{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,20]],"date-time":"2026-07-20T21:04:53Z","timestamp":1784581493005,"version":"3.55.0"},"reference-count":15,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"2","license":[{"start":{"date-parts":[[2026,7,1]],"date-time":"2026-07-01T00:00:00Z","timestamp":1782864000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,7,1]],"date-time":"2026-07-01T00:00:00Z","timestamp":1782864000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,7,1]],"date-time":"2026-07-01T00:00:00Z","timestamp":1782864000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"Institute of Information & Communications Technology Planning and Evaluation"},{"name":"Korea government","award":["RS-2024-00399936"],"award-info":[{"award-number":["RS-2024-00399936"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Comput. Arch. Lett."],"published-print":{"date-parts":[[2026,7]]},"DOI":"10.1109\/lca.2026.3703423","type":"journal-article","created":{"date-parts":[[2026,6,15]],"date-time":"2026-06-15T19:45:56Z","timestamp":1781552756000},"page":"267-270","source":"Crossref","is-referenced-by-count":0,"title":["On the Feasibility Boundaries of LLM Prefill Inference on Unified-Memory Edge SoCs"],"prefix":"10.1109","volume":"25","author":[{"ORCID":"https:\/\/orcid.org\/0009-0008-0947-7370","authenticated-orcid":false,"given":"Juyeon","family":"Lee","sequence":"first","affiliation":[{"name":"Computer Architecture and Parallel Processing Lab (CAPP), Department of Electrical and Computer Engineering, Seoul National University, Seoul, South Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jiayue","family":"Wang","sequence":"additional","affiliation":[{"name":"Computer Architecture and Parallel Processing Lab (CAPP), Department of Electrical and Computer Engineering, Seoul National University, Seoul, South Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-5228-9801","authenticated-orcid":false,"given":"Soojung","family":"Ryu","sequence":"additional","affiliation":[{"name":"Computer Architecture and Parallel Processing Lab (CAPP), Department of Electrical and Computer Engineering, Seoul National University, Seoul, South Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6811-9647","authenticated-orcid":false,"given":"Hyuk-Jae","family":"Lee","sequence":"additional","affiliation":[{"name":"Computer Architecture and Parallel Processing Lab (CAPP), Department of Electrical and Computer Engineering, Seoul National University, Seoul, South Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/MCHPC49590.2019.00014"},{"key":"ref2","article-title":"Solving entry-level edge AI challenges with NVIDIA Jetson Orin Nano","author":"Karumbunathan","year":"2022"},{"key":"ref3","first-page":"31094","article-title":"FlexGen: High-throughput generative inference of large language models with a single GPU","volume-title":"Proc. 40th Int. Conf. Mach. Learn.","author":"Sheng","year":"2023"},{"key":"ref4","article-title":"On-device language models: A comprehensive review","author":"Xu","year":"2024"},{"issue":"4","key":"ref5","first-page":"82","article-title":"Technical trends in on-device small language model technology development","volume":"39","author":"Kim","year":"2024","journal-title":"Electron. Telecommun. Trends"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/MCOM.001.2300550"},{"key":"ref7","first-page":"193","article-title":"DistServe: Disaggregating prefill and decoding for goodput-optimized large language model serving","volume-title":"Proc. USENIX Conf. Operating Syst. Des. Implementation","author":"Zhong","year":"2024"},{"key":"ref8","article-title":"FastDecode: High-throughput GPU-efficient LLM serving using heterogeneous pipelines","author":"He","year":"2024"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO61859.2024.00030"},{"key":"ref10","first-page":"293","article-title":"SARATHI: Efficient LLM inference by piggybacking decodes with chunked prefills","volume-title":"Proc. USENIX Annu. Techn. Conf.","author":"Agrawal","year":"2024"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1145\/3724420"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.acl-long.718"},{"key":"ref13","article-title":"On evaluating performance of LLM inference serving systems","author":"Agrawal","year":"2025"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/ICCD65941.2025.00043"},{"key":"ref15","article-title":"Unified splitting: Inference utilizing unified memory architecture","volume-title":"Proc. ACM Meas. Anal. Comput. Syst.","volume":"37","author":"Due\u00f1as","year":"2025"}],"container-title":["IEEE Computer Architecture Letters"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/10208\/11579484\/11563512.pdf?arnumber=11563512","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,20]],"date-time":"2026-07-20T20:18:57Z","timestamp":1784578737000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11563512\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7]]},"references-count":15,"journal-issue":{"issue":"2"},"URL":"https:\/\/doi.org\/10.1109\/lca.2026.3703423","relation":{},"ISSN":["1556-6056","1556-6064","2473-2575"],"issn-type":[{"value":"1556-6056","type":"print"},{"value":"1556-6064","type":"electronic"},{"value":"2473-2575","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7]]}}}