A technical paper on using AI-assisted thermal simulation to evaluate complex RDL geometries, compute temperature distributions, and support faster thermal design iteration in advanced semiconductor packages.
Abstract
We introduce Vinci-Thermal©, an AI-assisted tool to quickly compute the temperature distribution of complex electronic components. Trained with thousands of cases, we show that it can compute temperature distributions and effective conductivities over complex redistribution layer-inspired geometries in a few seconds. Additionally, Vinci-Thermal© incorporates a hierarchy of progressively accurate models to always return an answer with the desired accuracy. When used in conjunction with a tiling strategy, Vinci-Thermal© can compute temperature distributions over large components.
Asset Details
Type: Technical paper
Title: Towards AI-Assisted Design of Thermal Management Strategies
Authors: Gamal Refai-Ahmed, Md Malekkul Islam, Martia Shahsavan, Hoa Do, Hardik Kabaria, John L. Davenport, Joseph G. Kocheemoolayil, Nicholas Harrington, Sarah Osentoski, Sheik Dawood Beer Mohideen
Organizations: AMD Inc. and Vinci4D.ai Inc.
Topic: Thermal management, RDL design, AI-assisted simulation, effective thermal conductivity, semiconductor packaging
Format: PDF
Audience: Semiconductor packaging teams, thermal engineers, simulation engineers, advanced packaging researchers, and engineering leaders
Originally prepared for the IEEE Electronics Packaging Technology Conference, this paper reflects joint work between AMD and Vinci on AI-assisted thermal analysis for complex semiconductor package geometries.
