{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,6,14]],"date-time":"2025-06-14T04:03:35Z","timestamp":1749873815552,"version":"3.41.0"},"reference-count":28,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,5,18]],"date-time":"2025-05-18T00:00:00Z","timestamp":1747526400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,5,18]],"date-time":"2025-05-18T00:00:00Z","timestamp":1747526400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,5,18]]},"DOI":"10.1109\/imw61990.2025.11026974","type":"proceedings-article","created":{"date-parts":[[2025,6,12]],"date-time":"2025-06-12T17:40:04Z","timestamp":1749750004000},"page":"1-4","source":"Crossref","is-referenced-by-count":0,"title":["Analog AI Accelerators for Transformer-based Language Models: Hardware, Workload, and Power Performance"],"prefix":"10.1109","author":[{"given":"H.","family":"Tsai","sequence":"first","affiliation":[{"name":"IBM Research–Almaden,San Jose,CA,USA"}]},{"given":"H.","family":"Benmeziane","sequence":"additional","affiliation":[{"name":"IBM Research–Zurich,Rüschlikon,Switzerland"}]},{"given":"I.","family":"Boybat","sequence":"additional","affiliation":[{"name":"IBM Research–Zurich,Rüschlikon,Switzerland"}]},{"given":"J.","family":"B\u00fcchel","sequence":"additional","affiliation":[{"name":"IBM Research–Zurich,Rüschlikon,Switzerland"}]},{"given":"P.","family":"Narayanan","sequence":"additional","affiliation":[{"name":"IBM Research–Almaden,San Jose,CA,USA"}]},{"given":"M. Le","family":"Gallo","sequence":"additional","affiliation":[{"name":"IBM Research–Zurich,Rüschlikon,Switzerland"}]},{"given":"S.","family":"Jain","sequence":"additional","affiliation":[{"name":"IBM Research–T. J. Watson Research Center,Yorktown Heights,NY,USA"}]},{"given":"A.","family":"Vasilopoulos","sequence":"additional","affiliation":[{"name":"IBM Research–Zurich,Rüschlikon,Switzerland"}]},{"given":"W.","family":"Simon","sequence":"additional","affiliation":[{"name":"IBM Research–Zurich,Rüschlikon,Switzerland"}]},{"given":"K.","family":"Hosokawa","sequence":"additional","affiliation":[{"name":"IBM Tokyo Research Laboratory,Tokyo,Japan"}]},{"given":"M.","family":"Ishii","sequence":"additional","affiliation":[{"name":"IBM Tokyo Research Laboratory,Tokyo,Japan"}]},{"given":"Y.","family":"Kohda","sequence":"additional","affiliation":[{"name":"IBM Tokyo Research Laboratory,Tokyo,Japan"}]},{"given":"A.","family":"Chen","sequence":"additional","affiliation":[{"name":"IBM Research–Almaden,San Jose,CA,USA"}]},{"given":"C.","family":"Mackin","sequence":"additional","affiliation":[{"name":"IBM Research–Almaden,San Jose,CA,USA"}]},{"given":"K. El","family":"Maghraoui","sequence":"additional","affiliation":[{"name":"IBM Research–T. J. Watson Research Center,Yorktown Heights,NY,USA"}]},{"given":"A.","family":"Okazaki","sequence":"additional","affiliation":[{"name":"IBM Tokyo Research Laboratory,Tokyo,Japan"}]},{"given":"A. M.","family":"Friz","sequence":"additional","affiliation":[{"name":"IBM Research–Almaden,San Jose,CA,USA"}]},{"given":"J.","family":"Luquin","sequence":"additional","affiliation":[{"name":"IBM Research–Almaden,San Jose,CA,USA"}]},{"given":"A.","family":"Sebastian","sequence":"additional","affiliation":[{"name":"IBM Research–Zurich,Rüschlikon,Switzerland"}]},{"given":"V.","family":"Narayanan","sequence":"additional","affiliation":[{"name":"IBM Research–T. J. Watson Research Center,Yorktown Heights,NY,USA"}]},{"given":"G. W.","family":"Burr","sequence":"additional","affiliation":[{"name":"IBM Research–Almaden,San Jose,CA,USA"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"year":"2024","key":"ref2","article-title":"Granite 3.0 language models"},{"article-title":"Mixtral of experts","year":"2024","author":"Jiang","key":"ref3"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.emnlp-main.298"},{"article-title":"DeepSeek-V3 Technical Report","year":"2024","author":"Liu","key":"ref5"},{"article-title":"Transformers are SSMs: Generalized models and efficient algorithms through structured state space duality","year":"2024","author":"Dao","key":"ref6"},{"year":"2024","key":"ref7","article-title":"Bamba: Inference-efficient hybrid Mamba2 model"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/MSPEC.2021.9641759"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/TCSI.2021.3064189"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-022-04992-8"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/TED.2021.3115993"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1038\/s41928-023-01010-1"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-023-06337-5"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2019.2907886"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1109\/BioCAS58349.2023.10388585"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1126\/sciadv.aau5759"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1109\/VLSI-TSA51926.2021.9440071"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/JSSC.2021.3098671"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/TVLSI.2022.3221390"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2025.3550811"},{"key":"ref21","first-page":"1","article-title":"Design of Analog-AI hardware accelerators for Transformer-based language models","volume-title":"2023 IEDM","author":"Burr"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/IEDM50854.2024.10873479"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1063\/5.0168089"},{"article-title":"AIHWKIT-lightning: a scalable HW-aware training toolkit for analog in-memory computing","volume-title":"NeurIPS 2024 Workshop: Machine Learning with new Compute Paradigms","author":"B\u00fcchel","key":"ref24"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-023-40770-4"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.23919\/DATE64628.2025.10993217"},{"article-title":"Cerebras-GPT: Open compute-optimal language models trained on the cerebras wafer-scale cluster","year":"2023","author":"Dey","key":"ref27"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1038\/s43588-024-00753-x"}],"event":{"name":"2025 IEEE International Memory Workshop (IMW)","start":{"date-parts":[[2025,5,18]]},"location":"Monterey, CA, USA","end":{"date-parts":[[2025,5,21]]}},"container-title":["2025 IEEE International Memory Workshop (IMW)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11026918\/11026883\/11026974.pdf?arnumber=11026974","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,13]],"date-time":"2025-06-13T05:37:18Z","timestamp":1749793038000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11026974\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,5,18]]},"references-count":28,"URL":"https:\/\/doi.org\/10.1109\/imw61990.2025.11026974","relation":{},"subject":[],"published":{"date-parts":[[2025,5,18]]}}}