Survey of dynamic voltage scaling in power management

WANG Biao,WANG Xiao-ge

DOI: https://doi.org/10.3969/j.issn.1001-3695.2007.08.003

2007-01-01

Abstract:This paper surveyed related document released in recent 10 years,sorted and introduced various policies,refined their mathematic model for DVS problem,abstracted their solution's common thoughts,made comments on their strongpoint and shortcoming,and briefly introduced their typical implementation.

What problem does this paper attempt to address?

Probabilistic analysis of static voltage stability in power systems based on two-point estimate method

HOU Xue-yong,JU Ping,FU Hong-jun

DOI: https://doi.org/10.3876/j.issn.1000-1980.2010.03.022

2010-01-01

Abstract:For the quantitative analysis of voltage stability,especially for meeting the requirement of adaptation to its probabilistic analysis,a generalized margin index for static voltage stability (SVS) is proposed. The index can be employed to qualitatively judge the stability and quantitatively describe the degree of instability. The two-point estimate method was introduced into the probabilistic analysis of SVS. With regard to the uncertainty of parameters of power sources and load in power systems,the generalized index of the power difference margin was regarded as a stochastic variable to be solved. Its statistical characteristic values,such as the mean value and the standard deviation,were calculated so as to obtain the probability of the static voltage stability. The results of a calculation example were compared with those of the Monte-Carlo simulation method. It is concluded that the proposed method is valid,and that it can greatly save time while maintaining an equivalent precision.
A Review of Development in Dynamic Equivalent Technologies for Power System Modeling

ZHOU Hai-qiang,JU Ping,YANG Hui,CHEN Yan-ping

DOI: https://doi.org/10.3969/j.issn.1009-0665.2008.05.003

2008-01-01

Abstract:With the inceasing enlargement of the scale of modern power system the demands for memory and speed of the computation machines are becoming more and more.Even modern super computers are also too hard to qualify the analysis task with such a large and nonlinear artificial system.How to provide an accurate and speedy transient analysis service for consumers is always one of the most important concerns for power engineers.Dynamic equivalence can protrude dominant characters and analyze the internal system more speedy by system simplification.In this paper the development process of dynamic equivalence technology in recent years is introduced.The thinking and methods of different dynamic equivalent technologies are also discussed.Finally some suggestion with future equivalent study about power systems is given.
User- and Process-Driven Dynamic Voltage and Frequency Scaling.

Bin Lin,Arindam Mallik,Peter Dinda,Gokhan Memik,Robert Dick

DOI: https://doi.org/10.1109/ispass.2009.4919634

2011-01-01

Abstract:We describe and evaluate two new, independently-applicable power reduction techniques for power management on processors that support dynamic voltage and frequency scaling (DVFS): user-driven frequency scaling (UDFS) and process-driven voltage scaling (PDVS). In PDVS, a CPU-customized profile is derived offline that encodes the minimum voltage needed to achieve stability at each combination of CPU frequency and temperature. On a typical processor, PDVS reduces the voltage below the worst-case minimum operating voltages given in datasheets. UDFS, on the other hand, dynamically adapts CPU frequency to the individual user and the workload through direct user feedback. Our UDFS algorithms dramatically reduce typical operating frequencies and voltages while maintaining performance at a satisfactory level for each user. We evaluate our techniques independently and together through user studies conducted on a Pentium M laptop running Windows applications. We measure the overall system power and temperature reduction achieved by our methods. Combining PDVS and the best UDFS scheme reduces measured system power by 49.9% (27.8% PDVS, 22.1% UDFS), averaged across all our users and applications, compared to Windows XP DVFS. The average temperature of the CPU is decreased by 13.2degC. User trace-driven simulation to evaluate the CPU only indicates average CPU dynamic power savings of 57.3% (32.4% PDVS, 24.9% UDFS), with a maximum reduction of 83.4%. In a multitasking environment, the same UDFS+PDVS technique reduces the CPU dynamic power by 75.7% on average.
Power Reduction Through Measurement and Modeling of Users and CPUs: Summary

Bin Lin,Arindam Mallik,Peter A. Dinda,Gokhan Memik,Robert P. Dick

DOI: https://doi.org/10.1145/1269899.1254931

2007-01-01

ACM SIGMETRICS Performance Evaluation Review

Abstract:Dynamic Voltage and Frequency Scaling (DVFS) is one of the most commonly used power reduction techniques in high-performance processors. DVFS varies the frequency and voltage of a microprocessor in real-time according to processing needs. Although there are different versions of DVFS, at its core DVFS adapts power consumption and performance to the current workload of the CPU. Specifically, existing DVFS techniques in high-performance processors select an operating point (CPU frequency and voltage) based on the utilization of the processor. This approach integrates OS-level control, but such control is pessimistic. Existing DVFS techniques are pessimistic about the user. Indeed, they ignore the user, assuming that CPU utilization or the OS events prompting it are sufficient proxies. A high CPU utilization simply leads to a high frequency and high voltage, regardless of the user’s satisfaction or expectation of performance. Existing DVFS techniques are pessimistic about the CPU. They assume worst-case manufacturing process variation and operating temperature by basing their policies on loose worstcase bounds given by the processor manufacturer. A voltage level for a given frequency is set such that even the worst shipped processor of a given generation will be stable at the highest specified temperature. In response to these observations, we have developed, implemented, and evaluated the following two new power management techniques that can be readily employed independently or together. We elaborate on these techniques in detail elsewhere [4].
System-level Power Management Model Based on Dynamic Voltage Scaling Technology

ZHANG Sheng,WANG Nailong,ZHOU Runde

DOI: https://doi.org/10.3321/j.issn:1000-0054.2005.04.032

2005-01-01

Abstract:This paper presents a theoretical model for system-level power management based on adaptive voltage scaling, in which the voltage-scaling factor was used to indicate different operating states so that the delay and power consumption can be evaluated together more efficiently. Different fixed voltage strategies and dynamic voltage scaling strategies were compared theoretically and in simulations. The results indicate that the optimized dynamic voltage scaling strategy proves better speed-power accommodating ability and better performance with much lower power consumption.
An Enhanced Dynamic Voltage Scaling Scheme for Energy-Efficient Embedded Real-Time Control Systems

Feng Xia,Youxian Sun

DOI: https://doi.org/10.1007/11751632_59

2006-01-01

Abstract:Real-Time Dynamic Voltage Scaling (RT-DVS) has been one of the most important techniques for energy savings in battery-powered embedded systems. However, pure RT-DVS approaches rarely take into account the actual performance requirements of the target applications. With the primary goal of further reducing energy consumption while satisfying Quality of Control (QoC) requirements in real-time control systems, an enhanced dynamic voltage scaling (EDVS) scheme is suggested. Following the direct feedback scheduling methodology, EDVS exploits a QoC-aware adaptive resource allocation mechanism. It enables flexible timing constraints on control tasks, which facilitates further energy saving over pure RT-DVS. Simulation experiments argue that EDVS is highly cost-effective and can save much more energy over the optimal pure RT-DVS scheme, while providing comparable QoC.
Control-theoretic dynamic voltage scaling for embedded controllers

Feng Xia,Yu-Chu Tian,Youxian Sun,Jinxiang Dong

DOI: https://doi.org/10.48550/arXiv.0806.0132

2008-06-01

Abstract:For microprocessors used in real-time embedded systems, minimizing power consumption is difficult due to the timing constraints. Dynamic voltage scaling (DVS) has been incorporated into modern microprocessors as a promising technique for exploring the trade-off between energy consumption and system performance. However, it remains a challenge to realize the potential of DVS in unpredictable environments where the system workload cannot be accurately known. Addressing system-level power-aware design for DVS-enabled embedded controllers, this paper establishes an analytical model for the DVS system that encompasses multiple real-time control tasks. From this model, a feedback control based approach to power management is developed to reduce dynamic power consumption while achieving good application performance. With this approach, the unpredictability and variability of task execution times can be attacked. Thanks to the use of feedback control theory, predictable performance of the DVS system is achieved, which is favorable to real-time applications. Extensive simulations are conducted to evaluate the performance of the proposed approach.

Operating Systems
Online Evaluation and Deduction of Power System Dynamic Voltage Support Availability

Yawei Wei,Ying Lv,Zigan Wang,Hongyang Dai,Bing Wang

DOI: https://doi.org/10.1109/PowerCon58120.2023.10331383

2023-01-01

Abstract:The online utility renewable energy station has increased continuously in scale into the major power grid. The local power-load balance has shifted entirely into the UHVDC long-distance transfer and remote load center consumption mode. The new mode leads to a considerable amount of conventional units at both the sending and receiving end being retired. These retire trend results in a new challenge to the security of grid voltage on both sides. In this paper, the downward dynamic voltage support availability (DVSA) is first proposed to provide a quantitively judgement for the prominent transient voltage instability and power grid transient overvoltage at both ends. Thereafter, the relationship between the dynamic voltage support availability and the reactive power demand after disturbances is also discovered. The DVSA problem has a strong connection to their local reactive power reserve. Therefore, to distinguish the difference of multi-area power system DVSAs, the DVSA-based partition and weak partition screening methods are proposed as well. In the end, an online evaluation method of the area DVSA value is proposed considering the analysis time-window. This fast evaluation method for the power grid's DVSA weak moment and weak area is proposed in response to the grid's future operation mode. The rationality of the DVSA index proposed in this paper is verified by using industrial provincial power grid simulations.
Influence of CPU Frequency Upon Battery and Improvement of Dynamic Voltage Scaling Algorithms

WU Gang,TANG Zhi-qiang,LI Wen-rong,TU Si-liang

DOI: https://doi.org/10.3969/j.issn.1000-1220.2006.11.021

2006-01-01

Abstract:Long battery lifetime is an important target of battery-powered embedded systems.Kinds of DVS (Digital Voltage Scaling) algorithms can be used to prolong battery lifetime. Unfortunately, there is still no universal way to analyze the influence of CPU frequency upon battery, and known DVS algorithms are not flawless. A new way is presented in this paper to analyze the influence, whose conclusion is universal and verified by experiments. The conclusion is applied to a DVS algorithm to make it more suitable for battery-powered systems.
Analysis of Power Saving Effect for Dynamic Power Management

Wenzhi Chen,Huan Zheng

DOI: https://doi.org/10.1109/MESA.2006.296937

2007-01-01

Abstract:Dynamic power management (DPM) is a design methodology for dynamically reconfiguring systems to provide the requested services and performance levels with a minimum number of active components or a minimum load on such components. The principle of DPM selectively shuts down or depresses performance of some components that are idle. Many approaches have been exploited for DPM, different approach leads to different effects of power saving. Former analysis of power saving effect of DPM approach used to use theoretical computation or simulation which are difficult to reflect the real efficiency of power saving of the target DPM approach, this paper presents an analytic approach which could be embedded into the real implementation of DPM architecture to monitor the behavior of DPM approach, record information it needs for analysis, and analyze power saving effect offline, it could also support systems that have multi low power states. Experiment results shows that this analytic approach gets the same results as the simulated results
Performance-Aware Power Management in Embedded Controllers with Multiple-Voltage Processors

Feng Xia,Liping Liu,Longhua Ma,Youxian Sun,Jinxiang Dong

DOI: https://doi.org/10.3923/itj.2008.942.947

2008-01-01

Information Technology Journal

Abstract:The goal of this work is to minimize the energy dissipation of embedded controllers without jeopardizing the quality of control (QoC). Taking advantage of the dynamic voltage scaling (DVS) technology, this paper develops a performance-aware power management scheme for embedded controllers with processors that allow multiple voltage levels. The periods of control tasks are adapted online with respect to the current QoC, thus facilitating additional energy reduction over standard DVS. To avoid the waste of CPU resources as a result of the discrete voltage levels, a resource reclaiming mechanism is employed to maximize the CPU utilization and also to improve the QoC. Simulations are conducted to evaluate the performance of the proposed scheme. Compared with the optimal standard DVS scheme, the proposed scheme is shown to be able to save remarkably more energy while maintaining comparable QoC.
Optimal DVS algorithm for real-time systems with double voltage scalable processor

Huayong WANG,Yu CHEN,Shuo KANG,Yiqi DAI

DOI: https://doi.org/10.3321/j.issn:1000-0054.2005.10.029

2005-01-01

Abstract:The power consumption in real-time embedded systems can be improved with an optimal dynamic voltage scaling(DVS) algorithm for micro-processors with double scalable voltages.Partial frequency scaling was used for the real-time periodic task sets.A mathematical model of the scheduling system was used to prove the optimality.Analyses and simulations further prove the effectiveness of the algorithm.The test results show that the DVS algorithm significantly reduces the power consumptions.Although the power savings depends greatly on the task sets and processor states,it averages 20% or more,more than previous DVS algorithms.
Multi-Battery Scheduling for Battery-Powered DVS Systems.

Peng Ouyang,Shouyi Yin,Leibo Liu,Shaojun Wei

DOI: https://doi.org/10.1587/transcom.e95.b.2278

2012-01-01

IEICE Transactions on Communications

Abstract:More and more mobile devices adopt multi-battery and dynamic voltage scaling policy (DVS) to reduce the energy consumption and extend the battery runtime. However, since the nonlinear characteristics of the multi-battery are not considered, the practical efficiency is not good enough. In order to reduce the energy consumption and extend the battery runtime, this paper proposes an approach based on the battery characteristics to implement the co-optimization of the multi-battery scheduling and dynamic voltage scaling on multi-battery powered systems. In this work, considering the nonlinear discharging characteristics of the existing batteries, we use the Markov process to depict the multi-battery discharging behavior, and build a multi-objective optimal model to denote the energy consumption and battery states, then propose a binary tree based algorithm to solve this model. By means of this method, we get an optimal and applicable scheme about multi-battery scheduling and dynamic voltage scaling. Experimental results show that this approach achieves an average improvement in battery runtime of 17.5% over the current methods in physical implementation.
Power Aware Embedded Software and Hardware Design Driven by Battery Model

HUANG Jiang-wei,HU Wei,XIANG Ling-xiang,CHEN Tian-zhou

DOI: https://doi.org/10.3785/j.issn.1008-973x.2009.12.004

2009-01-01

Abstract:The energy of the battery was affected by the load current. A power-aware software and hard-ware design approach for the dynamic voltage scaling (DVS) enabled system was presented according to the relationship between the changes in the battery capacity and DVS. The approach took the battery char-acteristics into account and parameterized the impacts of DVS on the battery capacity. Then the power-a-ware embedded system design model was established in theory. According to the model, there are two power-efficient embedded system design principles: the optimal speedup factor and the optimal work ratio. Experimental results show that the approach is more accurate to reflect the trade-off between battery life and performance, and the speed derived from the approach is optimal.
Robust and adaptive dynamic power management for time varying system

Min Li,Xiaobo Wu,Menglian Zhao,Ping Li,Xiaolang Yan

DOI: https://doi.org/10.1007/11535409_65

2005-01-01

Abstract:Dynamic Power Management (DPM) is an effective power reduction technique to dynamically control power state of system components. Although there are already a lot of papers on DPM, few of them present robust and adaptive solution to handle the inherent time varying behavior of real world system. In this paper, we propose techniques for targeting DPM on the time varying behaviors. In our approach, we apply efficient policy optimization method to generate online power management policy to achieve adaptiveness. Moreover, in order to handle the small scale varying behavior (perturbation) that tends to significantly degrade the performance of DPM, we apply the Bounded Parameter Markov Decision Process and Interval Value Iteration to derive robust policy tolerant to perturbations. Simulation results show that the proposed techniques are effective.
Dynamic Voltage Scaling with Links for Power Optimization of Interconnection Networks

L Shang,LS Peh,NK Jha

DOI: https://doi.org/10.1109/l-ca.2002.10

2003-01-01

Abstract:Originally developed to connect processors and memories in multicomputers, prior research and design of interconnection networks have focused largely on performance. As these networks get deployed in a wide range of new applications, where power is becoming a key design constraint, we need to seriously consider power efficiency in designing interconnection networks. As the demand for network bandwidth increases, communication links, already a significant consumer of power now, will take up an ever larger portion of total system power budget. In this paper, we motivate the use of dynamic voltage scaling (DVS) for links, where the frequency and voltage of links are dynamically adjusted to minimize power consumption. We propose a history-based DVS policy that judiciously adjusts link frequencies and voltages based on past utilization. Our approach realizes up to 6.3X power savings (4.6X on average). This is accompanied by a moderate impact on performance (15.2% increase in average latency before network saturation and 2.5% reduction in throughput.) To the best of our knowledge, this is the first study. that targets dynamic power optimization of interconnection networks.
Optimization of Power Consumption Through Dynamic Voltage Scaling Based on Hardware Profiling

WU Hao,ZHOU Xuehai

DOI: https://doi.org/10.3969/j.issn.1000-3428.2007.12.084

2007-01-01

Abstract:Embedded systems are often severely energy-constrained. The need to meet real-time requirements while staying within energy constraint presents new design challenges. Dynamic voltage scaling (DVS) is an effective technique for reducing processor energy consumption. The technique uses a hardware profiling scheme to identify "hotspot" , adjusts processor voltage and speed in response to the amount of ILP observed. In comparison to a processor running at a fixed voltage and speed, the approach can improve energy consumption with less impact on the performance.
Hard Real-Time Dynamic Voltage Scaling Algorithm Based on Resource Operation

WU Gang,TANG Zhi-qiang,CHEN Zhang-long

DOI: https://doi.org/10.3969/j.issn.1000-1220.2006.10.042

2006-01-01

Abstract:Dynamic Voltage Scaling algorithms are used to save system energy in order to decrease CPU heat and/or prolong the lifetime of battery-powered systems. Unfortunately, known DVS algorithms assume that the processes never operate resources(requiring or releasing resources), which is almost impossible is actual applications and hence these algorithms are not suitable for actual applications. This paper presents a new hard real-time DVS algorithm. As compared with known algorithms, the new algorithm completes processes with less energy while the processes are not forbidden to operate resources. The additional function of the new algorithm is deadlock avoidance. It is valuable for actual applications.
Feedback Scheduling of Real-Time Control Tasks in Power-Aware Embedded Systems

F Xia,XH Dai,XD Wang,YX Sun

DOI: https://doi.org/10.1109/icess.2005.56

2005-01-01

Abstract:Power awareness has become a critical issue in real-time scheduling of embedded systems. In the context of control applications, the goal of high control performance and low energy consumption are at odds with each other. While dynamic voltage/frequency scaling (DVS) has proved to be promising in energy saving while preserving task schedulability, traditional DVS algorithms use either open loop or ad hoc solutions, and hence cannot perform well for dynamic systems where the workload varies significantly. By targeting these systems, a novel scheme, namely DVS-FS, which combines DVS and feedback scheduling, is suggested. The objective is to save CPU energy as much as possible, while still providing control performance guarantees, which largely depends on successful schedule of the control task set. DVS-FS exploits feedback control methodology, and facilitates tradeoffs between energy consumption and control performance through controlling the CPU utilization at a considerably high level. Simulation experiments demonstrate that DVS-FS can easily reduce significant energy consumption at the expense of only minor control performance degradation.
Dynamic Voltage and Frequency Scaling for Scientific Applications

Chung-Hsing Hsu,Ulrich Kremer

DOI: https://doi.org/10.1007/3-540-35767-x_6

2003-01-01

Abstract:Dynamic voltage and frequency scaling (DVFS) of the CPU has been shown to be one of the most effective ways to reduce energy consumption of a program. This paper discusses the benefit of dynamic voltage and frequency scaling for scientific applications under different optimization levels. The reported experiments show that there are still many opportunities to apply DVFS to the highly optimized codes, and the profitability is significant across the benchmarks. It is also observed that there are performance and energy consumption tradeoffs for different optimization levels in the presence of DVFS. While in general compiling for performance will improve energy usage as well, in some cases the less successful optimization lead to higher energy savings. Finally, a comparison of the benefits of operating system support versus compiler support for DVFS is discussed.

Survey of dynamic voltage scaling in power management

Probabilistic analysis of static voltage stability in power systems based on two-point estimate method

A Review of Development in Dynamic Equivalent Technologies for Power System Modeling

User- and Process-Driven Dynamic Voltage and Frequency Scaling.

Power Reduction Through Measurement and Modeling of Users and CPUs: Summary

System-level Power Management Model Based on Dynamic Voltage Scaling Technology

An Enhanced Dynamic Voltage Scaling Scheme for Energy-Efficient Embedded Real-Time Control Systems

Control-theoretic dynamic voltage scaling for embedded controllers

Online Evaluation and Deduction of Power System Dynamic Voltage Support Availability

Influence of CPU Frequency Upon Battery and Improvement of Dynamic Voltage Scaling Algorithms

Analysis of Power Saving Effect for Dynamic Power Management

Performance-Aware Power Management in Embedded Controllers with Multiple-Voltage Processors

Optimal DVS algorithm for real-time systems with double voltage scalable processor

Multi-Battery Scheduling for Battery-Powered DVS Systems.

Power Aware Embedded Software and Hardware Design Driven by Battery Model

Robust and adaptive dynamic power management for time varying system

Dynamic Voltage Scaling with Links for Power Optimization of Interconnection Networks

Optimization of Power Consumption Through Dynamic Voltage Scaling Based on Hardware Profiling

Hard Real-Time Dynamic Voltage Scaling Algorithm Based on Resource Operation

Feedback Scheduling of Real-Time Control Tasks in Power-Aware Embedded Systems

Dynamic Voltage and Frequency Scaling for Scientific Applications