EECS Researchers Awarded NSF Research Grant to Study Near-Threshold Computing   Bookmark and Share

NTC Researchers Blaauw, Mudge, and Sylvester

EECS Professors David Blaauw, Trevor Mudge, and Dennis Sylvester have received a National Science Foundation research grant to study ultra-energy-efficient computing. At just under $2M, the grant is the largest of 44 recently received from NSF by the University of Michigan.

The study is part of a joint research project funded by grants totaling nearly $3M, in which the U-M team will collaborate with peers at Arizona State University and Harvey Mudd College to address a problem of pressing significance to the semiconductor industry: how to continue increases in microchip transistor density while enhancing performance. Under current device designs and architectures, potential future gains in performance by increasing density will be negated by unacceptable levels of power use and heat dissipation.

The research team believes that the solution to this microchip "energy crisis" is the universal application of aggressive low-voltage operation across all computational devices and platforms. They intend to demonstrate this through near-threshold computing (NTC), under which operating voltage is lowered to near the threshold voltage of the devices where they pivot between being on and off. Any voltage excess over the turn-on point of the transistor is eliminated, thereby reducing power consumption.

The researchers believe that their work will yield novel methods to overcome the barriers that have historically relegated ultra-low voltage operation to niche markets. The research should be applicable to a broad range of computation needs, including:

  • High-performance platforms, typically targeted for use in data centers, which currently create large amounts of heat and require major investments in power and cooling infrastructure, resulting in major environmental and societal impact.

  • Personal computing platforms, which are increasingly wireless and miniaturized, and are limited by tradeoffs between battery lifetimes and computational requirements.

  • Sensor-based platforms, which critically depend on ultra-low power and miniturized form factors to unlock new semiconductor applications, such as medically implanted monitoring and actuation devices, as well as ubiquitous environmental monitoring, such as structural sensing within critical infrastructure elements such as bridges.

Read More:

To learn more about federal stimulus-package research funds awarded to U-M researchers, see this U-M News Service press release.

Related Topics:  Blaauw, David    Integrated Circuits and VLSI   Near-threshold computing   Sylvester, Dennis