"Computing is like potato chips - there's never enough!"

Giovanni De Micheli © Giovanni De Micheli / EPFL 2022

Giovanni De Micheli © Giovanni De Micheli / EPFL 2022

EPFL’s Professor Giovanni De Micheli was recently recognized for his strong impact on the electronic system design industry, and he sees another decade of rapid design innovation ahead.

Giovanni De Micheli is a man on the move! Director of EPFL’s Integrated Systems Laboratory, part of both the School of Computer and Communication Sciences (IC) and the School of Engineering (STI) for more than 40 years he has been breaking the boundaries of Computer Science and Electronic design. He is credited for the invention of the Network on Chip (NoC) design automation paradigm, enabling large scale chip design, such as those used in mobile phones and smart vehicles.

Now, he has been honoured with the 2022 Phil Kaufman Award for Distinguished Contributions to Electronic System Design, presented annually by the Electronic System Design Alliance and the Institute of Electrical and Electronics Engineers (IEEE) Council on Electronic Design Automation (CEDA). “I am thrilled because this award is given with strong input from the industry and so it relates very much to the impact of the work that I have done and on how this has influenced the evolution of micro-electronics,” De Micheli said.

During a career that has expanded more than three decades, Giovanni De Micheli reflects that the field has changed dramatically since his early days at IBM. “Most people didn’t believe that you could design hardware like software, writing a program and transforming that into a chip: now this is possible, and current design practice. Indeed, design automation is about creating complex systems from high-level models.”

As the electronics industry continues to evolve at pace, De Micheli sees no end to the challenges that need to be overcome. These include the need to design increasingly more complex systems, integrated within silicon chips. The complexity of design can be tamed by using components and by interconnecting them by means of a Network on Chip (NoC) that embodies some of the principles of networks, such as the Internet, that we ubiquitously use today.

De Micheli believes that, despite what many people think, the reduction of transistor sizes will continue for at least another decade, and new materials will enable us to build specific accelerators for computing. This poses new challenges for the design of chips, in terms of mastering the very large number of transistors (such as over a trillion in a recent realization) and the physical properties of nanometer-scale devices.

The trend towards a greener computation landscape has also led to a renewed vision of electronic design. The total energy for computing will range from 5% to 10% of the world total energy budget, and it is on the rise. “In the last 20 years there's been a large effort in the design automation community, as well as in my group, for low-power electronics. This entails choices at the logic design level, such as information representation, and at the device level. Overall, structured interconnection on chips, such as NoCs, have helped to reduce on-chip power consumption drastically.

After forty years of his life dedicated to research De Micheli says that the joy and challenge of solving strange new problems still gets him up in the morning! Some are problems that have never been solved and others are new. “A field that’s going to be extremely interesting is smart cars. Like a phone, the car is an object touched by everybody, and we'll have more electric cars, as well as assisted-driving vehicles. This means that accidents are less likely to happen,” he says. “The end goal is a fully self-driving car, but this requires overcoming incredible hurdles in design and technology.”

Looking beyond chip applications, electronics will enable us to redesign smart vehicles, where assisted/automated driving exploits real-time communication among sensors and other vehicles.

The requirements for designing a car, that is similar to a supercomputer on wheels, are severe and require the integration and networking of many hardware devices, such as lidars and radars in addition to monitors of the road, engine and car chassis. Part of De Micheli’s research is linked to bridging the gap between integrated electronic and sensors, and finding ways to designing sensors in a structured manner, so that modularity can increase their ease of interface and reduce design time.

The broader impact of De Micheli’s research affects systems for improving human health and monitoring the environment, as showcased the Swiss Nano-Tera.ch research program that he created and directed. This program aims at bettering society through the use of technology as well as attracting students to scientific research with a broad outreach. Whereas NoCs exploited networking in the microscopic domain, a key distinguishing factor of Nano-Tera.ch is the application of networking inside/around the human body for health monitoring, and in the environment for its protection.

And his advice to students entering the field today? “Be bold and take risks. If you just try to extend the ongoing state-of-the-art, your work will be incremental with small value. Don't be afraid to fail because in failure you learn a lot for the next stage of your career. At the same time, be open and interact with others because physical interaction in presence is key to our intellectual growth.

Author: Tanya Petersen

Source: Computer and Communication Sciences | IC

This content is distributed under a Creative Commons CC BY-SA 4.0 license. You may freely reproduce the text, videos and images it contains, provided that you indicate the author’s name and place no restrictions on the subsequent use of the content. If you would like to reproduce an illustration that does not contain the CC BY-SA notice, you must obtain approval from the author.