About
MOSAIC: Modular Optimization of System Architectures for Intelligent Computing explores how modular, heterogeneous system architectures — enabled by advances like 2D, 2.5D, and 3D stacking can be optimized to meet the growing computational demands of intelligent systems. By decoupling architectural components and co-optimizing across compute, memory, and interconnect layers, we aim to build computing systems that are more adaptable, efficient, and scalable.
Our mission is to advance the design of intelligent computing systems by developing modular, heterogeneous architectures that leverage 2D, 2.5D, and 3D stacking and cross-layer co-optimization— systems that can be flexibly extended and scaled, reducing the cost and complexity of building high-performance, adaptive computing platforms.
Combining modular decomposition with heterogeneous, stacked integration will yield greater overall performance, efficiency, and adaptability than monolithic, planar, or homogeneous designs.
Project Mosaic