CN101067758B - Energy consumption management method for inserting system - Google Patents
Energy consumption management method for inserting system Download PDFInfo
- Publication number
- CN101067758B CN101067758B CN200710028578A CN200710028578A CN101067758B CN 101067758 B CN101067758 B CN 101067758B CN 200710028578 A CN200710028578 A CN 200710028578A CN 200710028578 A CN200710028578 A CN 200710028578A CN 101067758 B CN101067758 B CN 101067758B
- Authority
- CN
- China
- Prior art keywords
- task
- scheduling
- module
- performance
- idle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Abstract
The invention provides an energy consumption management method for an embedded system, implemented by dynamic power supply management support module, comprising: policy management module, idle detection module, and performance detection module, where the policy management module makes effective power supply management at system layer so as to implement system performance and power saving, and gives attention to idle state and executed load of system, and self-adaptively regulates equipment idle time according to system task executing state and makes the system enter a corresponding low-power state when the idle time continues to a certain extent; the idle detection module provides idle information of CPU and operating system and monitors the idle thread or activity of a given working load in the operating system and provides policy-making parameter information for the policy management module; and the performance detection module monitors CPU utilization ratio and utilization conditions of main memory and provides accurate executed load information for the policy management module.
Description
Technical field
The invention belongs to embedded system power management technique field.The energy consumption management method that particularly relates to a kind of embedded system.
Background technology
Fast development and widespread use along with embedded technology, embedded technology is combined closely with daily life, no matter be civilian or military use product all proposes more and more higher requirement to the application of embedded system, but performance increase substantially the problem of also having brought system energy consumption to increase.
In existing embedded system low-power consumption realization technology, mainly contain and be divided into low static power consumption and dynamic Low-power Technology.The low static power consumption technology mainly is the angle design electronic devices and components from hardware; Dynamically Low-power Technology is that angle from software realizes, mainly is to realize optimal design to low-power consumption from operating system, compiler and three levels of built-in application program.Dynamic power management DPM (Dynamic Power, Management) technology provides a kind of operating system grade other power management capabilities, comprise CPU frequency of operation and voltage, external bus clock frequency, dynamic adjustments, the management function of aspects such as external unit clock/power supply.Generate strategy and kernel provides management function mutual by client layer, adjust power parameter in real time and satisfy the real-time demands of applications of system simultaneously, allow the power management parameter at free time of short time or task run during at low power supply requirement, can by continually, low lingeringly adjust, thereby realize meticulousr, more intelligent power management.
Yet be low static power consumption technology or dynamic Low-power Technology all not from the finiteness of terminal resource and the characteristics of application program itself, often have technical sophistication, shortcomings such as low-power consumption DeGrain.
Summary of the invention
The objective of the invention is to deficiency at existing dynamic Low-power Technology, designed the adaptive dynamic energy consumption management method of a kind of embedded system, it mainly is to have considered the finiteness of terminal resource and the characteristics of application program itself, low power dissipation design and optimisation strategy at processor, Application Parameters is provided with the dynamic change of variation to system resource requirements, realizes the low energy consumption of embedded system.
In order to realize the foregoing invention purpose, the technical scheme of employing is as follows:
A kind of energy consumption management method of embedded system realizes that by the dynamic power management support module described dynamic power management support module comprises policy management module, idle detecting module, performance detecting module; Described policy management module is carried out effective power management in system layer, thereby the performance and the Power Cutback of the system of realization, the idle condition of its attention location system and execution load, according to the adaptive adjustment device free time of system task executing state, make system enter corresponding low power supply status when lasting till to a certain degree between at one's leisure; Described idle detecting module provides the idle information of CPU and operating system, its monitor idle thread or the activity of given operating load in operating system, and the parameter information of making a strategic decision is provided to policy management module; Described performance detecting module monitors the operating position of cpu busy percentage and main memory, for policy management module provides accurate execution load information.
In the technique scheme, described dynamic power management support module also comprises Subscriber Interface Module SIM, and described Subscriber Interface Module SIM is between application layer and policy management module, and the interface that allows user's input parameter control power management policy is provided.
Described dynamic power management support module also comprises application interface module, in order to the device registration application program, registered the ID that comprises equipment, be used for to specific equipment carry out power managed processing function, equipment frequency of operation and to the sensitive information of state.
Embedded system energy consumption management method of the present invention, can adjust the operating voltage and the frequency of processor according to each equipment of embedded system dynamically to the state variation of resource requirement, the limited resources of maximum using system, the availability of raising system, enable efficiently, stably to run on embedded platform, can satisfy high performance demands, can realize that again the whole energy consumption of system is less embedded product.
Technical essential of the present invention comprises:
(1) global variable of increase DPM dynamic power management in the core code of embedded system, this global variable represents whether the energy consumption of current system is in optimum state, that is to say, system whether under limited available resources the best power consumption of performance of system also less.The present invention is provided with and has defined five kinds of parameters of operating system by dynamic power management global variable DPMOperatingPoint is set, and comprises state, processor core voltage, processor core frequency, system bus frequency and the memory bus frequency of processor.
(2) need write the application programming interfaces storehouse that meets DPM API, each device driver in the embedded system is registered in the DPM software, control the state of each device power supply (DPS) Admin Events by these interfaces.
(3) definition three function Init (void), Exit (void), GetState () carry out initialization, termination and the removing of idle detecting module, the operations such as idle information of obtaining current system respectively.
(4) application programming interfaces of performance detecting module are shared identical data structure DPMPMUInfo with DPM software, the information that exchanges mutually between record performance detecting module and the policy management module.
(5) policy management module has realized a kind of dynamic power management adaptive control algorithm, adjusts supply voltage and frequency adaptively according to the dynamic change of system's external resource, and the performance of maximization system makes the energy consumption minimum of system simultaneously.
So how gordian technique point of the present invention obtains the demand information of whole each application module of embedded system at the dynamic power management support module, and integrate these information and make corresponding power policies selection.
The present invention has mainly considered the finiteness of terminal resource and the characteristics of application program itself, and at the low power dissipation design and the optimisation strategy of processor, Application Parameters is provided with the dynamic change of variation to system resource requirements, realizes the low energy consumption of embedded system.
Description of drawings
Fig. 1 is a dynamic power supply management method general frame logical schematic;
Fig. 2 is a dynamic power supplies tube side method task control schematic flow sheet;
Fig. 3 is a dynamic power supplies tube side method self-adaptive processing schematic flow sheet;
Fig. 4 is that self-adaptation dynamic power management algorithm controls is carried out schematic flow sheet.
Embodiment
Below in conjunction with accompanying drawing, concrete enforcement of the present invention is described in further detail.
Accompanying drawing 1 has been showed the logical organization of DPM dynamic power supply management method general frame, the figure illustrates the change of kernel and driver being done for the power managed of realizing system.
(1) sequence number 11: offer the POSIX interface of application program in operating system, increased DPM API in order to the device registration application program, by registration driver can be responded to some extent when fulgurite director part takes place.It has mainly registered the ID that comprises equipment, be used for to specific equipment carry out power managed processing function, equipment frequency of operation and to the sensitive information of state.When power parameter changes, can offer driver to these information, driver can be got ready for unsettled conversion, can veto conversion in case of necessity.
(2) sequence number 12: because dynamic power management system mainly is to select the power management decision-making by system and application message that comprehensive other each modules of policy management module obtain, so must being provided, corresponding special system function finishes these processing, as long as DPM core support module comprises dpm_start, dpm_int, SET_SPM_INT, UNSET_SPM_INT, dpm_event, mach_update_init. wherein, dpm_start opens the DPM module, dpm_int initialization DPM module, SET_SPM_INT is provided with the DPM interrupt bit, UNSET_SPM_INT, remove the DPM interrupt bit, dpm_event sends CPU and interrupts, and mach_update_init responds interrupt request.
(3) sequence number 13:DPM client driver can directly or by the power management services of operating system be cooperated with the DPM policy management module, and this state is mapped to the electric source modes of processor.
Accompanying drawing 2 has been showed the task control execution flow process of dynamic power management system, has mainly described the interactive information between each core composition module of dynamic power management support module.
(1) sequence number 21: application program is finished registration in DPM, comprises that the frequency of operation of function, equipment is handled in ID number of equipment, power managed and to the sensitive information of state.
(2) sequence number 22: the definition Subscriber Interface Module SIM, be used for some input parameters of control strategy administration module, and defining the device driver interface of registering in some DPM softwares, policy management module realizes operating system is had the control of electric source modes by these interfaces.DPMRegisterDevice, CallBack, DPMAddDevice, DPMDeRegisterDevice, five functions of DPMRemoveDevice have been defined, wherein the DPMRegisterDevice function is the driver of each equipment of registration embedded system in the DPM module, CallBack call back function and DPM module are carried out communication, DPMAddDevice is interpolation equipment in DPM, DPMDeRegisterDevice cancels the registration of equipment in DPM, DPMRemoveDevice removes the equipment that is added by DPMAddDevice.
(3) sequence number 23: policy management module calls application interface and idle detecting module communication.Wherein call function DPMIdleProfilerInit finishes the initialization of idle detecting module; Call function DPMIdleProfileExit stops the operation of idle detecting module and finishes removing work, and policy management module was no longer received any message from idle detecting module after idle detecting module failed; Call function DPMIdleProfilerGetState obtains the Idle information of current system.
(4) sequence number 24: the performance detecting module mainly monitors the utilization factor of CPU and the operating position of main memory.Mainly defined three functions: the PP_START performance is surveyed beginning; The PP_STOP performance is visited and is finished; PP_SAMPLE unitary sampling incident.Interruptable controller is notified to device driver to the system information that the performance detector detects, driver is transferred to performance to the employing information of obtaining by interruptable controller and surveys controller, do as one likes can be surveyed controller these information are submitted to policy management module, carries out the adaptive decision-making of power management.
(5) sequence number 25: the self-adaptive processing module is to realize the key modules of resource system control, and is closely related with resource.Expense and dirigibility the present invention of considering system design at the corresponding self-adaptive processing module of a certain class resource (application program), carry out self-adaptive processing with certain class resource to system, as CPU self-adaptive processing module, network self-adapting processing module that the network related resource is handled etc.It is made up of two modules: self-adaptive processing task and observation mission.The self-adaptive processing task is as the controller of goal task, and the state of the observed goal task of task is according to the observation determined to distribute to the resource quantity of goal task by control algolithm.
Accompanying drawing 3 has been showed the structural drawing of self-adaptive processing module, obtains current available power information by the self-adaptive processing module, and makes corresponding adaptation and handle the ruuning situation of application and current available power are adapted.
(1) step 31: the state of goal task is the inner parameter of goal task, mainly reflects the behavioral characteristics of resource consumption in the goal task.
(2) step 32: the self-adaptive processing task is as the controller of goal task, and the state of the observed goal task of task is according to the observation determined to distribute to the resource quantity of goal task by control algolithm.
(3) step 33: observation mission mainly is the real-time status of object observing task.
Accompanying drawing 4 has been showed embedded system self-adaptation dynamic power management algorithm, this algorithm is based on the DPM algorithm (DLDPMA) of device level, but it is at the characteristics of the randomness of the extraneous services request of embedded system and the dynamic change of environmental resource (such as the of short duration signal attenuation of unlimited communication network and the variation of powered battery energy), utilize the processor historical information of free time, with and the relevant information of state in the future, the free time of adaptive prediction processor, and along with mode switch is carried out adaptive adjustment to the operating voltage of processor, thereby improved the utilization factor of processor and saved system power dissipation.
Control algolithm among the present invention mainly is a kind of like this design philosophy, is exactly in the self-adaptation dynamic power management system, and the goal task that algorithm is realized is asked certain resource and carried out sequence of operations according to input, to produce input information.If requested resource is met, claim that this request is received request, task successfully obtains this resource; If request temporarily can not be met, claim that then this request is unaccredited request, task or continuation execution under the situation that does not need this resource, or be in waiting status, can use up to resource.This algorithm has been realized in the cpu resource of multitask distributes, can guarantee the complete work that task is executed the task under the more rich situation of resource, also satisfy situation the closing time that can when inadequate resource, improve task, thereby obtain more stable system performance.
The main operation that realizes mainly contains:
(1) start the multitask embedded system and use after, by using a resource situation that needs (comprise to the demand of CPU, peripherals etc.) is sent to the power management task;
(2) the power management task is carried out frequency reducing or raising frequency processing to CPU, and unwanted equipment is reduced to the lowest power consumption pattern, and the equipment that needs is raised to normal mode of operation;
(3) obtain current available power information by the self-adaptive processing module of system, and make the ruuning situation of application and current available power adapt as corresponding self-adaptive processing.
If have N periodically independently can seize real-time task in the embedded system, the note task-set is TK={TKi|1≤i≤N}, task TKi tlv triple (T
i, C
i, D
i) expression, wherein, T
iBe the performance period, C
iBe the poorest execution time of task, D
iBe the time limit of expression task, T is arranged
i=D
iProcessor M scalable frequency { f arranged
1, f
2... f
M, task is with frequency f
kMove with speed k level with regard to the title task during operation.Other the poorest execution time of arbitrary runtime class and the energy consumption of supposing all tasks all are known, t
I, jExecution time when expression task TKi moves with the j level, e
I, jPower consumption when expression task TKi moves with the j level, vi shows repayment or the utilization factor of task TKi.The performance of system is weighed by the overall performance of all tasks of system's operation, and the target of algorithm is to find such task subclass
Make the performance of system
Reach maximization.
All task i that definition S concentrates are with s
iThe level operation, si ∈ 1,2, ..., M}.And the task of total system must satisfy 2 points: (1) time-constrain.Arbitrary carrying out of task that is scheduled all must be finished before deadline D; (2) energy constraint.The energy that operation task consumed can not surpass available gross energy E in the system
Max
Like this, S set and runtime class s
i, be exactly
s
i∈ 1,2 ..., M}, satisfy:
With
And make,
Obtain peaked S and s
iObviously this is the difficult problem of a NP-, and problem can be converted into the policing issue of seeking a near optimal.
For this class problem, former algorithm is always supposed and can be carried out task scheduling continuously always, does not consider the constraint of task time and system energy consumption, and like this, usage factor of system resource is not high.And the managing power consumption algorithm must consider to come because of the dynamic change of system resource the dispatching requirement of Adjustment System task, and the performance that makes system reaches maximization under Limited resources is supplied with.
Self-adaptation dynamic power management power consumption algorithm of the present invention is the variation according to system resource and services request, operates by add, task executions speed is adjusted in deletion etc. and upgrades the task that Real-Time Scheduling is concentrated dynamically, chooses optimal scheduling.Defined parameters is as follows: supposition
, j ∈ 1,2 ..., M}, v (i) expression system's return rate or task utilization factor; T (i, j) expression task working time; E (i, j) energy requirement of expression task; Two boolean's sequence: selected (i) and considered (i) represent the state storage information of scheduler task; Two variablees: the execution time that task scheduling is total
The scheduling power consumed is
With the set of V storage current task dispatching concentration task utilization rate,
SV stores the maximal value of current V, the highest scheduling of expression task utilization factor; Speed (i) represents the set of all task speed.
Algorithm steps is described:
The initialization system parameter.When initial, the scheduling queue of putting system is that (selected (i)=false) does not have scheduler task (considered (i)=false), energy=0, time=0, SV=0, V=0 to sky.A certain scheduler task for scheduler task is concentrated if satisfy the condition and the SV＜V of time-constrain, so, writes down this scheduling, and puts SV=V.
For the task-set in a certain scheduling, change in resources according to system, for the realization system does not only influence the scheduling of system and the maximizing performance utilization of the system of assurance under Limited resources is supplied with, want the task of real-time update dispatching concentration, increase can realize the task scheduling of system function optimization, influence the task of system performance in the deletion task-set, adjust the operating voltage or the frequency of processor as required, regulate task executions speed.Be mainly concerned with four operations in the realization of algorithm: preserve optimal scheduling, interpolation task, deletion task and governing speed, and can not be added to scheduler task again to the task of having deleted and concentrate.
Preserve optimal scheduling.Make among the SV that the task scheduling of V maximum is optimal scheduling.Preserve current optimal scheduling.Carry out the interpolation task then.
The interpolation task.Can add the task that current scheduling is concentrated to, generally be that the execution time is short and energy consumption is lower.The condition that must satisfy: (1) this task be not scheduled the failure or deleted; (2) current scheduling is schedulable; When (3) current scheduling is carried out with the speed of minimum, the energy maximal value E that energy consumption budget energy+e (i, the 1) system of being no more than can provide
Max(4) in satisfying the task of above condition, scheduling can make
Obtain peaked task; After finishing the interpolation task, forward the preservation optimal scheduling to.If not satisfying condition of task can be added, forward governing speed to.
Promote the travelling speed of scheduler task.Task is added and to be finished, if when not had task can be dispatched to current scheduling to concentrate, but improves the travelling speed of scheduler task, and task is finished, in order to the new scheduling of beginning.The executive condition that task satisfies is: must be that current scheduling is concentrated schedulable task; Can not move (s with peak value
i≠ M); Energy+e (i, s
i+ 1)-e (i, s
i)＜E
MaxUnit interval internal state conversion is least in power-consuming (promptly
The ratio minimum), then, forward the preservation optimal scheduling to, if not satisfying condition of task forwards the deletion task to.
The deletion task.If first two steps operation (interpolation task and raising task speed) not success, then this task of deletion in the task queue of operation.Then, forward step 2 operation to.Be performed the condition that the task of deletion action should satisfy: be the task that current scheduling is concentrated; In all scheduler tasks of having added, make
The task of ratio minimum.
Claims (4)
1. the energy consumption management method of an embedded system is realized by the dynamic power management support module, it is characterized in that described dynamic power management support module comprises policy management module, idle detecting module, performance detecting module;
Described policy management module is carried out effective power management in system layer, thereby the performance and the Power Cutback of the system of realization, the idle condition of its attention location system and execution load, according to the adaptive adjustment device free time of system task executing state, make system enter corresponding low power supply status when lasting till to a certain degree at one's leisure, adopted a kind of dynamic power management adaptive control algorithm, according to the dynamic change of system's external resource the CPU supply voltage and the frequency of Adjustment System adaptively, make the energy consumption minimum of system in the time of the performance of maximization system, described dynamic power management adaptive control algorithm, it is variation according to system resource and services request, by adding, deletion is adjusted task executions speed and is upgraded the task that Real-Time Scheduling is concentrated dynamically, choose optimal scheduling, the implementation procedure of dynamic power management adaptive control algorithm is as follows:
Make having N periodically independently can seize real-time task in the embedded system, the note task-set is TK={TKi|1≤i≤N}, task TKi tlv triple (T
i, C
i, D
i) expression, wherein, T
iBe the performance period, C
iBe the poorest execution time of task, D
iBe the time limit of expression task, T is arranged
i=D
iProcessor has M scalable frequency, is designated as { f
1, f
2... f
M, task is with frequency f
kMove with speed k level with regard to the title task during operation; It all is known making other the poorest execution time of arbitrary runtime class and the energy consumption of all tasks, t
I, jExecution time when expression task TKi moves with the j level, e
I, jPower consumption when expression task TKi moves with the j level, Vi represents repayment or the utilization factor of task TKi, and the performance of system is weighed by the overall performance of all tasks of system's operation, and the target of algorithm is to find such task subclass
Make the performance of system
Reach maximization;
All task i that definition S concentrates move with the Si level, si ∈ 1,2, ..., M}, and the task of total system, must satisfy 2 points:
(1) time-constrain, arbitrary carrying out of task that is scheduled all must be finished before deadline D;
(2) energy constraint, the energy that operation task consumed can not surpass available gross energy E in the system
MaxLike this, S set and runtime class Si are exactly
Si ∈ 1,2 ..., M}, satisfy:
With
And make,
Obtain peaked S and Si;
Order
J ∈ 1,2 ..., M}, v (i) expression system's return rate or task utilization factor; T (i, j) expression task working time; E (i, j) energy requirement of expression task; Selected (i) and considered (i) represent the state storage information of scheduler task; The execution time that task scheduling is total
The scheduling power consumed is
The set of V storage current task dispatching concentration task utilization rate,
SV stores the maximal value of current V, the highest scheduling of expression task utilization factor; Speed (i) represents the set of all task speed;
First step, the initialization system parameter:
The scheduling queue of putting system is empty selected (i)=false, does not have scheduler task considered (i)=false, energy=0, time=0, SV=0, V=0; A certain scheduler task for scheduler task is concentrated if satisfy the condition and the SV＜V of time-constrain, so, writes down this scheduling, and puts SV=V; For the task-set in a certain scheduling, change in resources according to system, for the realization system does not only influence the scheduling of system and the maximizing performance utilization of the system of assurance under Limited resources is supplied with, want the task of real-time update dispatching concentration, increase can realize the task scheduling of system function optimization, influence the task of system performance in the deletion task-set, adjust the operating voltage or the frequency of processor, regulate task executions speed;
Second step, preserve optimal scheduling:
Make among the SV that the task scheduling of V maximum is optimal scheduling, judge whether optimal scheduling of current scheduling, if, then preserve optimal scheduling, carry out third step then;
Third step, the interpolation task:
The task that current scheduling is concentrated can be added to, following condition must be satisfied:
(1) this task be not scheduled the failure or deleted;
(2) current scheduling is schedulable;
When (3) current scheduling is carried out with the speed of minimum, the energy maximal value E that energy consumption budget energy+e (i, the 1) system of being no more than can provide
Max
(4) in satisfying the task of above condition, scheduling can make
Obtain peaked task;
If the task of satisfying above-mentioned condition is arranged, then add current scheduling to and concentrate and repeat second step;
If do not satisfy the task of above-mentioned condition, then carry out the 4th step;
The 4th step, the travelling speed of lifting scheduler task:
Task is added and to be finished, if when not had task can be dispatched to current scheduling to concentrate, but improves the travelling speed of scheduler task, and task is finished, in order to the new scheduling of beginning;
The executive condition that task satisfies is: must be that current scheduling is concentrated schedulable task; Can not move with peak value, i.e. s
i≠ M; Energy+e (i, s
i+ 1)-e (i, s
i)＜E
MaxUnit interval internal state conversion is least in power-consuming, promptly
The ratio minimum,
If the task of satisfying above-mentioned condition is arranged, then improve the travelling speed of satisfying condition of task, re-execute second step then;
If do not satisfy the task of above-mentioned condition, then carry out the 5th step;
The 5th step, the deletion task:
If carry out the task of failure in third and fourth step, then this task of deletion in the task queue of operation then, forwards the operation of second step to;
Described idle detecting module provides the idle information of CPU and operating system, its monitor idle thread or the activity of given operating load in operating system, and the parameter information of making a strategic decision is provided to policy management module;
Described performance detecting module monitors the operating position of cpu busy percentage and main memory, for policy management module provides accurate execution load information.
2. the energy consumption management method of embedded system according to claim 1, it is characterized in that described dynamic power management support module also comprises Subscriber Interface Module SIM, described Subscriber Interface Module SIM is between application layer and policy management module, and the interface that allows user's input parameter control power management policy is provided.
3. the energy consumption management method of embedded system according to claim 1, it is characterized in that described dynamic power management support module also comprises application interface module, in order to the device registration application program, registered the ID that comprises equipment, be used for to specific equipment carry out power managed processing function, equipment frequency of operation and to the sensitive information of state.
4. the energy consumption management method of embedded system according to claim 1, it is characterized in that described dynamic power management support module also increases the global variable of dynamic power management in the core code of embedded system, described global variable has defined five kinds of parameters of operating system, whether the energy consumption that the state, processor core voltage, processor core frequency, system bus frequency and the memory bus frequency that comprise processor, the parameter by this global variable are provided with the current system of expression is in optimum state.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200710028578A CN101067758B (en) | 2007-06-14 | 2007-06-14 | Energy consumption management method for inserting system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200710028578A CN101067758B (en) | 2007-06-14 | 2007-06-14 | Energy consumption management method for inserting system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101067758A CN101067758A (en) | 2007-11-07 |
| CN101067758B true CN101067758B (en) | 2010-05-19 |
Family
ID=38880330
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200710028578A Expired - Fee Related CN101067758B (en) | 2007-06-14 | 2007-06-14 | Energy consumption management method for inserting system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101067758B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011150792A1 (en) * | 2010-11-29 | 2011-12-08 | 华为技术有限公司 | Power saving realization method and device for cpu |
| CN103345298A (en) * | 2013-07-16 | 2013-10-09 | 山东省计算中心 | Data center energy saving system and method based on virtual IT resource distribution technology |
Families Citing this family (34)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101344812B (en) * | 2008-08-19 | 2010-06-02 | 华南理工大学 | Magnetic disk dynamic power supply management method based on embedded system |
| CN101727171B (en) * | 2008-10-14 | 2011-10-19 | 上海摩波彼克半导体有限公司 | Method for reducing CPU power consumption in embedded system |
| CN101727172B (en) * | 2008-10-27 | 2012-12-19 | 联想(北京)有限公司 | Method and device for measuring process consumption of computer, and computer system |
| US8607075B2 (en) * | 2008-12-31 | 2013-12-10 | Intel Corporation | Idle duration reporting for power management |
| CN101477403B (en) * | 2009-01-22 | 2011-10-05 | 浪潮电子信息产业股份有限公司 | Automatic control method for system power consumption |
| CN101881994A (en) * | 2009-05-06 | 2010-11-10 | 鸿富锦精密工业（深圳）有限公司 | Energy-saving control system and control method |
| EP2388700A3 (en) * | 2010-05-18 | 2013-08-07 | Kaspersky Lab Zao | Systems and methods for policy-based program configuration |
| CN101943944B (en) * | 2010-09-15 | 2012-07-04 | 中国人民解放军国防科学技术大学 | Idle history information-based energy-saving method for computing array |
| TWI493332B (en) | 2011-11-29 | 2015-07-21 | Intel Corp | Method and apparatus with power management and a platform and computer readable storage medium thereof |
| CN103092744B (en) * | 2012-05-09 | 2015-06-03 | 华南理工大学 | Dynamic power management strategy performance assessment method |
| CN103324268A (en) * | 2013-05-29 | 2013-09-25 | 东南大学 | Low-power design method for wireless sensor network core chip |
| CN104866379B (en) * | 2014-02-24 | 2019-06-14 | 中兴通讯股份有限公司 | A kind of multi-core processor dispatching method, device and terminal |
| CN103955263B (en) * | 2014-05-16 | 2017-07-28 | 华为技术有限公司 | ARM method for managing power supply and processor |
| US9541988B2 (en) * | 2014-09-22 | 2017-01-10 | Western Digital Technologies, Inc. | Data storage devices with performance-aware power capping |
| CN105589544A (en) * | 2014-10-20 | 2016-05-18 | 中兴通讯股份有限公司 | Method, device and terminal reducing central processor energy consumption |
| CN106708238B (en) * | 2015-07-24 | 2019-11-29 | 浪潮电子信息产业股份有限公司 | It is a kind of for there are the energy consumption management methods that the high-performance calculation of a large amount of iteration is applied |
| KR102375925B1 (en) * | 2015-08-31 | 2022-03-17 | 삼성전자주식회사 | Method of operating cpu and method of operating system having the cpu |
| KR20180031322A (en) * | 2016-09-20 | 2018-03-28 | 한화테크윈 주식회사 | In-line idle production time estimation apparatus |
| US20180262991A1 (en) * | 2017-03-10 | 2018-09-13 | Futurewei Technologies, Inc. | Optimization of energy management of mobile devices based on specific user and device metrics uploaded to cloud |
| CN108108237B (en) * | 2017-12-27 | 2021-09-28 | 电子科技大学 | MILP-based heterogeneous multi-core mapping scheduling method for periodic associated tasks |
| WO2019153187A1 (en) * | 2018-02-08 | 2019-08-15 | Alibaba Group Holding Limited | Hybrid system-on-chip for power and performance prediction and control |
| CN108845873A (en) * | 2018-06-22 | 2018-11-20 | 复旦大学 | A kind of self adaptation stream waterline parallel method of low-power consumption |
| WO2020102929A1 (en) * | 2018-11-19 | 2020-05-28 | Alibaba Group Holding Limited | Unified power management |
| CN110320813B (en) * | 2019-07-29 | 2022-07-19 | 青岛海尔科技有限公司 | Power management method and device for Internet of things equipment |
| CN110940947B (en) * | 2019-12-19 | 2022-04-22 | 国网宁夏电力有限公司检修公司 | Self-adaptive ultra-long working time method of handheld polarity testing device |
| CN111124096B (en) * | 2019-12-27 | 2021-08-17 | 联想(北京)有限公司 | Data processing method, device, computer system and medium |
| CN111541451B (en) * | 2020-06-23 | 2021-10-29 | 深圳比特微电子科技有限公司 | Method and clock circuit for up-converting a clock signal |
| CN112399536A (en) * | 2020-11-26 | 2021-02-23 | 深圳北斗通信科技有限公司 | Gun positioning terminal power consumption management method and system and storage medium |
| CN113238648B (en) * | 2021-05-11 | 2023-05-09 | 成都海光集成电路设计有限公司 | Power consumption adjusting method and device |
| CN114003121B (en) * | 2021-09-30 | 2023-10-31 | 中国科学院计算技术研究所 | Data center server energy efficiency optimization method and device, electronic equipment and storage medium |
| CN114095970A (en) * | 2021-12-29 | 2022-02-25 | 诺领科技(南京)有限公司 | Method and system for reducing load peak value of cellular network/server |
| CN114138095B (en) * | 2022-01-29 | 2022-05-10 | 阿里巴巴（中国）有限公司 | Power consumption processing method and device for internet data center IDC and readable medium |
| CN114578950B (en) * | 2022-03-07 | 2023-11-07 | 成都九华圆通科技发展有限公司 | Low-power consumption management control system and method based on miniature electromagnetic sensing equipment |
| CN115185211B (en) * | 2022-06-17 | 2023-08-25 | 中北大学 | Dynamic power management system and method suitable for low-temperature vibration sensor |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1459692A (en) * | 2002-05-21 | 2003-12-03 | 联想(北京)有限公司 | Method of CPU frequency controlling based on the insertion system of computer platform |
| CN1508683A (en) * | 2002-12-17 | 2004-06-30 | 深圳市中兴通讯股份有限公司 | Method for detecting occupied rate of CPU in multi-task real-time operation system |
| CN1752896A (en) * | 2005-11-08 | 2006-03-29 | 北京大学 | Power source management method of embedded equipment under operation system cooperation and its system |
| CN1828481A (en) * | 2006-03-10 | 2006-09-06 | 浙江大学 | Energy-saving switching method for waiting overtime judge of flush type system outer apparatus |
-
2007
- 2007-06-14 CN CN200710028578A patent/CN101067758B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1459692A (en) * | 2002-05-21 | 2003-12-03 | 联想(北京)有限公司 | Method of CPU frequency controlling based on the insertion system of computer platform |
| CN1508683A (en) * | 2002-12-17 | 2004-06-30 | 深圳市中兴通讯股份有限公司 | Method for detecting occupied rate of CPU in multi-task real-time operation system |
| CN1752896A (en) * | 2005-11-08 | 2006-03-29 | 北京大学 | Power source management method of embedded equipment under operation system cooperation and its system |
| CN1828481A (en) * | 2006-03-10 | 2006-09-06 | 浙江大学 | Energy-saving switching method for waiting overtime judge of flush type system outer apparatus |
Non-Patent Citations (1)
| Title |
|---|
| 同上. |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011150792A1 (en) * | 2010-11-29 | 2011-12-08 | 华为技术有限公司 | Power saving realization method and device for cpu |
| CN103345298A (en) * | 2013-07-16 | 2013-10-09 | 山东省计算中心 | Data center energy saving system and method based on virtual IT resource distribution technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101067758A (en) | 2007-11-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101067758B (en) | Energy consumption management method for inserting system | |
| CN100407104C (en) | Power source management method of embedded equipment under operation system cooperation and its system | |
| US9026814B2 (en) | Power and load management based on contextual information | |
| CN101821997B (en) | Hierarchical reservation resource scheduling infrastructure | |
| CN102567106B (en) | Task scheduling method, system and device | |
| Goiri et al. | Matching renewable energy supply and demand in green datacenters | |
| CN102521021B (en) | Green via flow of event management calculates | |
| CN104239152B (en) | Method and apparatus for improving the turbine accelerating ability of event handling | |
| US8225119B2 (en) | Energy-aware server management | |
| US9391922B2 (en) | Computer workload redistribution schedule | |
| CN103561092A (en) | Method and device for managing resources under private cloud environment | |
| Buttazzo | Achieving scalability in real-time systems | |
| US20240004707A1 (en) | Methods and systems for energy-efficient scheduling of periodic tasks on a group of processing devices | |
| Wu et al. | CPU scheduling for statistically-assured real-time performance and improved energy efficiency | |
| CN106802822A (en) | A kind of cloud data center cognitive resources dispatching method based on moth algorithm | |
| Kong et al. | Guaranteed scheduling for (m, k)-firm deadline-constrained real-time tasks on multiprocessors | |
| Gelernter et al. | Piranha scheduling: Strategies and their implementation | |
| Chetto | Dynamic power management for fixed priority real-time systems with regenerative energy | |
| CN113590324A (en) | Heuristic task scheduling method and system for cloud edge-side collaborative computing | |
| Carriero et al. | Piranha scheduling: Strategies and their implementation | |
| Sun et al. | Power-aware virtual machine placement for mobile edge computing | |
| CN115454640B (en) | Task processing system and self-adaptive task scheduling method | |
| Masson et al. | RTSJ extensions: event manager and feasibility analyzer | |
| Chetto et al. | Main Results on Earliest Deadline First Scheduling for Energy Neutral Sensors | |
| Xavier et al. | Integrating Real-Time Main Controller with Power Prediction Interface and State Classifier in Energy Harvesting Systems: An Approach |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20100519 Termination date: 20120614 |