Yu Feng
Assistant Professor
Works on computer architecture for efficient mobile visual computing and on-device inference acceleration.
Yu Feng's lab at SJTU works on efficient training & inference and accelerators, chips & architecture. 47 papers since 2023, 21 at top venues, mostly ISCA and MICRO. 4 PhD students and 7 master's students. 1 alumnus. Applications close .
At a glance
- papers, 2023–2026
- 12/ year
- at T1 venues
- 45%
- PhD students
- 4
- MS students
- 7
- alumni listed
- 1
Papers by yearfrom the lab homepage, 2023–2026
- 22 papers2026 · T1 13 · collecting
- 17 papers2025 · T1 5
- 5 papers2024 · T1 2
- 3 papers2023 · T1 1
T1 venuesOtherVenues: ISCA, MICRO, DAC, WACV, ACL
Research focuspaper titles that use each topic's words
- Efficient training & inference13 papers
- Accelerators, chips & architecture6 papers
- visual computing2 papers
- inference acceleration0 papers
- on-device0 papers
Paper list
From the lab homepage.
202622 papers
- ICLR (T1)Astraea: A Token-wise Acceleration Framework for Video Diffusion Transformers
- ISCA (T1)CODO: An Automated Compiler for Comprehensive Dataflow Optimization
- MICRO (T1)DeGS: A Scalable 3DGS Architecture via Decoupled Workload Parsing and Reorganization
- MICRO (T1)Deltoris: Enabling Real-time VLA Inference in Embodied AI via Bit-level Sparsity and Speculative Inference
- MICRO (T1)DSTAR: Accelerating Diffusion Transformers via Spatial and Temporal Redundancy Reduction
- The International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS)EARTH: An Efficient MoE Accelerator with Entropy-Aware Speculative Prefetch and Result Reuse
- ISCA (T1)ELSA: An ELastic SNN Inference Architecture for Efficient Neuromorphic Computing
- MICRO (T1)Enhancing the Efficiency of Embodied AI System with Squeezed Analog Perception
- Proceedings of the IEEE International Symposium on High Performance Computer Architecture (HPCA)FlashFuser: Expanding the Scale of Kernel Fusion for Compute-Intensive operators via Inter-Core Connection
- MICRO (T1)GPU-Tile-Sim: A Tile-Centric GPU Simulation Framework for LLM Hardware-Software Co-Design
- MICRO (T1)Kaleido: Algorithm-Hardware Co-Design for Video Diffusion Transformers by Exploiting Latent Space Correlations
- The International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS)M2XFP: A Metadata-Augmented Microscaling Data Format for Efficient Low-bit Quantization
202517 papers
- ICML (T1)An Efficient Private GPT Never Autoregressively Decodes
- TCADAPU: Accelerate Point Cloud Neural Networks via Unified Processing-in-SRAM Architecture
- IEEE International Conference on Computer DesignAutoVCoder: A Systematic Framework for Automated Verilog Code Generation using LLMs
- NeurIPS (T1)ClusterFusion: Expanding Operator Fusion Scope for LLM Inference via Cluster-Level Collective Primitive
- Optics ExpressCo-designed pre-capture privacy optics for computer vision
- ACM Transactions on Architecture and Code OptimizationEDAS: Enabling Fast Data Loading for GPU Serverless Computing
- ACM Transactions on Architecture and Code OptimizationJUNO++: Optimizing ANNS and Enabling Efficient Sparse Attention in LLM via Ray Tracing Core
- ACM Transactions on Architecture and Code OptimizationJustitia: Fair and Efficient Scheduling for LLM Applications
- ISCA (T1)Lumina: Real-Time Neural Rendering by Exploiting Computational Redundancy
- Proceedings of the IEEE International Symposium on High Performance Computer Architecture (HPCA)MANT: Efficient Low-Bit Group Quantization for LLMs via Mathematically Adaptive Numerical Type
- Proceedings of the ACM International Conference on Architectural Support for Programming Languages and Operating SystemsMetaSapiens: Real-Time Neural Rendering with Efficiency-Aware Pruning and Accelerated Foveated Rendering
- WACVPrivate-Eye: In-Sensor Privacy Preservation Through Optical Feature Separation
20245 papers
- Proceedings of the ACM International Conference on Architectural Support for Programming Languages and Operating SystemsAmanda: Unified Instrumentation Framework for Deep Neural Networks
- ISCA (T1)BlissCam: Boosting Eye Tracking Efficiency with Learned In-Sensor Sparse Sampling
- ISCA (T1)Cicero: Real-Time Neural Rendering by Radiance Warping and Memory Optimizations
- Proceedings of the ACM International Conference on Architectural Support for Programming Languages and Operating SystemsJUNO: Optimizing High-Dimensional Approximate Nearest Neighbour Search with Sparsity-Aware Algorithm and Ray-Tracing Core Mapping
- ACM Transactions on Architecture and Code OptimizationPotamoi: Accelerating Neural Rendering via a Unified Streaming Architecture
20233 papers
- ISCA (T1)CamJ: Enabling System-level Energy Modeling and Architectural Exploration for In-sensor Visual Computing
- WACVFast and Accurate: Video Enhancement Using Sparse Depth
- IEEE/ACM International Conference on Computer Aided Design (ICCAD)Learned In-Sensor Visual Computing: From Compression to Eventification
Where alumni went1 listed
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
- USUCSD
Tajana Šimunić Rosing
- Accelerators, chips & architecture
- CNSJTU
Guohao Dai
- Efficient training & inference
- CNSJTU
Xiaofeng Hou
- Accelerators, chips & architecture
- GBCambridge
Robert Mullins
- Accelerators, chips & architecture
- CAToronto
Gennady Pekhimenko
- Accelerators, chips & architecture
- CHETH Zurich
Torsten Hoefler
- Distributed, cloud & edge systems
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