Abstract
Dynamic voltage and frequency scaling (DVFS) is an essential mechanism for power saving in smartphones and mobile devices. Central processing unit (CPU) load based DVFS algorithms are widely used due to their simplicity of implementation. However, such algorithms often lead to a poor response time, which is one of the most important factors of user experience, especially for interactive applications. In this paper, the response time is mathematically modeled by considering the CPU frequency and characteristics of the running applications based on the Linux kernel's completely fair scheduler (CFS), and a Response time constrained Frequency & Priority (RFP) control scheme for improved power efficiency of smartphones is proposed. In the RFP algorithm, the CPU frequency and priority of the interactive applications are adaptively adjusted by estimating the response time in real time. The experimental results show that RFP can save energy up to 24.23% compared to the ondemand governor and up to 7.74% compared to HAPPE while satisfying the predefined threshold of the response time in Android-based smartphones.
Original language | English |
---|---|
Pages (from-to) | 65-78 |
Number of pages | 14 |
Journal | IEICE Transactions on Information and Systems |
Volume | E100D |
Issue number | 1 |
DOIs | |
Publication status | Published - 2017 Jan |
Fingerprint
All Science Journal Classification (ASJC) codes
- Software
- Hardware and Architecture
- Computer Vision and Pattern Recognition
- Electrical and Electronic Engineering
- Artificial Intelligence
Cite this
}
Response time constrained cpu frequency and priority control scheme for improved power efficiency in smartphones. / Jo, Sung Woong; Ha, Taeyoung; Kyong, Taehyun; Chung, Jong Moon.
In: IEICE Transactions on Information and Systems, Vol. E100D, No. 1, 01.2017, p. 65-78.Research output: Contribution to journal › Article
TY - JOUR
T1 - Response time constrained cpu frequency and priority control scheme for improved power efficiency in smartphones
AU - Jo, Sung Woong
AU - Ha, Taeyoung
AU - Kyong, Taehyun
AU - Chung, Jong Moon
PY - 2017/1
Y1 - 2017/1
N2 - Dynamic voltage and frequency scaling (DVFS) is an essential mechanism for power saving in smartphones and mobile devices. Central processing unit (CPU) load based DVFS algorithms are widely used due to their simplicity of implementation. However, such algorithms often lead to a poor response time, which is one of the most important factors of user experience, especially for interactive applications. In this paper, the response time is mathematically modeled by considering the CPU frequency and characteristics of the running applications based on the Linux kernel's completely fair scheduler (CFS), and a Response time constrained Frequency & Priority (RFP) control scheme for improved power efficiency of smartphones is proposed. In the RFP algorithm, the CPU frequency and priority of the interactive applications are adaptively adjusted by estimating the response time in real time. The experimental results show that RFP can save energy up to 24.23% compared to the ondemand governor and up to 7.74% compared to HAPPE while satisfying the predefined threshold of the response time in Android-based smartphones.
AB - Dynamic voltage and frequency scaling (DVFS) is an essential mechanism for power saving in smartphones and mobile devices. Central processing unit (CPU) load based DVFS algorithms are widely used due to their simplicity of implementation. However, such algorithms often lead to a poor response time, which is one of the most important factors of user experience, especially for interactive applications. In this paper, the response time is mathematically modeled by considering the CPU frequency and characteristics of the running applications based on the Linux kernel's completely fair scheduler (CFS), and a Response time constrained Frequency & Priority (RFP) control scheme for improved power efficiency of smartphones is proposed. In the RFP algorithm, the CPU frequency and priority of the interactive applications are adaptively adjusted by estimating the response time in real time. The experimental results show that RFP can save energy up to 24.23% compared to the ondemand governor and up to 7.74% compared to HAPPE while satisfying the predefined threshold of the response time in Android-based smartphones.
UR - http://www.scopus.com/inward/record.url?scp=85008425654&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85008425654&partnerID=8YFLogxK
U2 - 10.1587/transinf.2016EDP7148
DO - 10.1587/transinf.2016EDP7148
M3 - Article
AN - SCOPUS:85008425654
VL - E100D
SP - 65
EP - 78
JO - IEICE Transactions on Information and Systems
JF - IEICE Transactions on Information and Systems
SN - 0916-8532
IS - 1
ER -