Issue No.02 - April-June (1997 vol.14)
Published by the IEEE Computer Society
How is Europe doing in microelectronics? This is a question Europeans traveling to other continents hear frequently. In the seventies, it looked as if IC fabrication would move completely out of Europe within a couple of years. Three years before the turn of the century, it turns out, that has not happened. Although not all branches of microelectronics are well represented in Europe, the Old World and its people have provided leading-edge solutions in areas such as telecommunications and consumer and automotive electronics.
Alcatel and Siemens Telecom, for example, with combined revenues exceeding $10 billion, are the leading telecommunications companies. Nokia is a specialist in mobile communications. The GSM international standard for digital mobile communications currently provides mobile communications to 27 million people in 43 countries, and 105 countries have adopted the standard. For cable-based communication, ISDN (Integrated Services Digital Network) is in widespread use. DECT, the European digital standard for cordless phones, provides secure communication within homes and offices.
Philips and SGS-Thomson are among the top five companies in consumer electronics. They have been involved in products and technologies such as digital TV sets, CD audio and video equipment (including the new digital video disk), digital tape recording (DAT), and digital audio broadcasting (DAB). Bosch is the world's revenue leader in automotive electronics.
In research, Europe has made significant contributions in CAD topics such as hardware description languages, high-level synthesis, formal verification, and floor planning. Other strong research areas are process technology and testing.
In the mid- to late eighties, Europe organized its first workshops and conferences on design, test, and design automation. The design community launched EuroASIC, which focused on actual designs, in 1985. The test community held its first European event, the European Test Conference, in 1989. The first CAD conference, the European Design Automation Conference (EDAC), took place in Glasgow in 1990. Since then, these events have merged, becoming the European Design and Test Conference and Exhibition (ED&TC). With 398 paper submissions for the 1996 event, ED&TC has clearly achieved the goal of the merger: providing a high-quality, internationally recognized event that covers all aspects of design, design automation, and test.
In this special issue of IEEE Design & Test, we bring you articles based on some of the finest papers presented at ED&TC 96. To make our selection, we consulted the program committee's list of best paper candidates and evaluation forms from conference attendees. We also considered articles from authors who responded to the call published in IEEE D&T's Spring 1996 issue. Three to four referees reviewed each article.
In the first article, Liem, Naçabal, Valderrama, Paulin, and Jerraya describe software design tools and methodologies they developed to meet the requirements of a European microelectronics system house. More and more products of such companies, like the videophone system described here, use embedded processors. Designing software for these processors requires tools, including compilers, debuggers, profilers, and simulators. The authors propose methods for generating such tools quickly. This work exemplifies the emerging field of retargetable compilation, a field stimulated by a growing interest in core processors.
CMOS RAMs are especially vulnerable to open defects. In our second article, M. Sachdev describes test and testability enhancements that became necessary due to field returns at another European systems house. The potential problems resulting from stuck-at-open errors have been known for some time, but conventional test procedures do not take them fully into account. The author proposes solutions that can be applied to most systems containing embedded CMOS RAMs.
Microelectromechanical system manufacturing is an area in which Europe is rather strong. Micromechanical systems such as sensors and actuators, for example, are important to embedded systems. Until now, most progress in this area has been in fabrication technology. Although their increasing complexity makes manual design extremely difficult, there has been little CAD support for microelectromechanical systems. Karam, Courtois, and Boutamine describe how CMP, a European wafer service, is bridging the gap between microsystem designers and foundry facilities by adapting existing tools to provide an integrated CAD environment.
We selected the next article, by Bergamaschi and Raje, from the non-European papers presented at the conference. High-level, or behavioral, synthesis is a very complex process, and its correctness by construction is not guaranteed. Hence, we need techniques for checking the correctness of automatically generated designs. Making one of the first contributions to this area, the authors present algorithms for efficiently checking the equivalence between a behavioral specification and its scheduled implementation.
Formal design verification is essential to achieving product quality and avoiding company losses. Until recently, formal verification of multipliers was hardly feasible. Recent improvements of data structures for function representation have changed this situation. Drechsler, Becker, and Ruppertz have developed an extended binary decision diagram, the K*BMD, that maintains all the power of standard BDDs yet still allows efficient multiplier verification.
The trend toward embedded systems has generated increased interest in fault-tolerant design. In our sixth article, Nicolaidis, Duarte, Manich, and Figueras describe techniques for securing arithmetic operations through the use of parity generation. They present designs that avoid the large overhead typical of some classical techniques.
We hope you find these articles enlightening. To learn more about ED&TC, visit our web pages. Full conference proceedings are available online. For ED&TC 96, see ls12-www.informatik.uni-dortmund.de/edtc/96/edtc.html. For ED&TC 97, see http://www.imec.be/edtc/97/.
As in previous years, next year's conference and exhibition will be held in Paris in the spring. However, after a merger with EuroDAC, its new name will be "Design, Automation, and Test in Europe," or DATE. Future events are scheduled for Munich.
Finally, we thank all the people who spent time reviewing submissions for this special issue, as well as all who helped make ED&TC a success.
Peter Marwedel is a professor in the Computer Science Department of the University of Dortmund, Germany, where he served as department dean for several years. His current research areas include hardware-software codesign, high-level test generation, high-level synthesis, and code generation for embedded processors. He received the Dr. Habil. degree for his work on high-level synthesis and retargetable code generation based on the hardware description language MIMOLA. He is a coeditor of the recently published Code Generation for Embedded Processors. Marwedel received his PhD in physics from the University of Kiel. He is a member of the IEEE Computer Society, the ACM, and the GI (Gesellschaft für Informatik).
Carlos A. López-Barrio is the director of innovation at Telefónica I+D and a professor of electronics technology at the Technical University of Madrid. At Telefónica, he is responsible for developing new telecommunications services, providing hardware and software technologies, and ensuring a long-term technological vision for the company. His research interests include VLSI design and test methodologies, EDA tools for design and test of ASICs and systems, and digital architectures. He served as the general chair of the 1996 European Design and Test Conference. He is author or coauthor of several papers and conference presentations and of five textbooks on electronic circuits. López-Barrio holds a PhD in telecommunications engineering from the Technical University of Madrid.