Schedule
All times are in Hong Kong time (HKT, UTC+8). Sessions sharing a time slot run in parallel.
Monday, November 16, 2026
| –Opening remarks |
| –Keynote: Prof. Youjip Won (KAIST) |
| Break and room reconfiguration |
|---|
Session 1: LLM Inference I – · Room 1
| ATP: Workload-Adaptive Scheduling for Goodput-Optimized Tensor-Parallel LLM Serving with PD Disaggregation Yinhe Chen (University of Science and Technology of China), Dongqi Tian (University of Science and Technology of China), Youhui Bai (University of Science and Technology of China), Chaoyi Ruan (National University of Singapore), Jiaan Zhu (University of Science and Technology of China), Long Zhao (Anhui University; Institute of Artificial Intelligence, Hefei Comprehensive National Science Center), Qinghe Wang (University of Science and Technology of China; Institute of Artificial Intelligence, Hefei Comprehensive National Science Center), Cheng Li (University of Science and Technology of China; Institute of Artificial Intelligence, Hefei Comprehensive National Science Center) |
| CacheSched: Efficient LLM Request Scheduling for Hierarchical Prefix Caching Rilin Huang (Peking University), Junyi Shu (UCLA), Yaofeng Tu (Nanjing Branch of ZTE Corporation), Xin Jin (Peking University) |
| KVLite: Efficient LLM Agent Scheduling via Head-Aware KV Sharing Jiaqi Si (National University of Defense Technology), Dezun Dong (National University of Defense Technology) |
| Requests of a Feather Must Flock Together: Batch Size vs. Prefix Homogeneity in LLM Inference Saksham Rathi (Indian Institute of Technology Bombay), Preeti (Indian Institute of Technology Bombay), Mythili Vutukuru (Indian Institute of Technology Bombay) |
| Rocks, Pebbles and Sand: Modality-aware Scheduling for Multimodal Large Language Model Inference Konstantinos Papaioannou (IMDEA Software Institute), Thaleia Dimitra Doudali (IMDEA Software Institute) |
Session 2: Network Measurement and Simulation – · Room 2
| UniSketch: A Unified Sketch for Efficient General Flow Spread Measurement He Huang (Soochow University), Yifan Han (Soochow University), Chen Lou (University of Science and Technology of China), Yang Du (Soochow University), Yu-E Sun (Soochow University), Jia Liu (Nanjing University), Zhongjun Qiu (Soochow University), Fu Xiao (Nanjing University of Posts and Telecommunications), Guoju Gao (Soochow University), Shigang Chen (University of Florida) |
| LifeSketch: Lifecycle-aware Active Flow Counting on Programmable Switches He Huang (Soochow University), Hanwen Zhang (Soochow University), Yang Du (Soochow University), Jia Liu (Nanjing University), Zhongjun Qiu (Soochow University), Yu-E Sun (Soochow University), Yao Tian (The Hong Kong University of Science and Technology), Xiaofang Zhou (The Hong Kong University of Science and Technology), Fu Xiao (Nanjing University of Posts and Telecommunications) |
| Medivh: Harnessing Compatible and Accurate Sketch-based Network Measurement with Sparsity-based Virtualization Xiang Chen (Zhejiang University), Feng Zhu (Independent Researcher), Hongmin Tan (Zhejiang University), Jiayu Li (Zhejiang University), Jiashuo Yu (Zhejiang University), Longlong Zhu (Zhejiang University), Xingyuan Li (Zhejiang University), Yuchen Xian (Zhejiang University), Jiajie Su (Zhejiang University), Xingwei Lin (Zhejiang University), Qiang Wu (Zhejiang University), Chunming Wu (Zhejiang University) |
| Rattan: An Extensible and Scalable Modular Internet Path Emulator Yixin Shen (Tsinghua University), Minhu Wang (Tsinghua University), Bo Wang (Tsinghua University), Haixuan Tong (Tsinghua University), Chenghao Li (Tsinghua University), Yutong Xie (Tsinghua University), Yixuan Gao (Tsinghua University), Mingwei Xu (Tsinghua University), Jianping Wu (Tsinghua University) |
| Clockwork: A Compiler for Composable Network Simulation Schedulers Hao Li (Xi'an Jiaotong University), Yihan Dang (Xi'an Jiaotong University), Danfeng Shan (Xi'an Jiaotong University), Peng Zhang (Xi'an Jiaotong University) |
Session 3: Software Reliability and Bugs – · Room 3
| Hardening Programs with Data Value Space Randomization of Semantically Arbitrary Values Jaeyoon Kim (Sungkyunkwan University), Hajeong Lim (Sungkyunkwan University), Kyuwon Cho (Sungkyunkwan University), Hojoon Lee (Sungkyunkwan University) |
| Dynamic Analysis of Unsafe Rust: Behaviors and Benchmarks Jiaye Fang (The George Washington University), Arsalan Bin Najeeb (The George Washington University), Dylan O'Neill (The George Washington University), Prince Noah Johnson (University of Notre Dame), Zhixiao Zhang (The George Washington University), Jie Zhou (The George Washington University) |
| RunaSan: A Rust-Native Sanitizer for Precise and Context-Sensitive Sanitization Jongyoon Kim (Sungkyunkwan University), Kyuwon Cho (Sungkyunkwan University), Hojoon Lee (Sungkyunkwan University) |
| SubSan: Efficient Memory Error Detection via Subtraction Checks Jeonghwan Kang (Pusan National University), Kyounghwan Kim (Pusan National University), Donghyun Kwon (Pusan National University) |
| Instrumentation Optimization for Practical Dynamic Race Detection Alexey Paznikov (National University of Singapore), Michael Schwarz (National University of Singapore), Andrey Kogutenko (Saint Petersburg Electrotechnical University "LETI"), Yaroslav Osipov (Saint Petersburg Electrotechnical University "LETI"), Umang Mathur (National University of Singapore) |
| –Lunch |
Session 4: Serverless and Elastic LLM Inference – · Room 1
| IgniteServe: Accelerating LLM Serving Engine Cold Starts via Fine-Grained Parallelization Wenbiao Cao (Huazhong University of Science and Technology), Hang Huang (Huazhong University of Science and Technology), Jia Rao (The University of Texas at Arlington), Zhuo Huang (Huazhong University of Science and Technology), Song Wu (Huazhong University of Science and Technology), Hai Jin (Huazhong University of Science and Technology) |
| Premature Hardware Binding is the Root of all Evil! Making Serverless Performance-Portable with Lazy Hardware Binding Serhii Ivanenko (INESC-ID, Instituto Superior Técnico, University of Lisbon), Carlos Segarra (Imperial College London), Rodrigo Bruno (INESC-ID, Instituto Superior Técnico, University of Lisbon) |
| HybridFaaS: Efficient GPU Serverless via Hybrid Autoscaling Hao Ming (Huazhong University of Science and Technology), Chencheng Ye (Huazhong University of Science and Technology), XIAOFEI LIAO (Huazhong University of Science and Technology), Hai Jin (Huazhong University of Science and Technology), Qiang-Sheng Hua (Huazhong University of Science and Technology), Yuanchao Xu (University of California, Santa Cruz) |
| Energy-Aware Scheduling for Serverless LLM Serving on Shared GPUs Tianyu Wang (University of Pittsburgh), Gourav Rattihalli (HPE Labs), Aditya Dhakal (HPE Labs), Longfei Shangguan (University of Pittsburgh & Google), Dejan Milojicic (HPE Labs) |
| gKSM: GPU Memory Deduplication at Driver Layer for Serverless LLM Inference Shuheng Gao (University of Science and Technology of China), Fan Guo (MinJiang University), Yongkun Li (University of Science and Technology of China), Ziye Yue (University of Science and Technology of China), Yinlong Xu (University of Science and Technology of China, MinJiang University), John C.S. Lui (The Chinese University of Hong Kong) |
Session 5: VM, Containers and Migration – · Room 2
| SwiftBaton: Dependency-Aware Staged Reconstruction for Live Migration of Stateful Containers Xianzhi Zhu (University of Science and Technology of China), Gonggu Chen (University of Science and Technology of China), Yancan Mao (National University of Singapore), Richard Ma (National University of Singapore), Lulu Yao (Ningbo University), Yongkun Li (University of Science and Technology of China), Yinlong Xu (University of Science and Technology of China), John C.S. Lui (The Chinese University of Hong Kong) |
| Chameleon: Adaptive Memory Overcommitment for Disaggregated Virtual Machines Yilong Luo (Peking University), Yulong Zhang (SKLP, Institute of Computing Technology, CAS; University of Chinese Academy of Sciences), Siyuan Xu (Tongji University), Siyi Zhang (SAIS, University of Chinese Academy of Sciences), Yanwen Xia (SKLP, Institute of Computing Technology, CAS; University of Chinese Academy of Sciences), Mingyu Wu (Shanghai Jiao Tong University), Diyu Zhou (Peking University), Tao Xie (Peking University), Chenxi Wang (SKLP, Institute of Computing Technology, CAS; University of Chinese Academy of Sciences) |
| FM2: Fast Live VM Migration in CXL Memory Pools Zezhou Wang (University of Washington), Jonggyu Park (University of Washington), Shihang (Vic) Li (University of Washington), Pedro Henrique Penna (Microsoft Research), Thomas E. Anderson (University of Washington), Simon Peter (University of Washington) |
| PAVE: Prebuilding Address Translations for Fast Nested MicroVM Restore Muhan Song (University of Macau), Xingguo Pang (University of Macau), Li Li (University of Macau), Xiaobo Zhou (University of Macau) |
| Joseph: Hypervisor–RDMA Co-Design for Kernel-Bypass Remote MicroVM SnapStart Xingguo Pang (Unversity of Macau), Muhan Song (University of Macau), Chuang Hu (University of Macau), Kanye Ye Wang (University of Macau), Dazhao Cheng (WuHan University), Xiaobo Zhou (University of Macau) |
Session 6: Storage and File Systems – · Room 3
| SeaStore: Redesigning Ceph’s Storage Backend for CPU-Efficient NVMe Performance Myoungwon Oh (Korea Aerospace University), Yingxin Cheng (DDN), Matan Breizman (IBM), Josh Durgin (IBM) |
| AccFS: An Accelerator-Aware File System for Non-Volatile Memory Sangjin Lee (Chung-Ang University), Taehwan Ahn (Chung-Ang University), Yongseok Son (Chung-Ang University) |
| REFlash: Reducing Total Endurance Cost in Flash-Based Tiering via Cross-Tier Write Balancing Sejun Kwon (Samsung Electronics and Seoul National University), Minwook Kim (Seoul National University), Jin-Soo Kim (Seoul National University) |
| FDP-Rocks: An Effective Use Case of FDP in RocksDB Jian Zhou (The University of Texas at Arlington), Inho Song (Virginia Tech), Chen Zhong (The University of texas at arlington), Peiran Du (The University of Texas at Arlington), Lingfeng Xiang (The University of Texas at Arlington), Sam H. Noh (Virginia Tech), Wenguang Wang (Broadcom Inc.), Jia Rao (The University of Texas at Arlington), Song Jiang (The University of Texas at Arlington) |
| SplineFS: Adaptive Learned Directory Indexing for Memory-Constrained Filesystems Jianqiang Zeng (Peking University), Shaocong Sun (Peking University), Chengying Huan (Nanjing University), Peng Wang (Huazhong University of Science and Technology), Ke Zhou (Huazhong University of Science and Technology), Yingwei Luo (Peking University), Xiaolin Wang (Peking University), Zhenlin Wang (Michigan Technological University), Diyu Zhou (Peking University), Yingdi Shan (Tsinghua University), Jie Zhang (Peking University) |
| Coffee break |
|---|
Session 7: Long-context and Multimodal LLMs – · Room 1
| SptcAttn: Accelerating Long-Context Attention with Diagonal Stripes using Sparse Tensor Core Chenpeng Wu (Shanghai Jiao Tong University), Qiqi Gu (Shanghai Jiao Tong University), Heng Shi (Shanghai Enflame Technology Co.Ltd; Shanghai Jiao Tong University), Jianguo Yao (Shanghai Jiao Tong University), Haibing Guan (Shanghai Jiao Tong University) |
| Mosaic: Efficient CPU–GPU Coordinated Block-Sparse Attention for Long-Context Decoding Feiyu Yao (Beijing University of Technology), Zhixiong Niu (Microsoft Research), Xiaqing Li (Beijing Jiaotong University), Yongqiang Xiong (Microsoft Research), Juan Fang (Beijing University of Technology), Qian Wang (Beijing University of Technology) |
| JetFire: Approximate-then-Correct Attention on Hybrid Memory for Long-Context LLM Inference Feng Yu (National University of Singapore), Hongshi Tan (National University of Singapore), Yao Chen (Huazhong University of Science and Technology), Gustavo Alonso (ETH Zürich), Weng-Fai Wong (National University of Singapore), Bingsheng He (National University of Singapore) |
| HeadFlow: Scaling Long-Context Training with Head-Group Activation Residency Yixin Song (Shanghai Jiao Tong University), Jianxiang Gao (Shanghai Jiao Tong University), Shuhuai Zhang (Zhejiang University), Zeyu Mi (Shanghai Jiao Tong University), Haibo Chen (Shanghai Jiao Tong University) |
| FlashCodec: Unleash All Decoding Resources for Low-Latency Video Decoding in MLLM Inference Lingxiao Zhao (Wuhan University), Haoran Zhou (Wuhan University), Yuezhi Che (Wuhan University), Rongyu Luo (Wuhan University), Dazhao Cheng (WuHan University) |
Session 8: Data Center Networking – · Room 2
| SAIL: Scalable In-Network-Assisted Loss Recovery for Cross-Datacenter RDMA Xiao Li (Tsinghua University), Haoyuan Yu (Tsinghua University), Yuan Yang (Tsinghua University), Mingwei Xu (Tsinghua University) |
| Detecting and Localizing Network Faults in RoCE-Based AI Clouds under Restricted Telemetry Sinuo Cao (National Key Laboratory for Multimedia Information Processing, School of Computer Science, Peking University), Weiqi Zhu (National Key Laboratory for Multimedia Information Processing, School of Computer Science, Peking University), Jie Gui (National Key Laboratory for Multimedia Information Processing, School of Computer Science, Peking University), Jintao He (National Key Laboratory for Multimedia Information Processing, School of Computer Science, Peking University), Qun Huang (National Key Laboratory for Multimedia Information Processing, School of Computer Science, Peking University), Hao Liang (Huawei Technologies Co. Ltd), Yang Wang (Huawei Technologies Co. Ltd), Bolong Cui (Huawei Technologies Co. Ltd), Tingfeng Jiang (Huawei Technologies Co. Ltd), Weiling Gu (Huawei Technologies Co. Ltd), Tao Peng (Huawei Technologies Co. Ltd), Xudong Liu (Huawei Technologies Co. Ltd), Ronghua Sun (Huawei Technologies Co. Ltd), Zhi Zhang (Huawei Technologies Co. Ltd), Xiangyang Wu (Huawei Technologies Co. Ltd), Wei Song (Huawei Technologies Co. Ltd) |
| CAVER: Enhancing RDMA Load Balancing by Hunting Less-Congested Paths Haotian Deng (Tsinghua University), Jue Zhang (Beihang University), Menghao Zhang (Beihang University), Yuan Yang (Tsinghua University), Xiao Li (Tsinghua University), Junhao Dong (Tsinghua University), Yi Liu (Tsinghua University), Mingwei Xu (Tsinghua University) |
| D-Pingmesh: Dynamic Probing and Fault Localization for Reconfigurable Datacenter Networks Yunze Wei (Tsinghua University), Kaicheng Yang (Huawei), Xiaohui Xie (Tsinghua University), Dunkai Mao (Tsinghua University), Peiying Tao (Huawei), Yikai Zhao (Huawei), Bolong Cui (Huawei), Chen Sun (Huawei), Bingyang Liu (Huawei), Kun Tan (Unaffiliated), Kai Chen (Hong Kong University of Science and Technology), Yong Cui (Tsinghua University) |
| Resi-DINC: Resilient Distributed In-Network Computing with Closed-Loop Runtime Assurance Xiaoying Huang (Pengcheng Laboratory), Kun Huang (Pengcheng Laboratory), Mengqian Pang (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences; Pengcheng Laboratory; University of Chinese Academy of Sciences) |
Session 9: Memory Tiering and Management – · Room 3
| Starfish: Fault-Tolerant Far Memory with Low Resource and Performance Overhead Yanwen Xia (SKLP, Institute of Computing Technology, CAS; University of Chinese Academy of Sciences), Benyong Deng (Huazhong University of Science and Technology), Hong Huang (SKLP, Institute of Computing Technology, CAS; University of Chinese Academy of Sciences), Yuzheng Wang (Peking University), Quanxi Li (SKLP, Institute of Computing Technology, CAS; University of Chinese Academy of Sciences), Xingda Wei (Shanghai Jiao Tong University), Xiaobing Feng (SKLP, Institute of Computing Technology, CAS; University of Chinese Academy of Sciences), Huimin Cui (SKLP, Institute of Computing Technology, CAS; University of Chinese Academy of Sciences), Chenxi Wang (SKLP, Institute of Computing Technology, CAS; University of Chinese Academy of Sciences) |
| Zswap Boost: Accelerating Transparent Memory Compression in the Cloud Zheng Liu (Zhejiang University, Alibaba Group), Teng Ma (Alibaba Group), Chengkun Wei (Zhejiang University), Ning Zhang (Alibaba Group), Xingyu Liu (Zhejiang University), Gang Deng (Alibaba Group), Shangming Cai (Alibaba Group), Wenzhi Chen (Zhejiang University) |
| A^2: Characterizing and Mitigating Imbalanced Reclaim in MGLRU via Active Aging at Production Scale Fei Zeng (Central South University), Zicheng Wang (Honor Device Co., Ltd), Ligeng Chen (Honor Device Co., Ltd), Yinggang Guo (Honor Device Co.,Ltd), Hao Wu (Nanjing University), Lu Liu (Honor Device Co.,Ltd), Tao Wang (Honor Device Co.,Ltd), Jun Xiao (Honor Device Co.,Ltd), Borui Li (Southeast University), Feng Lyu (Central South University) |
| Spectra: Bandwidth-Faithful Data Placement for Tiered Memory Jiahui Zhang (Huazhong University of Science and Technology), Qiang Cao (Huazhong University of Science and Technology), Hong Jiang (UT Arlington) |
| Building Memory Efficient Key-Value Stores in the CXL Era Theodore Michailidis (University of California, San Diego), Steven Swanson (UCSD), Jishen Zhao (UCSD) |
| –Poster setup |
| –Poster session 1 |
Tuesday, November 17, 2026
Session 10: LLM Inference II – · Room 1
| Chorus: GPU–CPU Cooperative LLM Serving for Length-Skewed Decode Dong Chen (Hong Kong University of Science and Technology), Yukun Zhou (Hong Kong University of Science and Technology), Sheng Yao (Hong Kong University of Science and Technology), Ruichuan Chen (Nokia Bell Labs), Binhang Yuan (Hong Kong University of Science and Technology), Shuai Wang (Hong Kong University of Science and Technology), Luping Wang (Alibaba Group), Cheng Huang (Alibaba Group), Guodong Yang (Alibaba Group), Liping Zhang (Alibaba Group), Wei Wang (Hong Kong University of Science and Technology) |
| InfiniLoRA: Disaggregated Multi-LoRA Serving for Large Language Models Hongyu Chen (Shanghai Jiao Tong University), Letian Ruan (Shanghai Jiao Tong University), Zilin Xu (Bytedance), Yuchen Li (Singapore Management Univerisity), Xinyu Chen (The Hong Kong University of Science and Technology (Guangzhou)), Bingsheng He (National University of Singapore), Jingwen Leng (Shanghai Jiao Tong University), Jieru Zhao (Shanghai Jiao Tong University), Shixuan Sun (Shanghai Jiao Tong University), Minyi Guo (Shanghai Jiao Tong University) |
| ELDR: Expert-Locality-Aware Decode Routing for PD-Disaggregated MoE Serving Sangjin Choi (KAIST), Sukmin Cho (KAIST), Yifan Xiong (Microsoft Research), Ziyue Yang (Shanghai Xingyunzhili Artificial Intelligence Institute), Youngjin Kwon (KAIST), Peng Cheng (Microsoft Research) |
| HACK: Homomorphic Acceleration via Compression of the Key-Value Cache for Disaggregated LLM Inference Zeyu Zhang (University of Virginia), Haiying Shen (University of Virginia), Shay Vargaftik (VMware Research by Broadcom), Ran Ben Basat (University College London and Broadcom), Michael Mitzenmacher (Harvard University), Minlan Yu (Harvard University) |
| MoE-Lego: A Plane-Grained System for Dynamic-Precision MoE Offloading Haozhe Wei (Shanghai Jiao Tong University), Xiaofeng Hou (Shanghai Jiao Tong University), Wenfeng Wang (Shanghai Jiao Tong University), Cheng Xu (Shanghai Jiao Tong University), Jiacheng Liu (Shanghai Jiao Tong University), Chao Li (Shanghai Jiao Tong University), Minyi Guo (Shanghai Jiao Tong University) |
Session 11: Search, Recommender and Graph Systems – · Room 2
| GPipeANN: Taming Throughput-Latency Tradeoff in GPU-SSD ANNS Puqing Wu (Nanjing University and Renmin University of China), Minhui Xie (Nanjing University), Yiheng Tong (Renmin University of China), Jie Yin (Tencent Inc.), Sen Yang (Tencent Inc.), Yunpeng Chai (Renmin University of China) |
| UBASE: An AI Search Engine for Trillion-Scale Vector Data Management Yao Tian (Bytedance), Yuncheng Lu (Bytedance), Liyao Xiong (Bytedance), Yuming Xu (ByteDance), Hao Zhang (Bytedance), Weichen Zhao (Bytedance), Xi Zhao (ByteDance), Bo Kuang (ByteDance), Dongyu Wang (ByteDance), Jiehui Li (ByteDance), Yakun Li (ByteDance), Lei Zhang (ByteDance) |
| DiGiT: Co-Designing GNN Algorithms and Storage I/O for Out-of-Core Training Acceleration Qian Wei (Shandong University), Yi Li (The University of Texas at Dallas), Junchi Ren (Shandong University), Zehao Chen (Shandong University), Zhaoyan Shen (Shandong University), Dongxiao Yu (Shandong University), Bingzhe Li (University of Texas at Dallas) |
| Skip it, don't speed it. Accelerating Graph Pattern Matching through Disjoint Intersection Bypassing Haotian Ma (The Chinese University of Hong Kong, Shenzhen), Wei-Chung Hsu (The Chinese University of Hong Kong, Shenzhen), Yeh-Ching Chung (The Chinese University of Hong Kong, Shenzhen) |
| EMDedup: Accelerating DLRM via In-Network EMbedding Layer Deduplication Wenjie Li (SKLP, Institute of Computing Technology, Chinese Academy of Sciences; Zhengzhou University), Ke Liu (SKLP, Institute of Computing Technology, Chinese Academy of Sciences), Zhao Zhang (SKLP, Institute of Computing Technology, Chinese Academy of Sciences), Ning Liu (Huawei), Hui Yuan (Huawei), Zixuan Guan (Huawei), Xiaolong Zheng (Huawei), Xu Zhang (SKLP, Institute of Computing Technology, Chinese Academy of Sciences), Kefan Liu (SKLP, Institute of Computing Technology, Chinese Academy of Sciences), Guanghui Zhang (Shandong University), Yungang Bao (ICT, CAS), Mingyu Chen (SKLP, Institute of Computing Technology, Chinese Academy of Sciences), Chenxi Wang (SKLP, Institute of Computing Technology, CAS; University of Chinese Academy of Sciences) |
Session 12: Security and Privacy – · Room 3
| Sharing is caring: Attestable and Trusted Cloud Pipelines out of Distrustful Components Amir Al Sadi (Imperial College London), Sina Abdollahi (Imperial College London), Adrien Ghosn (Microsoft Azure Research), Hamed Haddadi (Imperial College London), Marios Kogias (Imperial College London) |
| FLASH: Fault-Led, Attested, Secure Hybrid-Key Checkpoint for Confidential GPU Jonghyun Lee (University of Southern California), Taewoon Kang (Korea University), Pratheek B (NVIDIA), Yongqin Wang (Roblox), Rachit Rajat (NVIDIA), Mengyuan Li (USC), Gunjae Koo (Korea University), Daniel Wong (University of California, Riverside), Murali Annavaram (USC) |
| Flitter: Exploiting Voltage Glitches and Frequency Jitter to Create Cross-core Covert Channels on Modern CPUs Yucheng Ji (Tsinghua University), Chang Liu (National University of Singapore), Trevor E. Carlson (National University of Singapore, Google), Dongsheng Wang (Tsinghua University), Weidong Liu (Tsinghua University) |
| SwiftMPC: Accelerating ML under MPC with Secure-Chip-Assisted One-Way Communication Tianze Song (Zhejiang University), Chengkun Wei (Zhejiang University), Yichen Tu (Alibaba Cloud), Yangming Zhang (Alibaba Cloud), Heng Cai (Alibaba Cloud), Haiyuan Gao (Alibaba Cloud), Yuyang Cui (Tianjin University), Minghong Tang (Zhejiang University), Jin Shi (Zhejiang University), Tong Chen (Zhejiang University), Yin Zhang (Zhejiang University), Wei Wang (Alibaba Group), Wenzhi Chen (Zhejiang University) |
| Janus: Practical Load Balancing for Confidential Cloud Services Qi Guo (Max Planck Institute for Informatics), Alice Dethise (Nokia Bell Labs), Ruichuan Chen (Nokia Bell Labs), Istemi Ekin Akkus (Nokia Bell Labs), Ivica Rimac (Nokia Bell Labs), Lieven Trappeniers (Nokia Bell Labs) |
| Break |
|---|
Session 13: Mobile, Edge and Embedded systems – · Room 1
| ATLAS: Expert-Guided Sustainable Satellite Backhaul for Off-Grid Wildlife Monitoring Hao Fang (Simon Fraser University), Miao Zhang (Simon Fraser University), Haoyuan Zhao (Simon Fraser University), Long Chen (Simon Fraser University), Yi Ching Chou (Simon Fraser University), Chi Xu (Simon Fraser University), Cheng Pan (Simon Fraser University), Jiangchuan Liu (Simon Fraser University) |
| CoreCapture: Anomaly-Aware System Trace Sampling for Resource-Constrained Mobile Devices Yongqian Sun (Nankai University), Xiaotian Wang (Nankai University), Yicheng Sui (Nankai University), Jintao Feng (Nankai University), Yue Zhang (Nankai University), Shenglin Zhang (Nankai University), Bowen Zheng (Huawei), Qiang Li (Huawei), Wei Liang (Huawei), Dan Pei (Tsinghua University) |
| Bypassing Android System Services for Fast UI Responsiveness: A Kernel-Application Co-Design Approach Zizhan Chen (Chongqing University), Renhai Chen (Huawei Technologies Co., Ltd), Zili Shao (The Chinese University of Hong Kong) |
| ESROS2: Application Sandboxes for Secure ROS2 Dev Tejas Gandhi (Indian Institute of Science Bangalore), Vinod Ganapathy (Indian Institute of Science Bangalore) |
Session 14: Networking I – · Room 2
| Lean: Saving OSS Network Cost by Application-Infrastructure Coordination Qiankun Yue (Alibaba Cloud), Kun Qian (Alibaba Cloud), Jiashi Lu (Alibaba Cloud), Xue Li (Alibaba Group), Qingchao Luo (Alibaba Cloud), Weibin Yan (Alibaba Cloud), Zhihong Li (Alibaba Cloud), Jianyao Sun (Alibaba Cloud), Junping Liang (Alibaba Cloud), Yuyao Chen (Alibaba Cloud), Liefeng Rong (Alibaba Cloud), Jing Zhang (Alibaba Cloud), Ensheng Xu (Alibaba Cloud), Boyu Ying (Alibaba Cloud), Fanghu Zheng (Alibaba Cloud), Guangren Chai (Alibaba Cloud), Erci Xu (Alibaba Cloud), Yuanchao Su (Alibaba Cloud), Dennis Cai (Alibaba Cloud) |
| Courier: Host-Transparent Direct Network I/O for Secure Containers in Cloud VMs Hang Huang (Huazhong University of Science and Technology), Zhucan Ye (Huazhong University of Science and Technology), Zhiheng Tao (Ant Group), Hui Lu (The University of Texas at Arlington), Wenlong Hou (Ant Group), Wei Guo (Alibaba Group), Quan Xu (Alibaba Group), Jiang Liu (Alibaba Group), Jiangshan Lai (Ant Group), Jia Rao (The University of Texas at Arlington), Zhuo Huang (Huazhong University of Science and Technology), Song Wu (Huazhong University of Science and Technology), Chen Yu (Huazhong University of Science and Technology), Hai Jin (Huazhong University of Science and Technology) |
| Towards High-Throughput Collective Communication via Harnessing PCIe and Dedicated Interconnects Yuezheng Liu (Beihang University), Menghao Zhang (Beihang University), Xuebin Song (Beihang University), Juner Shen (Beihang University), Chunming Hu (Beihang University), Xudong Liu (Beihang University) |
| NosRacer: Dynamic Detection of Race Conditions in On-Device Network Operating Systems Runze Wu (Nanjing University), Jingbo Zhai (Nanjing University), Shanming Ping (Huawei Technologies Co., Ltd), Lingzhi Ouyang (Huawei Technologies Co., Ltd), Hua Duan (Huawei Technologies Co., Ltd), Qin Zou (Huawei Technologies Co., Ltd), Chengcheng Huang (Huawei Technologies Co., Ltd), Bingshe Liu (Huawei Technologies Co., Ltd), Xudong Lang (Huawei Technologies Co., Ltd), Xiaoxing Ma (Nanjing University), Yu Huang (Nanjing University) |
Session 15: Remote Memory and RDMA – · Room 3
| SymRDMA: Enabling In-Situ Remote Virtual Memory Sharing Huijun Shen (Hunan University), Bojie Li (Pine AI), Haixin Pang (Hunan University), Peihao Huang (Hunan university), Xingtong Lin (Hunan University), Xingyu Zhang (Hunan University), Guo Chen (Hunan University) |
| Plurio: Scaling Microkernel-based Operating System via Service Localization and Memory Pooling Zhenlin Qi (Shanghai Jiao Tong University), Shengan Zheng (Shanghai Jiao Tong University), Bowen Zhang (Shanghai Jiao Tong University), Linpeng Huang (Shanghai Jiao Tong University) |
| Everything you always wanted to know about RDMA-based distributed Key-Value Stores Paul Breuil (Inria and University of Sydney), Sumit Kumar Monga (Inria), Baptiste Lepers (Inria) |
| fRAC: Remote Accelerator Calls on Hostless FPGAs Ziyi Yang (KAUST), Krishnan B. Iyer (KAUST), Yixi Chen (KAUST), Ran Shu (Microsoft Research), Zsolt István (TU Darmstadt), Marco Canini (KAUST), Suhaib Fahmy (KAUST) |
| –Lunch |
Session 16: AI Agents – · Room 1
| Scaling LLM Agent Tool Access: The MCP Gateway Mingxin Li (Nanjing University and Alibaba Cloud), Enge Song (Alibaba Cloud), Yueshang Zuo (Alibaba Cloud), Xiaodong Liu (Alibaba Cloud), Rong Wen (Fudan University and Alibaba Cloud), Qiang Fu (RMIT University), Gianni Antichi (Politecnico di Milano), Jian He (Alibaba Cloud), Jing Tie (Alibaba Cloud), Zhou Shao (Alibaba Cloud), Xiaobo Xue (Alibaba Cloud), Xiong Xiao (Alibaba Cloud), Luyao Zhong (Alibaba Cloud), Shaokai Zhang (Alibaba Cloud), Jiangu Zhao (Alibaba Cloud), Jianyuan Lu (Alibaba Cloud), Shize Zhang (Alibaba Cloud), Xiaoqing Sun (Alibaba Cloud), Changgang Zheng (Alibaba Cloud), Zihao Fan (Alibaba Cloud), Haonan Li (Alibaba Cloud), Tian Pan (Alibaba Cloud), Xiaomin Wu (Alibaba Cloud), Yang Song (Alibaba Cloud), Xing Li (Zhejiang University and Alibaba Cloud), Biao Lyu (Alibaba Cloud), Meng Li (Nanjing University), Haipeng Dai (Nanjing University), Guihai Chen (Nanjing University), Shunmin Zhu (Alibaba Cloud) |
| Cloudless Tomorrow: A Memory-Centric System for Self-Improving Device-Cloud LLM Agents Zeyu Qiu (Huazhong University of Science and Technology), Yaohui Wang (Huazhong University of Science and Technology), Shuyi Wang (Huazhong University of Science and Technology), Yu Kuang (Huazhong University of Science and Technology), Qinbin Li (Huazhong University of Science and Technology), XIAOFEI LIAO (Huazhong University of Science and Technology), Hai Jin (Huazhong University of Science and Technology) |
| Make Graphs Great Again: Rethinking LLM Agent Architecture via Structure-Semantic Separation Jingyi Jia (Huazhong University of Science and Technology), Jiale Wang (Huazhong University of Science and Technology), Haoyi Zhang (Huazhong University of Science and Technology), Qinbin Li (Huazhong University of Science and Technology), Xiaofei Liao (Huazhong University of Science and Technology), Hai Jin (Huazhong University of Science and Technology) |
| TraJudger: Structure-Guided Failure Attribution for LLM-Powered Multi-Agent Systems Yongxin Zhao (Nankai University), Shenglin Zhang (Nankai University), Wenwei Gu (Nankai University), Zijing Que (Nankai University), Weiwen Qian (Nankai University), Yongqian Sun (Nankai University), Li Shi (Alibaba Group), Wei Li (Alibaba Group), Liping Zhang (Alibaba Group), Dan Pei (Tsinghua University) |
| AgentMesh: Toward Redundancy- and Barrier-Free Dataflow in Multi-Agent Systems Yuchen Xu (Peking University), Yang Li (Huawei), Yitao Yuan (Peking University), Nengyuan Zhang (Peking University), Yaojing Wang (Huawei), Haoyu You (Peking University), Xiaohan Hou (Peking University), Shaoke Fang (Peking University), Mingrui Liu (Peking University), Mufeng Liu (Peking University), Chen Sun (Huawei), Xia Zhu (Huawei), Bingyang Liu (Huawei), Wenfei Wu (Peking University) |
Session 17: Scheduling and Serverless – · Room 2
| Faastion: Elastic and Scalable Native Library Isolation for High-Density Serverless Platforms Miguel Lourenço (INESC-ID, Instituto Superior Técnico, University of Lisbon), Rahul Kondlapudi (Purdue University), Serhii Ivanenko (INESC-ID, Instituto Superior Técnico, University of Lisbon), Rodrigo Cidra (INESC-ID, Instituto Superior Técnico, University of Lisbon), Pedro Fonseca (Purdue University), Rodrigo Bruno (INESC-ID, Instituto Superior Técnico, University of Lisbon) |
| Unipool: Efficient Management of Pre-Provisioned Resources for Container Platforms Hanfei Yu (Stevens Institute of Technology), Enrique Saurez (Microsoft Azure Research - Systems), Íñigo Goiri (Microsoft Azure Research - Systems), Hao Wang (Stevens Institute of Technology), Rodrigo Fonseca (Microsoft Azure Research - Systems) |
| XCG: Decoupling Control from Configuration to Improve Cgroup Scalability with Flat Mapping Ruofan Xiong (Xiamen University), Li Wang (KylinSoft), Zhihao Zhang (Tsinghua University), Xinbiao Gan (NUDT), Yiming Zhang (Shanghai Jiao Tong University) |
| Efficient Strict Serializability for Stateful Serverless Workflows via Workflow Reconciliation Shenghao Ao (Department of Computer Science and Engineering, Shanghai Jiao Tong University), Zijun Li (Department of Computer Science and Engineering, Shanghai Jiao Tong University), Chuhao Xu (Department of Computer Science and Engineering, Shanghai Jiao Tong University), Quan Chen (Department of Computer Science and Engineering, Shanghai Jiao Tong University), Shuo Quan (Cloud Computing Research Institute, China Telecom), Yanan Yang (Cloud Computing Research Institute, China Telecom), Jie Wu (China Telecom Cloud Computing Research Institute; Temple University), Meiqi Zhu (China Telecom Cloud Technology Co., Ltd.), Yujun Wang (China Telecom Cloud Technology Co., Ltd.), Minyi Guo (Department of Computer Science and Engineering, Shanghai Jiao Tong University; Guizhou University) |
| Beyond Target Placement: Unified Multi-Objective Path-Aware Offline VM Scheduling at Scale Zihan Xu (Alibaba Cloud Computing), Zhidong Hu (Alibaba Cloud Computing), Zhengtong Zhang (Alibaba Cloud Computing), Fei Peng (Alibaba Cloud Computing), Kaiyuan Shen (Alibaba Cloud Computing), Feng Wang (Alibaba Cloud Computing), Xiangyun Kong (Alibaba Cloud Computing), Binda Ma (Alibaba Cloud Computing), Jian Zhou (Alibaba Cloud Computing) |
Session 18: Programming Languages and Compilers – · Room 3
| A Novel AI-Native GPU Kernel Programming Language for Autonomous Kernel Tuning Heng Shi (Shanghai Enflame Intelligence Technologies Co.,Ltd.; Shanghai Jiao Tong University), Xiaofeng Guan (Shanghai Enflame Intelligence Technologies Co.,Ltd.; Shanghai Jiao Tong University), Sijie Wu (Shanghai Enflame Intelligence Technologies Co.,Ltd.), Enming Fan (Shanghai Enflame Intelligence Technologies Co.,Ltd.), Chenpeng Wu (Shanghai Jiao Tong University), Qiqi Gu (Shanghai Jiao Tong University), Jiahang Xie (Shanghai Enflame Intelligence Technologies Co.,Ltd.), Guoqing Bao (Shanghai Enflame Intelligence Technologies Co.,Ltd.), Jianguo Yao (Shanghai Jiao Tong University) |
| EvoFuzz: IR-Aware Fuzzing for JIT Compilers through the Lens of Node Evolution Jinze Si (Shanghai Jiao Tong University), Yuchen Sun (Shanghai Jiao Tong University), Mingyu Wu (Shanghai Jiao Tong University), Haibo Chen (Shanghai Jiao Tong University) |
| Cold-Patching for Faster Edit-Compile-Run Cycles Matthew Sotoudeh (Stanford University) |
| EBFuzz: Breaking Hidden Structural Assumptions in Compilers via Extended Assembly Injection Jiangchang Wu (State Key Laboratory for Novel Software Technology, Nanjing University), Yibiao Yang (State Key Laboratory for Novel Software Technology, Nanjing University) |
| DESAN: READ/WRITE-Aware Redundant Sanitizer Check Elimination Yunlong Xing (Michigan Technological University; George Mason University), Shiyu Sun (George Mason University), Kun Sun (George Mason University) |
| Coffee break |
|---|
Session 19: Model Training – · Room 1
| PrefixFlow: Training-Time KV Caching for Schedule-Level Prefix Reuse in LLM RL Training Pengbo Li (Hong Kong University of Science and Technology), Feiyuan Zhang (Hong Kong University of Science and Technology), Guangming Sheng (Tencent), Guangxin He (Hong Kong University of Science and Technology), Di Chai (Shanghai University of Finance and Economics), Ziniu Li (Tencent), Taiqiang Wu (The University of Hong Kong), Han Tian (Hong Kong University of Science and Technology), Wenyu Mao (Tencent), Binhang Yuan (Hong Kong University of Science and Technology), Kai Chen (Hong Kong University of Science and Technology) |
| Efficient Training on Multiple Consumer GPUs with RoundPipe Yibin Luo (Tsinghua University), Shiwei Gao (Tsinghua University), Huichuan Zheng (Tsinghua University), Youyou Lu (Tsinghua University), Jiwu Shu (Tsinghua University) |
| FlexMARL: Graph-Driven Rollout-Training Co-Design for LLM-Based Multi-Agent Reinforcement Learning Zhida Jiang (JD.com), Zhaolong Xing (JD.com), Jiawei Lu (JD.com), Yipei Niu (Huawei), Qingyuan Sang (JD.com), Liangxu Zhang (JD.com), Wenquan Dai (Huawei), Junhua Shu (JD.com), Jiaxing Wang (JD.com), Qiangyu Pei (Huawei), Qiong Chen (Huawei), Xinyu Liu (Huawei), Fangming Liu (Huazhong University of Science and Technology), Ai Han (JD.com), Zhen Chen (JD.com), Ke Zhang (JD.com) |
Session 20: Quantum Computing – · Room 2
| SynDebug: A Framework to Debug Logic Bugs in Fault-Tolerant Quantum Computing Programs Nicholas S. DiBrita (Rice University), Jack Lange (Oak Ridge National Lab), Tirthak Patel (Rice University) |
| HybridQC: Hardware-Grounded Simulation of Tightly Integrated Hybrid Quantum-Classical Systems Panayiotis Christou (Fordham University), Shuwen Kan (Fordham University), Ying Mao (Fordham University) |
| EntangledEdge: A Distributed System for Orbital Quantum Sensing and Learning Mohammad Abrarul Hasanat (University of Utah), Amit Samanta (University of Utah), Jason Ludmir (Rice University), Tirthak Patel (Rice University), Rohan Basu Roy (University of Utah) |
Session 21: New: Verification and Recovery of AI Systems – · Room 3
| DeltaBox: Millisecond-Scale Sandbox Checkpoint and Rollback for Stateful AI Agents Yunpeng Dong (Shanghai Jiao Tong University), Jingkai He (Shanghai Jiao Tong University), Shiqi Liu (Shanghai Jiao Tong University), Yuze Hou (Shanghai Jiao Tong University), Dong Du (Shanghai Jiao Tong University), Zhonghu Xu (Huawei Technologies Co., Ltd.), Si Yu (Huawei Technologies Co., Ltd.), Baochuan Yang (Huawei Technologies Co., Ltd.), Yubin Xia (Shanghai Jiao Tong University), Haibo Chen (Shanghai Jiao Tong University) |
| MoECatch: Efficient Checkpointing and Consistent Recovery for Mixture-of-Experts Models Guorui Xu (Southeast University), Xiang Liu (Southeast University), Weiwei Wu (Southeast university, P. R. China), Keni Qiu (Capital Normal University), Guang Kou (Defense Innovation Institute), Chenchen Fu (Southeast University), Jianping Wang (City University of Hong Kong) |
| BMV: Breaking Monolithic Verification for Fast Speculative LLM Inference Dongxu Liu (Huazhong University of Science and Technology), Bing Zhu (Huazhong University of Science and Technology), Xinyu Zhu (Huazhong University of Science and Technology), Long Zheng (Huazhong University of Science and Technology), Xiaofei Liao (Huazhong University of Science and Technology), Hai Jin (Huazhong University of Science and Technology), Jingling Xue (UNSW Sydney) |
| –Poster setup |
| –Poster session 2 |
Wednesday, November 18, 2026
Session 22: Large-scale AI – · Room 1
| Model-as-a-Service in the Wild: Characterizing Production LLM Inference Clusters at Scale Yiyun Zheng (The Chinese University of Hong Kong, Shenzhen), Lingyun Yang (Alibaba Group & Shanghai Jiao Tong University), Yinghao Yu (Alibaba Group), Guodong Yang (Alibaba Group), Liping Zhang (Alibaba Group), Minchen Yu (The Chinese University of Hong Kong, Shenzhen) |
| Axon: Asynchronous Global KV-Aware Scheduling Across Distributed Instances for Agentic Workloads Yuqi Qiu (University of Macau), Jinting Zou (University of Macau), Zizhao Mo (Great Bay University and University of Macau), Junlin Chen (University of Macau), Liao Chen (University of Macau, Macau SAR, China), Huanle Xu (University of Macau), Guoyao Xu (Alibaba Group), Guodong Yang (Alibaba Group), Liping Zhang (Alibaba Group), ChengZhong Xu (University of Macau) |
| CacheFS: Peer-to-Peer Cooperative Caching for LLM Weight Distribution at Cloud Scale Zheng Liu (Zhejiang University, Alibaba Group), Eryu Guan (Alibaba Group), Shang Zhao (Alibaba Group), Chengkun Wei (Zhejiang University), Teng Ma (Alibaba Group), Jin Shi (Zhejiang University), Ye Cao (Alibaba Group), Chen Cao (Alibaba Group), Wenpeng Li (Alibaba Group), Tao He (Alibaba Group), Wenyuan Yu (Alibaba Group), Tao Ma (Alibaba Group), Wenzhi Chen (Zhejiang University) |
| The GPU Changes Everything: Rethinking GPU I/O Stack at Massive Scale Guanyi Chen (The Hong Kong University of Science and Technology (Guangzhou)), Jian Zhang (The Hong Kong University of Science and Technology (Guangzhou)) |
| SOLO: Making Single-Local-Replica Storage Practical for Geo-Distributed Recommendation Model Training Tao Li (University of Science and Technology of China), Long Deng (University of Science and Technology of China), Xinyuan Li (ByteDance China), Xukun Chen (ByteDance China), Zheyu Lin (ByteDance China), Shibo Chen (ByteDance China), Shiru Ren (ByteDance China), Jingwei Xu (ByteDance China), Wei Zhang (ByteDance China), Jianhui Lu (ByteDance China), Yong Tian (ByteDance China), Bin Zhang (ByteDance China), Wei Feng (ByteDance China), Yang Guo (ByteDance China), Hang Cheng (ByteDance China), Jixi Shan (ByteDance China), Jun Song (ByteDance China), Wenzhe Zhu (University of Science and Technology of China), Yongkun Li (University of Science and Technology of China), Yinlong Xu (University of Science and Technology of China) |
Session 23: Storage and OS Policies – · Room 2
| Workload Analysis of a Production BigData-Storage Cluster for ML and Data Analytics Runhua Bian (Tsinghua University, ByteDance), Jinxin Liu (ByteDance), Hao Guo (Tsinghua University), Shaoyu Zhang (ByteDance), Liqiang Zhang (ByteDance), Zhongbao Zhang (ByteDance), Jiwu Shu (Tsinghua University), Youyou Lu (Tsinghua University) |
| Storage Evolution for Exascale Supercomputers: Lessons from Sunway TaihuLight to OceanLight Bin Yang (Tianjin University), Cheng Su (Tianjin University), Hao Wei (Tsinghua University), Wenze Du (Tianjin University), Jiwei Yan (Tianjin University), Shihan Zhang (Tianjin University), Wei Xue (Tsinghua University), Ce Yu (Tianjin University) |
| Breaking Host-Wide Blocking: Boundary-Closure SCSI Error Handling Dongdong Hao (Xiamen University), Ye Xu (LeapIO Inc.), Jiazheng Zhang (LeapIO Inc.), Xiaoyang Wang (LeapIO Inc.), Shuo Wang (LeapIO Inc.), Jiwu Shu (Xiamen University, Tsinghua University) |
| Towards Robust Cache Hits with Scalability via a Middle Layer Representing Workload Preferences Kunpeng Ding (University of Science and Technology of China), Jiangping Han (University of Science and Technology of China), Kaiping Xue (University of Science and Technology of China), Qiangqiang Wei (University of Science and Technology of China) |
| Operating System Policies in the Wild Yushan Qin (UW-Madison), Ananya Virmani (University of Wisconsin--Madison), Sungkar Bolat (student), Ayushi Chaudhary (University of Wisconsin-Madison), Nida Tanveer Ul Hasan (University of Wisconsin--Madison), Mai Zheng (Iowa State University), Sebastian Angel (University of Pennsylvania), Christopher J. Rossbach (UT Austin and Microsoft), Michael Swift (University of Wisconsin--Madison) |
| Break |
|---|
Session 24: Kernel Interfaces and Service Reliability – · Room 1
| Fuzz Testing on IPC Based OS Services with Resource State Awareness Yuheng Shen (Tsinghua University), Yifei Chu (Tsinghua University), Fan Ding (Central South University), Jianzhong Liu (Tsinghua University), Qiang Zhang (Jiangxi University of Finance and Economics), Heyuan Shi (Central South University), Yu Jiang (Tsinghua University) |
| BPFFoundary: Scalable and Practical Multi eBPF Program Composition and Testing at Hyperscale Prankur Gupta (Meta), Takshak Chahande (Meta), Miao Xu (Meta), Prashanth Kannan (Meta), Tejas Birajdar (Meta), Lilian Liu (Meta), Rajiv Krishnamurthy (Meta), Theophilus A. Benson (Carnegie Mellon University) |
| BPF-Ext: Safely Extending the eBPF Compilation Pipeline with Native Operations Yusheng Zheng (UC Santa Cruz), Zhengjie Ji (Virginia Tech), Weichen Tao (Telecom Paris), Hao Sun (ETH Zurich), Wei Zhang (University of Connecticut), Dan Williams (Virginia Tech), Andi Quinn (University of California Santa Cruz) |
| IBTpoline: Flexible, Robust, and Intel IBT-compliant System Call Hooking Alexander J. Gaidis (Brown University), Ioulios Patmanidis (TU Delft), Isabelle Shapiro (Brown University), Georgios Portokalidis (IMDEA Software Institute), Alexios Voulimeneas (TU Delft), Vasileios P. Kemerlis (Brown University) |
| Detecting API Misuses Across Context Boundaries in Linux Kernel Drivers Thomas Fourier (Inria), Julia Lawall (Inria), Baptiste Lepers (Inria) |
Session 25: Formal Methods and Verification – · Room 2
| MBCheck: Practical Static Analysis Framework for Detecting Memory Ordering Violations under WMM Jaidev Shastri (Virginia Tech), Binoy Ravindran (Virginia Tech) |
| Boomslang: A Modular Framework for Database Transaction Isolation Checking Jian Zhang (Northeastern University), Shuai Mu (Stony Brook University), Cheng Tan (Northeastern University) |
| brg-pk: A Practical Verified High-Throughput Kernel Lock for OpenBSD David Gwynne (University of Queensland), Robert Colvin (University of Queensland), Brijesh Dongol (University of Surrey), Roger Su (Australian National University), Daniel Wright (University of Surrey) |
| UniMM: Enabling Automated and Unified Memory Model Description Chen Gao (Lanzhou University), Xiangwei Meng (Lanzhou University), Fengyuan Ren (Lanzhou University / Tsinghua University) |
| Understanding and Detecting Atomicity Violation Bugs in Distributed Database Transaction Mechanisms Zhiyong Wu (School of Software, Tsinghua University), Jie Liang (Beihang University), Jingzhou Fu (Tsinghua University), Chi Zhang (Tsinghua University), Heyuan Shi (Central South University), Yu Jiang (Tsinghua University) |
| –Lunch |
Session 26: LLM Inference III – · Room 1
| TensorBridge: Bridging Microscaled FP4 Weights to FP8 Tensor Cores for LLM Serving Jiaxiang Zou (The Hong Kong University of Science and Technology (Guangzhou)), Yonghao Chen (The Hong Kong University of Science and Technology (Guangzhou)), Rui Meng (The Hong Kong University of Science and Technology (Guangzhou)), Ruilong WU (The Hong Kong University of Science and Technology (Guangzhou)), Xinyu Chen (The Hong Kong University of Science and Technology (Guangzhou)) |
| AutoTailor: Automatic and Efficient Adaptive Model Deployment for Diverse Edge Devices Mengyang Liu (Southeast University, University of Macau), Chenyu Lu (HKUST (GZ)), Haodong Tian (Tsinghua University), Fang Dong (Southeast University), Ruiting Zhou (Southeast University), Wei Wang (HKUST), Dian Shen (Southeast University), Guangtong Li (Southeast University), Ye Wan (Southeast University), Li Li (University of Macau) |
| RS-Infer: A Fast and Secure LLM Inference System for Mobile Devices with Recallable Resource Isolation Yinpeng Wu (Shanghai Jiao Tong University), Yitong Cheng (Shanghai Jiao Tong University), Lixiang Wang (Shanghai Jiao Tong University), Jinyu Gu (Shanghai Jiao Tong University), Zhichao Hua (Shanghai Jiao Tong University), Yubin Xia (Shanghai Jiao Tong University) |
| STRIDE: Efficient Secure Transformer Inference via Inter-Operator Protocol and Operator Co-Design Zhengyi Li (Shanghai Jiao Tong University), Yangjie Zhou (National University of Singapore), Zihan Liu (Shanghai Jiao Tong University), Ziyu Xiao (Shanghai Jiao Tong University), Kang Yang (State Key Laboratory of Cryptology), Jiaping Gui (Shanghai Jiao Tong University), Yu Feng (Shanghai Jiao Tong University), Yun Lin (Shanghai Jiao Tong University), Jin Song Dong (National University of Singapore), Yu Yu (Shanghai Jiao Tong University), Minyi Guo (Shanghai Jiao Tong University), Ning Liu (Shanghai Jiao Tong University), Jingwen Leng (Shanghai Jiao Tong University) |
| Schedules Are Solvable Symbols: Tuning-Free Compilation of Tile Programs on Dataflow Architectures Heru Wang (National University of Singapore), Zhenyu Bai (National University of Singapore), Wei Li (National University of Singapore), Tulika Mitra (National University of Singapore) |
Session 27: Performance Diagnosis and Operational Repair – · Room 2
| Understanding and Detecting Performance Regressions in Modern Open-Source Server Applications Jinze Si (Shanghai Jiao Tong University), Xiaoqian Ma (Shanghai JiaoTong University), Mingyu Wu (Shanghai Jiao Tong University), Haibo Chen (Shanghai Jiao Tong University) |
| Lightweight Memory Tracing via General Binary Analysis and Offline Reconstruction Ziheng Liu (University of California, San Diego), Steven Swanson (University of California, San Diego) |
| HeapLENS: Heap Layout Evaluation & Navigation Suite Trevor Brown (University of Waterloo), Sean Ovens (University of Waterloo), Gautam Pathak (University of Waterloo) |
| Dasheng: LLM-powered Reliability Engineering Assistant for Cloud Overlay Networks Yining Qi (Zhejiang University, Alibaba Cloud), Chongrong Fang (Shanghai Jiao Tong University), Xiaoqing Sun (Alibaba Cloud), Tong Wu (Zhejiang University), Jiaxiang Chen (Shanghai Jiao Tong University), Xiaobo Xue (Alibaba Cloud), Shize Zhang (Alibaba Group), Xiaonan Shi (Alibaba Cloud), Yangyang Shi (Alibaba Cloud), Shuai Yang (Alibaba Cloud), Xiong Xiao (Alibaba Cloud), Peng Cheng (Zhejiang University), Jiming Chen (Zhejiang University), Biao Lyu (Alibaba Cloud) |
| Astragalus: Automatic Configuration Repair for Production Networks Zhenrong Gu (Xi'an Jiaotong University), Peng Zhang (Xi'an Jiaotong University), Xing Feng (Xi'an Jiaotong University), Xu Liu (Xi'an Jiaotong University) |
| Coffee break |
|---|
Session 28: LLM Inference IV – · Room 1
| DittoCache: From CPU-Orchestrated to GPU-Centric KVCache for Efficient LLM Decoding Jia He (University of Science and Technology of China; Institute of Artificial Intelligence, Hefei Comprehensive National Science Center), Youhui Bai (University of Science and Technology of China), Jiawei Yi (University of Science and Technology of China), Ping Gong (University of Science and Technology of China), Zewen Jin (University of Science and Technology of China), Shannon Wang (Institute of Artificial Intelligence, Hefei Comprehensive National Science Center), Jiaqi Ruan (University of Science and Technology of China), Jiaan Zhu (University of Science and Technology of China), Pengcheng Wang (Huawei), Haibo Wang (Huawei), Weiguang Wang (Huawei), Xia Zhu (Huawei Technologies Co., Ltd), Cheng Li (University of Science and Technology of China; Institute of Artificial Intelligence, Hefei Comprehensive National Science Center) |
| Octopus: Unlocking SSD Parallelism for Sparse KVCache Retrieval Tuowei Wang (Tsinghua University), Liyun Chu (Tsinghua University), Ruwen Fan (Tsinghua University), Ju Ren (Tsinghua University) |
| DAG-KV: DAG-Guided KV Propagation for Structure-Aware Knowledge Reuse Wenbin Zhu (Shandong University), Zhaoyan Shen (Shandong University), Qian Wei (Shandong University), Shujian Zhao (Shandong University), Xin YAO (Huawei Technologies Co., Ltd), Xi Peng (Huawei Technologies Co., Ltd), Wei Zhang (Huawei Technologies Co., Ltd), Renhai Chen (Huawei Technologies Co., Ltd), Gong Zhang (Huawei Technologies Co., Ltd), Hongjun Dai (Shandong University), Dongxiao Yu (Shandong University) |
| HYPIC: Position-Independent Caching for Hybrid-Attention LLM Serving Yifei Liu (Xiaohongshu Inc.), Juntong Wu (Peking University), Yang Liu (Shanghai Jiao Tong University), Junhao Hu (Peking University), Minghao Li (Xiaohongshu Inc.), Xiaoxu Chen (Xiaohongshu Inc.), Weihang Chen (Xiaohongshu Inc.) |
| TwinCache: A Retrieval- and Reranking-Aware Caching System for Efficient RAG Serving Youxuan Xu (Beijing University of Posts and Telecommunications), Rongtian Fu (Beijing University of Posts and Telecommunications), Jinliang Shi (Beijing University of Posts and Telecommunications), Tong Wu (Beijing University of Posts and Telecommunications), Zhi Ma (Beijing University of Posts and Telecommunications), Shigang Li (Beijing University of Posts and Telecommunications) |
Session 29: Networking II – · Room 2
| LiveART: Scaling WebRTC Ingestion Jing Liu (ByteDance), Jianlin Xu (ByteDance), Bo Wang (ByteDance), Sandesh Dhawaskar Sathyanarayana (ByteDance), Wei Cao (ByteDance), Liang Yin (ByteDance), Feng Qian (ByteDance), Shu Shi (ByteDance) |
| Fine-Grained Loss Recovery via Packet-Level Content-Awareness for Ultra-low-latency Video Communications Wei Li (Hong Kong University of Science and Technology), Zhewen Yang (Northeastern University), Hua Meng (Hong Kong University of Science and Technology), Zili Meng (Hong Kong University of Science and Technology) |
| OpenNetLab: An Open Platform for AI-Based Bandwidth Estimation in Real-Time Communications Wenxue Cheng (Microsoft Research), Jeongyoon Eo (Seoul National University & Microsoft Research), Zhixiong Niu (Microsoft Research), Francis Y. Yan (University of Illinois Urbana-Champaign), Rui Gao (Microsoft Research), Ze Gan (Microsoft), Scott Inglis (Microsoft), Vishak Gopal (Microsoft), Byung-Gon Chun (Seoul National University), Ran Shu (Microsoft Research), Peng Cheng (Microsoft Research), Yongqiang Xiong (Microsoft Research) |
| Demystifying the Performance of Kernel-Bypass TCP Stacks Minhu Wang (Tsinghua University), Mingwei Xu (Quan Cheng Laboratory), Jianping Wu (Tsinghua University) |