April 5-7, 2022|Santa Clara Convention Center| Santa Clara, CA

Welcome to the DesignCon 2022 agenda and presentation download site. Here you can view and download conference, Chiphead Theater, and other event presentations before, during, and after the event. If you're looking for a presentation from a specific session that you're unable to find here, it is likely because the presenter has not provided permission for external use or has not yet shared their presentation with us. Please check back after the event for a more complete catalog of available presentations.

Boot Camp – AI & Deep Learning for SI/PI


Chris Cheng  (Distinguished Technologist, HP Enterprise)

Paul D. Franzon, Ph.D., Fellow IEEE  (Cirrus Logic Distinguished Professor of Electrical and Computer Engineering, NC State University)

Osama Waqar Bhatti  (P.h.D Candidate, Georgia Institute of Technology)

YongJin Choi  (Master Technologist, Storage Division, Hewlett-Packard Enterprise)

Madhavan Swaminathan  (John Pippin Chair in Microsystems Packaging & Electromagnetics Director - 3D Systems Packaging Research Cente, Georgia Institute of Technology)

Location: Ballroom GH

Date: Tuesday, April 5

Time: 9:00 am - 4:30 pm

Track: 14. Machine Learning for Microelectronics, Signaling & System Design, 01. Signal & Power Integrity for Single-Multi Die, Interposer & Packaging

Format: Boot Camp

Education Level: All

Pass Type: All Access Pass

Vault Recording: TBD

Audience Level: All

This full-day boot camp session will focus on using the latest AI and Deep Learning technologies and appling them to signal integrity and power integrity problems.

The boot camp is a 5-hour session: 9:00-11:30 am and 2:00-4:30 pm. With the mid-day break to attend the keynote and enjoy a networking lunch.

The two focus areas for this boot camp will be digital twins and generative adversarial networks (GAN).

The boot camp will be organized by the following sections:

a) Introduction
We will start with a brief recap of neural networks and generative surrogate models, which are the foundations of digital twins and GAN.

b) Digital Twins (generative surrogate models)
Building on the concept of generative models (digital twins), we will introduce forward and reverse neural network models in design space exploration. Using a channel design project as an example, we will see how these digital twins can automatically set the design parameters to generate the desirable performance outcome for a given channel.

c) Deep learning GAN engine
We will use the powerful deep learning GAN engine to replace the modelling process of SerDes transceivers. The underlying GAN engine has no concept of SerDes and generates results by deep learning measured waveforms and eye diagrams. This form of deep learning can fundamentally change our SI/PI modelling methodology.

We will show step by step how to train the recurrent neural network to decode the posterior to the latent space that will encode the generator model output. We will also discuss the design trade-offs for setting the critical inputs for the discriminator. The complete results will show how a complex channel and SerDes design can be successfully modelled by the GAN engine.

d) Hands on lab
In the hands-on lab, attendees can experiment with the trained GAN engine to explore changing channel condition impacts on a trained GAN engine and compare the results with actual measurements