CN105808355A - Binary linear regression equation-based dynamic frequency modulation method - Google Patents
Binary linear regression equation-based dynamic frequency modulation method Download PDFInfo
- Publication number
- CN105808355A CN105808355A CN201610139773.7A CN201610139773A CN105808355A CN 105808355 A CN105808355 A CN 105808355A CN 201610139773 A CN201610139773 A CN 201610139773A CN 105808355 A CN105808355 A CN 105808355A
- Authority
- CN
- China
- Prior art keywords
- linear regression
- regression equation
- binary linear
- cpu
- linux
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/26—Power supply means, e.g. regulation thereof
- G06F1/32—Means for saving power
- G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
- G06F1/3234—Power saving characterised by the action undertaken
- G06F1/329—Power saving characterised by the action undertaken by task scheduling
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Debugging And Monitoring (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a binary linear regression equation-based dynamic frequency modulation method. The method comprises the following steps: selecting a binary linear regression equation as a system load prediction algorithm and applying the binary linear regression equation in a CPUFreq module in a Linux operation system; taking a CPU utilization rate and system executable progress number information of a system as two independent variables of the binary linear regression equation and solving a fitted binary linear regression equation by taking obtained system load information as a dependent variable. According to the equation, the independent variables are predicted through the dependent variable. According to the method, the correct prediction of the system load information is realized so that the CPU operation frequency can be set in time according to the load information; the response speed of system frequency setting is improved and the transition time when the system load achieves a steady state is eliminated; meanwhile, the CPU operation frequency is correctly set, so that the problems of performance deficiency and performance surplus of the system can be avoided and then the aims of giving full play to the system performance and reducing the system power consumption are achieved.
Description
Technical field
The invention belongs to Computer Applied Technology field, particularly relate to a kind of dynamic frequency method based on binary linear regression equation.
Background technology
Dynamic frequency modulation technology is the focus of computer power consumption area research, is also computer function module important component part.Dynamic frequency modulation technology is while ensureing computer process ability, it is possible to effectively reduce the power consumption of system.The quality of dynamic frequency modulation technology, will directly affect the power consumption of computer system.Computer system is in normal course of operation, and system load can constantly change.Such as, the requirement of cpu performance is higher than when carrying out high-performance calculation when listening music.The performance of CPU and the running frequency linear correlation of CPU, the more high calculated performance of running frequency is more strong.Otherwise, the more low calculated performance of running frequency is more weak.When the CPU running frequency of system and system load are not mated, arise that the problem that performance is not enough or performance is superfluous, and then cause unnecessary power wastage.When system is in high capacity and CPU is in low-frequency operation, computer system arises that the phenomenons such as program is run slowly, operation is not smooth.When CPU is in high frequency operation when system is in low-load, arises that and the phenomenon that CPU calculated performance is superfluous cause system CPU to cause that energy consumption is wasted because flogging a dead horse more.Therefore, system load information is accurately, obtain timely, it is ensured that the accurate setting of CPU running frequency, while ensureing computer CPU computing capability, effectively eliminates unnecessary power wastage, reduces the purpose of system power dissipation.The promptness that the accuracy of system load information prediction and CPU running frequency are arranged, by the reducing effect of decision systems power consumption.Core concept based on the AIBDDVS algorithm of timeslice: adopt the system CPU utilization rate of the system CPU usage forecast subsequent time in system front four moment.Front four timeslice cpu busy percentages are designated as respectively: p1、p2、p3、p4, and take the cpu busy percentage weights of front four timeslices respectively 0.1,0.2,0.3 and 0.4.The formula of the cpu busy percentage p of prediction current system is as follows:
P=0.1*p1+0.2*p2+0.3*p3+0.4*p4；
If p=max, then set CPU and perform maximum running frequency；If p=min, then set CPU and currently perform minimum operation frequency；Otherwise being compared with threshold value by p, if p>0.8, then CPU performs one class of frequency upgrading, if p<0.2, then CPU performs frequency one class of decline.During the cpu busy percentage of AIBD algorithm predicts system, rely on the system CPU utilization rate information of front four time points, the cpu busy percentage of cause four time points before this is the measured value of system itself, and its percentage ratio is 100%, is initially the cpu busy percentage of AIBD algorithm predicts system from the 5th time point.AIBD algorithm accuracy in prognoses system cpu busy percentage is relatively low, sometimes 94.00% can be reached, sometimes only have 47.10%, sometimes may be up to again 184.00%, therefore adopt the prediction algorithm of unitary variant, in the accuracy of prognoses system load, there is very big unstability, want the Accurate Prediction realizing system load, need to select the parameter of more influential system load, the load of system is predicted.
Summary of the invention
It is an object of the invention to provide a kind of dynamic frequency method based on binary linear regression equation, it is intended to the problem such as length system load prediction accuracy is low, system load arrives steady statue transit time solving that the AIBD prediction algorithm of ondemand strategy that existing algorithm such as (SuSE) Linux OS carries and the single independent variable that Kang Qing proposes exists.
The present invention is namely based on the dynamic frequency method of binary linear regression equation, the described dynamic frequency method choice binary linear regression equation based on binary linear regression equation is as the prediction algorithm of system load, and is applied in the CPUFreq module of (SuSE) Linux OS；Using the system CPU utilization rate obtained and system can executive process number as binary linear regression equation y=a1x1+a2x2Two independent variable x of+b1、x2, using the load information of system as dependent variable y.By dependent variable independent variable solved the load information doping processor accurately and in time in the system of subsequent time, and the running frequency of CPU can be set.
Further, the described dynamic frequency method based on binary linear regression equation comprises the following steps:
The process sequence of (SuSE) Linux OS is sampled, obtain ready queue can executive process number, gather the cpu busy percentage information of the system of synchronization simultaneously；
The cpu busy percentage of system and can executive process number as the independent variable of binary linear regression equation, it was predicted that system load value as the dependent variable of binary linear regression equation；
By independent variable information, dependent variable is predicted, accurately provides the system load information of subsequent time, and then the running frequency of CPU is set.
Further, utilize Regress Forecast algorithm, using cpu busy percentage and system can executive process number as the independent variable x of model1And x2, utilize Regress Forecast equation y=a1x1+a2x2+ b, is predicted the cpu load information of subsequent time.Wherein x1Represent the cpu busy percentage under (SuSE) Linux OS, x2Representing the process the run number of system under (SuSE) Linux OS, y represents the prediction load information of system under (SuSE) Linux OS；x1=us+sy, us represent the percent of operating system cpu busy percentage shared by (SuSE) Linux OS User space, and sy represents the percent of cpu busy percentage shared by operating system system space, can run process number runningtask and parameter x in binary linear regression equation2Corresponding, system load information loadaverage is corresponding with parameter y in binary linear regression equation.
Further, binary linear regression equation is: y=0.026083x1+0.63251x2-0.62255。
Further, the described process sequence to (SuSE) Linux OS is sampled, obtain ready queue can executive process number, the cpu busy percentage information of the system simultaneously gathering synchronization farther includes:
The process sequence of (SuSE) Linux OS is sampled, obtain ready queue can executive process number runningtask, gather cpu busy percentage information us, the sy of the system of synchronization simultaneously, and obtain system load information loadaverage；Wherein, us represents the percent of operating system cpu busy percentage shared by (SuSE) Linux OS User space, and sy represents the percent of cpu busy percentage shared by operating system system space；The interval of sampled point is set to every 10ms and gathers once；For the collection of CPU information, the CPU related data information collected is taked not conversation strategy immediately, first the data message collected is stored in relief area, when the data of relief area are full, data are intensively exported preservation.
Further, the cpu busy percentage of described system and can executive process number as the independent variable of binary linear regression equation, it is known that system load value farther include as the dependent variable of binary linear regression equation:
Utilize Regress Forecast algorithm, using cpu busy percentage and system can executive process number as model and Regress Forecast equation y=a1x1+a2x2The independent variable x of+b1And x2.Wherein x1Represent the cpu busy percentage under (SuSE) Linux OS, x2Representing the process the run number of system under (SuSE) Linux OS, y represents the prediction load information of system under (SuSE) Linux OS；Cpu busy percentage=the us+sy of operating system；By the data obtained solving equation, obtain the coefficient a of binary regression predictive equation1、a2And the value of b.
Further, described according to gained equation, by independent variable information, dependent variable is predicted, accurately provides the system load information of subsequent time, and then the running frequency arranging CPU farther includes:
According to equation y=a1x1+a2x2+ b, when known subsequent time cpu busy percentage and system can executive process number, measurable go out subsequent time system load the value of information, and then the running frequency of CPU is set.
Dynamic frequency method based on binary linear regression equation provided by the invention, arranges the stage in system load, directly arranges the load value of system, saves system load and arrives the transient process of steady statue, reaches to realize the purpose that system power dissipation reduces；In (SuSE) Linux OS CPUFreq dynamic frequency submodule, the change of system load has a buffering course, after a period of time, can be only achieved the steady load of system.In transient process, the system load of different time sections, corresponding to different CPU running frequencies, in order to reach to shorten system load transient process, reduce the purpose of system power dissipation, the present invention, on the basis researched and analysed that common prediction algorithm is carried out, selects binary linear regression equation as the prediction algorithm of system load, and applies it in the CPUFreq module in (SuSE) Linux OS.The present invention with the cpu busy percentage of system and system can executive process number information, as two independent variables of binary linear regression equation, simultaneously using the load information of system as dependent variable.By independent variable, dependent variable is predicted, it is possible to obtain the load information of system accurately and in time.The AIBD prediction algorithm of the single independent variable that the ondemand strategy carried compared to (SuSE) Linux OS and Kang Qing propose, Regress Forecast algorithm is provided with the load information of system in time while realizing the Accurate Prediction of linux system load, save system load and arrive the transit time of steady statue, achieve the reduction of system power dissipation at software level.The present invention proposes a kind of dynamic frequency method based on binary linear regression equation, compared to the ondemand chirping strategies that AIBDDVS algorithm and linux system carry, the method can not only the load information of prognoses system exactly, moreover it is possible to arrange the load information of system in time；By the cpu busy percentage of (SuSE) Linux OS and system can executive process number information, binary linear regression equation can realize system load information ground Accurate Prediction, directly obtain the load information of (SuSE) Linux OS, the running frequency of CPU is set in time according to load information, improve the response speed that system frequency sets, save system load simultaneously and reach the transit time of steady statue；The running frequency of CPU is accurately set, it is possible to avoids system the not enough problem superfluous with performance of performance occur, reaches give full play to systematic function and reduce the purpose of system power dissipation.
Accompanying drawing explanation
Fig. 1 is the dynamic frequency method flow diagram based on binary linear regression equation that the embodiment of the present invention provides.
Fig. 2 is that the CPU running frequency that the embodiment of the present invention provides arranges frame diagram.
Fig. 3 is the fitting a straight line schematic diagram of first equation of linear regression that the embodiment of the present invention provides.
Fig. 4 is prediction load and the actual loading change curve comparison diagram of embodiment of the present invention offer.
Detailed description of the invention
In order to make the purpose of the present invention, technical scheme and advantage clearly understand, below in conjunction with embodiment, the present invention is further elaborated.Should be appreciated that specific embodiment described herein is only in order to explain the present invention, is not intended to limit the present invention.
The problem that the Regress Forecast algorithm of the present invention mainly considers two aspects: the promptness that the accuracy of load information prediction, load information are arranged.Wherein, the prediction of load information is the premise that load information is arranged, and it is related to the accuracy of load information prediction, will directly influence the correctness of CPU running frequency.In order to obtain system load information predictive value accurately, binary linear regression algorithm is combined by the present invention with the dynamic frequency strategy of (SuSE) Linux OS, it is ensured that the accuracy that system load information obtains.
Below in conjunction with accompanying drawing, the application principle of the present invention is explained in detail.
As it is shown in figure 1, the dynamic frequency method based on binary linear regression equation of the embodiment of the present invention comprises the following steps:
S101: the process sequence of (SuSE) Linux OS is sampled, obtain ready queue can executive process number, gather the cpu busy percentage information of the system of synchronization simultaneously；
S102: the cpu busy percentage of system and can executive process number as the independent variable of binary linear regression equation, system load value is as the dependent variable of binary linear regression equation；
S103: dependent variable is predicted by independent variable information, is accurately provided the system load information of subsequent time, and then arranges the running frequency of CPU.
The basic thought of the dynamic frequency method based on binary linear regression equation of the present invention: the process sequence of (SuSE) Linux OS is sampled, what obtain ready queue can executive process number, gather the cpu busy percentage information of the system of synchronization simultaneously, the cpu busy percentage of system and can executive process number as the independent variable of binary linear regression equation, system load value is as the dependent variable of binary linear regression equation, by independent variable information, dependent variable is predicted, accurately provide the system load information of subsequent time, and then the running frequency of CPU is set.Its flow process framework such as Fig. 2:
Under (SuSE) Linux OS, the frequency modulation of CPU dynamic frequency strategy is according to the resource utilisation information being system, and the factor affecting cpu load information is the current executable process number of system.The embodiment of the another one aspect of linux system resource utilisation information is the utilization rate of system CPU, and namely program run duration takies the percentage ratio of CPU.Realizing (SuSE) Linux OS dynamic frequency and reduce the purpose of power consumption, the acquisition of operating system CPU operation information is most important.Operating system cpu busy percentage and the executable process number of system can intuitively embody the load information of system.
By on the research affecting (SuSE) Linux OS load information factor, it has been found that cpu busy percentage and system can executive process number be two main factors of influential system load information.Utilize Regress Forecast algorithm, using cpu busy percentage and system can executive process number as the independent variable x of model1And x2, it is possible to subsequent time system load information is predicted exactly, directly obtains the load information of (SuSE) Linux OS.Using this load estimation value as the foundation of the running frequency arranging CPU, set the running frequency of CPU timely and accurately, it is possible to while obtaining CPU operational performance, also can reduce the power consumption of system.
Regress Forecast equation y=a1x1+a2x2+ b, x1Represent the cpu busy percentage under (SuSE) Linux OS, x2Representing the process the run number of system under (SuSE) Linux OS, y represents the load information of system under (SuSE) Linux OS.Data are set up about independent variable x by experiment1(cpu busy percentage) and independent variable x2Regress Forecast algorithm equation between (Linux can run process number) and dependent variable y (system load) information.The experimental data obtained based on Thinkcentrem8400t desktop computer is as shown in table 1:
Table 1 system load information tables of data
| Us | sy | running task | load average |
| 0.0 | 0.0 | 1 | 0.00 |
| 12.6 | 0.0 | 2 | 0.99 |
| 87.6 | 0.1 | 8 | 7.03 |
| 37.5 | 0.0 | 4 | 3.00 |
| 87.5 | 0.1 | 9 | 7.36 |
| 62.5 | 0.0 | 6 | 4.52 |
| 75.0 | 0.0 | 7 | 5.32 |
| 25.0 | 0.0 | 3 | 2.00 |
| 87.5 | 0.0 | 8 | 6.46 |
| 100 | 0.0 | 9 | 8.00 |
| 50.0 | 0.0 | 5 | 4.00 |
Us represents the percent of operating system cpu busy percentage shared by (SuSE) Linux OS User space, sy represents the percent of cpu busy percentage shared by operating system system space, runningtask represent in operating system system can executive process number, loadaverage represents the load information of system.Cpu busy percentage=the us+sy of operating system, this value and parameter x in binary linear regression equation1Corresponding, process number runningtask and parameter x in binary linear regression equation can be run2Corresponding, loadaverage is corresponding with parameter y in binary linear regression equation.Converse solved acquisition relevant parameter is carried out as shown in table 2 by MATLAB:
The Converse solved binary linear regression equation tables of data of table 2
That b represents is the parameter x of binary linear regression equation1、x2With the point estimate of the regression coefficient of b, the b numerical value obtaining binary linear regression equation is the matrix of three row string.Its value is :-0.62255,0.026083,0.63251 wherein at y=α1x1+α2x2In+b, corresponding b is-0.62255 respectively, α1It is 0.026083, α2It is 0.63251.
The data representation of bint is the interval estimation of binary linear regression equation regression coefficient, and the data representation of r is residual error, the confidence interval of the binary linear regression equation under this data acquisition system of the data representation of bint.The data of stats are for checking the accuracy of regression equation, and its data value has four, represent that correlation coefficient, F check Probability p value, the error variance that numerical value is corresponding with F respectively.
The value of stats: 0.99199,495.19,4.1226e-09,0.072721.It is 0.99199 close to 1 for the correlation coefficient of binary linear regression equation, regression equation very notable is described；F value is 495.19 simultaneously, is a very big number, and regression equation highly significant is also described；The value of p be 4.1226e-09 close to zero, also illustrate that this binary linear regression equation is significant.Can obtain binary linear regression equation from the data above is:
Y=0.026083x1+0.63251x2-0.62255
The fitting a straight line of binary linear regression equation is as shown in Figure 3.
Below in conjunction with experiment, the application effect of the present invention is explained in detail.
Experimental data mainly arranges the promptness of CPU running frequency after the accuracy of Regress Forecast algorithm predicts system load, algorithm predicts, and the two aspect illustrates the advantage that Regress Forecast algorithm is applied to dynamic frequency strategy.In order to embody the advantage of this algorithm intuitively, the ondemand dynamic frequency strategy carried by binary linear regression dynamic frequency algorithm homologous ray, the AIBDDVS algorithm based on timeslice that Kang Qing proposes contrasts.
Table 3 data are based on binary linear regression equation y=0.026083x1+0.63251x2The system load predictive value of-0.62255 and system ondemand chirping strategies are about the contrast of system load data.Can be drawn intuitively by curve, the accuracy rate of more than 95% is remain between system load experiment value under the system load predictive value of binary linear regression equation acquisition and (SuSE) Linux OS ondemand strategy, rejecting the impact of Individual testwas data point, binary linear regression equation maintains very high forecasting accuracy.Illustrating can after executive process number what obtain (SuSE) Linux OS cpu busy percentage information and system, and binary linear regression equation can the load of prognoses system exactly.
Table 3 system load predictive value and experiment value contrast table
Fig. 4 can be seen that under the ondemand strategy that (SuSE) Linux OS carries, the change of load needs the buffering of a period of time, gets to certain stable system load, and this process needs the time of 5 minutes.For the dynamic frequency strategy based on binary linear regression equation, then eliminate this transient process.This algorithm can not only the load information of prognoses system exactly, moreover it is possible to the load information of direct access systems, saves the transient process of system load change, thus reducing the purpose of system power dissipation.
The present invention proposes a kind of dynamic frequency method based on binary linear regression equation.Compared to the ondemand chirping strategies that AIBDDVS algorithm and linux system carry, the method can not only the load information of prognoses system exactly, moreover it is possible to arrange the load information of system in time.By the cpu busy percentage of (SuSE) Linux OS and system can executive process number information, binary linear regression equation can realize system load information ground Accurate Prediction, directly obtain the load information of (SuSE) Linux OS, the running frequency of CPU is set in time according to load information, improve the response speed that system frequency sets, save system load simultaneously and reach the transit time of steady statue.The running frequency of CPU is accurately set, it is possible to avoids system the not enough problem superfluous with performance of performance occur, reaches give full play to systematic function and reduce the purpose of system power dissipation.
The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all any amendment, equivalent replacement and improvement etc. made within the spirit and principles in the present invention, should be included within protection scope of the present invention.
Claims (7)
1. the dynamic frequency method based on binary linear regression equation, it is characterized in that, the described dynamic frequency method choice based on the binary linear regression equation binary linear regression equation prediction algorithm as system load, and it is applied to the CPUFreq module in (SuSE) Linux OS；Using the system CPU utilization rate obtained and system can executive process number as binary linear regression equation y=a1x1+a2x2Two independent variable x of+b1、x2, and using the load information of system as dependent variable y；By dependent variable independent variable solved the system load information doping processor accurately and in time at subsequent time, the running frequency of CPU is set in time.
2. the dynamic frequency method based on binary linear regression equation as claimed in claim 1, it is characterised in that the described dynamic frequency method based on binary linear regression equation comprises the following steps:
The process sequence of (SuSE) Linux OS is sampled, obtain ready queue can executive process number, gather the system CPU utilization rate information of synchronization simultaneously；
The cpu busy percentage of system and can executive process number as the independent variable of binary linear regression equation, it is known that system load value, as the dependent variable of binary linear regression equation, solves the binary linear regression equation of matching；
Utilize the binary linear regression equation tried to achieve, by independent variable information, dependent variable is predicted, accurately provides the load information of subsequent time system, and then the running frequency of CPU is set in time.
3. the dynamic frequency method based on binary linear regression equation as claimed in claim 2, it is characterised in that utilize Regress Forecast algorithm, using cpu busy percentage and system can executive process number as the independent variable x of model1And x2, Regress Forecast equation y=a1x1+a2x2In+b, x1Represent the cpu busy percentage under (SuSE) Linux OS, x2The system represented under (SuSE) Linux OS can run process number, and y represents the load information of system under (SuSE) Linux OS；Us, sy represent operating system cpu busy percentage shared by (SuSE) Linux OS User space and the percent of cpu busy percentage shared by operating system system space, the cpu busy percentage=us+sy of operating system, this value and parameter x in binary linear regression equation respectively1Corresponding, process number runningtask and parameter x in binary linear regression equation can be run2Corresponding, system load information loadaverage is corresponding with parameter y in binary linear regression equation.
4. the dynamic frequency method based on binary linear regression equation as claimed in claim 2, it is characterised in that binary linear regression equation is: y=0.026083x1+0.63251x2-0.62255。
5. the dynamic frequency method based on binary linear regression equation as claimed in claim 2, it is characterized in that, the described process sequence to (SuSE) Linux OS is sampled, and what obtain ready queue can executive process number, meanwhile, the system CPU utilization rate information gathering synchronization farther includes:
The process sequence of (SuSE) Linux OS is sampled, obtain ready queue can executive process number runningtask, meanwhile, gather cpu busy percentage information us, the sy of system of synchronization, and obtain system load information loadaverage；The interval of sampled point is set to every 10ms and gathers once.
6. as claimed in claim 2 based on binary linear regression equation dynamic frequency method, it is characterized in that, described system CPU utilization rate and can executive process number as the independent variable of binary linear regression equation, system load value farther includes as the dependent variable of binary linear regression equation:
Utilize Regress Forecast algorithm, using cpu busy percentage and system can executive process number as the independent variable x of model1And x2, i.e. Regress Forecast equation y=a1x1+a2x2+ b, wherein x1Represent the cpu busy percentage under (SuSE) Linux OS, x2Representing the process the run number of system under (SuSE) Linux OS, y represents the load information of system under (SuSE) Linux OS；Cpu busy percentage=the us+sy of operating system；By solving, obtain the coefficient a of binary linear regression equation1、a2And the value of b.
7. the dynamic frequency method based on binary linear regression equation as claimed in claim 2, it is characterized in that, the binary linear regression equation that described basis is tried to achieve, by independent variable information, dependent variable is predicted, accurately provide the load information of subsequent time system, and then the running frequency arranging CPU farther include:
According to equation y=a1x1+a2x2+ b, when known subsequent time cpu busy percentage and system can executive process number, measurable go out subsequent time system load the value of information, and then the running frequency of CPU is set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610139773.7A CN105808355B (en) | 2016-03-10 | 2016-03-10 | Dynamic frequency modulation method based on binary linear regression equation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610139773.7A CN105808355B (en) | 2016-03-10 | 2016-03-10 | Dynamic frequency modulation method based on binary linear regression equation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105808355A true CN105808355A (en) | 2016-07-27 |
| CN105808355B CN105808355B (en) | 2019-12-10 |
Family
ID=56467193
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610139773.7A Active CN105808355B (en) | 2016-03-10 | 2016-03-10 | Dynamic frequency modulation method based on binary linear regression equation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105808355B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110275773A (en) * | 2018-10-30 | 2019-09-24 | 湖北省农村信用社联合社网络信息中心 | Paas resource circulation utilization index system based on truthful data models fitting |
| CN110703898A (en) * | 2019-09-06 | 2020-01-17 | 无锡江南计算技术研究所 | Dynamic management system and method for processor power consumption based on periodic query and interrupt |
| CN110868330A (en) * | 2018-08-28 | 2020-03-06 | 中国移动通信集团浙江有限公司 | Evaluation method, device and evaluation system for CPU resources which can be divided by cloud platform |
| CN114236232A (en) * | 2021-12-16 | 2022-03-25 | 广州城市理工学院 | Small hydropower frequency prediction method considering frequency change trend |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH06258113A (en) * | 1993-03-08 | 1994-09-16 | Yamatake Honeywell Co Ltd | Electromagnetic flowmeter |
| US5954820A (en) * | 1997-09-26 | 1999-09-21 | International Business Machines Corporation | Portable computer with adaptive demand-driven power management |
| CN103488539A (en) * | 2013-09-23 | 2014-01-01 | 北京交通大学 | Data center energy saving method based on central processing unit (CPU) dynamic frequency modulation technology |
| CN105323111A (en) * | 2015-11-17 | 2016-02-10 | 南京南瑞集团公司 | Operation and maintenance automation system and method |
-
2016
- 2016-03-10 CN CN201610139773.7A patent/CN105808355B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH06258113A (en) * | 1993-03-08 | 1994-09-16 | Yamatake Honeywell Co Ltd | Electromagnetic flowmeter |
| US5954820A (en) * | 1997-09-26 | 1999-09-21 | International Business Machines Corporation | Portable computer with adaptive demand-driven power management |
| CN103488539A (en) * | 2013-09-23 | 2014-01-01 | 北京交通大学 | Data center energy saving method based on central processing unit (CPU) dynamic frequency modulation technology |
| CN105323111A (en) * | 2015-11-17 | 2016-02-10 | 南京南瑞集团公司 | Operation and maintenance automation system and method |
Non-Patent Citations (2)
| Title |
|---|
| 崔慎杰: "《集群系统中动态节能调度方法研究》", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 李登: "《分布式系统负载均衡策略研究》", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110868330A (en) * | 2018-08-28 | 2020-03-06 | 中国移动通信集团浙江有限公司 | Evaluation method, device and evaluation system for CPU resources which can be divided by cloud platform |
| CN110868330B (en) * | 2018-08-28 | 2021-09-07 | 中国移动通信集团浙江有限公司 | Evaluation method, device and evaluation system for CPU resources which can be divided by cloud platform |
| CN110275773A (en) * | 2018-10-30 | 2019-09-24 | 湖北省农村信用社联合社网络信息中心 | Paas resource circulation utilization index system based on truthful data models fitting |
| CN110275773B (en) * | 2018-10-30 | 2020-08-28 | 湖北省农村信用社联合社网络信息中心 | Paas resource recycling index system based on real data model fitting |
| CN110703898A (en) * | 2019-09-06 | 2020-01-17 | 无锡江南计算技术研究所 | Dynamic management system and method for processor power consumption based on periodic query and interrupt |
| CN114236232A (en) * | 2021-12-16 | 2022-03-25 | 广州城市理工学院 | Small hydropower frequency prediction method considering frequency change trend |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105808355B (en) | 2019-12-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Xu et al. | A balanced virtual machine scheduling method for energy-performance trade-offs in cyber-physical cloud systems | |
| Qi et al. | A QoS-aware virtual machine scheduling method for energy conservation in cloud-based cyber-physical systems | |
| Zhou et al. | Minimizing SLA violation and power consumption in Cloud data centers using adaptive energy-aware algorithms | |
| US9158351B2 (en) | Dynamic power limit sharing in a platform | |
| CN108021451B (en) | Self-adaptive container migration method in fog computing environment | |
| CN104636197B (en) | A kind of evaluation method of data center's virtual machine (vm) migration scheduling strategy | |
| Zhou et al. | A novel virtual machine deployment algorithm with energy efficiency in cloud computing | |
| CN105808355A (en) | Binary linear regression equation-based dynamic frequency modulation method | |
| CN104038392A (en) | Method for evaluating service quality of cloud computing resources | |
| CN104239153A (en) | Method and device for balancing multi-core CPU load | |
| CN102087619A (en) | Method and apparatus to improve turbo performance for events handling | |
| CN103037109B (en) | Multicore equipment energy consumption management method and device | |
| CN111026548A (en) | Power communication equipment test resource scheduling method for reverse deep reinforcement learning | |
| Fu et al. | Traffic prediction-enabled energy-efficient dynamic computing resource allocation in cran based on deep learning | |
| CN101819459B (en) | Heterogeneous object memory system-based power consumption control method | |
| CN114895773B (en) | Energy consumption optimization method, system and device for heterogeneous multi-core processor and storage medium | |
| CN114240019A (en) | Flexible resource value evaluation method and device suitable for new energy power system | |
| CN110795238A (en) | Load calculation method and device, storage medium and electronic equipment | |
| CN115150471A (en) | Data processing method, device, equipment, storage medium and program product | |
| Du et al. | An energy-aware resource allocation method for avionics systems based on improved ant colony optimization algorithm | |
| Li et al. | Dynamic virtual machine consolidation algorithm based on balancing energy consumption and quality of service | |
| CN102520785A (en) | Energy consumption management method and system for cloud data center | |
| Li et al. | Unified accelerator for attention and convolution in inference based on FPGA | |
| CN103593304B (en) | The quantization method of effective use based on LPT device model caching | |
| Aiwen et al. | Energy-optimal task offloading algorithm of resources cooperation in mobile edge computing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |