论文信息 - The Danger of Interval-Based Power Efficiency Metrics: When Worst Is Best

The Danger of Interval-Based Power Efficiency Metrics: When Worst Is Best

This paper shows that if the execution of a program is divided into distinct intervals, it is possible for one processor or configuration to provide the best power efficiency over every interval, and yet have worse overall power efficiency over the entire execution than other configurations. This unintuitive behavior is a result of a seemingly intuitive use of power efficiency metrics, and can result in suboptimal design and execution decisions. This behavior may occur when using the energy-delay product and energy-delay product metrics but not with the energy metric.

Dean M. Tullsen | Rakesh Kumar | Yiannakis Sazeides | Theofanis Constantinou

[1] Anantha P. Chandrakasan,et al. Low Power Digital CMOS Design , 1995 .

[2] Anantha P. Chandrakasan,et al. Low-power CMOS digital design , 1992 .

[3] Norman P. Jouppi,et al. Processor Power Reduction Via Single-ISA Heterogeneous Multi-Core Architectures , 2003, IEEE Computer Architecture Letters.

[4] Norman P. Jouppi,et al. Single-ISA Heterogeneous Multi-Core Architectures: The Potential for Processor Power Reduction , 2003, MICRO.

[5] Margaret Martonosi,et al. Wattch: a framework for architectural-level power analysis and optimizations , 2000, Proceedings of 27th International Symposium on Computer Architecture (IEEE Cat. No.RS00201).

[6] Manish Gupta,et al. Power-Aware Microarchitecture: Design and Modeling Challenges for Next-Generation Microprocessors , 2000, IEEE Micro.

[7] M. Horowitz,et al. Low-power digital design , 1994, Proceedings of 1994 IEEE Symposium on Low Power Electronics.