Hierarchical modeling and scalable algorithms for in-situ analysis of integrated circuit packages

Abstract

The objective of this work is to investigate high-resolution and high-performance computational methods for the first-principles analysis of in-situ product-level integrated circuit (IC) packages. The novelties and key technical approaches of the proposed work include: (i) a scalable geometry-based domain decomposition (DD) method to conquer the geometric complexity of physical domain, which leads to quasi-optimal convergence that is provably scalable for multi-scale objects. Moreover, it results in parallel and scalable computational algorithms to reduce the time complexity via high performance computing facilities; (ii) a hierarchical multi-scale simulator for high-definition IC package systems, in which the technical ingredients include a skeleton-based multi-region multi-solver method and a variational macro-micro analysis for multi-scale modeling. The capability and benefits of the algorithms are explored and illustrated through several real-world 3D IC package applications.