{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,5]],"date-time":"2026-03-05T08:00:04Z","timestamp":1772697604853,"version":"3.50.1"},"reference-count":54,"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\/501100012166","name":"National Key R&D Program of China","doi-asserted-by":"publisher","award":["2022YFB4501400"],"award-info":[{"award-number":["2022YFB4501400"]}],"id":[{"id":"10.13039\/501100012166","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\/hpca68181.2026.11408460","type":"proceedings-article","created":{"date-parts":[[2026,3,4]],"date-time":"2026-03-04T20:47:22Z","timestamp":1772657242000},"page":"1-15","source":"Crossref","is-referenced-by-count":0,"title":["Towards Compute-Aware In-Switch Computing for LLMs Tensor-Parallelism on Multi-GPU Systems"],"prefix":"10.1109","author":[{"given":"Chen","family":"Zhang","sequence":"first","affiliation":[{"name":"Shanghai Jiao Tong University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Qijun","family":"Zhang","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhuoshan","family":"Zhou","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yijia","family":"Diao","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Haibo","family":"Wang","sequence":"additional","affiliation":[{"name":"Huawei Technologies Co. Ltd."}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhe","family":"Zhou","sequence":"additional","affiliation":[{"name":"Huawei Technologies Co. Ltd."}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhipeng","family":"Tu","sequence":"additional","affiliation":[{"name":"Huawei Technologies Co. Ltd."}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhiyao","family":"Li","sequence":"additional","affiliation":[{"name":"Huawei Technologies Co. Ltd."}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Guangyu","family":"Sun","sequence":"additional","affiliation":[{"name":"Peking University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhuoran","family":"Song","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhigang","family":"Ji","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jingwen","family":"Leng","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Minyi","family":"Guo","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA52012.2021.00034"},{"key":"ref2","article-title":"Gpt-4 technical report","author":"Achiam","year":"2023","journal-title":"arXiv preprint"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1145\/3140659.3080231"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1145\/3620666.3651379"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1145\/3458817.3476178"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2010.11929"},{"key":"ref7","article-title":"The llama 3 herd of models","author":"Dubey","year":"2024","journal-title":"arXiv e-prints"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1145\/3452296.3472904"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/MM.2017.37"},{"key":"ref10","article-title":"In-network aggregation for shared machine learning clusters","volume-title":"Proceedings of Machine Learning and Systems (MLSys)","author":"Gebara","year":"2021"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/COMHPC.2016.006"},{"key":"ref12","article-title":"The llama 3 herd of models","author":"Grattafiori","year":"2024","journal-title":"arXiv preprint"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1145\/3577193.3593739"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1145\/3650200.3656616"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/3575693.3575708"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.48550\/arxiv.1811.06965"},{"key":"ref17","article-title":"Nvswitch and dgx-2","author":"Ishii","year":"2018","journal-title":"Hot Chips"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/SC41406.2024.00032"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1145\/3503222.3507778"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/ISPASS.2013.6557149"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1145\/3132747.3132764"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO50266.2020.00086"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA45697.2020.00047"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA45697.2020.00085"},{"key":"ref25","first-page":"341","article-title":"Reducing activation recomputation in large transformer models","volume-title":"Proceedings of Machine Learning and Systems","volume":"5","author":"Korthikanti","year":"2023"},{"key":"ref26","first-page":"741","article-title":"{ATP}: In-network aggregation for multi-tenant learning","volume-title":"18th USENIX Symposium on Networked Systems Design and Implementation (NSDI 21)","author":"Lao","year":"2021"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.acl-long.134"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1145\/3458817.3476145"},{"key":"ref29","doi-asserted-by":"crossref","first-page":"279","DOI":"10.1145\/3307650.3322259","article-title":"Accelerating distributed reinforcement learning with in-switch computing","volume-title":"Proceedings of the 46th International Symposium on Computer Architecture","author":"Li","year":"2019"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/HCS66204.2025.11154421"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1145\/3037697.3037731"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1145\/3582016.3582037"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00986"},{"key":"ref34","doi-asserted-by":"crossref","first-page":"123","DOI":"10.1145\/3123939.3124534","article-title":"Beyond the socket: Numa-aware gpus","volume-title":"Proceedings of the 50th Annual IEEE\/ACM International Symposium on Microarchitecture","author":"Milic","year":"2017"},{"key":"ref35","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3458817.3476209","article-title":"Efficient large-scale language model training on gpu clusters using megatron-lm","volume-title":"Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis","author":"Narayanan","year":"2021"},{"key":"ref36","volume-title":"Nvidia h100 tensor core gpu","year":"2022"},{"key":"ref37","volume-title":"Upgrading multi-gpu interconnectivity with the third-generation nvidia nvswitch","year":"2022"},{"key":"ref38","volume-title":"Cuda templates for linear algebra subroutines","year":"2024"},{"key":"ref39","volume-title":"Introduction to nvidia dgx h100\/h200 systems","year":"2024"},{"key":"ref40","volume-title":"Nvidia collective communication library (nccl) documentation","year":"2024"},{"key":"ref41","volume-title":"Nvidia gb200 nv172","year":"2024"},{"key":"ref42","volume-title":"Cuda: New features and beyond","year":"2025"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1145\/3620665.3640410"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1109\/SC41406.2024.00094"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/SC41405.2020.00024"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA52012.2021.00049"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1145\/3394486.3406703"},{"key":"ref48","first-page":"785","article-title":"Scaling distributed machine learning with {In-Network} aggregation","volume-title":"18th USENIX Symposium on Networked Systems Design and Implementation (NSDI 21)","author":"Sapio","year":"2021"},{"key":"ref49","article-title":"Megatron-lm: Training multi-billion parameter language models using model parallelism","author":"Shoeybi","year":"2019","journal-title":"arXiv preprint"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1145\/3577193.3593711"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1145\/3567955.3567959"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1145\/3544216.3544262"},{"key":"ref54","first-page":"683","article-title":"Unlocking the power of inline {Floating-Point} operations on programmable switches","volume-title":"19th USENIX Symposium on Networked Systems Design and Implementation (NSDI 22)","author":"Yuan","year":"2022"}],"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\/11408460.pdf?arnumber=11408460","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,5]],"date-time":"2026-03-05T06:59:06Z","timestamp":1772693946000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11408460\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1,31]]},"references-count":54,"URL":"https:\/\/doi.org\/10.1109\/hpca68181.2026.11408460","relation":{},"subject":[],"published":{"date-parts":[[2026,1,31]]}}}