{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,11]],"date-time":"2026-08-11T19:28:01Z","timestamp":1786476481692,"version":"build-2736575974"},"publisher-location":"New York, NY, USA","reference-count":54,"publisher":"ACM","license":[{"start":{"date-parts":[[2026,8,11]],"date-time":"2026-08-11T00:00:00Z","timestamp":1786406400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/legalcode"}],"funder":[{"name":"Hong Kong RGC TRS","award":["T41-603\/20R"],"award-info":[{"award-number":["T41-603\/20R"]}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2026,8,17]]},"DOI":"10.1145\/3789240.3829130","type":"proceedings-article","created":{"date-parts":[[2026,8,11]],"date-time":"2026-08-11T18:33:22Z","timestamp":1786473202000},"page":"60-73","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["TurboBus: Pooling PCIe Bandwidth for LLM Workloads via Scale-Up Fabrics"],"prefix":"10.1145","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8512-1922","authenticated-orcid":false,"given":"Xinyu","family":"Yang","sequence":"first","affiliation":[{"name":"Hong Kong University of Science and Technology, Hong Kong, Hong Kong"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0501-5968","authenticated-orcid":false,"given":"Kaiqiang","family":"Xu","sequence":"additional","affiliation":[{"name":"Hong Kong University of Science and Technology, Hong Kong, Hong Kong"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2587-6028","authenticated-orcid":false,"given":"Kai","family":"Chen","sequence":"additional","affiliation":[{"name":"Hong Kong University of Science and Technology, Hong Kong, Hong Kong"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2026,8,11]]},"reference":[{"key":"e_1_3_2_1_1_1","unstructured":"2026. Azure Dedicated Host. Microsoft Azure Documentation. https:\/\/azure.microsoft.com\/en-us\/products\/virtual-machines\/dedicated-host\/ Accessed 2026-01-26."},{"key":"e_1_3_2_1_2_1","unstructured":"2026. Dedicated Instances. Amazon Web Services Documentation. https:\/\/docs.aws.amazon.com\/AWSEC2\/latest\/UserGuide\/dedicated-instance.html Accessed 2026-01-26."},{"key":"e_1_3_2_1_3_1","volume-title":"Technical Report: A Highly Capable Language Model Locally on Your Phone. arXiv preprint arXiv.2404.14219","author":"Marah Abdin","year":"2024","unstructured":"Marah Abdin et al. 2024. Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone. arXiv preprint arXiv.2404.14219 (2024). https:\/\/arxiv.org\/abs\/2404.14219"},{"key":"e_1_3_2_1_4_1","unstructured":"Advanced Micro Devices Inc. 2020. AMD Infinity FabricTM Link Installation Guide. Technical Report. AMD. https:\/\/www.amd.com\/content\/dam\/amd\/en\/documents\/instinct-tech-docs\/other\/56978.pdf"},{"key":"e_1_3_2_1_5_1","volume-title":"PipeSwitch: Fast Pipelined Context Switching for Deep Learning Applications. In 14th USENIX Symposium on Operating Systems Design and Implementation (OSDI 20)","author":"Bai Zhihao","year":"2020","unstructured":"Zhihao Bai, Zhen Zhang, Yibo Zhu, and Xin Jin. 2020. PipeSwitch: Fast Pipelined Context Switching for Deep Learning Applications. In 14th USENIX Symposium on Operating Systems Design and Implementation (OSDI 20). 499\u2013514."},{"key":"e_1_3_2_1_6_1","doi-asserted-by":"publisher","DOI":"10.1145\/850657.850659"},{"key":"e_1_3_2_1_7_1","doi-asserted-by":"publisher","DOI":"10.1145\/3731569.3764843"},{"key":"e_1_3_2_1_8_1","volume-title":"Memory Harvesting in Multi-GPU Systems with Hierarchical Unified Virtual Memory. In 2022 USENIX Annual Technical Conference (USENIX ATC 2022","author":"Choi Sangjin","year":"2022","unstructured":"Sangjin Choi, Taeksoo Kim, Jinwoo Jeong, Rachata Ausavarungnirun, Myeongjae Jeon, Youngjin Kwon, and Jeongseob Ahn. 2022. Memory Harvesting in Multi-GPU Systems with Hierarchical Unified Virtual Memory. In 2022 USENIX Annual Technical Conference (USENIX ATC 2022), Carlsbad, CA, USA, July 11\u201313, 2022. USENIX Association, 625\u2013638. https:\/\/www.usenix.org\/conference\/atc22\/presentation\/choi-sangjin"},{"key":"e_1_3_2_1_9_1","unstructured":"UALink Consortium. 2025. UALink Technical Overview. https:\/\/nand-research.com\/research-note-ualink-alliance-accelerator-interconnect-specification\/ Accessed: 2025-12-19."},{"key":"e_1_3_2_1_10_1","unstructured":"NVIDIA Corporation. 2017. NVIDIA DGX-1 With Tesla V100 System Architecture. https:\/\/images.nvidia.com\/content\/pdf\/dgx1-v100-system-architecture-whitepaper.pdf Accessed: 2025-12-19."},{"key":"e_1_3_2_1_11_1","volume-title":"NVIDIA DGX A100 System Architecture. https:\/\/images.nvidia.com\/aem-dam\/en-zz\/Solutions\/data-center\/dgx-a100\/dgxa100-system-architecture-white-paper.pdf White Paper WP-10083-001","author":"NVIDIA Corporation","year":"2026","unstructured":"NVIDIA Corporation. 2020. NVIDIA DGX A100 System Architecture. https:\/\/images.nvidia.com\/aem-dam\/en-zz\/Solutions\/data-center\/dgx-a100\/dgxa100-system-architecture-white-paper.pdf White Paper WP-10083-001; Accessed: 2026-06-04."},{"key":"e_1_3_2_1_12_1","unstructured":"NVIDIA Corporation. 2023. NVIDIA Grace Hopper Superchip. https:\/\/www.nvidia.com\/en-us\/data-center\/grace-hopper-superchip\/ Accessed: 2026-06-03."},{"key":"e_1_3_2_1_13_1","unstructured":"NVIDIA Corporation. 2024. CUDA C++ Best Practices Guide. https:\/\/docs.nvidia.com\/cuda\/cuda-c-best-practices-guide\/index.html Accessed: 2026-06-26."},{"key":"e_1_3_2_1_14_1","unstructured":"NVIDIA Corporation. 2024. Introduction to NVIDIA DGX H100\/H200 Systems \u2014 NVIDIA DGX H100\/H200 User Guide. https:\/\/docs.nvidia.com\/dgx\/dgxh100-user-guide\/introduction-to-dgxh100.html Accessed: 2025-12-19."},{"key":"e_1_3_2_1_15_1","unstructured":"NVIDIA Corporation. 2024. Introduction to the NVIDIA DGX A100 System \u2014 NVIDIA DGX A 100 User Guide. https:\/\/docs.nvidia.com\/dgx\/dgxa100-user-guide\/introduction-to-dgxa100.html Accessed: 2025-12-19."},{"key":"e_1_3_2_1_16_1","unstructured":"NVIDIA Corporation. 2024. NVIDIA GB200 NVL72. https:\/\/www.nvidia.com\/en-us\/data-center\/gb200-nvl72\/ Accessed: 2026-01-28."},{"key":"e_1_3_2_1_17_1","unstructured":"NVIDIA Corporation. 2025. Multi-Process Service (MPS). https:\/\/docs.nvidia.com\/deploy\/mps\/index.html Accessed: 2026-02-04."},{"key":"e_1_3_2_1_18_1","unstructured":"NVIDIA Corporation. 2025. NVIDIA Management Library (NVML). https:\/\/developer.nvidia.com\/management-library-nvml Accessed: 2025-01-22."},{"key":"e_1_3_2_1_19_1","unstructured":"Advanced Micro Devices. 2024. AMD Instinct MI300A APU. https:\/\/www.amd.com\/en\/products\/accelerators\/instinct\/mi300\/mi300a.html Accessed: 2026-06-15."},{"key":"e_1_3_2_1_20_1","doi-asserted-by":"publisher","DOI":"10.1145\/3575693.3575703"},{"key":"e_1_3_2_1_21_1","volume-title":"ServerlessLLM: Low-Latency Serverless Inference for Large Language Models. In 18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 24)","author":"Fu Yao","year":"2024","unstructured":"Yao Fu, Leyang Xue, Yeqi Huang, Andrei-Octavian Brabete, Dmitrii Ustiugov, Yuvraj Patel, and Luo Mai. 2024. ServerlessLLM: Low-Latency Serverless Inference for Large Language Models. In 18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 24). 135\u2013153."},{"key":"e_1_3_2_1_22_1","volume-title":"Cost-Efficient Large Language Model Serving for Multi-turn Conversations with CachedAttention. In 2024 USENIX Annual Technical Conference (USENIX ATC 24)","author":"Gao Bin","year":"2024","unstructured":"Bin Gao, Zhuomin He, Puru Sharma, Qingxuan Kang, Djordje Jevdjic, Junbo Deng, Xingkun Yang, Zhou Yu, and Pengfei Zuo. 2024. Cost-Efficient Large Language Model Serving for Multi-turn Conversations with CachedAttention. In 2024 USENIX Annual Technical Conference (USENIX ATC 24). 111\u2013126."},{"key":"e_1_3_2_1_23_1","unstructured":"Georgi Gerganov and contributors. 2023. llama.cpp. https:\/\/github.com\/ggml-org\/llama.cpp LLM inference in C\/C++."},{"key":"e_1_3_2_1_24_1","unstructured":"Aaron Grattafiori et al. 2024. The Llama 3 Herd of Models. arXiv preprint arXiv:2407.21783 (2024). https:\/\/arxiv.org\/abs\/2407.21783"},{"key":"e_1_3_2_1_25_1","doi-asserted-by":"publisher","DOI":"10.1145\/3627703.3629567"},{"key":"e_1_3_2_1_26_1","volume-title":"Analysis of Large-Scale Multi-Tenant GPU Clusters for DNN Training Workloads. In 2019 USENIX Annual Technical Conference (USENIX ATC '19)","author":"Jeon Myeongjae","year":"2019","unstructured":"Myeongjae Jeon, Shivaram Venkataraman, Amar Phanishayee, Junjie Qian, Wencong Xiao, and Fan Yang. 2019. Analysis of Large-Scale Multi-Tenant GPU Clusters for DNN Training Workloads. In 2019 USENIX Annual Technical Conference (USENIX ATC '19). USENIX Association, 947\u2013960. https:\/\/www.usenix.org\/conference\/atc19\/presentation\/jeon"},{"key":"e_1_3_2_1_27_1","doi-asserted-by":"publisher","DOI":"10.1145\/3552326.3567508"},{"key":"e_1_3_2_1_28_1","doi-asserted-by":"publisher","DOI":"10.1145\/3772356.3772415"},{"key":"e_1_3_2_1_29_1","doi-asserted-by":"publisher","DOI":"10.1145\/3600006.3613165"},{"key":"e_1_3_2_1_30_1","volume-title":"InfiniGen: Efficient Generative Inference of Large Language Models with Dynamic KV Cache Management. In 18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 24)","author":"Lee Wonbeom","year":"2024","unstructured":"Wonbeom Lee, Jungi Lee, Junghwan Seo, and Jaewoong Sim. 2024. InfiniGen: Efficient Generative Inference of Large Language Models with Dynamic KV Cache Management. In 18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 24). 155\u2013172."},{"key":"e_1_3_2_1_31_1","doi-asserted-by":"publisher","DOI":"10.1145\/3651890.3672228"},{"key":"e_1_3_2_1_32_1","volume-title":"LMCache: An Efficient KV Cache Layer for Enterprise-Scale LLM Inference. arXiv preprint arXiv:2510.09665","author":"Liu Yuhan","year":"2025","unstructured":"Yuhan Liu, Jiayi Yao, Yihua Cheng, Yuwei An, Xiaokun Chen, Shaoting Feng, Yuyang Huang, Samuel Shen, Rui Zhang, Kuntai Du, and Junchen Jiang. 2025. LMCache: An Efficient KV Cache Layer for Enterprise-Scale LLM Inference. arXiv preprint arXiv:2510.09665 (2025). https:\/\/arxiv.org\/abs\/2510.09665"},{"key":"e_1_3_2_1_33_1","unstructured":"Meta AI. 2025. Llama \u2014 Industry Leading Open-Source AI. https:\/\/www.llama.com\/."},{"key":"e_1_3_2_1_34_1","doi-asserted-by":"publisher","DOI":"10.1145\/3458817.3476205"},{"key":"e_1_3_2_1_35_1","volume-title":"ZeRO-Offload: Democratizing Billion-Scale Model Training. In 2021 USENIX Annual Technical Conference (USENIX ATC 21)","author":"Ren Jie","year":"2021","unstructured":"Jie Ren, Samyam Rajbhandari, Reza Yazdani Aminabadi, Olatunji Ruwase, Shuangyan Yang, Minjia Zhang, Dong Li, and Yuxiong He. 2021. ZeRO-Offload: Democratizing Billion-Scale Model Training. In 2021 USENIX Annual Technical Conference (USENIX ATC 21). 551\u2013564."},{"key":"e_1_3_2_1_36_1","volume-title":"19th USENIX Symposium on Operating Systems Design and Implementation (OSDI 25)","author":"Ren Zhenghang","year":"2025","unstructured":"Zhenghang Ren, Yuxuan Li, Zilong Wang, Xinyang Huang, Wenxue Li, Kaiqiang Xu, Xudong Liao, Yijun Sun, Bowen Liu, Han Tian, Junxue Zhang, Mingfei Wang, Zhizhen Zhong, Guyue Liu, Ying Zhang, and Kai Chen. 2025. Enabling Efficient GPU Communication over Multiple NICs with FuseLink. In 19th USENIX Symposium on Operating Systems Design and Implementation (OSDI 25). 91\u2013108."},{"key":"e_1_3_2_1_37_1","volume-title":"International Conference on Machine Learning (ICML","author":"Sheng Ying","year":"2023","unstructured":"Ying Sheng, Lianmin Zheng, Binhang Yuan, Zhuohan Li, Max Ryabinin, Daniel Y. Fu, Zhiqiang Xie, Beidi Chen, Clark Barrett, Joseph E. Gonzalez, Percy Liang, Christopher R\u00e9, Ion Stoica, and Ce Zhang. 2023. FlexGen: High-Throughput Generative Inference of Large Language Models with a Single GPU. In International Conference on Machine Learning (ICML 2023). PMLR, 31094\u201331116. https:\/\/proceedings.mlr.press\/v202\/sheng23a.html"},{"key":"e_1_3_2_1_38_1","unstructured":"Hugo Touvron et al. 2023. Llama 2: Open Foundation and Fine-Tuned Chat Models. arXiv preprint arXiv:2307.09288 (2023). https:\/\/arxiv.org\/abs\/2307.09288"},{"key":"e_1_3_2_1_39_1","unstructured":"Ubicloud Team. 2025. Virtualizing NVIDIA HGX B200 GPUs with Open Source. Ubicloud Blog. https:\/\/www.ubicloud.com\/blog\/virtualizing-nvidia-hgx-b200-gpus-with-open-source Accessed 2026-01-26."},{"key":"e_1_3_2_1_40_1","doi-asserted-by":"publisher","DOI":"10.1145\/3676641.3715983"},{"key":"e_1_3_2_1_41_1","unstructured":"Ben Wang and Aran Komatsuzaki. 2021. GPT-J-6B: A 6 Billion Parameter Autoregressive Language Model. https:\/\/github.com\/kingoflolz\/mesh-transformer-jax."},{"key":"e_1_3_2_1_42_1","unstructured":"Guanhua Wang Olatunji Ruwase Bing Xie and Yuxiong He. 2024. FastPersist: Accelerating Model Checkpointing in Deep Learning. arXiv:2406.13768 [cs]"},{"key":"e_1_3_2_1_43_1","volume-title":"Colocating ML Inference and Training with Fast GPU Memory Handover. In 2025 USENIX Annual Technical Conference (USENIX ATC '25)","author":"Wang Jiali","year":"2025","unstructured":"Jiali Wang, Yankui Wang, Mingcong Han, and Rong Chen. 2025. Colocating ML Inference and Training with Fast GPU Memory Handover. In 2025 USENIX Annual Technical Conference (USENIX ATC '25). USENIX Association. https:\/\/www.usenix.org\/conference\/atc25\/presentation\/wang-jiali"},{"key":"e_1_3_2_1_44_1","doi-asserted-by":"publisher","DOI":"10.1145\/3600006.3613145"},{"key":"e_1_3_2_1_45_1","doi-asserted-by":"publisher","unstructured":"Hao Wu Junxiao Deng Minchen Yu Yue Yu Yaochen Liu Hao Fan Song Wu and Wei Wang. 2024. FaaSTube: Optimizing GPU-oriented Data Transfer for Serverless Computing. arXiv:2411.01830 10.48550\/arXiv.2411.01830","DOI":"10.48550\/arXiv.2411.01830"},{"key":"e_1_3_2_1_46_1","volume-title":"2025 USENIX Annual Technical Conference (USENIX ATC 2025","author":"Wu Tianyuan","year":"2025","unstructured":"Tianyuan Wu, Wei Wang, Yinghao Yu, Siran Yang, Wenchao Wu, Qinkai Duan, Guodong Yang, Jiamang Wang, Lin Qu, and Liping Zhang. 2025. Greyhound: Hunting Fail-Slows in Hybrid-Parallel Training at Scale. In 2025 USENIX Annual Technical Conference (USENIX ATC 2025), Boston, MA, USA, July 2025. USENIX Association, 731\u2013747. https:\/\/www.usenix.org\/conference\/atc25\/presentation\/wu-tianyuan"},{"key":"e_1_3_2_1_47_1","volume-title":"18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 2024","author":"Xiang Lingfeng","year":"2024","unstructured":"Lingfeng Xiang, Zhen Lin, Weishu Deng, Hui Lu, Jia Rao, Yifan Yuan, and Ren Wang. 2024. Nomad: Non-Exclusive Memory Tiering via Transactional Page Migration. In 18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 2024), Santa Clara, CA, USA, July 10\u201312, 2024. USENIX Association, 19\u201335. https:\/\/www.usenix.org\/conference\/osdi24\/presentation\/xiang"},{"key":"e_1_3_2_1_48_1","doi-asserted-by":"publisher","DOI":"10.1145\/3731569.3764815"},{"key":"e_1_3_2_1_49_1","doi-asserted-by":"publisher","DOI":"10.1145\/3669940.3707266"},{"key":"e_1_3_2_1_50_1","volume-title":"22nd USENIX Symposium on Networked Systems Design and Implementation (NSDI 2025","author":"Yang Lingyun","year":"2025","unstructured":"Lingyun Yang, Yongchen Wang, Yinghao Yu, Qizhen Weng, Jianbo Dong, Kan Liu, Chi Zhang, Yanyi Zi, Hao Li, Zechao Zhang, Nan Wang, Yu Dong, Menglei Zheng, Lanlan Xi, Xiaowei Lu, Liang Ye, Guodong Yang, Binzhang Fu, Tao Lan, Liping Zhang, Lin Qu, and Wei Wang. 2025. GPU-Disaggregated Serving for Deep Learning Recommendation Models at Scale. In 22nd USENIX Symposium on Networked Systems Design and Implementation (NSDI 2025), Philadelphia, PA, USA, April 28\u201330, 2025. USENIX Association, 847\u2013863. https:\/\/www.usenix.org\/conference\/nsdi25\/presentation\/yang"},{"key":"e_1_3_2_1_51_1","volume-title":"Resource-Efficient Inference. In 2025 USENIX Annual Technical Conference (USENIX ATC 25)","author":"Yu Minchen","year":"2025","unstructured":"Minchen Yu, Ao Wang, Dong Chen, Haoxuan Yu, Xiaonan Luo, Zhuohao Li, Wei Wang, Ruichuan Chen, Dapeng Nie, Haoran Yang, and Yu Ding. 2025. Torpor: GPU-Enabled Serverless Computing for Low-Latency, Resource-Efficient Inference. In 2025 USENIX Annual Technical Conference (USENIX ATC 25). 597\u2013612."},{"key":"e_1_3_2_1_52_1","volume-title":"OPT: Open Pre-trained Transformer Language Models. arXiv preprint arXiv:2205.01068","author":"Susan Zhang","year":"2022","unstructured":"Susan Zhang et al. 2022. OPT: Open Pre-trained Transformer Language Models. arXiv preprint arXiv:2205.01068 (2022). https:\/\/arxiv.org\/abs\/2205.01068"},{"key":"e_1_3_2_1_53_1","volume-title":"Shiyi Cao, Christos Kozyrakis, Ion Stoica, Joseph E. Gonzalez, Clark W. Barrett, and Ying Sheng.","author":"Zheng Lianmin","year":"2024","unstructured":"Lianmin Zheng, Liangsheng Yin, Zhiqiang Xie, Chuyue Sun, Jeff Huang, Cody Hao Yu, Shiyi Cao, Christos Kozyrakis, Ion Stoica, Joseph E. Gonzalez, Clark W. Barrett, and Ying Sheng. 2024. SGLang: Efficient Execution of Structured Language Model Programs. In Advances in Neural Information Processing Systems 37 (NeurIPS 2024). https:\/\/arxiv.org\/abs\/2312.07104"},{"key":"e_1_3_2_1_54_1","volume-title":"Managing Memory Tiers with CXL in Virtualized Environments. In 18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 2024","author":"Zhong Yuhong","year":"2024","unstructured":"Yuhong Zhong, Daniel S. Berger, Carl A. Waldspurger, Ryan Wee, Ishwar Agarwal, Rajat Agarwal, Frank Hady, Karthik Kumar, Mark D. Hill, Mosharaf Chowdhury, and Asaf Cidon. 2024. Managing Memory Tiers with CXL in Virtualized Environments. In 18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 2024), Santa Clara, CA, USA, July 10\u201312, 2024. USENIX Association, 37\u201356. https:\/\/www.usenix.org\/conference\/osdi24\/presentation\/zhong-yuhong"}],"event":{"name":"SIGCOMM '26: ACM SIGCOMM 2026 Conference","location":"Colorado Convention Center Denver CO USA","acronym":"SIGCOMM '26","sponsor":["SIGCOMM ACM Special Interest Group on Data Communication"]},"container-title":["Proceedings of the ACM SIGCOMM 2026 Conference"],"original-title":[],"link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3789240.3829130","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,11]],"date-time":"2026-08-11T18:42:09Z","timestamp":1786473729000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3789240.3829130"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,11]]},"references-count":54,"alternative-id":["10.1145\/3789240.3829130","10.1145\/3789240"],"URL":"https:\/\/doi.org\/10.1145\/3789240.3829130","relation":{},"subject":[],"published":{"date-parts":[[2026,8,11]]},"assertion":[{"value":"2026-08-11","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}