parallel, and Cluster Computing **arXiv:2608.24307** (cs) ## Title:An HPC Approach to Accelerate Tensor Decompositions [View PDF](https://arxiv.org/pdf/2608.24307) [HTML (experimental)](https://arxiv.org/html/2608.24307v1) > Abstract:Quantum systems grow in complexity so rapidly that even modest models become difficult to simulate, creating a strong need for methods that can handle high-dimensional data, also known as tensors. In this work, we investigate a novel Jacobi-type tensor algorithm for tensor decomposition and develop a CUDA-based algorithm that supports tensors of arbitrary order on a single GPU. We test the implementation on NVIDIA H100 GPUs and show that the algorithm converges correctly for diagonalizable tensors up to nine dimensions, with runtime scaling in a predictable way as tensor order grows. Finally, our general algorithm outperforms the original MATLAB reference by more than two orders of magnitude. | Subjects: | Distributed, Parallel, and Cluster Computing (cs.DC); Mathematical Software (cs.MS) | | --- | --- | | Cite as: | [arXiv:2608.24307](https://arxiv.org/abs/2608.24307) \[cs.DC\] | | | (or [arXiv:2608.24307v1](https://arxiv.org/abs/2608.24307v1) \[cs.DC\] for this version) | | | [https://doi.org/10.48550/arXiv.2608.24307](https://doi.org/10.48550/arXiv.2608.24307) Focus to learn more arXiv-issued DOI via DataCite (pending registration) | ## Submission history From: Roman Iakymchuk \[[view email](https://arxiv.org/show-email/6e027a00/2608.24307)\] **\[v1\]** Tue, 25 Aug 2026 09:36:54 UTC (47 KB) ## Bibliographic and Citation Tools Bibliographic Explorer *([What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))* Connected Papers *([What is Connected Papers?](https://www.connectedpapers.com/about))* Litmaps *([What is Litmaps?](https://www.litmaps.co/))* scite Smart Citations *([What are Smart Citations?](https://www.scite.ai/))* ## Code, Data and Media Associated with this Article alphaXiv *([What is alphaXiv?](https://alphaxiv.org/))* CatalyzeX Code Finder for Papers *([What is CatalyzeX?](https://www.catalyzex.com/))* DagsHub *([What is DagsHub?](https://dagshub.com/))* Gotit.pub *([What is GotitPub?](http://gotit.pub/faq))* Hugging Face *([What is Huggingface?](https://huggingface.co/huggingface))* ScienceCast *([What is ScienceCast?](https://sciencecast.org/welcome))* ## Demos Replicate *([What is Replicate?](https://replicate.com/docs/arxiv/about))* Hugging Face Spaces *([What is Spaces?](https://huggingface.co/docs/hub/spaces))* TXYZ.AI *([What is TXYZ.AI?](https://txyz.ai/))* ## Recommenders and Search Tools Influence Flower *([What are Influence Flowers?](https://influencemap.cmlab.dev/))* CORE Recommender *([What is CORE?](https://core.ac.uk/services/recommender))* ## arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? [**Learn more about arXivLabs**](https://info.arxiv.org/labs/index.html). [Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.24307) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))