{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,12,12]],"date-time":"2025-12-12T13:50:57Z","timestamp":1765547457366,"version":"3.44.0"},"reference-count":30,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,6,22]],"date-time":"2025-06-22T00:00:00Z","timestamp":1750550400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,6,22]],"date-time":"2025-06-22T00:00:00Z","timestamp":1750550400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/100006190","name":"Research and Development","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100006190","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,6,22]]},"DOI":"10.1109\/dac63849.2025.11133274","type":"proceedings-article","created":{"date-parts":[[2025,9,15]],"date-time":"2025-09-15T17:35:41Z","timestamp":1757957741000},"page":"1-7","source":"Crossref","is-referenced-by-count":2,"title":["HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference"],"prefix":"10.1109","author":[{"given":"Shuzhang","family":"Zhong","sequence":"first","affiliation":[{"name":"Institute for Artificial Intelligence, Peking University,Beijing,China"}]},{"given":"Yanfan","family":"Sun","sequence":"additional","affiliation":[{"name":"Beihang University,School of Computer Science and Engineering,Beijing,China"}]},{"given":"Ling","family":"Liang","sequence":"additional","affiliation":[{"name":"Peking University,School of Integrated Circuits,Beijing,China"}]},{"given":"Runsheng","family":"Wang","sequence":"additional","affiliation":[{"name":"Peking University,School of Integrated Circuits,Beijing,China"}]},{"given":"Ru","family":"Huang","sequence":"additional","affiliation":[{"name":"Peking University,School of Integrated Circuits,Beijing,China"}]},{"given":"Meng","family":"Li","sequence":"additional","affiliation":[{"name":"Institute for Artificial Intelligence, Peking University,Beijing,China"}]}],"member":"263","reference":[{"key":"ref1","article-title":"Outrageously large neural networks: The sparsely-gated mixture-of-experts layer","author":"Shazeer","year":"2017","journal-title":"arXiv preprint arXiv:1701.06538"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1007\/s10462-012-9338-y"},{"key":"ref3","article-title":"Fast inference of mixture-of-experts language models with offloading","author":"Eliseev","year":"2023","journal-title":"arXiv preprint arXiv:2312.17238"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1109\/isca59077.2024.00078"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1145\/3676536.3676741"},{"key":"ref6","article-title":"Promoe: Fast moe-based 11 m serving using proactive caching","author":"Song","year":"2024","journal-title":"arXiv preprint arXiv:2410.22134"},{"key":"ref7","article-title":"Hobbit: A mixed precision expert offloading system for fast moe inference","author":"Tang","year":"2024","journal-title":"arXiv preprint arXiv:2411.01433"},{"key":"ref8","first-page":"633","article-title":"Ship compute or ship data? why not both?","volume-title":"18th USENIX Symposium on Networked Systems Design and Implementation (NSDI 21)","author":"You"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1145\/3656019.3676949"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1145\/3694715.3695964"},{"key":"ref11","article-title":"Caraserve: Cpu-assisted and rank-aware lora serving for generative 1 lm inference","author":"Li","year":"2024","journal-title":"arXiv preprint arXiv:2401.11240"},{"key":"ref12","article-title":"Fiddler: Cpu-gpu orchestration for fast inference of mixture-of-experts models","author":"Kamahori","year":"2024","journal-title":"arXiv preprint arXiv:2402.07033"},{"year":"2024","key":"ref13","article-title":"Ktransformers: A flexible framework for experiencing cutting-edge 1 lm inference optimizations"},{"issue":"120","key":"ref14","first-page":"1","article-title":"Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity","volume":"23","author":"Fedus","year":"2022","journal-title":"Journal of Machine Learning Research"},{"key":"ref15","article-title":"No language left behind: Scaling human-centered machine translation","author":"Costa-juss\u00e0","year":"2022","journal-title":"arXiv preprint arXiv:2207.04672"},{"key":"ref16","article-title":"Qwen2 technical report","volume-title":"arXiv preprint arXiv:2407.10671","author":"Yang","year":"2024"},{"key":"ref17","article-title":"Mixtral of experts","author":"Jiang","year":"2024","journal-title":"arXiv preprint arXiv:2401.04088"},{"key":"ref18","article-title":"Deepseek-v2: A strong, economical, and efficient mixture-of-experts language model","author":"Liu","year":"2024","journal-title":"arXiv preprint arXiv:2405.04434"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.70"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/SC41404.2022.00051"},{"key":"ref21","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"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.emnlp-main.217"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/sc41406.2024.00096"},{"key":"ref24","article-title":"Examining post-training quantization for mixture-of-experts: A benchmark","author":"Li","year":"2024","journal-title":"arXiv preprint arXiv:2406.08155"},{"key":"ref25","article-title":"Moe-infinity: Activationaware expert offloading for efficient moe serving","author":"Xue","year":"2024","journal-title":"arXiv preprint arXiv:2401.14361"},{"article-title":"Expertflow: Optimized expert activation and token allocation for efficient mixture-of-experts inference","year":"2024","author":"He","key":"ref26"},{"article-title":"ggerganov\/llama.cpp: Port of facebook\u2019s llama model in c\/c++","year":"2023","author":"Gerganov","key":"ref27"},{"key":"ref28","article-title":"Judging llm-as-a-judge with mt-bench and chatbot arena","volume":"36","author":"Zheng","year":"2024","journal-title":"Advances in Neural Information Processing Systems"},{"article-title":"Judging 1lm-as-a-judge with mt-bench and chatbot arena","year":"2023","author":"Zheng","key":"ref29"},{"year":"2023","key":"ref30","article-title":"Chatgpt-prompts"}],"event":{"name":"2025 62nd ACM\/IEEE Design Automation Conference (DAC)","start":{"date-parts":[[2025,6,22]]},"location":"San Francisco, CA, USA","end":{"date-parts":[[2025,6,25]]}},"container-title":["2025 62nd ACM\/IEEE Design Automation Conference (DAC)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11132383\/11132091\/11133274.pdf?arnumber=11133274","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,16]],"date-time":"2025-09-16T05:25:17Z","timestamp":1758000317000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11133274\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,6,22]]},"references-count":30,"URL":"https:\/\/doi.org\/10.1109\/dac63849.2025.11133274","relation":{},"subject":[],"published":{"date-parts":[[2025,6,22]]}}}