Automatic-Optimal-Via-Placement-In-Package-Design

Project Overview

• Duration: August 2022 - May 2023
• Institution: University of Illinois Urbana-Champaign
• Supervisor: Professor Jose E Schutt-Aine

The "Automatic Optimal Via Placement in Package Design" project embarked on a comprehensive journey to optimize Via placements within packaging designs using high-level computations and simulations. Over the course of this project, we utilized tools like Scipy and Pytorch to evaluate the efficacy of various optimizer algorithms, creating a pivotal shift in optimizing package designs.

Objective

The project aimed at analyzing and determining the optimal performance ranges of various optimizer algorithms. It took a deep dive into theoretical analysis to develop an automated router for Via placement, bringing innovation and efficiency to the forefront. The collaborative efforts with the IBM project team allowed for a detailed analysis of antenna signal data and Via configurations, crafting a landscape for efficient and effective package design.

Highlights

1. Optimizer Performance Assessment

   • Utilized Scipy and Pytorch to simulate random functions and adjust test variables, aiding in the identification of optimal operating intervals for various optimizers.

2. Automated Via Placement Router

   • Conducted an in-depth theoretical analysis of optimizers and developed code to facilitate automatic Via placement, paving the way for enhanced design efficiency.

3. Output Bound Evaluation

   • Employed output bounds to evaluate methodologies aligning with the output constraints, ensuring a consistent data distribution. This facet of the project found its application and reference in several other teams' projects.

4. Antenna Signal & Via Configuration Analysis

   • In collaboration with the IBM project team, analyzed antenna signal data and Via configurations. This phase was central to simulating efficiency, understanding fitting degrees, and assessing the gravity of output bound violations.

5. Performance Visualization

   • Crafted comprehensive visual reports that encapsulated the performance outcomes, providing a clear and concise view of the project's results.

Technologies Used

• Scipy: For high-level computations and simulations in optimizing the performance ranges.
• Pytorch: Assisted in assessing the efficacy of different optimizer algorithms.

Conclusion

This project not only deepened the understanding of package design optimization but also fostered a platform for collaboration and knowledge sharing. The findings and developments in this project are expected to influence further advancements in the field of package design optimization.

This README is subject to updates as the project progresses towards its completion in May 2023.

Thank you for visiting this repository.