The vast majority of the research on power management to date has focused on energy conservation. However, many researchers are now starting to address the next key challenge: limiting power consumption to a less-than-peak power budget with minimal performance degradation. In this talk, I will first motivate approaches for limiting power consumption and contrast them with energy conservation approaches. As an example of research on limiting power, I will detail our recent work on limiting memory power by dynamically adjusting the power states of the memory devices. As part of this work, we found that memory power can be limited with very low performance degradation. We also found that limiting power is at least as effective for energy conservation as state-of-the-art techniques explicitly designed for performance-aware energy conservation. These latter results represent a departure from current energy conservation research and practice.
Bio: Ricardo Bianchini earned his PhD from the University of Rochester. He is an Associate Professor at Rutgers University in the Computer Science Department.