Building Intelligent Systems from Algorithms to Silicon
At the Composable Silicon (CompSi) Lab, we believe the future of computing requires innovation across the entire technology stack. We design systems where algorithms, architectures, circuits, memory technologies, and packaging solutions are co-optimized to unlock unprecedented efficiency for artificial intelligence and emerging workloads.
Our researchers tackle some of the most pressing challenges in modern computing: reducing data movement, overcoming memory bottlenecks, enabling scalable AI infrastructure, and delivering intelligence at the edge. Through interdisciplinary collaboration, advanced prototyping, and silicon realization, we transform fundamental research into practical technological impact.
Research Highlights
AI Accelerators: Specialized architectures that dramatically improve the efficiency and performance of modern AI workloads.
Compute-in-Memory: Novel computing paradigms that move computation closer to data and reduce the energy cost of memory access.
Chiplet Systems: Composable, modular architectures that enable scalable heterogeneous integration for future computing systems.
Edge Intelligence: Ultra-efficient architectures that bring advanced AI capabilities to power-constrained devices.
Our Vision
Driving Innovation Across Academia and Industry
Our work bridges fundamental research and real-world deployment. From architecture exploration to silicon implementation, CompSi advances technologies that influence future data centers, intelligent edge devices, semiconductor platforms, and AI infrastructure. Through partnerships with government agencies, industry leaders, and academic collaborators, we pursue research that addresses both scientific challenges and societal needs.