November 30, 2009
Please note that contributed articles, blog entries, and comments posted on EDACafe.com are the views and opinion of the author and do not necessarily represent the views and opinions of the management and staff of Internet Business Systems and its subsidiary web-sites.
We learn very early in life through the story of the little boy who cried wolf that credibility is important. Personal credibility is not something you are born with. It can not be inherited or bought. It must be earned. As someone once said, credibility is like virginity, once lost it can not be regained. The same is true with the credibility of a firm. There are exceptions like the Japanese car companies whose early products in the US had a poor reputation.
EDA firms in particular need credibility. Despite being involved in leading edge products, the consumers of EDA products are generally conservative. Due to the fast pace of technology and short windows of market opportunity, EDA customers have very tight and short schedules for product development. They do not have the luxury of trying out a new way of doing something. The risk of missing the proverbial market window is large under the best of circumstances and the penalty for falling short is great for all involved. The tendency is to stick with proven methodologies and tools until the pain becomes too great, i.e. when the ability of those methodologies and tools to deliver product with required QoR within time and dollar budgets fails. To change to a new tool much less a new methodology requires a large leap of faith. There is no real option to carry out two parallel developments. Benchmarks can only go so far in reducing the uncertainty and in many instances require considerable time and resource to carry out.
I had an opportunity to discuss this topic among other things with John Sanguinetti, founder and now CTO of Forte.
I like to start out by asking for a brief background.
I am not originally an EE. In school I was in Computer Science. I was interested in operating systems and analysis of software systems. I ended up working for a couple of computer manufacturers, Digital Equipment and Amdahl and then a succession of smaller companies making hardware; ELXSI, Ardent and NeXT. Then I started Chronologic Simulation which was the company that developed VCS (Verilog Compiled Simulator). That was my foray into EDA. I had actually been doing design verification at two companies before that. I was really a practicing design verification engineer before I started Chronologic. I made that product for myself. I was the target customer. That made market research really easy. I was always convinced that this was one of the reasons the company and the product was successful. Since then, Chronologic was sold to ViewLogic and eventually was acquired by Synopsys. VCS is still very much in the market. I then started the company Forte Design System. Its predecessor was CynApps. We started that in 1998 and merged with Chronology in 2001 to form Forte. We have been prgoressing ever since, slogging it out. I was the CEO of the company when it was founded but now and for some time I am the CTO.
I recently interviewed a gentleman that had been CTO of several companies. Often the scenario is the one here. After some time the investors suggest that the founding CEO, typically a technical guy, be replaced by an experienced CEO with a more marketing and sales background.
We went through that process. When we got the second round of funding, the lead investor made it very clear that he expected that we would get "professional management" in the very near future. We did. We hired a CEO. Since then we hired another CEO. A year later he left and we promoted our VP of Engineering to CEO. That was far more successful than our previous attempts. EDA is a different industry. You can not just bring someone in from the outside and expect them to lead a company, an EDA company, particularly a small, single product EDA company. I’ve always been convinced that this industry is about credibility. If you do not have it, there is nothing you can do to influence your customers or your employees. It is very easy to lose credibility, just say something stupid. I am convinced that it is important for the CEO to be a real technical person, not necessarily a heavyweight, but somebody who really fundamentally understands the technology of the company and the products you are trying to produce and sell. You are trying to convince customers that this is something they should take a risk on. If you can not do that, if you can not speak knowledgeably, communicate your vision to people to buy this thing, you are just not going to be successful. I think we have seen that in this industry over and over. It’s important to people outside the company but it is also important to people inside the company. You bring someone in from the outside and he’s a financial guy, he says you should do this and this and this. They want it. That’s not what our vision was. We started the company to do this and this is what the technology is supposed to do. But he wants to do that. It does not match very well. He can not take the company in the direction he wants. That’s my riff on credibility. It is fundamental to this industry.
There is the example of Scully at Apple.
That’s an example but computer makers are a little different form EDA. They can get away with someone not quite as technical leading an Apple.
Or an IBM.
Or an IBM. The first startup I joined was a company called Ardent Computer. The guy who was CEO of that company was not a technical guy, he was a salesman, a management guy. But people did not expect him to know the details of the technology we were making. It was a reasonably large company, lots of moving parts. EDA companies, particularly EDA startups, typically have one or two moving parts. You have to understand those.
Will you tell me a little bit about Forte? What is Forte’s vision?
Our vision is really very consistent from when we started the company in 1998, high level design and high level synthesis to support it. We were convinced eleven years ago that hardware design would move up a level of abstraction. It was the time to do that. There were alternative ways of doing it but we decided to use C++ as the design language to make a synthesis product that could take a design in a higher level language like C++ or some variant of that which would produce RTL and would go into the standard design flow. That was the vision. It is still the vision. It has taken a long time to realize it. In the intervening years we have gone through a number of variations of trying to produce other products to support this higher level design environment. We have basically pared all of that stuff away now. It is just a synthesis product now called Cynthesizer. We put all of our effort into that. Its purpose is to take high level design input in SystemC and to produce Verilog RTL as output that will go into the standard design flow and produce good working silicon. So, that’s what we do. We have been shipping that product since 2002. It’s matured a great deal in the last seven years. That’s our vision. As we get more successful, we may add things to it, probably in the realm of IP. I do not see us adding other products to it necessarily and certainly not anything that is not very related to high level synthesis.
You have a vision that you can explain to investors and to the CEOs and CFOs of companies that you wish to sell to. So how come Forte is not one of the Big Three?
It has taken a long time for the product to mature to the point where it realizes the vision. When we started the company I made the analogy that RTL was the assembly language of hardware design. We just needed to move up the comparable level of abstraction which was then in SystemC or something similar. That was something like the FORTRAN of the 1960s. Compiler technology has advanced a great deal in the 50 years it has been around. That is what we started with, a vectorizing, parallelizing compiler that we had the rights to from Ardent Computer where I had worked in the eighties. That was the starting point for synthesis. As it turns out there is very little of the original code left. It is a very hard problem. We clearly underestimated how hard it was. If I had known how hard it was, I doubt that I would have gotten it funded. I doubt that I would have the enthusiasm to do the company in the first place. The reality is that today there are over 20 products that have chips in them that were done by Cynthesiszer. These are products that you can buy at BestBuy. There have been over 100 tapeouts. There are over 500 users. It is in regular use now, very successful. There are really dramatic success stories over the past year or so. It is not widely known. This technology is still viewed with some suspicion by most of the market. People are skeptical that it can support the claims we have made for it. We are fairly comfortable now that we have enough experience that we can back up the claims.
You can find the full EDACafe event calendar here.
To read more news, click here.
-- Jack Horgan, EDACafe.com Contributing Editor.