EDA Confidential: Entrepreneur, Engineer, Energy ...
A Conversation: Dr. Mar Hershenson
by Peggy Aycinena
May 21, 2010
She will be receiving the award at the Workshop for Women in Design Automation on June 14th in Anaheim, and again on June 15th during the opening session of the Design Automation Conference.
Hershenson is a remarkable person. It’s not just her intelligence, her entrepreneurial spirit, and her openness to new ideas -- it’s her energy, plain and simple.
Over the last 10 years, Hershenson finished her Ph.D. at Stanford, co-founded Barcelona Design, a company based on the technology at the core of her doctoral thesis, moved beyond Barcelona to co-found Sabio Labs, sold Sabio Labs to Magma, and for the last two years has been serving as Vice President of product development in the Custom Design Business Unit at Magma as a highly-committed advocate for analog design automation.
For those entrepreneurs among you who think their CVs would match Hershenson’s, accomplishment for accomplishment, you might be right.
But now match this: She’s had three children over the last 8 years, their ages now 8, 6, and 18 months.
Mar Hershenson and I spoke by phone during the week her Marie R. Pistilli Award was announced. Ours was a wide-ranging conversation that touched on numerous topics, but in particular the acquisition of Sabio Labs by Magma.
“Magma Design Automation decided to move from digital to mixed-signal four years ago,” Hershenson said. “Over that time, Magma has put a lot of money and people on the mixed-signal platform. They acquired some of that technology, but most of it was developed internally.
“Our team from Sabio arrived two years ago, just as Magma was putting the whole thing together. The basic platform was almost ready, but not complete. The tool we had developed at Sabio was for optimization on the electrical and physical side, and was compatible with the Cadence flow.
“When we came to Magma, however, we were asked to integrate our technology with the Magma flow. We had almost complete freedom on how to do the integration, and how to improve the platform to take advantage of our technology."
Hershenson noted that the timing of the acquisition could not have been better. The Sabio technology was still at a point where it could be modified for the integration, and the Magma platform was not yet finalized. “The integration would have been much harder if the Magma platform were fully mature when we were acquired,” she said.
“Sabio was mostly a technology company, yet we were the antithesis of a venture-backed EDA company. Ours was a very technical acquisition, with the whole team coming over to Magma. The acquisition process went extremely smoothly.”
Hershenson, like many accomplished technologists, is very aware of the business side; it’s all about selling products and succeeding in the commercial space. She observed, “The customer base has grown significantly for our tools over the last two years, especially now as people are beginning to open their wallets again. We believe we’ll see even more market demand in the next quarter.”
When asked about competitors to the Magma offerings, Hershenson replied, “Specifically for the Sabio part [of the toolset], there are no real direct competitors. Although, if I use the word 'optimization' to describe our tools, you would think NeoLinear.
“But, we’re also trying to solve the mixed-signal circuit. We’re not just selling a replacement for something that already exists. That’s why our true competition is both the inertia of the designers and the EDA vendors out there. If we were selling a new SPICE perhaps we could talk about competitors, but we are not. We’re selling a brand new tool and technology!”
Hershenson said new technology is needed, because the demographic and mind set of analog and mixed-signal designers is changing.
“When I first started selling analog tools, 10 years ago at Barcelona,” she said, “The analog designer would commonly say: ‘I’m an artist!'
“There was a very macho culture surrounding analog design. It was a ‘my OpAmp versus your OpAmp’ kind of culture. But that has changed, for two reasons.
“First, today it’s more fashionable to say, ‘My system is better than your system.’ Designers don’t want to waste their time at a basic level designing a block. They don’t want be repeatedly doing bandgap designs.
“Second, there’s a whole new generation out there who knows MATLAB and computers, and aren’t afraid to say, ‘I’ll just write a script and do things differently.’ We understand this, and are meeting the needs of the changing market.”
Hershenson understands the digital designer as well, however. She said, “Digital designers are artists too, but digital design is more structured than analog. The optimization process is completely different, and involves a well-established set of tools.”
“In analog design, however, it depends a lot more on a gut feeling. The last guy did it this way, it’s always been done this way, if we don’t do it that way, we won’t succeed. Both analog and digital designers are afraid of making mistakes, but the analog designers have traditionally had far fewer tools.”
Nonetheless, Hershenson said it’s not always easy to convince established analog designers to begin to use the tools. “It’s always a question of risk and reward,” she said. “Obviously, you can continue to do traditional design, but mistakes are costly.”
She noted that if the tools can be proven to help, analog designers understand the rewards and are open to change.
Speaking of risk, I asked Hershenson about the risk of developing EDA tools. She said, “That’s a very interesting question. You have to determine how much the research is going to cost, and how much it’s going to cost to develop the support for the tool. Then you have to determine how much you can sell it for.”
She chuckled, and added, “Of course, that’s the ideal. In reality, it’s rarely done that way. Somebody up above decides the company is going to go after a particular market. Or, more commonly, somebody below comes up with something really good.
“The EDA market is very technology driven, so ideas are [crucial]. On the digital side, it’s fairly mature at this point. If you think you can develop a better tool, you have to decide if it’s enough better to gain market share. The answer is not always yes. It has to be a at least 10x better [to succeed], because a lot of tools are bundles up these days.”
Does Hershenson think the practice of bundling tools is damaging? She said, “On the digital side, you want to go to the customers and say, 'Look, the tools are commoditized.' They think this way and understand.
"On the analog side, however, it’s just the opposite. Analog designers think differently. They don't think of themselves as consumers. But, as Penny Herscher used to say, When you sell a pain killer in EDA, the sale happens. There’s far more pain on the analog side of design today, and far less medicine available.”
So, what tools should analog designers be using today? Hershenson chuckled, “I’d like to say use just Magma tools, and you’ll be done. But, this is what really motivates me. The idea that we are producing tools that will help. There’s a lot more to be done, of course. We continue to look at what engineers need, and continue to work to meet those needs.”
Hershenson added, “There are less than 5000 analog designers in the world today, and about 2000 out of those are really good. When I teach at Stanford, I encourage young engineering students to consider their options. I say to them, ‘How many MBAs are produced each year? Do something different, and become an analog designer.'
“Everything is very global today, and there's a need. Analog companies are hugely profitable, and engineers share in those profits when a chipset is successful. It’s a very attractive option [for young engineers], and I continue to encourage them to consider it.”
She added, “The bigger question, however, is how to keep good people coming into all areas of technology, not just analog design. I with I had an answer to that question, but I do know there’s a very good feeling from knowing that what we do is at the crux of progress [for society].”
Does Hershenson hope to direct her own children into technology? She responded, “My sons are very different. The younger one is an engineer at heart. He has a bunch of dead cellphones he’s collected from friends and neighbors. He takes them apart to try to understand how they work. This kid is going to make things when he grows up. My other son is more of a reader.
“As a parent, it is not your job to impose on your children, however, but to teach them by exposing them to ideas and people. I try to remember not to act like a CAD manager in raising my children. Those guys always say no first to everything, before they say yes. I try to remember to say yes more than I say no.”
Given that positive attitude, how would Hershenson advise young high school students in the process of sorting out education and career options? She said, “There are so many exciting things happening in technology.
“You can get a great education at many schools, but engineering is a wonderful option and can lead to so many different careers. It’s rare today that you study a technology, go into a career, and just do that technology. There are so many fascinating things going on -- in clean tech, in EDA, and so many other fields. There are things out there that are hot, the things VCs are investing in, things like biotech.
“I would tell high school students to be broad minded, to keep themselves open to ideas. But most importantly, I would encourage them to learn to work in teams. That’s such an important concept, and one that should be embraced even before high school!
“But I would also remind them to master the basics, whether they’re going into economics, or chemistry, or whatever. Take those basic classes again and again until you really understand them. Master physics, and chemistry, and Maxwell’s equations, and calculus. When you’ve really come to know it and to understand it, then there are so many resources out there, you’ll be able to do anything.
“You need to build your basic toolset, to keep growing. Education absolutely doesn’t end in college. You need to be learning all your life. It’s amazing how many courses there are out there for everything, materials for leaning online, in blogs, and everywhere. I’m not ashamed to say that I’m always learning new things!”
And that, quite simply, is why Mar Hershenson is my hero!
If you attend the Workshop for Women in Design Automation, Mar will be there as well. She will receive the 2010 Marie R. Pistilli Award at the culmination of that event from DAC General Chair Sachin Saptanekar.
She will be honored again on Tuesday morning, June 15th, during the opening Plenary Session at DAC.
See you all in Anaheim!
Peggy Aycinena owns and operates EDA Confidential:
Copyright (c) 2010, Peggy Aycinena. All rights reserved.