EPIC aces spice with new simulator

Electronic News, Dec 2, 1996 by Judy Erkanat

San Jose, Calif.--A new approach to mixed-signal design simulation from electronic design automation (EDA) provider EPIC Design Technology, Inc., called ACE (analog circuit engine), will launch in early 1997, according to Sang Wang, EPIC's chairman and CEO. The new simulation engine, which just finished development and is currently going through beta testing, will address mixed-signal designs of SRAMs and DRAMs. Dr. Wang feels the new simulator will revolutionize simulation test for IC designers and become the base for what he called his company's domination of this segment of the EDA market.

"We have overcome the limitations of Spice and created a tremendous amount of efficiency in this new core engine," said Dr. Wang, who joined Electronic News recently with EPIC's president, Bernard Aronson, for a roundtable breakfast interview.

"This is one of the most exciting products we've ever brought out," said Mr. Aronson. "We start production in January for Unix workstations like IBM, Sun, HP. We're calling the new products TimeMillACE and PowerMillACE."

Besides the new simulator, the EPIC executives spoke on a number of diverse issues, from the need for industry standards and factors driving future products, to the importance of product simulation and issues facing DRAM houses.

Naming its main competitor as Avant! Corp., through its acquisition of Anagram, Inc. (EN, Aug. 19), Mr. Aronson addressed the importance of analog and mixed-signal circuit design.

"I think a mixed-signal solution has been sought for a long time, with many companies, like Mentor Graphics and Cadence, trying to provide a mixed-signal solution," he said. "For some reason, analog tools found only a small market, but a tool that will do digital and analog will find a big market. Multimedia and telecommunication are heating up, so the numbers of silicon design starts will increase demand for this function. Right now, we might just have the best product in this particular area of mixed-signal."

Mr. Aronson said mixed-signal's most difficult challenge is the interface between the digital and analog. "That interface always causes trouble, and if you try to merge two entities together, like a Verilog engine with Spice, there is always trouble due to Spice's limitations," he explained. "So we developed an engine with Spice-like accuracy, but based on a totally different approach, allowing us to handle pretty large analog circuits with no problem. At the same time, the digital engine is merging to simulate errors simultaneously, providing a much better interface between the digital and the analog; a co-engine. Next year you will hear more about EPIC's ability and our customers' acceptance of it. We are very confident in this."

EN: Will digital and analog still be done separately and then integrated, or will they be developed together?

Dr. Wang: "The best way is for digital and analog to be designed together. Typically, an analog designer will have no problem designing digital circuitry, because it is simpler. Before, every company had an analog design group and specialized in designing analog without thinking about digital. And the digital people didn't think about analog, either. But they want to design bigger, more efficient digital circuits and now the requirements are that we put analog and digital together. Either the two need to work more closely together or the analog design people will have to start to think about designing these circuits with digital portions better. We will see some kind of collaboration in the future. You cannot think about just digital or analog separately anymore. We will see some movement toward co-design of the mixed-signal chip in the future, an area where our new tool will, of course, help."

EN: How do you see the growth pattern of the EDA industry?

Dr. Wang: "The EDA industry used to grow slowly. Two years ago, it was growing 8-9 percent a year, but last year it began to grow more rapidly. According to EDAC, it grew something like 15 percent. This year, it is anticipated to grow 20 percent, although the slowdown of the overall semiconductor industry may affect EDA's annual growth rate. Within the EDA segment, the so-called deep-submicron solution segment is growing much faster than the rest of the industry. Dataquest has the growth numbers for this segment at 35-40 percent. EPIC is right in the middle of this deep submicron new market demand, with the IC industry designing for 0.35 micron and below needing a new solution. Last year, we grew 121 percent and this fiscal year, finishing in September, we grew 76 percent, much faster than the overall EDA industry and faster than this segment of deep submicron. The whole industry is moving ahead with more momentum than two or three years ago."

Mr. Aronson: "EDA, just like any other industry, keeps changing. We were very lucky. When Sang started the company, and really over the last five years, there was a lot of room for small companies as the industry went through the change from 1.5 to 1.2 to 0.8 micron. The semiconductor industry needed new tools, so there was a lot of opportunity for start-ups and point tool companies. Right now, we're going through a consolidation phase for whatever reasons, partly because the customers really don't want to buy from six or seven different EDA companies. They would like to buy from a smaller number, or even a single vendor, so we are seeing consolidation. That is also one of the reasons everybody's beefing up and buying companies; they want to give the total solutions the customers are demanding. It will be a little harder for point tool EDA companies in the future as the competitive nature of our business is probably going to accelerate."