Forte Design Systems' Forte
Do you provide the service as part of the software or do you charge separately for it?
It depends. For some of our larger customers, who have given us millions of dollars for licenses, that is something that just comes along with it. Having said that, the way the packaging is usually done, we want to make sure that our customers are successful and have done that for 4 or 5 years now. It just depends upon the deal. They can certainly buy more training and things like that along the way. The majority of anything we sell, upwards of 90% of our overall business, is software licensing.
Is there any sweet spot (SoC, microprocessors, ) for Cynthesizer?
Brett: This is a good question. I have noticed in some of the other articles you have written that we get attacked pretty heavily in this area. I will kind of give you the straight story. We tended to do best as far as applications go in the consumer electronics area. The reason is that these guys are building high volume, complex ASICs. They are also the people in the industry making the money. It is the biggest profitable subsegment of the semiconductor industry. They happen to be very data path oriented or algorithm based designs. What happens is that there are vendors out there that try to label us as only being able to handle data path and algorithms. It turns out that this is not true. We can handle complex control. If you write an if statement, you have written control. If-then-else is a control statement. A case statement is control. We handle this as well if not better than any of the other vendors out there. But where the ROI is for these customers is taking a very complex C algorithm that may have run on a processor at one point and turning it into hardware. When you start writing a design that 90% control based, where you are writing if-then-else, the code you are writing looks a lot like Verilog code. For those customers advancing immediately to a SystemC based design flow or a C based design flow may not provide the same ROI level as the customer who has a mostly algorithmic description. Having said that every design we do is a mixture of algorithm and control. But they do tend to be a little bit more algorithm based than control based.
In terms of company names who do you see as your competition?
David. In terms of competition the biggest things we have inhibiting us is the reluctance of a lot of people to move out of the RTL Verilog flow. People still revert back to try to do that when they are under pressure. When it comes to the synthesizer space, I guess the number one competitor would be Mentor Graphics.
In the software world when high level languages started to emerge, many preferred to stay with the assembly language because they felt they could get better performance. They were more comfortable. However, these days there are very few people writing in assembly language.
Brett: The founder of our company once said in the early days of our company that our job was to turn RTL Verilog into the assemble language of hardware design. I think the analogy is absolutely right.
On a go forward basis how do you envision your product evolving 2 to 3 years out?
Brett: The number one thing in this area is the Quality of Results. We have been the leader in QoR. We will make sure that we will continue to be the leader. If you look at our roadmap, the number one item along the way is generally based on QoR. Having said that, there are other areas where we have invested heavily for the last couple of years and will continue to invest. One of these areas is TLM. What's involved going forward is we have already added a number of IP libraries, IP blocks into a TLM library, for all the interfaces that are out there. This is something we have been releasing to our customers. We will keep investing in that space. The other place where we have spent a lot of time and that has paid off for us is this area of completing the ESL flow. One of the things our customers will see and people who visit us at DAC will see is that we have taken the time to do very complete and thorough integration with not only the popular RTL tools that are below us in the flow out there. Customers that are using Cynthesizer can build a flow that incorporates the best debug and analysis tool in the market, the formal verification technologies, the various simulators and things that go along with them tools that take RTL back to C. We have done integrations with all of these different technologies. It really provides our customers critical mass as far as the ESL flow. These are some of the big things. But we also have behavioral synthesis oriented technology that we have been working on. Some of that includes automated partitioning of design, incorporating more and more of the C and SystemC language constructs into the synthesizable subset.
I see that you are on the board of OSCI. How does OSCI or other standards organizations help Forte? What value do you see?
We've been invested in the SystemC community for about 5 years. You may remember way back when we used to have our own library called SimLib. We realized along the way that it made much more sense for the industry to organize together around one central standard. We dropped our language or C++ library. We went off to joined OSCI and integrated best features of our technology in with SystemC library. We have invested heavily within the company to the better part of two people full time but the aspect of the committee. We are involved heavily. Not only do we have a seat on the board, we are the president of the organization. What we have found is that by promoting OSCI, making sure to keep to an industry standard, we have been able to allow the ecosystem to grow much more quickly than if we all had our own technology based upon our own libraries, open subsets and everything else.
In your articles over the last two months two of the vendors who claim to be leaders in this space and said that they were releasing an open version of an extension to C that you can off and use. The reason they are doing this is they finally figured out that C doesn't have all the necessary extensions that something like SystemC has. The problem with it was that they have now released things that they have called open but they are non-standard, they don't fit in the flow, you can't use formal verification tools, and you can't use another synthesis tool. There are all sorts of problems with this. But by investing in OSCI, by being a participating member, and also by making sure we are compliant that alleviates the problem for us. Anybody writing SystemC code can synthesize through our Cynthesizer, can simulate with the OSCL simulator and verify it with your favorite formal verification tool. It's really been a good investment for us. Competitively it has paid off big time. Other vendors are trying to catch up because they don't support SystemC flow.
I know you will be shocked to hear that most firms I interview claim to be number 1.
I think that's the marketing guy's job. None of us will come out and say we're number 5. The proof is being able to back it up. One of the vendors you wrote about recently told the whole world they were number one but couldn't name any of their customers or refused to. We don't have that problem. We have 7 out of the top 12 semiconductor companies, over 20 customers.
In Japan Toshiba, Cannon and OKI. In Europe Philips.
Brett: You now how the marketing game goes. Customers are not willing to let you use their name or come out and aggressively say they have had a success with your technology unless it is true and that they feel there is some benefit to them to go off and say it. Toshiba and OKI are two of our published success stories. These guys want to let the world know that they are leading in this space, that they are leading the charge for SystemC based design and succeeding with it.
Be the first to review this article