A couple of months before the start of DAC I promised myself that I would find time to attend some of the technical program sessions. Then, the phone calls from marketing and public relations people begun, and next thing I knew, my time was all booked up with interviews, booth visits, and meetings with senior executives and financial analysts. The latter meetings are a peculiarity of mine. In addition to my technical degree, I also have a degree in Business Administration and Economics, and during my career as an executive at various EDA companies I acquired an appreciation for the economics side of the industry.
My "take-aways" from this year's event can be summarized by stating that the industry is at both a technical and a financial inflexion point. Some of what went on around DAC has made me think that the next year or two will require decisions that have the potential of reshaping the EDA industry.
Organizers can declare themselves satisfied with the outcome. The number of attendees was down from previous years, the reported preliminary total was 6001, but an overwhelming number of exhibitors stated that the quality of their leads was very good. Less visitors, but more qualified visitors. Those who came wanted to be there. A more telling figure of the recession and the diminishing funds for academic research is the number of full conference attendees. This number has been steadily decreasing every year, at least since 2006, regardless of the location of the conference, and this year the number was 1554. Still respectable but almost half the number of registrant in the same category in 2006.
The lack of support by the industry of traditional communication channels was also evident at DAC. In an effort to save money, many companies now rely on "social media" channels to deliver their message. DAC tried to address the divergence between professional journalists and bloggers by having two different areas dedicated to them. In addition to the traditional Press Room and interviews area, there was an area on the exhibit floor dedicated to bloggers. I am not sure how successful it was, although it looked attractive. The number of journalists attending DAC continues to decrease. Attendance from both US based professionals and international reporters was less than last year. The lack of professional attendance made the Press Room seem underutilized.
The Financial Stuff
I do not have the official number of financial analysts attendees, but it seemed to me that there were a few less than in previous years. Richard Valera, of Needham and Company, issued a report on his finding at DAC. In his summary he stated: " We attended the 47th annual Design Automation Conference in Anaheim, California this week. Not surprisingly, the tone of the show was meaningfully improved from last year. Much of the buzz/chatter on the floor was around the recent M&A activity in the space--especially the Virage (VIRL, Hold) and Denali transactions-and what the implications might be for other smaller EDA and IP companies. Another popular topic was Carl Icahn's recent involvement with Mentor Graphics (MENT, Buy), as some speculated whether it might be a catalyst for some form of restructuring or M&A with Mentor or the industry. Overall, while we gleaned plenty of interesting information at the show, our opinions and ratings remain unchanged."
I agree that Synopsys two acquisitions just before DAC, Virage Logic and Synfora, were topics of discussion. In my opinion, the Virage Logic acquisition is in part a reaction to the acquisition by Cadence of Denali. Synopsys sees itself as the provider of all things required to produce an IC, and the amount of memory blocks used by designers is increasing. The rest of the IP obtained from the acquisition is less important to Synopsys. How it will manage it and how it will integrate some of the IP into its DesignWare line will be telling. In particular, is there a future for the ARC processor? The Denali deal, instead, is specifically focused on memory: both IP and tools. Some analysts question the amount paid by Cadence stating that it was too high. But they forget that Denali was the leading privately funded company in EDA. That it had one year's worth of revenue in cash. That it was very profitable, and that is was the recognized leader in memory IP. Denali seamlessly complements the existing offerings by Cadence in the IP market.
Synoposys' acquisition of the Synfora assets diminishes the probability that a small company in the same market will now be acquired. In particular Forte's Cynthesizer may now face increased competition if Synopsys is successful in integrating the Synphony C compiler technology into its own flow. There is no doubt in my mind that the technology acquisition has strengthened Synopsys presence in the ESL market.
As far as the Mentor Graphics situation is concerned, I think that the company will benefit from a discussion with Carl Icahn. To me Mentor corporate structure seems to have redundant functionalities scattered among the operational divisions. Its operating mode also looks like it favors divisional success at the potential expense of corporate success. The latter is seen as a natural derivative of divisional success, but the result is that Mentor looks at times as a conglomerate of smaller specialized companies whose mission can be explained in far easier terms than explaining the corporate mission. Mentor continues to make good progress technologically, with the majority of its products judged to be very competitive. It seems that Carl Icahn is thinking that the company is under performing not because it is loosing its technological edge, but because it needs to evolve its corporate structure.
John Bruggeman of Cadence continued his mission to explain his EDA360 document to the industry. It was one of the subjects discussed by attendees and the opinions ranged from "We had that idea ten years ago but nobody paid attention" to "It is just marketing hype". The truth, of course is in the middle. The message is clear, at least to me. The industry must expand its market and change its way of doing business if it wants to grow. It is not a matter of survival: that is assured. But it is a matter of reaping returns commensurate to the contributions the industry makes to the semiconductor and systems industries.
It is true that the document does not contain a road map, or even specifics of how such a strategy can be implemented. These are company proprietary matters with high competitive value. Cadence would be foolish to disclose them. In my opinion the major message given is "it's the system, stupid", same thing I have said in various venues a number of times. What Bruggeman adds is the notion that applications, or "Apps" as Apple likes to call them, are key elements of a system and have significant commercial value. Whether an application takes the form of firmware, software, or hardware, is immaterial. But if the EDA industry can find a way to be part of this aftermarket, it will have found a significant way to expand its revenues.
During DAC I wrote that Formal Verification is a misnomer and that the market segment should be called Formal Methods. At the same time, and quite independent from each other, Gary Smith during his opening remarks in the DAC Pavilion stated that designers should be using formal analysis tools during the design phase.
Formal Verification tools can be divided into two classes: equivalence checking and formal analysis. Equivalence checking used to be much more useful when some type of manual transformation was necessary to proceed from one abstraction level to another in the same design. This requirement is becoming rare as tools are now available to proceed from one abstraction level to the next. Even at the circuit level it is less common to make manual changes, since the complexities and numbers of design rules make such attempts very risky.
Thus equivalence checking is very useful for tools developers as part of the quality assurance process when they develop synthesis tools, for example. For tools users equivalence checking is a way to prove that the tool they are using is trustworthy. In this case, again, it is mostly a matter of quality assurance, not development.