earticle

논문검색

Power Estimation for Alpha 21264 Using Performance Events and Impact of Ambient Temperature

원문정보

초록

영어

Higher power has become one of primary obstacles for improving the performance of processors. To manage the power of processor dynamically, it is necessary to obtain the power timely and accurately. Power estimation using performance events is a better way in terms of timeliness and accuracy. We analyzed the correlation between performance events and powers of microprocessors. Only one performance event was selected for the power estimation of one functional unit based on the correlation analysis. The power of each functional unit was estimated based on simple regression analysis. In addition, the impact of ambient temperature on the accuracy for power estimation was presented. It can be seen from the experimental results that (1) the power of each functional unit has close linear correlation with a performance event and can be estimated accurately; (2) the ambient temperature has the negative impacts on the accuracy for power estimation.

목차

Abstract
 1. Introduction
 2. Related Work
 3. Experimental Methodology
 4. Selection of Performance Events
 5. Regression Analysis
 6. Impact of Ambient Temperature on Estimation Accuracy
 7. Conclusion
 References

저자정보

  • Biying Zhang Department of Computer Science and Technology, Harbin Institute of Technology, Harbin, 150001, China, College of Computer and Information Engineering, Harbin University of Commerce, Harbin, 150028, China
  • Gang Cui Department of Computer Science and Technology, Harbin Institute of Technology, Harbin, 150001, China

참고문헌

자료제공 : 네이버학술정보

    함께 이용한 논문

      ※ 원문제공기관과의 협약기간이 종료되어 열람이 제한될 수 있습니다.

      0개의 논문이 장바구니에 담겼습니다.