VLSI System Computing Lab (VSCLAB)
Advancing reliable, intelligent chip design through AI and physics-based computation
Directed by Prof. Sheldon Tan, VSCLAB develops AI- and LLM-powered methods for VLSI reliability, thermal modeling, and accelerator design—bridging electronic design automation with high-performance computing.
VSCLAB introduction
The VLSI System and Computation Lab (VSCLAB) is directed by Prof. Sheldon Tan, Professor in the Department of Electrical and Computer Engineering at the Bourns College of Engineering, and a cooperating faculty member of the Computer Science and Engineering department.
Dr. Tan served as Editor-in-Chief of Integration, the VLSI Journal from 2016 to 2025 — a primary journal in VLSI design and CAD/EDA (CiteScore 5.2, IF 2.5) — and currently serves as a subject editor. The lab’s work spans reliability modeling, thermal management, and parallel computation on accelerator-rich platforms.
Curriculum Vitae Short Biography Publication List Google Scholar
Prof. Sheldon Tan
Director & Principal Investigator
- Professor, Electrical & Computer Engineering — UC Riverside
- Co-faculty, Computer Science and Engineering — UC Riverside
- IEEE Fellow, 2025 (CEDA)
- Fellow, Asia-Pacific Artificial Intelligence Association (AAIA), 2025
- Past Editor-in-Chief, Integration, the VLSI Journal (2016–2025)
- 30+ years of research in VLSI design & EDA
Research Highlights
- Thermal Map Database Full-chip thermal maps for commercial CPU, GPU, and TPU multi/many-core processors.
- EMSpice 3 Full-chip, temperature-aware multiphysics framework coupling electromigration, thermomigration, and IR-drop — with realistic spatial thermal maps, Joule self-heating, iterative resistance feedback, and Monte Carlo lifetime prediction.
- GLU v3.0 GPU-accelerated sparse LU factorization solver for circuit simulation — a hybrid CPU-symbolic / GPU-numeric pipeline delivering up to 13× faster factorization than GLU 2.0, with a SciPy-style Python API (pyglu).
- Physics-based EM Models MATLAB electromigration models with documentation and reproducible examples.
- LLM for VLSI Reliability AI- and LLM-powered approaches for reliability modeling and design optimization.
- ChipletTherm Fast, FEM-grade static and transient thermal analysis for 2.5D and 3D chiplet stacks — a spectral solver delivering FEM-accurate temperature maps (0.21 K RMSE) up to ~1400× faster than 3D FEM, fast enough for real-time thermal management and agentic EDA flows.
- Chiplet Digital Twins Digital twins for advanced package and chiplet designs for thermal and reliability integrity.
Current research areas
- AI and LLM-powered approaches for VLSI reliability modeling and optimization
- Low power hardware accelerator design for machine and deep learnings
- Digital twins for advanced package and chiplet designs for thermal and reliability integrity
- Machine learning based thermal modeling, optimization and dynamic thermal management emerging package in system and chiplet design
- Parallel computing and analysis on heterogeneous and accelerator-rich (GPUs) platforms
Latest News from VSCLAB
We appreciate support from the following sponsors
The job openings in VSCLAB
My group now has 2-3 Ph.D. openings for the Fall 2025. Students with electrical engineering, computer science, physics and applied mathematics backgrounds are welcome to apply. Students with M.S. degrees are preferred. Full financial supports will be provided for qualified students.