Sophia Shao
Associate Professor
Designs domain-specific accelerators for deep learning, such as the Gemmini generator, and agile hardware methods.
Sophia Shao's lab at UC Berkeley works on accelerators, chips & architecture. 16 papers since 2023. 12 PhD students. 7 alumni, most often to NVIDIA and Microsoft. Applications close .
At a glance
- papers, 2023–2026
- 4/ year
- at T1 venues
- 0%
- PhD students
- 12
- MS students
- 0
- alumni listed
- 7
Papers by yearfrom the lab homepage, 2023–2026
- 0 papers2026 · T1 0 · collecting
- 6 papers2025 · T1 0
- 4 papers2024 · T1 0
- 6 papers2023 · T1 0
T1 venuesOther
Research focuspaper titles that use each topic's words
- Accelerators, chips & architecture4 papers
- accelerators3 papers
- domain-specific accelerators0 papers
- agile0 papers
- domain-specific0 papers
Paper list
From the lab homepage.
20256 papers
- International Symposium on Workload Characterization (IISWC), OctoberCharacterizing and Optimizing Real-Time Optimal Control for Embedded SoCs
- International Symposia on VLSI Technology and Circuits (VLSI), JuneCygnus: A 1 GHz Heterogeneous Octa-Core RISC-V Vector Processor for DSP
- International Symposia on VLSI Technology and Circuits (VLSI), JuneMAVERIC: A 16nm 72 FPS, 10 mJ/frame Heterogeneous Robotics SoC with 4 Cores and 13 INT8/FP32 Accelerators
- International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS) , AprilSuperNoVA: Algorithm-Hardware Co-Design for Resource-Aware SLAM
- International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS) , AprilVirgo: Cluster-level Matrix Unit Integration in GPUs for Scalability and Energy Efficiency
- International Symposium on Workload Characterization (IISWC), OctoberXRSight: An End-to-End Hardware–Software Co-Design Platform for XR SoC Evaluation
20244 papers
- Conference on Robot Learning (CoRL), NovemberDiffuseLoco: Real-Time Legged Locomotion Control with Diffusion from Offline Datasets
- International Symposium on Computer Architecture (ISCA), JuneFireAxe: Partitioned FPGA-Accelerated Simulation of Large-Scale RTL Designs
- Neural Information Processing Systems (NeurIPS), DecemberKVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantization
- International Symposium on Microarchitecture (MICRO), NovemberStellar: An Automated Design Framework for Dense and Sparse Spatial Accelerators
20236 papers
- International Symposium on Microarchitecture (MICRO), OctoberAuRORA: Virtualized Accelerator Orchestration for Multi-Tenant Workloads
- International Symposium on Computer Architecture (ISCA), JuneCDPU: Co-designing Compression and Decompression Processing Units for Hyperscale Systems
- International Symposium on Microarchitecture (MICRO), OctoberDOSA: Differentiable Model-Based One-Loop Search for DNN Accelerators
- IEEE International Symposium on High-Performance Computer Architecture (HPCA) , MarchMoCA: Memory-Centric, Adaptive Execution for Multi-Tenant Deep Neural Networks
- ISCA@50 Retrospective: - , JuneRETROSPECTIVE: Aladdin: A Pre-RTL, Power-Performance Accelerator Simulator Enabling Large Design Space Exploration of Customized Architectures
- International Symposium on Computer Architecture (ISCA), JuneRoSÉ: A Hardware-Software Co-Simulation Infrastructure Enabling Pre-Silicon Full-Stack Robotics SoC Evaluation
Where alumni went7 listed, 57% to industry
Industry 4Academia 3
Most often to
- NVIDIA
- Microsoft
- OpenAI
Also to
- Cornell University
- Purdue University
- University of Michigan
Apply to UC BerkeleyFall 2027
UC Berkeley EECS Computer Science PhD
- Round
- Single round
- Opens
- —
- GRE
- not required
- English
- TOEFL/IELTS (non-English-country degrees)
- Fee
- See program page
Confirmed on the program page.
Similar labsnearest by research summary
- CNPeking
Zhong Sun
0 T1 - CNTsinghua
Hongwei Chen
- Accelerators, chips & architecture
- CNZhejiang
Liqiang Lu
- Accelerators, chips & architecture
- USUSC
Christopher Torng
- Accelerators, chips & architecture
- USUW
C. J. Richard Shi
- Accelerators, chips & architecture
- CNSJTU
Zhezhi He
- 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.
UC Berkeleycloses