editorial

Spring Changes The more things change, the more they don't stay the same--in publishing just as in semiconductors and EDA. by Tets Maniwa

The ongoing changes in the semiconductor and EDA industries refute a claim that change is the remaining constant. It seems to me that the rate of change is increasing: The Semiconductor Industry Association Roadmap, for example, seems to be out of date within a year of its publication. The 1994 edition predicted the 0.25-µm process becoming available in 2000, but last year as much as a quarter of new production starts were in 0.25-µm processes. The 1997 Roadmap slated 0.10-µm devices for 2003, and yet Motorola has recently announced that the first 0.10-µm devices--under development for PowerPCs--will go into production sometime next year.

Though they're moving full speed ahead, nanometer technologies have created a growing design gap between what's available in silicon and what the EDA tools can handle. Designers will, for the short term, have to rely more on internal EDA tool development until the standard commercial tools can catch up. After looking carefully at their EDA requirements, the large design houses and semiconductor companies are already starting to develop more tools internally. They want tools that can solve their problems--that is, analyze the proper parts of the problem at appropriate times--without requiring the designer to be an expert in multiple fields of analysis and in EDA tool interoperability.

The apparent increase in the rate of change for technology development and adoption strains all areas of electronics and EDA. It creates a need for tools that can simultaneously facilitate higher levels of abstraction while encompassing the design with greater detail and accuracy. The growing design gap is symptomatic of ensuing discontinuities and asynchronicities in the development processes.

One interesting bit of information I obtained from talking with people in various areas in the industry is that almost everyone--whether in design, manufacturing and process development, or in EDA--sees their development cycle staying at about 18 to 24 months from start of the development to completion. Development time is a source of contention among the three parties: The designers want tools capable of designing tomorrow's ICs with technology that's still under development, the semiconductor companies don't want to share their development information for fear of their competitors knowing what they're doing, and the EDA companies can't finish the tools until the final device and process models are set.

But like the software developers, who don't start the hardware-software integration until the prototypes are running in the boards, EDA companies, receiving partial data late in the process design cycle, still don't have a good reason for the egregious errors and bugs that ship with many tools. They also lack a good reason to keep their tools and formats separate in their respective Towers of Babel.

The ability to design an IC with millions of gates now rests on having tools and processes complete well before the designer needs them. Early availability, then, with very close correlation between the models and physical devices, should dictate much closer interaction between the semiconductor and EDA communities. If the individual companies aren't going to work together to develop the new tools and processes, then the industry needs the big consortia, like Sematech, to step in and coordinate activities. When the tools and processes come together in a way that the design community needs them, we will have done something useful to reduce the design gap.

Speaking of changes, Spring seems to be the time for them at Integrated System Design. Last year we changed the design of the pages, and this year we're changing management. Thus I want to introduce myself as the new editor-in-chief. I'm looking forward to the challenge of following and documenting the progress of the industry and the underlying technologies that make systems on a chip possible. If you'd like to share your thoughts concerning this editorial or other features of ISD, please write to me at tets@isdmag.com.

In my new duties, I'm displacing--but not entirely replacing--Jonah McLeod. Following Miller Freeman's acquisition of ISD, Jonah has become the editorial director and is focusing his talents and energies on developing new publishing opportunities. We wish him well in his latest endeavors, and we continue to look to him for his insight and leadership.

To voice an opinion on this or any Integrated System Design article, please email your message to miker@isdmag.com.

integrated system design  March 1999