{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,23]],"date-time":"2026-10-23T22:50:46Z","timestamp":1792795846992,"version":"4.5.0"},"reference-count":46,"publisher":"IEEE","license":[{"start":{"date-parts":[[2026,1,31]],"date-time":"2026-01-31T00:00:00Z","timestamp":1769817600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,1,31]],"date-time":"2026-01-31T00:00:00Z","timestamp":1769817600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100010418","name":"Institute for Information & communications Technology Promotion","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100010418","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,1,31]]},"DOI":"10.1109\/hpca68181.2026.11408603","type":"proceedings-article","created":{"date-parts":[[2026,3,4]],"date-time":"2026-03-04T20:47:22Z","timestamp":1772657242000},"page":"1-14","source":"Crossref","is-referenced-by-count":2,"title":["V-Rex: Real-Time Streaming Video LLM Acceleration via Dynamic KV Cache Retrieval"],"prefix":"10.1109","author":[{"given":"Donghyuk","family":"Kim","sequence":"first","affiliation":[{"name":"KAIST,Daejeon,Republic of Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sejeong","family":"Yang","sequence":"additional","affiliation":[{"name":"KAIST,Daejeon,Republic of Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wonjin","family":"Shin","sequence":"additional","affiliation":[{"name":"KAIST,Daejeon,Republic of Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Joo-Young","family":"Kim","sequence":"additional","affiliation":[{"name":"KAIST,Daejeon,Republic of Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","volume-title":"Bg6 series (m.2) \u2014 kioxia \u2013 japan","year":"2025"},{"key":"ref2","volume-title":"Jetson agx orin for next-gen robotics \u2013 nvidia","year":"2025"},{"key":"ref3","volume-title":"Nvidia a100 tensor core gpu","year":"2025"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01742"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA61900.2025.00084"},{"key":"ref6","article-title":"Streaming video question-answering with in-context video KV-cache retrieval","volume-title":"The Thirteenth International Conference on Learning Representations","author":"Di","year":"2025"},{"key":"ref7","article-title":"Model tells you what to discard: Adaptive KV cache compression for LLMs","volume-title":"The Twelfth International Conference on Learning Representations","author":"Ge","year":"2024"},{"key":"ref8","author":"Grattafiori","year":"2024","journal-title":"The Llama 3 Herd of Models"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA61900.2025.00034"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO56248.2022.00051"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/HOTCHIPS.2016.7936170"},{"key":"ref12","author":"Kim","year":"2025","journal-title":"Infinipot-v: Memoryconstrained kv cache compression for streaming video understanding"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1145\/3695053.3731019"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1145\/3695053.3731073"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1109\/IISWC.2017.8167762"},{"key":"ref16","first-page":"155","article-title":"{InfiniGen}: Efficient generative inference of large language models with dynamic {KV} cache manage-ment","volume-title":"18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 24)","author":"Lee","year":"2024"},{"key":"ref17","article-title":"CommVQ: Commutative vector quantization for KV cache compression","volume-title":"Forty-second International Conference on Machine Learning","author":"Li","year":"2025"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/LCA.2020.2973991"},{"key":"ref19","article-title":"Minicache: Kv cache compression in depth dimension for large language models","volume-title":"Proceedings of the 38th International Conference on Neural Information Processing Systems, ser. NIPS \u201924. Red Hook","author":"Liu","year":"2025"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1145\/3651890.3672274"},{"key":"ref21","article-title":"Kivi: a tuning-free asymmetric 2bit quantization for kv cache","volume-title":"Proceedings of the 41st International Conference on Machine Learning, ser. ICML\u201924. JMLR.org","author":"Liu","year":"2024"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1145\/3695053.3731057"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/MM.2024.3420728"},{"key":"ref24","author":"Ning","year":"2025","journal-title":"Livevlm: Efficient online video understanding via streaming-oriented kv cache and retrieval"},{"key":"ref25","volume-title":"NVIDIA System Management Interface (nvidiasmi)","year":"2025"},{"key":"ref26","year":"2025","journal-title":"NVIDIA DRIVE OS 5.1 Linux SDK Documentation"},{"issue":"18\u201324","key":"ref27","first-page":"8748","article-title":"Learning transferable visual models from natural language supervision","volume-title":"Proceedings of the 38th International Conference on Machine Learning, ser. Proceedings of Machine Learning Research","volume":"139","author":"Radford"},{"key":"ref28","author":"Schneider","year":"2025","journal-title":"Quickvideo: Real-time long video understanding with system algorithm co-design"},{"key":"ref29","article-title":"LongVU: Spatiotemporal adaptive compression for long video-language understanding","volume-title":"Forty-second International Conference on Machine Learning","author":"Shen","year":"2025"},{"key":"ref30","first-page":"31094","article-title":"Flexgen: High-throughput generative inference of large language models with a single gpu","volume-title":"International Conference on Machine Learning. PMLR","author":"Sheng","year":"2023"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00130"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1145\/3706598.3714224"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1109\/TCSVT.2025.3566695"},{"key":"ref34","author":"Tao","year":"2025","journal-title":"Plug-and-play 1. x-bit kv cache quantization for video large language models"},{"key":"ref35","first-page":"49","article-title":"{MQSim}: A framework for enabling realistic studies of modern {Multi-Queue}{SSD} devices","volume-title":"16th USENIX Conference on File and Storage Technologies (FAST 18)","author":"Tavakkol","year":"2018"},{"key":"ref36","article-title":"VL-cache: Sparsity and modality-aware KV cache compression for vision-language model inference acceleration","volume-title":"The Thirteenth International Conference on Learning Representations","author":"Tu","year":"2025"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.52202\/079017-3488"},{"key":"ref38","article-title":"Streaming video understanding and multi-round interaction with memory-enhanced knowledge","volume-title":"The Thirteenth International Conference on Learning Representations","author":"Xiong","year":"2025"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1145\/3695053.3730996"},{"key":"ref40","volume-title":"Qwen3 technical report","author":"Yang","year":"2025"},{"key":"ref41","first-page":"3258","article-title":"Pyramidinfer: Pyramid kv cache compression for high-throughput lm inference","author":"Yang","year":"2024","journal-title":"ACL (Findings)"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1093\/nsr\/nwae403"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO61859.2024.00015"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01100"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.emnlp-demo.49"},{"key":"ref46","volume-title":"Unifying kv cache compression for large language models with leankv","author":"Zhang","year":"2025"}],"event":{"name":"2026 IEEE International Symposium on High Performance Computer Architecture (HPCA)","location":"Sydney, Australia","start":{"date-parts":[[2026,1,31]]},"end":{"date-parts":[[2026,2,4]]}},"container-title":["2026 IEEE International Symposium on High Performance Computer Architecture (HPCA)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11408404\/11408433\/11408603.pdf?arnumber=11408603","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,5]],"date-time":"2026-03-05T20:41:35Z","timestamp":1772743295000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11408603\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1,31]]},"references-count":46,"URL":"https:\/\/doi.org\/10.1109\/hpca68181.2026.11408603","relation":{},"subject":[],"published":{"date-parts":[[2026,1,31]]}}}