Semiconductor Engineering published a list of the latest technical research papers added to its research library, covering prominent developments in chip design, memory, and advanced packaging, in addition to security and thermal challenges.
The new research included advanced studies conducted by academic institutions, research organizations, and major tech companies, targeting complex engineering solutions in the semiconductor sector.
Key Research and Technical Topics
- HBM Power Consumption: Experimental analysis and power modeling in modern high-bandwidth memories using real HBM2 dies, supervised by researchers from ETH Zürich and New York University.
- GAAFET Leakage Paths: Uncovering new leakage paths in Gate-All-Around FET (GAAFET) transistors using predictive first-principles simulation in collaboration with Sandia National Laboratories and the Luxembourg Institute of Science and Technology.
- Agentic AI: Employing intelligent agents in heterogeneous system design across die and scaling domains, alongside their application in Register-Transfer Level (RTL) verification and debugging with the participation of universities such as Columbia, Illinois, and Google DeepMind.
- Nano-electromechanical eDRAM Memories: Developing embedded dynamic memory cells supported by nano-electromechanical gates to enhance energy efficiency, in a collaboration between UC Berkeley, Samsung Electronics, and Intel.
- Thermal Modeling for 3D Circuits: Introducing an AI-accelerated iterative thermal solver for three-dimensional integrated circuits (3D-ICs), jointly developed by Cadence Design Systems and the University of Notre Dame.
- Security Vulnerabilities in Autonomous Vehicles: A survey paper discussing thermal side-channel attacks and attacks on perception systems in autonomous vehicles, prepared by researchers from Missouri University of Science and Technology.
What Does This Report Represent?
These research papers reflect current industry trends toward addressing thermal bottlenecks and power consumption in dense memories, as well as attempting to automate complex design and verification stages using agentic AI algorithms to shorten development cycles and improve the efficiency of advanced chips.