Jones looks at the big picture for very small computing systems
By Susan Kantor, ECE ILLINOIS
February 22, 2010
- The Multiscale Systems Research Center works to create the technologies all semiconductor companies need.
- ECE Professor Doug Jones will lead research on small-scale systems and very low-power devices.
- Jones's research will lead to ultra-low power devices that are more efficient than current devices.
Since 1998, the Semiconductor Research Corporation, along with industry and the Defense Advanced Research Projects Agency (DARPA), have supported the Focus Center Research Program. This program consists of multi-university centers that work to create the technologies all companies need to advance.
But these centers aren’t physical places. University faculty from across the country work on different themes for each center and collaborate during monthly meetings online and an annual meeting in person.
"The idea is that they’re supposed to involve some of the best faculty across the country, focusing on specific topics of great interest to the semiconductor industry," Jones said.
Originally, five centers each focused on a different level of research. Last year, a sixth center was created to look at an even higher level--multiscale systems. A core group from the Gigascale Systems Research Center, which Jones and several other ECE faculty worked on, is now involved with the Multiscale Systems Center.
"Even industry started to realize that they have to design their chips and their products in light of the larger system that they live and work in," Jones said. "This center is going to look at fundamental issues."
Jones is leading research in the small-scale systems theme of the new center.
"We’re still looking at multiscale systems--things that involve several chips or several devices. But we’re looking at extreme, ultra-low power," Jones said.
Jones’ challenge is to build very small computing systems that push the envelope in terms of ultra-low power. Brain-machine interfaces could be implanted and help paralyzed people move. Cochlear implants could help the deaf hear, and computer vision chips could help the blind see by stimulating optic nerves.
This research doesn’t just have health applications. It could make buildings more energy-efficient by having sensors that turn off lights in vacant rooms and regulate temperature. Soldiers could carry more life-saving electronic devices if battery weight was minimal.
"There are applications, really viable things we can do, but we can only do it if we can get the power consumption of these devices down and the size way down, and make these things much more efficient than they are today," Jones said.
The two other themes in MuSyC deal are large-scale systems, which look at making data centers function more efficiently, and distributed sense and control systems, which looks at devices that include small to large-scale systems, like an airplane.
"It’s important that every piece of it is as efficient as possible," Jones said. "But designing every piece alone without thinking about the whole system and what it has to do is not going to get us too far."
The system doesn’t have to always be precise. A brain implant needs to be accurate when a person is writing; if the person is waving goodbye, the implant can be less accurate and save energy.
"We’re trying to design hugely scalable systems," Jones said. "Systems today are pretty much efficient only at one operating point." Instead, systems should be efficient at all operating points.
One of the challenges of this center will be to make sure that the individual components work well together as a whole.
"More so than any of the other centers, MuSyC will succeed or fail on how well it is able to make those connections across scale, across people, across institutions, across different aspects of things. That’s really the challenge, but also the opportunity," Jones said.
Editor's note: media inquiries should be directed to Brad Petersen, Director of Communications, at firstname.lastname@example.org or (217) 244-6376.