Breakfast Bytes Blogs

Cadence's Alessandra Nardi, software engineering group director, is the 2020 recipient of the Marie R. Pistilli Women in Electronic Design Award, an annual honor that recognizes individuals who have significantly helped advance women in electronic design. The award is named for the late Marie R. Pistilli. Although she had retired, she was one of the driving forces in the creation of the Design Automation Conference (DAC) and all that it has become. Her name lives on in the award and in MP Associates, the company that still runs DAC.

I called up Alessandra to find out a bit more about her background. I knew that she heads up our engineering for automotive functional safety. In fact, just a couple of weeks ago, she gave the keynote at this year's European Test Conference (ETS2020). I wrote about that in my post ETS2020: Functional Safety. I also attended a presentation that she made at, I think, OIP a couple of years ago and covered that in a post where I stole the title from her since it sums up automotive functional safety: Make Sure Your Car Doesn't Break Too Often...When It Does, Make Sure You Catch It.

Background

With the name Alessandra Nardi, I don't think you'll be surprised to learn that she is Italian. She grew up 20 kilometers from Venice in a city called Padova. It is a university city, home to the sixth oldest university in the world (Bologna is the oldest, founded in 1080). But, relevant for this award, maybe, Padova is the first university where a woman was awarded a PhD...in 1678.

She studied in Italy, and then started a PhD in 1997 in her hometown university, which was strong on microelectronics reliability. But she got an opportunity to do an internship in the CAD group at ST Microelectronics in Milan on something new: statistical worst-case modeling of VLSI circuits. So she ended up doing part of her PhD in Milan, But then in 1999, she came to UC Berkeley as an exchange program. Eventually, she went back to Italy for one week, graduated, and then returned to Berkeley.

She joined Alberto Sangiovanni-Vincentelli's (you can read his story in my post Alberto and the Origins of the EDA Industry) group and stayed as a post-doc with the full intention of becoming a professor. However, she got an offer from Magma and thought it would be good to do one or two years of industrial experience and then go back. She enjoyed it a lot, since it was a research style. She spearheaded a statistical timing engine from scratch.

After Magma was acquired, she decided to stay in industry, and came to Cadence in 2013, seven years ago. She started work on power integrity (Voltus and Voltus-Fi) and her background until then was digital signoff.

At the end of 2015, she was at a meeting with an automotive semi talking about reliability. The guy said that Cadence "has no idea what OEMs and tier-1s ask us to do". He explained what it took to satisfy the automotive segment. Alessandra talked to Anirudh Devgan (then running some of the product groups, now president of Cadence) and he said "go figure it out".

Or, as it says more formally in the DAC press release announcing her award:

Dr. Nardi created her own initiative at Cadence based upon her vision for a holistic approach to functional safety, which involved embedding functional safety capabilities directly into the semiconductor development flow. In support of the initiative, Dr. Nardi developed a patent for the automatic design and verification of safety-critical electronic systems, and she managed cross-functional collaborations on the development of a functional safety flow for digital, analog, and mixed-signal semiconductors.

She found it a different type of experience because until then she had worked mostly on one product at a time, but automotive is more of a complete solution, and so is much closer to the actual application. She ended up moving to the foundry team, since it was about the only team in Cadence in those days that was cross-functional.

When TSMC asked about automotive and wanted a presentation in 5 days, we put it together. It wasn’t perfect but it was good (153 slides). It covered reliability, quality, and a core of functional safety.  I started working on a story that connected all pieces together and how a functional safety solution should look like.

That was four years ago. I took the TÜV SÜD class and did our first presentation at DAC (just 20 slides this time!). She told me:

It is really difficult to go back to a single product after you have tasted the fun of a holistic product-centric solution connected to the final applications. Besides, revealing that I'm working on enabling autonomous driving earned me a whole new level of respect from my kids!

But functional safety is in many ways a green field still to be defined. There is a small set of people who have done it forever and for whom it is almost a religion. But now there is lots of automation that can be done and that we already have in EDA, that can be made safety aware: capture information, propagate information, define the direction. Plus automotive silicon doesn't just mean simple analog parts in decade-old process technology.

We need to leverage the work that we have already done in EDA.

I asked her about functional safety in the future. She emphasized that it is not just automotive but other stuff that is safety-critical: industrial, aerospace, medical.

Years ago, in manufacturing, people used to just throw requirements over the wall to a foundry or fab and not worry about manufacturability, but that has changed. Modern electronics cannot be implemented over the wall. What happened in manufacturing is now happening in the automotive supply chain, where everyone has to work with and talk to each other.

Accellera

I had planned to attend a presentation of the Accellera group on functional safety at DVCon, but DVCon got semi-canceled. I asked her how the working group got started:

Years ago, I had the idea of a safety language. I talked to NXP, Arm, AMD, Qualcomm, ST, TI, Mentor, and others. Everyone wanted a standard so we proposed an Accellera working group. That kicked off in December last year to gather industry requirements, and then present to the Accellera board with companies who are supportive. There was overwhelming support. Cadence is the chair, but it is just getting started. We want a data model to capture safety info, similar to the way in which UPF captures power intent. The actual technical effort just started in February.

Actually, Alessandra was being modest — she is the chair of the working group.

More information about functional safety in Accellera is on the Functional Safety Working Group page. As is obvious from the previous paragraph, the work is just beginning so if you are interested you can still participate. There is an overview set of slides of the group available.

My Life at Cadence

Alessandra was one of the women featured in the recent My Life at Cadence video series. 2' 20" long:

Press Release

You can read the DAC Press Release making the announcement.

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