EEtimes'dan bir yazı {EDA can shine beyond IC borders}

[gsever]

When I teach my graduate seminar in computer engineering, I often start with the question "What are the 10 most important engineering problems today?" This question never fails to wake up the class, and with some prompting and discussion, several topics tend to emerge: world hunger, clean air and water, infrastructure for developing countries, global warming, eradicating diseases. Most of us would agree that tackling these problems is important. The next question is harder: "Why aren't you working on one of these?"

Looking at our field of electronic design automation, we might be tempted to think that we, as a discipline, have nothing to do with solving these important problems. After all, what do logic minimization, placement, routing, simulation and so on have to do with, say, clean water?

But I believe that our work has made a difference, and that EDA has the potential to have an even more direct and positive impact on these issues. Already our industry has enabled the creation of the powerful, small, low-power, pervasive computing and communication systems that have transformed our world in the last half century. And those systems are used every day to help solve many important problems.

This year, the Design Automation Conference (DAC) will feature several examples of the impact of EDA outside the traditional chip environment. In addition to this year's automotive theme, which provides an excellent means for exploring how design relates to the full system, we also have special sessions on "wild and crazy ideas," nanoscale circuits and architectures, synthetic biology, and design and manufacturing for emerging technologies.

On Thursday, June 5, professor Jan M. Rabaey will speak about a related topic in his keynote "Design without borders," where he will present some of the exciting work in bio design automation that is going on at the University of California, Berkeley.

EDA methodologies, techniques and tools are unique in that they approach problems in terms of levels of abstraction, which gives us the power to work on complex problems from a high-level representation. We see complex systems as hierarchies of interacting components. This is reflected in our second strength: the modeling of very large systems by extracting their key behaviors in efficient representations for simulation.

The EDA industry also excels at developing techniques to solve large, complex optimization problems. And we are willing to tackle the black art of synthesis.

These key techniques of abstraction, extraction, optimization and synthesis from the EDA toolbox can be used in a broader context than just for electronic systems design. They are also applicable to other problems characterized by the complex behavior of large numbers of interacting components, in fields as diverse as routing systems for vehicular traffic, drug design, biology and health care. I believe that EDA has a great opportunity in the next decade to apply the tools we have developed for electronic systems design to these and other complex problems.

In my research at the University of Pittsburgh, I am looking at modeling "mixed-signal multidomain" systems based on multiple technologies including electronics, optics, mechanics and fluidics. What I have learned in the modeling and simulation of such systems is that we (the tool makers) need to reach out into these other domains to educate ourselves to the language and representations used by the design engineers (tool users) who work in these other domains. That is the next great challenge for EDA.

Steven Levitan, the chairman of DAC, is the John A. Jurenko professor of computer engineering at the University of Pittsburgh.