Research Group: High Efficiency, Low Power Systems (HELPS)
- Eddie Pettis
- Changjiu Xian
- Yamini Nimmagadda
HAPPI: A Homogeneous Architecture for Power Policy Integration
A significant volume of research has concentrated on operating-system directed power management (OSPM). The primary focus of previous research has been the development of OSPM policies. Under different conditions, one policy may outperform another and vice versa. Hence, it is difficult, or even impossible, to design the ``best'' policy for all computers. In this paper, we explain how to select the best policies at run-time without user or administrator intervention. We present a software framework called the Homogeneous Architecture for Power Policy Integration (HAPPI). This architecture is portable across different platforms running Linux. HAPPI specifies common requirements for policies and provides an interface to simplify the implementation of policies in a commodity OS. We describe how policies may be improved using HAPPI. Our experiments indicate that many policies published in the literature require tuning to achieve significant energy savings. Our approach allows these policies to be compared simultaneously and improved iteratively without rebooting the system. By using HAPPI, we are able to improve the energy savings of several policies by up to 41 percent. HAPPI can select the best policy among a set of distinct policies at run-time. This mechanism provides a critical step towards the acceptance of sophisticated power management policies in commodity OSs.
We have posted two sample policies for those interested in how policies may be implemented in HAPPI.
Cross-Layer Energy Management by Architectures, Operating Systems, and Application Program
Energy conservation has become a major challenge in system design. Most existing techniques focus on individual layers: devices, circuits, architecture, or software. The proposed research conserves energy through a collaborative and coherent approach across different layers. In the hardware layer, special registers report cumulative energy consumption of different hardware modules. These registers are called energy counters, similar to performance counters in modern processors. Operating systems use the information from energy counters and follow a set of accounting rules to determine the amount of energy consumed by each process. These processes can adjust their behavior to use energy efficiently. For portable systems powered by batteries, some processes may off-load computation to a remote server that is grid-powered. A programming environment will be developed to assist programmers in improving energy efficiency. This proposal has broader impacts on improving the energy efficiency of electronic systems to provide services for longer periods of time for mobile users.
These papers are protected by copyright laws.
- Nathaniel Pettis and Yung-Hsiang Lu. "Implementation Guides for a Homogeneous Architecture for Power Policy Integration in Operating Systems." School of Electical and Computer Engineering, Purdue University, Tech. Rep. ECE07-10, March 2007.
- Jason Ridenour, Jianghai Hu, Nathaniel Pettis, and Yung-Hsiang Lu. "Low-Power Buffer Management for Streaming Data." IEEE Transactions on Circuits and Systems for Video Technology, 17 (2), p. 143-157, February 2007.
- Le Cai, Nathaniel Pettis, and Yung-Hsiang Lu. "Joint Power Management of Memory and Disk Under Performance Constraints." IEEE Transactions on Computer-Aided Design of Integrated Circuits, 25 (12), p. 2697-2711, December 2006.
- Le Cai and Yung-Hsiang Lu. "Power Reduction of Multiple Disks Using Dynamic Cache Resizing and Speed Control." International Symposium on Low Power Electronics and Design, p. 186-190, August 2006.
- Nathaniel Pettis, Le Cai, and Yung-Hsiang Lu. "Statistically Optimal Dynamic Power Management for Streaming Data." IEEE Transactions on Computers, 55 (7), p. 800-814, July 2006.
- Changjiu Xian and Yung-Hsiang Lu. "Dynamic Voltage Scaling for Multitasking Real-Time Systems with Uncertain Execution Time." Great Lakes Symposium on VLSI, p. 392-397, April 2006.
- Nathaniel Pettis, Jason Ridenour, and Yung-Hsiang Lu. "Automatic Run-Time Selection of Power Policies for Operating Systems." Design Automation and Test in Europe, p. 508-513, March 2006.
- Changjiu Xian and Yung-Hsiang Lu. "Energy Reduction by Workload Adaptation in a Multi-Process Environment." Design Automation and Test in Europe, p. 514-519, March 2006.
- Le Cai and Yung-Hsiang Lu. "Joint Power Management of Memory and Disk." Design Automation and Test in Europe, p. 86-91, March 2005.
- Le Cai and Yung-Hsiang Lu. "Energy Management Using Buffer Memory for Streaming Data." IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, p. 141-152, February 2005.
- Nathaniel Pettis, Le Cai, and Yung-Hsiang Lu. "Dynamic Power Management for Streaming Data." International Symposium on Low-Power Electronics and Design, p. 62-65, August 2004.
- Le Cai and Yung-Hsiang Lu. "Dynamic Power Management Using Data Buffers." Design Automation and Test in Europe, p. 526-531, March 2004.
Any opinions, findings and conclusions or recomendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation (NSF).