LLM 解码的 Token-能耗定律：活跃字节与实际时间解释每 token 焦耳数

Computer Science > Machine Learning arXiv:2610.28576v1 (cs) [Submitted on 20 Oct 2026] Title:The Token-Energy Law of LLM Decode: Active Bytes and Realized Time Account for Joules per Token Authors:Chenxu Niu, Batuhan Sencer, Wei Zhang, Yong Chen View a PDF of the paper titled The Token-Energy Law of LLM Decode: Active Bytes and Realized Time Account for Joules per Token, by Chenxu Niu and 3 other authors View PDF HTML (experimental) Abstract:Energy-efficient deployment of large language models requires understanding how model architecture, hardware, and serving settings shape energy per generated token. We introduce the Token-Energy Law, a compact, interpretable model of GPU-board energy per output token for dense, full-history decoding. It relates token energy to two core dimensions: active bytes and realized decode time. Active bytes describe logical byte demand per output, separating dense weights shared within a decode iteration from private key--value (KV) histories read and updated for each request. Realized decode time is the elapsed decode duration divided by the number of output tokens produced. We combine the two byte coordinates and time in a common affine form, with coefficients calibrated for each model and hardware/runtime configuration. This formulation explains how batching amortizes shared-weight work, longer attended histories increase private KV demand, and throughput determines the time-associated cost per output. Our experiments cover three models under BF16 decoding on NVIDIA Hopper and Blackwell GPUs, including two-GPU tensor parallelism on Hopper. Using fixed coefficients and observed decode time, held-out condition-mean energy estimates differ from measurements by at most 3.73%. Subjects: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC); Performance (cs.PF) Cite as: arXiv:2610.28576 [cs.LG] (or arXiv:2610.28576v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2610.28576 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Chenxu Niu [view email] [v1] Tue, 20 Oct 2026 17:32:08 UTC (765 KB) Full-text links: Access Paper: View a PDF of the paper titled The Token-Energy Law of LLM Decode: Active Bytes and Realized Time Account for Joules per Token, by Chenxu Niu and 3 other authors View PDF HTML (experimental) TeX Source view license Additional Features Audio Summary Current browse context: cs.LG new | recent | 2026-10 Change to browse by: cs cs.DC cs.PF References & Citations SciX Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv (What is alphaXiv?) Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) DagsHub Toggle DagsHub (What is DagsHub?) GotitPub Toggle Gotit.pub (What is GotitPub?) Huggingface Toggle Hugging Face (What is Huggingface?) ScienceCast Toggle ScienceCast (What is ScienceCast?) Demos Demos Replicate Toggle Replicate (What is Replicate?) Spaces Toggle Hugging Face Spaces (What is Spaces?) Spaces Toggle TXYZ.AI (What is TXYZ.AI?) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower (What are Influence Flowers?) Core recommender toggle CORE Recommender (What is CORE?) IArxiv recommender toggle IArxiv Recommender (What is IArxiv?) Author Venue Institution Topic About arXivLabs 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. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)