The AI Hardware Summit 2022

This year's AI Hardware Summit is coming up next week from 13th to 15th September at the Santa Clara Marriott. Cadence is a platinum sponsor of the event, and is participating in various ways.

Here's a summary of why you should attend the summit:

With over 30 ML practitioner speakers, the content is far from being all about hardware!

We treat optimizing machine learning as a full-stack effort.

We’re not a tech vendor promoting our wares at our event, so you can meet the whole industry in one place.

You won’t find this mix of people at another commercial show—ML practitioners, developers, systems builders, researchers, and investors under one roof.

You’ll take home unique perspectives and contacts, and get exposure to a different sector of the ML world.

Who knows, your input and feedback might just make AI computing better, and make your future easier...

Lip-Bu Tan

At 9.00am on Wednesday 14th, Lip-Bu Tan, Cadence's Executive Chairman, and also Chairman of Walden International, will present How AI Applications Are Changing Compute: The Most Promising Domains for AI in Industry.

Hear Lip-Bu talk about opportunities in data center cloud infrastructure with innovative startup companies in the areas of storage, security, photonics, etc.

Sriraman Chari

At 11.30am on Thursday 15th, Sriraman Chari, who is a Cadence fellow and Head of AI Accelerator IP Solution, will present Hardware & Systems: Design and Deployment of AI Engines.

The abstract says:

We have witnessed a big paradigm shift in how AI has affected our daily lives. While AI model training is typically done in a cloud infrastructure setting, model inferencing has grown enormously on power-, area-, bandwidth-, and memory-constrained edge devices.

These inferencing workloads have varying computational and memory needs, stringent power and silicon area requirements that can be very challenging to meet. AI-led innovation is affecting the next generation of embedded hardware and software design alike. This talk will illustrate the design philosophies and challenges around designing best-in-class AI hardware accelerators.

The focus of the presentation is the ease of bringing up AI solutions on the broad range of Tensilica processors.These days we have grouped the cores into three categories based on performance:

• AI Base: ISA-extended DSPs
• AI Boost: DSPs with hardware acceleration for AI
• AI Max: Turnkey inference engines

Cadence Workshop

At 1.45pm to 2.45pm on Tuesday the 13th, Cadence is presenting a workshop on Deploying On-Device AI. I attended the dry run, so here's a sneak peak at a few of the slides. To see the rest, be there!

Intel Pathfinder

It's not directly to do with the hardware summit, but it is to do with Cadence's Tensilica products. Intel made an announcement Intel Pathfinder for RISC-V Delivers New Capabilities for Pre-Silicon Deveopment. This is being delivered in two versions, the Starter Edition targeted at the hobbyist, academia, and the research community. The Professional Edition for...well...professionals. It comes with broad ecosystem support and targets organizations involved in commercial RISC-V silicon and software. Cadence's Tensilica Vision Q7 DSP is part of the Professional Edition as a vision and AI accelerator.

The full matrix for the two editions is below:

AI Hardware Summit

Full details of everything to do with the summit are on the conference website.

Sign up for Sunday Brunch, the weekly Breakfast Bytes email.

.