{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,15]],"date-time":"2026-06-15T07:56:21Z","timestamp":1781510181316,"version":"3.54.1"},"reference-count":43,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/tdm\/userlicense\/1.0\/"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/legal\/tdmrep-license"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-017"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-012"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-004"}],"content-domain":{"domain":["elsevier.com","sciencedirect.com"],"crossmark-restriction":true},"short-container-title":["Neurocomputing"],"published-print":{"date-parts":[[2026,10]]},"DOI":"10.1016\/j.neucom.2026.134247","type":"journal-article","created":{"date-parts":[[2026,6,8]],"date-time":"2026-06-08T16:35:45Z","timestamp":1780936545000},"page":"134247","update-policy":"https:\/\/doi.org\/10.1016\/elsevier_cm_policy","source":"Crossref","is-referenced-by-count":0,"special_numbering":"C","title":["HCAttention: Extreme KV cache compression via heterogeneous attention computing for LLMs"],"prefix":"10.1016","volume":"697","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-5200-8581","authenticated-orcid":false,"given":"Dongquan","family":"Yang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yifan","family":"Yang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8428-4650","authenticated-orcid":false,"given":"Xiaotian","family":"Yu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xianbiao","family":"Qi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Rong","family":"Xiao","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","reference":[{"key":"10.1016\/j.neucom.2026.134247_bib0005","author":"Achiam"},{"key":"10.1016\/j.neucom.2026.134247_bib0010","author":"Bai"},{"key":"10.1016\/j.neucom.2026.134247_bib0015","author":"Liu"},{"key":"10.1016\/j.neucom.2026.134247_bib0020","author":"Touvron"},{"key":"10.1016\/j.neucom.2026.134247_bib0025","author":"Ni"},{"key":"10.1016\/j.neucom.2026.134247_bib0030","author":"Grattafiori"},{"key":"10.1016\/j.neucom.2026.134247_bib0035","author":"Team"},{"key":"10.1016\/j.neucom.2026.134247_bib0040","author":"Jiang"},{"key":"10.1016\/j.neucom.2026.134247_bib0045","author":"Zhang"},{"key":"10.1016\/j.neucom.2026.134247_bib0050","author":"Li"},{"key":"10.1016\/j.neucom.2026.134247_bib0055","author":"Li"},{"key":"10.1016\/j.neucom.2026.134247_bib0060","author":"Zhang"},{"key":"10.1016\/j.neucom.2026.134247_bib0065","first-page":"132208","article-title":"Chain of agents: large language models collaborating on long-context tasks","volume":"37","author":"Zhang","year":"2024","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"10.1016\/j.neucom.2026.134247_bib0070","author":"Mao"},{"key":"10.1016\/j.neucom.2026.134247_bib0075","series-title":"Proceedings of the 31st International Conference on Neural Information Processing Systems","first-page":"6000","article-title":"Attention is all you need","author":"Vaswani","year":"2017"},{"key":"10.1016\/j.neucom.2026.134247_bib0080","author":"Zhang"},{"key":"10.1016\/j.neucom.2026.134247_bib0085","author":"Shazeer"},{"key":"10.1016\/j.neucom.2026.134247_bib0090","author":"Ainslie"},{"key":"10.1016\/j.neucom.2026.134247_bib0095","article-title":"Beyond kv caching: shared attention for efficient LLMs","volume":"648","author":"Liao","year":"2025","journal-title":"Neurocomputing"},{"key":"10.1016\/j.neucom.2026.134247_bib0100","doi-asserted-by":"crossref","DOI":"10.1016\/j.neucom.2023.127063","article-title":"Roformer: enhanced transformer with rotary position embedding","volume":"568","author":"Su","year":"2024","journal-title":"Neurocomputing"},{"key":"10.1016\/j.neucom.2026.134247_bib0105","first-page":"34661","article-title":"H2o: heavy-hitter Oracle for efficient generative inference of large language models","volume":"36","author":"Zhang","year":"2023","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"10.1016\/j.neucom.2026.134247_bib0110","author":"Xiao"},{"key":"10.1016\/j.neucom.2026.134247_bib0115","series-title":"Proceedings of the 41st International Conference on Machine Learning","first-page":"32332","article-title":"Kivi: a tuning-free asymmetric 2bit quantization for kv cache","author":"Liu","year":"2024"},{"key":"10.1016\/j.neucom.2026.134247_bib0120","first-page":"3304","article-title":"Kv cache is 1 bit per channel: efficient large language model inference with coupled quantization","volume":"37","author":"Zhang","year":"2024","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"10.1016\/j.neucom.2026.134247_bib0125","author":"He"},{"key":"10.1016\/j.neucom.2026.134247_bib0130","series-title":"Proceedings of the ACM SIGCOMM 2024 Conference","first-page":"38","article-title":"Cachegen: kv cache compression and streaming for fast large language model serving","author":"Liu","year":"2024"},{"key":"10.1016\/j.neucom.2026.134247_bib0135","series-title":"Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing","first-page":"18724","article-title":"Transformers are multi-state RNNS","author":"Oren","year":"2024"},{"key":"10.1016\/j.neucom.2026.134247_bib0140","series-title":"The Twelfth International Conference on Learning Representations","article-title":"Efficient streaming language models with attention sinks","author":"Xiao","year":"2023"},{"key":"10.1016\/j.neucom.2026.134247_bib0145","series-title":"Proceedings of Machine Learning and Systems","first-page":"114","article-title":"Keyformer: kv cache reduction through key tokens selection for efficient generative inference","volume":"vol. 6","author":"Adnan","year":"2024"},{"key":"10.1016\/j.neucom.2026.134247_bib0150","series-title":"The Thirteenth International Conference on Learning Representations","article-title":"SqueezeAttention: 2D management of kv-cache in LLM inference via layer-wise optimal budget","author":"Wang","year":"2025"},{"key":"10.1016\/j.neucom.2026.134247_bib0155","series-title":"Proceedings of the International Conference on Machine Learning (ICML)","article-title":"QUEST: query-aware sparsity for efficient long-context LLM inference","author":"Tang","year":"2024"},{"key":"10.1016\/j.neucom.2026.134247_bib0160","series-title":"Proceedings of the 31st International Conference on Computational Linguistics","first-page":"2316","article-title":"Asymkv: enabling 1-bit quantization of kv cache with layer-wise asymmetric quantization configurations","author":"Tao","year":"2025"},{"key":"10.1016\/j.neucom.2026.134247_bib0165","series-title":"Proceedings of the 38th International Conference on Neural Information Processing Systems","first-page":"68287","article-title":"Zipcache: accurate and efficient kv cache quantization with salient token identification","author":"He","year":"2024"},{"key":"10.1016\/j.neucom.2026.134247_bib0170","series-title":"International Conference on Machine Learning","first-page":"57355","article-title":"ShadowKV: KV cache in shadows for high-throughput long-context LLM inference","author":"Sun","year":"2025"},{"key":"10.1016\/j.neucom.2026.134247_bib0175","series-title":"18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 24)","first-page":"155","article-title":"{InfiniGen}: efficient generative inference of large language models with dynamic {KV} cache management","author":"Lee","year":"2024"},{"key":"10.1016\/j.neucom.2026.134247_bib0180","series-title":"The 39th Annual Conference on Neural Information Processing Systems","first-page":"47561","article-title":"HIFC: high-efficiency flash-based kv cache swapping for scaling LLM inference","author":"Jeong","year":"2025"},{"key":"10.1016\/j.neucom.2026.134247_bib0185","author":"Pan"},{"key":"10.1016\/j.neucom.2026.134247_bib0190","author":"Jiang"},{"key":"10.1016\/j.neucom.2026.134247_bib0195","series-title":"Proceedings of the 5th Workshop on Machine Learning and Systems","first-page":"56","article-title":"FlexInfer: breaking memory constraint via flexible and efficient offloading for on-device LLM inference","author":"Du","year":"2025"},{"key":"10.1016\/j.neucom.2026.134247_bib0200","series-title":"Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","first-page":"3119","article-title":"LongBench: a bilingual, multitask benchmark for long context understanding","author":"Bai","year":"2024"},{"key":"10.1016\/j.neucom.2026.134247_bib0205","author":"Kamradt"},{"key":"10.1016\/j.neucom.2026.134247_bib0210","series-title":"Proceedings of the 19th International Conference on World Wide Web","first-page":"1177","article-title":"Web-scale K-means clustering","author":"Sculley","year":"2010"},{"key":"10.1016\/j.neucom.2026.134247_bib0215","first-page":"2825","article-title":"Scikit-Learn: machine learning in Python","volume":"12","author":"Pedregosa","year":"2011","journal-title":"J. Mach. Learn. Res."}],"container-title":["Neurocomputing"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S0925231226016450?httpAccept=text\/xml","content-type":"text\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S0925231226016450?httpAccept=text\/plain","content-type":"text\/plain","content-version":"vor","intended-application":"text-mining"}],"deposited":{"date-parts":[[2026,6,15]],"date-time":"2026-06-15T07:50:44Z","timestamp":1781509844000},"score":1,"resource":{"primary":{"URL":"https:\/\/linkinghub.elsevier.com\/retrieve\/pii\/S0925231226016450"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10]]},"references-count":43,"alternative-id":["S0925231226016450"],"URL":"https:\/\/doi.org\/10.1016\/j.neucom.2026.134247","relation":{},"ISSN":["0925-2312"],"issn-type":[{"value":"0925-2312","type":"print"}],"subject":[],"published":{"date-parts":[[2026,10]]},"assertion":[{"value":"Elsevier","name":"publisher","label":"This article is maintained by"},{"value":"HCAttention: Extreme KV cache compression via heterogeneous attention computing for LLMs","name":"articletitle","label":"Article Title"},{"value":"Neurocomputing","name":"journaltitle","label":"Journal Title"},{"value":"https:\/\/doi.org\/10.1016\/j.neucom.2026.134247","name":"articlelink","label":"CrossRef DOI link to publisher maintained version"},{"value":"article","name":"content_type","label":"Content Type"},{"value":"\u00a9 2026 Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.","name":"copyright","label":"Copyright"}],"article-number":"134247"}}