Chairs
Jason Eshraghian, UC Sata Cruz, USA
Darshika G. Perera, University of Colorado, USA
The semiconductor industry is undergoing a significant transformation, shifting from traditional monolithic processors to multi-core and many-core architectures. These advanced designs, which can be integrated onto a single chip or within emerging chiplet-based systems, leverage cutting-edge packaging technologies and standardized interfaces. This innovation offers greater flexibility, accelerated time-to-market, and enhanced cost-efficiency, marking a pivotal change in client devices, server processors, and a broad range of emerging applications.
As generative AI models grow increasingly complex, with Large Language Model (LLM) parameters reaching billions, the demand for computational power intensifies across the software-hardware ecosystem. Chiplet-based architectures present an exciting solution to address these challenges, providing scalable, energy-efficient, and cost-effective processing capabilities. This advancement has the potential to redefine the future of cognitive systems and unlock new opportunities for AI-driven applications.
The topics of interest of the track include, but are not limited to:
- Chiplet-based architectures and designs for multicore processors.
- Multi-chiplet packages and standards.
- Advanced packaging for 2.5D/3D integration.
- Embedded Multi-die Interconnect Bridge (EMIB).
- 3D interconnects and memory design.
- Innovation in packaging and integration of third-party IP.
- Emerging monolithic 3D logic and memory devices.
- Architectural innovations for energy-efficient 3D heterogeneous integration (HI).
- Thermal cooling and management techniques.
- FPGA-based high-level synthesis design.
- EDA tools for 3D system integration.
- Advanced process technologies and foundries for chiplets.
Paper Submission
The paper submission for this session is OPEN. Please access the submission page for more information about the submission deadline and instructions.
PATRON&HOST, SPONSORS (TBC), PARTNERS & TS , SISTER CONFERENCES
