Chao Li
Professor
Works on scalable and energy-efficient intelligent computer architecture.
Chao Li's lab at SJTU works on accelerators, chips & architecture and efficient training & inference. 76 papers since 2023. 6 PhD students and 9 master's students. 31 alumni, most often to Huawei and Alibaba. Applications close .
At a glance
- papers, 2023–2026
- 20/ year
- at T1 venues
- 0%
- PhD students
- 6
- MS students
- 9
- alumni listed
- 31
Papers by yearfrom the lab homepage, 2023–2026
- 20 papers2026 · T1 0 · collecting
- 21 papers2025 · T1 0
- 21 papers2024 · T1 0
- 14 papers2023 · T1 0
T1 venuesOther
Research focuspaper titles that use each topic's words
- Efficient training & inference26 papers
- Accelerators, chips & architecture11 papers
- energy-efficient1 paper
- scalable1 paper
- computer architecture0 papers
Paper list
From the lab homepage.
202620 papers
- Proc. the International Conference on Parallel Processing (ICPP), SeptemberAdaptive Multi-Stream Data Offloading for Efficient Management of Disaggregated Memory Pool (opens in a new tab)
- Proc. the IEEE International Symposium on High-Performance Computer Architecture (HPCA), Feb.AUM: Unleashing the Efficiency Potential of Shared Processors with Accelerator Units for LLM Serving (opens in a new tab)
- Proc. the ACM/IEEE International Symposium on Computer Architecture (ISCA), JuneCODO： An Automated Compiler for Comprehensive Dataflow Optimization (opens in a new tab)
- ACM Transactions on Architecture and Code OptimizationDGS: A GPU-based Adaptive Graph Sampling Framework
- IEEE Transactions on Cloud ComputingDPServe: Dynamic-Precision LLM Serving for Multi-Tier SLO Preservation
- IEEE Transactions on ComputersEnabling Learning-Based Efficiency Optimizer with Shadow Cycles in Resource-Constrained Autonomous Embedded Systems
- Proc. the IEEE/ACM International Symposium on Microarchitecture (MICRO), OctoberEnhancing the Efficiency of Embodied AI System with Squeezed Analog Perception (opens in a new tab)
- Proc. the IEEE International Conference on Computer Design (ICCD), NovemberForwarding Vision Encoding to Computational Sensors for Efficient Edge VLM Inference (opens in a new tab)
- Proc. the IEEE/ACM International Symposium on Quality of Service (IWQoS), JuneLocMore: Locating More Bursty Latency-critical Jobs on Resource-constrained Nodes (opens in a new tab)
- Proc. the International Symposium on Advanced Parallel Processing Technology (APPT), JulyMOCAP: Wafer-Scale-Chip-Oriented Memory-Orchestrated Chunked Pipelining Framework for Prefill-Only LLM Inference (opens in a new tab)
- Proc. the International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS), MarchMoE-APEX: An Efficient MoE Inference System with Adaptive Precision Expert Offloading (opens in a new tab)
- Proc. the ACM SIGOPS Annual Technical Conference (ATC), NovemberMoE-Lego: A Plane-Grained System for Dynamic-Precision MoE Offloading (opens in a new tab)
202521 papers
- Proc. the International Symposium on Microarchitecture (MICRO), Oct.3D-PATH: A Hierarchy LUT Processing-in-memory Accelerator with Thermal-aware Hybrid Bonding Integration (opens in a new tab)
- Frontiers of Computer ScienceA Comprehensive Survey of Micro Datacenter: Current Technologies and Future Possibilities (opens in a new tab)
- Proc. the International Symposium on Advanced Parallel Processing Technology (APPT), JulyAccelerating Large-Scale Out-of-GPU-Core GNN Training with Two-Level Historical Caching (opens in a new tab)
- Proc. the International Symposium on Advanced Parallel Processing Technology (APPT), JulyAsymServe: Demystifying and Optimizing LLM Serving Efficiency on CPU Acceleration Units (opens in a new tab)
- Proc. the IFIP International Conference on Network and Parallel Computing (NPC), Nov.CGO: Cloud Game Orchestration via Resource Preception and CODEC Optimization (opens in a new tab)
- Proc. the ACM/IEEE International Symposium on Computer Architecture (ISCA), JuneCramming a Data Center into One Cabinet, a Co-Exploration of Computing and Hardware Architecture of Waferscale-Chip (opens in a new tab)
- ACM Transactions on Architecture and Code OptimizationEnhancing High-Throughput GPU Random Walks Through Multi-Task Concurrency Orchestration
- Proc. the International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS), MarchEXIST: Enabling Extremely Efficient Intra-Service Tracing Observability in Datacenters (opens in a new tab)
- Frontiers of Computer ScienceFLAPS: Fluctuation-Aware Power Auction Strategy for Reducing the Power Overload Probability (opens in a new tab)
- Proc. the ACM SIGMOD/PODS International Conference on Management of Data (SIGMOD), JuneGES: High-Performance Graph Processing Engine and Service in Huawei (opens in a new tab)
- IEEE Transactions on ComputersImproving Efficiency in Multi-modal Autonomous Embedded Systems through Adaptive Gating (opens in a new tab)
- IEEE Transactions on Sustainable ComputingImproving Energy Efficiency of Graph Processing on Shared-Memory Systems
202421 papers
- Proc. the ACM/IEEE International Symposium on Computer Architecture (ISCA), JulyA Tale of Two Domains: Exploring Efficient Architecture Design for Truly Autonomous Things (opens in a new tab)
- ACM Transactions on Architecture and Code OptimizationA2: Towards Accelerator Level Parallelism for Autonomous Micromobility Systems (opens in a new tab)
- Proc. the International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS), AprilAmanda: Unified Instrumentation Framework for Deep Neural Networks (opens in a new tab)
- Proc. the IEEE International Conference on Computer Design (ICCD), Nov.AutoVCoder: A Systematic Framework for Automated Verilog Code Generation using LLMs (opens in a new tab)
- IEEE Transactions on Parallel and Distributed SystemsBayesian-Driven Automated Scaling in Stream Computing with Multiple QoS Targets (opens in a new tab)
- Proc. Int. Conference for High Performance Computing, Networking, Storage, and Analysis (SC), Nov.Boosting Data Center Performance via Intelligently Managed Multi-backend Disaggregated Memory (opens in a new tab)
- Proc. the International Parallel and Distributed Processing Symposium (IPDPS), MayCoCG: Fine-grained Cloud Game Co-location on Heterogeneous Platform (opens in a new tab)
- Proc. the IEEE/ACM International Symposium on Quality of Service (IWQoS), JuneCPM: A Cross-Layer Power Management Facility to Enable Highly-efficient Real-time AIoT Systems (opens in a new tab)
- IEEE Transactions on Circuits and Systems for Artificial IntelligenceEfficient Orchestrated AI Workflows Execution on Scale-out Spatial Architecture (opens in a new tab)
- Proc. the ACM/IEEE International Symposium on Computer Architecture (ISCA), JulyExploiting Similarity Opportunity of Emerging AI Models on 3D Hybrid Bonding Architecture (opens in a new tab)
- Proc. the International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS), AprilFaaSGraph: Enabling Scalable, Efficient, and Cost-Effective Graph Processing with Serverless Computing (opens in a new tab)
- Proc. the International Conference on Very Large Databases (VLDB), Aug.FlowWalker: A Memory-efficient and High-performance GPU-based Dynamic Graph Random Walk Framework (opens in a new tab)
202314 papers
- Proc. the ACM/IEEE International Symposium on Computer Architecture (ISCA), JuneArchitecting Efficient Multi-modal AIoT Systems (opens in a new tab)
- Proc. Int. Conference for High Performance Computing, Networking, Storage, and Analysis (SC), Nov.BLAD: Adaptive Load Balanced Scheduling and Operator Overlap Pipeline for Accelerating the Dynamic GNN Training (opens in a new tab)
- Proc. the ACM SIGPLAN Annual Symposium on Principles and Practices of Parallel Programming (PPoPP), Feb.CoWalker: High-Throughput GPU Random Walk with Fine-tuned Concurrent Query Processing (opens in a new tab)
- ACM Transactions on Design Automation of Electronic SystemsDRAGON: Dynamic Recurrent Accelerator for Graph Online Convolution (opens in a new tab)
- Journal of Parallel and Distributed ComputingFargraph+: Excavating the Parallelism of Graph Processing Workload on RDMA-based Far Memory System (opens in a new tab)
- Proc. the International Parallel and Distributed Processing Symposium (IPDPS), MayFIRST: Exploiting the Multi-Dimensional Attributes of Functions for Power-Aware Serverless Computing (opens in a new tab)
- Frontiers of Computer ScienceFPGA Sharing in the Cloud: A Comprehensive Analysis (opens in a new tab)
- Proc. the IEEE International Symposium on Workload Characterization (IISWC), Oct.MMBench: Benchmarking End-to-End Multi-modal DNNs and Understanding Their Hardware-Software Implications (opens in a new tab)
- Proc. the International European Conference on Parallel and Distributed Computing (Euro-Par), Sep.MMExit: Enabling Fast and Efficient Multi-modal DNN Inference with Adaptive Network Exits (opens in a new tab)
- Proc. the ACM Symposium on Cloud Computing (SoCC), Nov.Not All Resources are Visible: Exploiting Fragmented Shadow Resources in Shared-State Scheduler Architecture (opens in a new tab)
- IEEE Transactions on ComputersOptimizing GPU-based Graph Sampling and Random Walk for Efficiency and Scalability
- Proc. International Conference for High Performance Computing, Networking, Storage, and Analysis (SC), Nov.ParLeiden: Boosting Parallelism of Distributed Leiden Algorithm on Large-scale Graphs (opens in a new tab)
Where alumni went31 listed, 62% to industry
Industry 16Academia 10
Most often to
- Huawei
- Alibaba
- NVIDIA
Also to
- Shanghai Jiao Tong University
- Carnegie Mellon University
- Chengdu No.7 High School
Apply to SJTUFall 2027
Computer Science PhD (international applicants)
- Round
- Self-funded
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- USUT Austin
Neeraja Yadwadkar
- Efficient training & inference
- USUCLA
Harry Xu
- Distributed, cloud & edge systems
- USUSC
Ibrahim Sabek
- Data management & big data
- SGNTU
Chen Wang
- Distributed, cloud & edge systems
- CHETH Zurich
Ana Klimovic
- Efficient training & inference
- CNTsinghua
Yu Wang
- Accelerators, chips & architecture
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
SJTUcloses