datasheets.com EBN.com EDN.com EETimes.com Embedded.com PlanetAnalog.com TechOnline.com   UBM Tech
UBM Tech

Sage words of advice from Forte Design Systems founder John Sanguinetti

-April 20, 2012

John Sanguinetti, the founder and chief technology officer of Forte Design Systems, is an inventor, an entrepreneur, a cancer survivor, and an interesting guy. EDN spoke with him during the recent 2012 Design and Verification Conference in San Jose, CA. A portion of that discussion follows.

What is it about the EDA industry that you love, or do you still love it?

A: I do still have a lot of affection for it, but “love” is too strong a word. The intellectual challenge still really appeals to me. What I don’t like is the way in which the business has stagnated. It has a dysfunctional business model. But it is still a good industry for a start-up. You don’t need to take too much investment money, and, as long as it doesn’t take you too long to develop a product, you can then get bought out.

That part of the business is still fun. It has gotten harder. I have seen a product recently where it took five or six guys a couple of years to develop a product, and then Synopsys bought them, but that is not typical. All of the easy things have been done, and a product has to interface to so many things and has to fit into a complex flow. So, even if there are no unforeseen elements, it takes a long time.

If I were an entrepreneur starting up an EDA company, what advice would you give me?

A: The biggest piece of advice is: Don’t take too much money. Don’t start spending money on marketing or sales until you have something working. If you think you can raise money, hire a couple of guys, and be done in a couple of years, then you are probably being unrealistic. If you [take that approach], then you have to take more money, and the initial investors get diluted or wiped out, and that [scenario] is no fun for anybody.

With Forte, it took a long time to get the product really working, and the investors had to be patient. I have been on the boards of companies where it is clear that the investors are completely fed up with it and just want to get some kind of return. In EDA, you can get bought, but you can’t expect to sell a company. Who are you going to sell it to? If investors start to think [that they are] going to sell a company, they end up giving it away.

How much money did Chronologic [Verilog Compiler Simulator] take?

A: Zero. We didn’t take any money, but you really can’t do that anymore. We did it because there were only two of us, and my wife could support me. My partner had put away some money as a design-and-verification consultant before we got started, and that [money] tided him over. We picked up a couple of other people along the way who didn’t need to get paid immediately. It took us 15 months from start to first customer ship.

There are a lot of start-ups in the high-level synthesis space. Why do you think that it still doesn’t seem to be taking off in the way that register-transfer-level synthesis did?

A: It hasn’t taken off in the way that register-transfer-level synthesis did, but it is growing. We just reported that we grew 30% last year, and we expect that growth to continue.

Originally, adoption was centered in Japan. Is that situation broadening?

A: It is. It started in Japan for a couple of reasons. Mainly, it is because consumer devices were the sweet spot for high-level synthesis. The kinds of designs that high-level synthesis initially worked well with were image manipulation, and digital TV was one of the areas in which it [found use]. Putting image-manipulation algorithms into hardware is something that high-level synthesis has been good at, and most of these products were developed in Japan.

There were some other reasons, and [the following] is more opinion. People have been doing hardware design from C for a long time. A survey we did about 15 years ago asked if people were using C as the starting point for their design, and about 50% said yes. People in the United States managed to figure out how to go from C to Verilog, even though it was mostly a manual process.

Companies in Japan didn’t start with C models. In the late ’90s, Japan realized that it needed to [increase] abstraction and readily adopted SystemC because it didn’t have any legacy [software]. But now, it is really growing in other parts of the world, including the United States. There were a number of early adopters in Europe, but Europe’s [use of high-level synthesis] seems to have faded. In the United States, people knew that they had to do something, and nothing better than SystemC has come along.

Did you ever finish writing your memoirs?

A: No, I didn’t. I am not sure how much sanitizing I would have to do. [I’ve told] the main story, and there are all of the details, but I am not sure it would be a good idea to have those published.

—interview conducted and edited by Brian Bailey

Loading comments...

Write a Comment

To comment please Log In

FEATURED RESOURCES