The DesignCon conference provided opportunities for thought, learning, and analysis. It also showed that we have yet to solve important problems in methods and training that have faced us for some time.
The Tech Forums at DesignCon are similar to embedded tutorials. They were held on Monday and I was interested in the one covering ESL topics. I arrived at the Santa Clara convention center a bit late and he sessions were already on-going. So, although about two dozen people were waiting in line to register, I could not get a good idea of the turnout. The ESL session had three panelists: Grant Martin, from Tensilica, Brian Bailey, independent consultant, and Andrew Piziali of Cadence. In case you do not know these gentlemen, let me assure you that were I to put together a panel of the subject, I would invite all three of them. There were only about two dozen attendees, but they stayed all three hours. At lunchtime, all, or most, of those registered were offered a free lunch and a keynote speech on multi-core architecture. The room was set up for 300 persons. It was standing room only and thus approximately 50 people eventually left without eating. The bottom line thus is that the Monday sessions were successful and that the majority of engineers are still more interested in implementation topics than in ESL topics. This is not good news for the American engineer that will see continuously increasing competition from offshore for implementation jobs, while architectural and design positions will remain in the US longer.
The lack of attention to architectural design in colleges and on the job was reinforced by a panel moderated by Ann Steffora Mutschler that posed the question: "Why are we still designing like it is 1992?". Of course we are not, but it may seem like it because all of the panelist focused on implementation, no one ever mentioned that we need to train student on skills that would allow them to be true designers and not just implementers, and no one pointed out that since the greater payoff is working silicon most of the focus and money in EDA R&D goes to back end problems. You can go to the news section to find a more detailed coverage of this panel.
On Tuesday morning the Business Forum track offered a panel dealing with thermal performance. It is not clear what the subject had to do with business, and the audience got it right, since it was composed mostly by design engineers, with the usual sprinkling of press editors. I came away from the session with two important messages. Thermal problems require an interdisciplinary team to develop a satisfactory solution. The electronic design industry is not very good at assembling such a team with the exception of very large companies who have staff trained in all the required disciplines. The second message was less direct, but also relevant to an ongoing debate within the EDA industry: all the panelists addressed the issue of power and no one mentioned the need for a standard, or expressed the wish that existing competing proposals would soon be harmonized. So, are we arguing for nothing? Do customers really care? I guess that if Si2 has over 100 members and only 16 have joined the Open Power Coalition, I already know the answer.
Another panel in the Business Forum track had the title: "Putting the "D" back into DFM". Yes I know, it is not clear what this has to do with business either, but it is an interesting topic, and one I have written about and care about. To my surprise, and that of some of the members of the audience, we were told that the D was always there, or, in the case of Cadence, had just been put back, and thus the world was a peace. I thought that in order to clarify the discussion it would be a good idea to define what D for design meant. Peter Rabkin, Program Director at Xilinx, was happy to answer me. Design is everything, he said. And went on to explain that every thing an design engineer will do in planning, developing, optimizing, and verifying a chip is design. OK, then why are we so concerned? My problem remains. A term that is everything is not a useful term. And there are people in the industry who feel that DFM tools are mostly targeting layout and fabrication issues. They try to identify a problem and fix it. We, the "D" team would like to see tools that help designers avoid the problems to begin with. That is the reason "D" is missing in DFM. But, no one of this persuasion was on the panel, and the moderator, through no fault of his own, could not fix this oversight.
On Wednesday Dr. Leah Jamieson keynote speech also sent a wakeup message to the US sector of our industry by reporting on a study of the higher educational system. Fewer and fewer high school graduates are choosing to enter engineering disciplines, especially electrical engineering. Colleges and Universities are studying the causes of this behavior and also trying to anticipate industry requirements for the next twelve years or so. The result is that there is a need to modify the curriculum to provide students with new skills, like managing change, become innovative, and entrepreneurial attitudes. And this is not just an internal problem. How do we improve the figure of the engineer in society, how do we motivate engineers to be more involved in society? Pressing questions that have no simple answers. One obvious problem in my opinion is that in many cases colleges and universities deal with "damaged goods". Students that in high school have not received the training and knowledge to allow them to blossom into mature individual capable of embarking on successful and well-rounded careers. The entire educational system in the United States needs to change in order to become competitive with those of Western Europe and leading Asian countries.
In conclusion, I am glad I attended DesignCon. There were, to be sure, lots of opportunities to deepen one's technical knowledge, but there were also opportunities to address some of today's important questions in our industry and through reflection and dialogue, learn something!