TW201316252A - Systems and methods for measuring the effectiveness of a workload predictor on a mobile device - Google Patents

Systems and methods for measuring the effectiveness of a workload predictor on a mobile device Download PDF

Info

Publication number
TW201316252A
TW201316252A TW100136777A TW100136777A TW201316252A TW 201316252 A TW201316252 A TW 201316252A TW 100136777 A TW100136777 A TW 100136777A TW 100136777 A TW100136777 A TW 100136777A TW 201316252 A TW201316252 A TW 201316252A
Authority
TW
Taiwan
Prior art keywords
workload
mobile device
error signal
logic
predictor
Prior art date
Application number
TW100136777A
Other languages
Chinese (zh)
Inventor
Brian J Salsbery
Norman S Gargash
Original Assignee
Qualcomm Inc
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Qualcomm Inc filed Critical Qualcomm Inc
Priority to TW100136777A priority Critical patent/TW201316252A/en
Publication of TW201316252A publication Critical patent/TW201316252A/en

Links

Landscapes

  • Debugging And Monitoring (AREA)

Abstract

Systems and methods for measuring the effectiveness of a workload predictor operative on a mobile device are disclosed. A load manager includes a workload predictor, a sensor, an error generator and a controller. The workload predictor generates an estimate of the workload on a processor core operative on the mobile device. The sensor generates a measure of the actual workload on the processor core. The error generator receives the estimate of the workload and the measure of the actual workload on the processor core and generates an error signal. The controller receives the error signal and determines the effectiveness of the workload predictor as a function of the error signal over time.

Description

用於量測行動裝置上工作負載預測器之效益之系統及方法System and method for measuring the benefit of a workload predictor on a mobile device

無線通信協定、能量儲存密度、處理器效能及效率之技術進步及蜂巢式傳輸器之密度的增加使得能夠在九十年代引入無線數位網路。隨著第二代行動裝置之使用變得更為普遍,變得明顯的是,對包括對網際網路之存取之資料服務的需求正在增長。藉由第三代無線通信系統針對資料傳輸引入封包交換而非電路交換使得能夠增加將串流媒體遞送至行動裝置所需之資料傳輸速率。Technological advances in wireless communication protocols, energy storage density, processor performance and efficiency, and increased densities of cellular transmitters have enabled the introduction of wireless digital networks in the 1990s. As the use of second generation mobile devices becomes more prevalent, it becomes apparent that the demand for data services including access to the Internet is growing. Introducing packet switching instead of circuit switching for data transmission by third generation wireless communication systems enables an increase in the data transmission rate required to deliver streaming media to mobile devices.

現今之具備資料網路能力之行動裝置包括用以將大量應用程式提供至終端使用者之高效能處理器。然而,對效能之需求導致電力消耗之相應增加。隨著給定實施所需之時脈頻率及電晶體之數目兩者增加,現代處理器中之電力耗散迅速增加。不幸地是,電池技術尚不能跟上此等具備資料網路能力之無線裝置的能量要求,此狀況導致應用程式效能與電池壽命之間的設計取捨。即使在判定可為了全效能而犧牲電池壽命時,對於常為手般大小之外殼內之電池及電子電路兩者,熱管理仍可變得成問題。Today's data-networking mobile devices include high-performance processors for delivering a large number of applications to end users. However, the demand for performance leads to a corresponding increase in power consumption. As both the clock frequency and the number of transistors required for a given implementation increase, the power dissipation in modern processors increases rapidly. Unfortunately, battery technology is not yet able to keep up with the energy requirements of these data-network-capable wireless devices, which leads to design trade-offs between application performance and battery life. Even when it is determined that battery life can be sacrificed for full performance, thermal management can become problematic for both battery and electronic circuits in a hand-sized enclosure.

對提供高效能及低電力消耗之處理器的需求已導致在處理器設計中使用動態電壓及頻率按比例調整(DVFS)。DVFS實現電力消耗與效能之間的取捨。經設計以利用DVFS之處理器允許藉由電壓之相應調整來調整處理器之時脈頻率。單獨減少時脈頻率並非有用的,此係因為任何電力節省皆被執行時間之增加所抵消,從而導致無所消耗之總能量的淨減少。然而,操作電壓之減少導致所消耗電力之比例性節省。The need for processors that provide high performance and low power consumption has led to the use of dynamic voltage and frequency scaling (DVFS) in processor designs. DVFS achieves a trade-off between power consumption and performance. A processor designed to utilize DVFS allows the clock frequency of the processor to be adjusted by corresponding adjustments in voltage. Reducing the clock frequency alone is not useful because any power savings are offset by an increase in execution time, resulting in a net reduction in total energy consumed. However, the reduction in operating voltage results in a proportional savings in the power consumed.

具備DVFS能力之處理器的主要問題為如何控制效能與電力節省之間的平衡。習知DVFS控制演算法使用閒置時間之量測來識別修改效能等級,該修改效能等級准許在時間上將預計工作負載延展至截止時間。亦即,藉由以較低電壓及較慢時脈頻率操作處理器而減少閒置時間。雖然此方法可適用於連續且已知之工作負載,但許多行動裝置應用程式係交互式的及/或對處理器提出不規則或不可預測之需求。因此,顯而易見的是,對於所有行動裝置應用程式、應用程式組合及使用,不存在最佳演算法。The main problem with DVFS capable processors is how to balance the balance between performance and power savings. The conventional DVFS control algorithm uses the measurement of the idle time to identify the modified performance level, which allows the expected workload to be extended to the deadline in time. That is, the idle time is reduced by operating the processor at a lower voltage and a slower clock frequency. While this approach is applicable to continuous and known workloads, many mobile device applications are interactive and/or impose irregular or unpredictable demands on the processor. Therefore, it is obvious that there is no optimal algorithm for all mobile device applications, application combinations and uses.

一種用於量測在一行動裝置上操作之一工作負載預測器之效益的方法包括以下步驟:產生該行動裝置上之實際工作負載的一即時量測;使用一工作負載預測器產生該行動裝置上之工作負載的一估計;比較該工作負載之該即時量測與該工作負載之該估計以產生一誤差信號;隨時間記錄該誤差信號;及產生該誤差信號之一函數作為該工作負載預測器之該效益的一量測。A method for measuring the benefit of operating a workload predictor on a mobile device includes the steps of: generating an instant measurement of an actual workload on the mobile device; generating the mobile device using a workload predictor An estimate of the workload; comparing the immediate measurement of the workload with the estimate of the workload to generate an error signal; recording the error signal over time; and generating a function of the error signal as the workload prediction A measure of the benefit of the device.

在該用於量測一工作負載預測器之該效益之方法的一替代實施例中,該方法進一步包括將該誤差信號施加至一控制器之步驟,該控制器回應於該誤差信號而調整至少一參數。In an alternate embodiment of the method for measuring the benefit of a workload predictor, the method further includes the step of applying the error signal to a controller, the controller adjusting at least in response to the error signal One parameter.

在一些實例實施例中,該用於量測一工作負載預測器之該效益的方法係實施於以下各者中之一者上:一蜂巢式電話、一傳呼器、一個人數位助理、一閱讀器、一衛星電話、一導航裝置,或一具有一無線連接之電腦裝置。In some example embodiments, the method for measuring the benefit of a workload predictor is implemented on one of: a cellular phone, a pager, a number of assistants, a reader , a satellite phone, a navigation device, or a computer device with a wireless connection.

在該用於量測一工作負載預測器之該效益之方法的一些其他實例實施例中,該方法進一步包括計算該誤差信號之平方在一選擇週期上之一積分。In some other example embodiments of the method for measuring the benefit of a workload predictor, the method further includes calculating an integral of the square of the error signal over a selection period.

在該用於量測一工作負載預測器之該效益之方法的又一替代實施例中,該方法進一步包含使該誤差信號與一作用中效能按比例調整演算法與至少一參數之一組合相關聯的步驟。In still another alternative embodiment of the method for measuring the benefit of a workload predictor, the method further includes correlating the error signal with an active performance scaling algorithm in combination with one of the at least one parameter The steps of the joint.

在該用於量測一工作負載預測器之該效益之方法的另一替代實施例中,該方法進一步包括以下步驟:使用該誤差信號之交叉相關的一量測以判定該行動裝置之一當前操作狀態是否類似於一已知使用狀況,且在該行動裝置經如此組態時應用一效能按比例調整演算法與一或多個參數之一適當組合。In another alternative embodiment of the method for measuring the benefit of a workload predictor, the method further includes the step of determining a current one of the mobile devices using a cross-correlation of the error signal Whether the operational state is similar to a known use condition, and applying a performance scaling algorithm in appropriate combination with one or more parameters when the mobile device is so configured.

根據另一態樣,揭示一種用於在一行動裝置上執行之負載管理器。該負載管理器包括一工作負載預測器、至少一感測器、一誤差產生器,及一控制器。該工作負載預測器產生在該行動裝置上操作之一處理器核心上之工作負載的一估計。該一或多個感測器產生在該行動裝置上操作之該處理器核心上之實際工作負載的一或多個量測。該誤差產生器接收該工作負載之該估計及該實際工作負載之該量測且產生一誤差信號。該控制器接收該誤差信號且調整該處理器核心之操作。According to another aspect, a load manager for execution on a mobile device is disclosed. The load manager includes a workload predictor, at least one sensor, an error generator, and a controller. The workload predictor generates an estimate of the workload on one of the processor cores operating on the mobile device. The one or more sensors generate one or more measurements of the actual workload on the processor core operating on the mobile device. The error generator receives the estimate of the workload and the measurement of the actual workload and generates an error signal. The controller receives the error signal and adjusts the operation of the processor core.

在該負載管理器之一實例實施例中,該工作負載預測器係對一效能按比例調整演算法作出回應。In one example embodiment of the load manager, the workload predictor is responsive to a performance scaling algorithm.

在該負載管理器之另一實例實施例中,該誤差信號之平方之量值用以判定該工作負載預測器預測該實際工作負載之準確度。In another example embodiment of the load manager, the magnitude of the square of the error signal is used to determine the accuracy of the workload predictor predicting the actual workload.

在該負載管理器之一替代實施例中,包括一時間序列儲存器以隨時間記錄該誤差信號。該負載管理器之此替代實施例可進一步包括耦接至該時間序列儲存器之一或多個交叉相關器。該一或多個交叉相關器產生該即時誤差信號與相關聯於該行動裝置之一經識別使用之誤差信號的相似性的各別量測。In an alternate embodiment of the load manager, a time series store is included to record the error signal over time. This alternate embodiment of the load manager can further include coupling to one or more cross correlators of the time series storage. The one or more cross correlators generate respective measurements of the similarity of the immediate error signal to an identified error signal associated with one of the mobile devices.

此外,一種用於管理一行動裝置上之一工作負載的系統包括:用於量測在該行動裝置上操作之一處理器核心之一實際工作負載的機構;用於產生在該行動裝置上操作之該處理器核心的該工作負載之一估計的機構;用於隨時間產生一誤差信號的機構;用於隨時間儲存該誤差信號的機構;用於使一隨時間所儲存之誤差信號與該行動裝置之一當前使用狀態相關聯的機構;及用於回應於該隨時間所儲存之誤差信號而選擇性地調整該行動裝置的機構。Additionally, a system for managing a workload on a mobile device includes: means for measuring an actual workload of one of the processor cores operating on the mobile device; for generating an operation on the mobile device Means for estimating one of the workloads of the processor core; means for generating an error signal over time; means for storing the error signal over time; for causing an error signal stored over time a mechanism associated with the current state of use of the mobile device; and means for selectively adjusting the mobile device in response to the error signal stored over time.

在用於管理一行動裝置上之一工作負載之該系統的一實施例中,用於產生在該行動裝置上操作之一處理器核心的該工作負載之一估計的該等機構包括一工作負載預測器。In an embodiment of the system for managing a workload on a mobile device, the mechanisms for generating an estimate of the workload of one of the processor cores operating on the mobile device include a workload Predictor.

在用於管理一行動裝置上之一工作負載之該系統的另一實施例中,用於回應於該誤差信號而選擇性地調整該行動裝置的該等機構包括一交叉相關器及一偵測器。In another embodiment of the system for managing a workload on a mobile device, the mechanisms for selectively adjusting the mobile device in response to the error signal include a cross correlator and a detection Device.

此外,在又一替代實施例中,揭示一種電腦程式產品,其包括可執行指令,該等可執行指令在執行時實施用於識別在一行動裝置上操作之一工作負載預測器之效益的方法。該電腦程式產品包括:經組態以產生該行動裝置上之實際工作負載之一即時量測的邏輯;經組態以產生該行動裝置上之工作負載之一估計的邏輯;經組態以比較該工作負載之該即時量測與該工作負載之該估計以產生一誤差信號的邏輯;經組態以隨時間記錄該誤差信號的邏輯;經組態以計算該誤差信號之一函數的邏輯;及經組態以回應於該誤差信號之該函數而判定該工作負載預測器之該效益的邏輯。Moreover, in yet another alternative embodiment, a computer program product is disclosed that includes executable instructions that, when executed, implement a method for identifying the benefit of operating a workload predictor on a mobile device . The computer program product includes: logic configured to generate an instant measurement of an actual workload on the mobile device; logic configured to generate an estimate of a workload on the mobile device; configured to compare The logic of the workload and the estimate of the workload to generate an error signal logic; logic configured to record the error signal over time; logic configured to calculate a function of the error signal; And logic configured to determine the benefit of the workload predictor in response to the function of the error signal.

在該電腦程式產品之一替代實施例中,該電腦程式產品進一步包括經組態以回應於該誤差信號而調整至少一參數的邏輯。In an alternate embodiment of the computer program product, the computer program product further includes logic configured to adjust the at least one parameter in response to the error signal.

在該電腦程式產品之另一替代實施例中,該電腦程式產品進一步包括經組態以使該誤差信號與一效能按比例調整演算法與至少一參數之一組合相關聯的邏輯。In another alternative embodiment of the computer program product, the computer program product further includes logic configured to associate the error signal with a performance scaling algorithm in combination with one of the at least one parameter.

在該電腦程式產品之又一替代實施例中,該電腦程式產品進一步包括經組態以產生交叉相關之一量測以判定該行動裝置之一當前操作狀態是否類似於一已知使用狀況的邏輯,已針對該已知使用狀況識別了一效能按比例調整演算法與一或多個參數之一適當組合以達成一所要目標。In still another alternative embodiment of the computer program product, the computer program product further includes logic configured to generate a cross-correlation measurement to determine whether a current operational state of one of the mobile devices is similar to a known usage condition A performance scaling algorithm has been identified for this known usage condition to properly combine one of the one or more parameters to achieve a desired goal.

在該電腦程式產品之又一替代實施例中,該電腦程式產品進一步包括經組態以儲存一效能按比例調整演算法與至少一參數之最類似組合之一指示的邏輯,及經組態以在該行動裝置經如此組態時應用該組合的邏輯。In still another alternative embodiment of the computer program product, the computer program product further includes logic configured to store an indication of one of a most similar combination of a performance scaling algorithm and at least one parameter, and configured to The combined logic is applied when the mobile device is so configured.

可參看以下諸圖較佳地理解用於量測在行動裝置上操作之工作負載預測器之效益的系統及方法。諸圖內之組件未必按比例繪製,而是著重於清楚地說明以下操作之原理:量測一工作負載;藉由一工作負載預測器來估計一工作負載;依據實際工作負載及工作負載之估計而產生一誤差信號;及將該誤差信號之一函數作為工作負載預測器之效益的量測。此外,在諸圖中,相同參考數字貫穿不同視圖表示相應部分。Systems and methods for measuring the benefits of a workload predictor operating on a mobile device can be better understood with reference to the following figures. The components in the figures are not necessarily drawn to scale, but rather to clearly illustrate the principles of the following operations: measuring a workload; estimating a workload with a workload predictor; estimating the actual workload and workload An error signal is generated; and one of the error signals is used as a measure of the benefit of the workload predictor. In addition, in the figures, like reference numerals refer to the

詞「例示性」在本文中用以意謂「充當一實例、例子或說明」。未必將本文中描述為「例示性」之任何態樣解釋為比其他態樣較佳或有利。The word "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any aspect described herein as "exemplary" is not necessarily to be construed as preferred or advantageous.

在此描述中,術語「應用程式」可包括具有可執行內容之檔案,諸如:目標碼、指令碼、位元組碼、標記語言檔案及修補程式。另外,如在本文中所提及,「應用程式」亦可與實質上不可執行之檔案(諸如,可能需要開啟之文件或需要存取之其他資料檔案)交互。In this description, the term "application" may include files having executable content such as object code, instruction code, byte code, markup language file, and patch. In addition, as mentioned herein, an "application" may also interact with files that are substantially unexecutable, such as files that may need to be opened or other data files that need to be accessed.

在此描述中,術語「行動裝置」用以描述在有限容量之電源供應器(諸如,電池)上操作之任何裝置。儘管電池操作之行動裝置已使用了數十年,但結合第三代及***(「3G」或「4G」)無線技術之出現的可再充電電池之技術進步已實現具有多種能力之眾多行動裝置。因此,行動裝置可為(連同其他裝置)蜂巢式電話、衛星電話、傳呼器、PDA、智慧型電話、導航裝置、智慧筆記型電腦(smartbook)或閱讀器、媒體播放器、前述裝置之組合、具有無線連接之膝上型電腦。In this description, the term "mobile device" is used to describe any device that operates on a limited capacity power supply, such as a battery. Although battery-operated mobile devices have been in use for decades, technological advances in rechargeable batteries that combine the emergence of third- and fourth-generation ("3G" or "4G") wireless technologies have enabled many capabilities. Mobile device. Thus, the mobile device can be (along with other devices) a cellular phone, a satellite phone, a pager, a PDA, a smart phone, a navigation device, a smartbook or a reader, a media player, a combination of the foregoing, A laptop with a wireless connection.

用於在行動裝置上實施各種功能之多核心處理器及專用硬體資源的引入及對改良電池壽命同時確保令人滿意之終端使用者體驗的持續需要已進一步使動態調整行動裝置以符合此等時常衝突之目標的任務變得複雜。The introduction of multi-core processors and dedicated hardware resources for implementing various functions on mobile devices and the continuing need to improve battery life while ensuring a satisfactory end-user experience have further enabled dynamic adjustment of mobile devices to comply with such The task of the goal of frequent conflicts becomes complicated.

如以上所解釋,習知DVFS控制演算法使用閒置時間之量測來識別修改效能等級,該修改效能等級准許在時間上將預計工作負載延展至截止時間。亦即,藉由以較低電壓及較慢時脈頻率操作處理器而減少閒置時間。雖然此方法可適用於連續且已知之工作負載,但許多行動裝置應用程式係交互式的及/或將不規則或不可預測之需求強加於處理器。此外,可用之不同行動裝置的數目、用以與裝置通信之各種通信協定或標準、可用之應用程式之不同類型、使用裝置之不同方式等使得以下事實為明顯的:對於所有可能之操作組合,無單一方法或解決方案可經調整或最佳化。因此,顯而易見,對於所有行動裝置應用程式、應用程式組合及使用,不存在最佳或適當演算法。As explained above, the conventional DVFS control algorithm uses the measurement of the idle time to identify the modified performance level, which allows the expected workload to be extended to the deadline in time. That is, the idle time is reduced by operating the processor at a lower voltage and a slower clock frequency. While this approach is applicable to continuous and known workloads, many mobile device applications are interactive and/or impose irregular or unpredictable demands on the processor. In addition, the number of different mobile devices available, the various communication protocols or standards used to communicate with the device, the different types of applications available, the different ways in which the devices are used, etc., make the following facts obvious: for all possible combinations of operations, No single method or solution can be adjusted or optimized. Therefore, it is obvious that there is no optimal or appropriate algorithm for all mobile device applications, application combinations and uses.

作為回應,揭示一種用於判定工作負載預測器之效益的改良行動裝置及方法。工作負載預測器之輸出可用以評估工作負載預測器之效能。當關於行動裝置之各種使用的資料可用時,工作負載預測器之輸出可用以識別行動裝置之即時使用。此外,可將工作負載偵測器之輸出轉遞至控制器以按多種方式調整行動裝置。舉例而言,可將工作負載預測器之輸出施加至控制器,該控制器選擇待應用於選擇演算法(諸如,在行動裝置上操作之效能按比例調整演算法)之操作參數集合。在另一實例中,將自工作負載預測器之輸出產生的誤差信號施加至交叉相關器,該交叉相關器提供即時誤差信號與先前儲存之誤差信號的相似性之量測。當交叉相關器指示該等誤差信號在某確定性程度上匹配且先前儲存之信號與行動裝置上之已知使用狀況相關聯時,控制器可選擇效能按比例調整演算法與參數集合之先前最佳化組合,以管理跨越可用於行動裝置上之多個處理核心的工作負載。In response, an improved mobile device and method for determining the effectiveness of a workload predictor is disclosed. The output of the workload predictor can be used to evaluate the performance of the workload predictor. The output of the workload predictor can be used to identify the immediate use of the mobile device when available information about the various uses of the mobile device is available. In addition, the output of the workload detector can be forwarded to the controller to adjust the mobile device in a variety of ways. For example, the output of the workload predictor can be applied to a controller that selects a set of operational parameters to be applied to a selection algorithm, such as a performance scaling algorithm operating on a mobile device. In another example, an error signal generated from the output of the workload predictor is applied to a cross-correlator that provides a measure of the similarity of the instantaneous error signal to the previously stored error signal. When the cross correlator indicates that the error signals match to a certain degree of certainty and the previously stored signal is associated with a known usage condition on the mobile device, the controller may select the performance prior to scaling the algorithm and the previous set of parameters. A combination of management to manage workloads across multiple processing cores that can be used on mobile devices.

用於量測在行動裝置上操作之工作負載預測器之效益的系統及方法大體上係經由軟體或韌體與硬體之組合來實施。該軟體及韌體可儲存於耦接至可在行動裝置上操作之一或多個處理器的非揮發性記憶體元件中。儲存於非揮發性記憶體元件中之軟體經傳達至合適之指令執行系統(微處理器)且藉由該指令執行系統(微處理器)執行。行動裝置上之硬體實施可包括以下技術(皆為此項技術中熟知的)中之任一者或組合:離散電子組件、具有用於對資料信號實施邏輯功能之邏輯閘的離散邏輯電路、具有適當邏輯閘之特殊應用積體電路、可程式化閘陣列(PGA)、場可程式化閘陣列(FPGA)等。Systems and methods for measuring the benefits of workload predictors operating on mobile devices are generally implemented via software or a combination of firmware and hardware. The software and firmware can be stored in a non-volatile memory component that is coupled to one or more processors that can operate on the mobile device. The software stored in the non-volatile memory element is communicated to a suitable instruction execution system (microprocessor) and executed by the instruction execution system (microprocessor). The hardware implementation on the mobile device can include any one or combination of the following techniques (all well known in the art): discrete electronic components, discrete logic circuits with logic gates for performing logic functions on data signals, Special application integrated circuits with appropriate logic gates, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

參看圖1,此圖為呈無線電話之形式的行動裝置100之一例示性、非限制性態樣的圖。如所展示,行動裝置100包括系統單晶片102,該系統單晶片102包括耦接在一起之數位信號處理器110及類比信號處理器126。如圖1中所說明,顯示控制器128及觸控螢幕控制器130耦接至數位信號處理器110。系統單晶片102外部之觸控螢幕顯示器132耦接至顯示控制器128及觸控螢幕控制器130。Referring to Figure 1, this figure is an illustrative, non-limiting view of one of the mobile devices 100 in the form of a radiotelephone. As shown, the mobile device 100 includes a system single chip 102 that includes a digital signal processor 110 and an analog signal processor 126 coupled together. As shown in FIG. 1 , the display controller 128 and the touch screen controller 130 are coupled to the digital signal processor 110 . The touch screen display 132 external to the system single chip 102 is coupled to the display controller 128 and the touch screen controller 130.

圖1進一步指示視訊編碼器134(例如,相位準線(「PAL」)編碼器、順序傳送與儲存彩色電視系統(「SECAM」)編碼器、國家電視系統委員會(「NTSC」)編碼器或任何其他視訊編碼器)耦接至數位信號處理器110。另外,視訊放大器136耦接至視訊編碼器134及觸控螢幕顯示器132。視訊埠138耦接至視訊放大器136。如圖1中所描繪,通用串列匯流排(「USB」)控制器140耦接至數位信號處理器110。又,USB埠142耦接至USB控制器140。記憶體112及用戶識別模組(SIM)卡146亦可耦接至數位信號處理器110。另外,如圖1中所展示,數位相機148可耦接至數位信號處理器110。在一例示性態樣中,數位相機148為電荷耦接裝置(「CCD」)相機或互補金屬氧化物半導體(「CMOS」)相機。Figure 1 further indicates a video encoder 134 (e.g., a phase alignment ("PAL") encoder, a sequential transmission and storage color television system ("SECAM") encoder, a National Television System Committee ("NTSC") encoder, or any The other video encoders are coupled to the digital signal processor 110. In addition, the video amplifier 136 is coupled to the video encoder 134 and the touch screen display 132. The video port 138 is coupled to the video amplifier 136. As depicted in FIG. 1, a universal serial bus ("USB") controller 140 is coupled to the digital signal processor 110. Also, the USB port 142 is coupled to the USB controller 140. Memory 112 and Subscriber Identity Module (SIM) card 146 may also be coupled to digital signal processor 110. Additionally, as shown in FIG. 1 , digital camera 148 can be coupled to digital signal processor 110 . In an exemplary aspect, digital camera 148 is a charge coupled device ("CCD") camera or a complementary metal oxide semiconductor ("CMOS") camera.

如圖1中進一步說明,立體聲音訊CODEC 150可耦接至類比信號處理器126。此外,音訊放大器152可耦接至立體聲音訊CODEC 150。在一例示性態樣中,第一立體聲揚聲器154及第二立體聲揚聲器156耦接至音訊放大器152。圖1展示麥克風放大器158亦可耦接至立體聲音訊CODEC 150。另外,麥克風160可耦接至麥克風放大器158。在一特定態樣中,頻率調變(「FM」)無線電調諧器162可耦接至立體聲音訊CODEC 150。又,FM天線164耦接至FM無線電調諧器162。另外,立體聲頭戴式耳機166可耦接至立體聲音訊COEDC 150。As further illustrated in FIG. 1 , the stereo audio CODEC 150 can be coupled to the analog signal processor 126 . In addition, the audio amplifier 152 can be coupled to the stereo audio CODEC 150. In an exemplary aspect, the first stereo speaker 154 and the second stereo speaker 156 are coupled to the audio amplifier 152. FIG. 1 shows that the microphone amplifier 158 can also be coupled to the stereo audio CODEC 150. Additionally, the microphone 160 can be coupled to the microphone amplifier 158. In a particular aspect, a frequency modulation ("FM") radio tuner 162 can be coupled to the stereo audio CODEC 150. Also, the FM antenna 164 is coupled to the FM radio tuner 162. Additionally, the stereo headset 166 can be coupled to the stereo audio COEDC 150.

圖1進一步指示射頻(「RF」)收發器168可耦接至類別信號處理器126。RF開關170可耦接至RF收發器168及RF天線172。如圖1中所展示,小鍵盤174可耦接至類比信號處理器126。又,具有麥克風之單聲道耳機176可耦接至類別信號處理器126。另外,振動器裝置178可耦接至類比信號處理器126。圖1亦展示電源供應器180(例如,電池)耦接至系統單晶片102。在一特定態樣中,電源供應器包括可再充電DC電池或自連接至交流電(「AC」)電源之AC至DC變換器導出的DC電源供應器。FIG. 1 further indicates that a radio frequency ("RF") transceiver 168 can be coupled to the class signal processor 126. The RF switch 170 can be coupled to the RF transceiver 168 and the RF antenna 172. As shown in FIG. 1, keypad 174 can be coupled to analog signal processor 126. Also, a mono headset 176 having a microphone can be coupled to the category signal processor 126. Additionally, the vibrator device 178 can be coupled to the analog signal processor 126. FIG. 1 also shows that a power supply 180 (eg, a battery) is coupled to the system single wafer 102. In a particular aspect, the power supply includes a rechargeable DC battery or a DC power supply derived from an AC to DC converter connected to an alternating current ("AC") power source.

圖1亦展示行動裝置100可包括監視器模組114。該監視器模組114與遍及系統單晶片102而散佈之多個操作感測器且與行動裝置100之數位信號處理器110通信。如圖1中所描繪,觸控螢幕顯示器132、視訊埠138、USB埠142、相機148、第一立體聲揚聲器154、第二立體聲揚聲器156、麥克風160、FM天線164、立體聲頭戴式耳機166、RF開關170、RF天線172、小鍵盤174、單聲道耳機176、振動器178及電源供應器180係在系統單晶片322外部。然而,應理解,監視器模組114亦可借助於類比信號處理器126及數位信號處理器110自此等外部裝置中之一或多者接收一或多個指示,以輔助可在行動裝置100上操作之資源的即時管理。FIG. 1 also shows that the mobile device 100 can include a monitor module 114. The monitor module 114 is in communication with a plurality of operational sensors distributed throughout the system single wafer 102 and with the digital signal processor 110 of the mobile device 100. As depicted in FIG. 1 , touch screen display 132 , video cassette 138 , USB port 142 , camera 148 , first stereo speaker 154 , second stereo speaker 156 , microphone 160 , FM antenna 164 , stereo headset 166 , RF switch 170, RF antenna 172, keypad 174, mono headset 176, vibrator 178, and power supply 180 are external to system single die 322. However, it should be understood that the monitor module 114 can also receive one or more indications from one or more of the external devices by means of the analog signal processor 126 and the digital signal processor 110 to assist in the mobile device 100. Instant management of resources on the operation.

在一特定態樣中,本文中所描述之方法步驟中的一或多者可由儲存於記憶體112中之可執行指令及參數來實施。此等指令可由數位信號處理器110、類比信號處理器126或另一處理器執行以執行本文中所描述之方法。另外,處理器110、126、記憶體112、儲存於記憶體112中之指令,或其組合可充當用於執行本文中所描述之方法步驟中之一或多者的構件。In a particular aspect, one or more of the method steps described herein can be implemented by executable instructions and parameters stored in memory 112. Such instructions may be executed by digital signal processor 110, analog signal processor 126, or another processor to perform the methods described herein. Additionally, processors 110, 126, memory 112, instructions stored in memory 112, or a combination thereof, can serve as a means for performing one or more of the method steps described herein.

圖2為說明圖1之行動裝置100之特徵的示意圖。如圖2中所指示,數位信號處理器110經由匯流排211耦接至記憶體112。數位信號處理器110為具有N個核心處理器之多核心處理器。亦即,數位信號處理器110包括第一核心222、第二核心224及第N核心230。如所已知,第一核心222、第二核心224及第N核心230中之每一者可用於支援專用應用程式或程式。或者,可跨越可用核心中之兩者或兩者以上而散佈一或多個應用程式或程式以用於處理。FIG. 2 is a schematic diagram illustrating features of the mobile device 100 of FIG. 1. As indicated in FIG. 2, the digital signal processor 110 is coupled to the memory 112 via the bus bar 211. Digital signal processor 110 is a multi-core processor with N core processors. That is, the digital signal processor 110 includes a first core 222, a second core 224, and an Nth core 230. As is known, each of the first core 222, the second core 224, and the Nth core 230 can be used to support a dedicated application or program. Alternatively, one or more applications or programs may be distributed for processing across two or more of the available cores.

第一核心222、第二核心224至第N核心230可整合於單一積體電路晶粒上,或其可整合或耦接於多電路封裝中之單獨晶粒上。設計者可經由一或多個共用快取記憶體耦接第一核心222、第二核心224至第N核心230,且其可經由諸如匯流排、環狀、網狀或縱橫式拓撲之網路拓撲來實施訊息或指令傳遞。The first core 222, the second core 224 to the Nth core 230 may be integrated on a single integrated circuit die, or they may be integrated or coupled to separate dies in a multi-circuit package. The designer can couple the first core 222, the second core 224 to the Nth core 230 via one or more shared cache memories, and can pass through a network such as a bus, ring, mesh or crossbar topology Topology to implement message or instruction delivery.

在所說明之實施例中,RF收發器268係經由數位電路元件實施且包括諸如核心處理器210(標示為「核心」)之至少一處理器。在此數位實施中,RF收發器268經由匯流排213耦接至記憶體112。In the illustrated embodiment, RF transceiver 268 is implemented via digital circuit elements and includes at least one processor, such as core processor 210 (labeled "core"). In this digital implementation, the RF transceiver 268 is coupled to the memory 112 via the bus bar 213.

如此項技術中所已知,匯流排211及匯流排213中之每一者可包括經由一或多個有線或無線連接之多個通信路徑。匯流排211及匯流排213可具有用以實現通信之額外元件(為了簡單起見,可省略該等額外元件),諸如,控制器、緩衝器(快取記憶體)、驅動器、中繼器及接收器。另外,匯流排211及匯流排213可包括用以實現前述組件間之適當通信的位址、控制及/或資料連接。As is known in the art, each of bus bar 211 and bus bar 213 can include multiple communication paths via one or more wired or wireless connections. The bus bar 211 and the bus bar 213 may have additional components for communication (for simplicity, such additional components may be omitted), such as controllers, buffers (cache memory), drivers, repeaters, and receiver. Additionally, bus 211 and bus 213 may include address, control, and/or data connections to enable proper communication between the aforementioned components.

如圖2中所展示,當以軟體來實施由行動裝置100所使用之邏輯時,應注意,起動邏輯250、管理邏輯260、DVFS介面邏輯270、應用程式儲存器280中之應用程式及檔案系統290之部分中的一或多者可儲存於任何電腦可讀媒體上以供任何電腦相關系統或方法使用或結合任何電腦相關系統或方法使用。As shown in FIG. 2, when the logic used by the mobile device 100 is implemented in software, it should be noted that the startup logic 250, the management logic 260, the DVFS interface logic 270, the application and file system in the application storage 280, One or more of the portions of 290 can be stored on any computer readable medium for use by or in connection with any computer related system or method.

在此文件之上下文中,電腦可讀媒體為可含有或儲存供電腦相關系統或方法使用或結合電腦相關系統或方法使用之電腦程式的電子、磁性、光學或其他實體裝置或構件。各種邏輯元件及資料儲存器可體現於任何電腦可讀媒體中以供指令執行系統、設備或裝置(諸如,基於電腦之系統、含有處理器之系統)使用或結合指令執行系統、設備或裝置(諸如,基於電腦之系統、含有處理器之系統或可自指令執行系統、設備或裝置提取指令且執行該等指令之其他系統)使用。在此文件之上下文中,「電腦可讀媒體」可為可儲存、傳達、傳播或輸送程式以供指令執行系統、設備或裝置使用或結合指令執行系統、設備或裝置使用的任何構件。In the context of this document, a computer-readable medium is an electronic, magnetic, optical, or other physical device or component that can contain or store a computer program for use by or in connection with a computer-related system or method. The various logic components and data storage can be embodied in any computer readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a processor-containing system) or in conjunction with an instruction execution system, apparatus, or device ( Use, for example, in a computer-based system, a processor-containing system, or other system that can extract instructions from a command execution system, device, or device and execute such instructions. In the context of this document, a "computer-readable medium" can be any component that can store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

電腦可讀媒體可為(例如,但不限於)電子、磁性、光學、電磁、紅外線或半導體系統、設備、裝置或傳播媒體。電腦可讀媒體之更特定實例(非詳盡清單)將包括以下各者:具有一或多個導線之電連接(電子)、攜帶型電腦磁片(磁性)、隨機存取記憶體(RAM)(電子)、唯讀記憶體(ROM)(電子)、可抹除可程式化唯讀記憶體(EPROM、EEPROM或快閃記憶體)(電子)、光纖(光學),及攜帶型緊密光碟唯讀記憶體(CDROM)(光學)。應注意,電腦可讀媒體可甚至為紙張或上面列印有程式之另一合適媒體,此係因為程式可以電學方式捕獲(例如,經由紙張或其他媒體之光學掃描)、接著以合適方式(在必要時)編譯、解譯或以其他方式處理,且接著儲存於電腦記憶體中。The computer readable medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, device, or communication medium. More specific examples (non-exhaustive list) of computer readable media will include the following: electrical connections (electronics) with one or more wires, magnetic disk (magnetic), portable memory (RAM) ( Electronic), read-only memory (ROM) (electronic), erasable programmable read-only memory (EPROM, EEPROM or flash memory) (electronic), optical (optical), and portable compact discs Memory (CDROM) (optical). It should be noted that the computer readable medium may even be paper or another suitable medium on which the program is printed, as the program can be electrically captured (eg, optically scanned via paper or other media), and then in a suitable manner (in Compile, interpret, or otherwise process as necessary, and then store it in computer memory.

在起動邏輯250、管理邏輯260及(可能)DVFS介面邏輯270中之一或多者係以硬體實施的替代實施例中,各種邏輯可藉由各自為此項技術中熟知之以下技術中之任一者或組合來實施:具有用於對資料信號實施邏輯功能之邏輯閘的離散邏輯電路、具有適當組合邏輯閘之特殊應用積體電路(ASIC)、可程式化閘陣列(PGA)、場可程式化閘陣列(FPGA)等。In an alternate embodiment in which one or more of boot logic 250, management logic 260, and (possibly) DVFS interface logic 270 are implemented in hardware, the various logics may be by the following techniques each well known in the art. Either or a combination of discrete logic circuits with logic gates for performing logic functions on data signals, special application integrated circuits (ASICs) with appropriate combinational logic gates, programmable gate arrays (PGAs), fields Programmable gate array (FPGA), etc.

記憶體112為諸如快閃記憶體或固態記憶體裝置之非揮發性資料儲存裝置。儘管經描繪為單一裝置,但記憶體112可為具有耦接至數位信號處理器及/或RF收發器268中之核心210(或額外處理器核心)之單獨資料儲存器的分散式記憶體裝置。Memory 112 is a non-volatile data storage device such as a flash memory or solid state memory device. Although depicted as a single device, the memory 112 can be a decentralized memory device having separate data stores coupled to the core signal 210 (or additional processor core) in the digital signal processor and/or RF transceiver 268 .

起動邏輯250包括一或多個可執行指令,該一或多個可執行指令用於選擇性地識別、載入及執行選擇程式以用於管理或控制可用核心(諸如,第一核心222、第二核心224至第N核心230)中之一或多者的效能。選擇程式可見於嵌入式檔案系統290之程式儲存器296中且由效能按比例調整演算法297與參數298之集合的特定組合定義。當由數位信號處理器之核心處理器及RF收發器268中之核心210中的一或多者執行時,選擇程式可根據由監視器模組114所提供之一或多個信號操作以按比例調整各別處理器核心之效能。就此而言,監視器模組114可提供事件、程序、應用程式、資源狀態條件、耗用時間、溫度等之一或多個指示符。The startup logic 250 includes one or more executable instructions for selectively identifying, loading, and executing a selection program for managing or controlling an available core (such as the first core 222, The performance of one or more of the second core 224 to the Nth core 230). The selection program can be found in the program store 296 of the embedded file system 290 and is defined by a particular combination of the performance scaling algorithm 297 and the set of parameters 298. When executed by one or more of the core processor of the digital signal processor and the core 210 of the RF transceiver 268, the selection program can be scaled according to one or more signals provided by the monitor module 114. Adjust the performance of each processor core. In this regard, the monitor module 114 can provide one or more indicators of events, programs, applications, resource status conditions, elapsed time, temperature, and the like.

管理邏輯260包括一或多個可執行指令,該一或多個可執行指令用於終止各別處理器核心中之一或多者上的操作效能按比例調整程式,以及選擇性地識別、載入及執行用於管理或控制可用核心中之一或多者之效能的更合適之替換程式。管理邏輯260經配置以在執行階段或在行動裝置經供電且由裝置之操作者使用時執行此等功能。替換程式可見於嵌入式檔案系統290之程式儲存器296中,且由效能按比例調整演算法297與參數298之集合的特定組合定義。如圖2中所說明,程式儲存器296可以階層方式配置於核心儲存器294及DVFS儲存器292內。當由數位信號處理器中之核心處理器或RF收發器268中之核心210中的一或多者執行時,該替換程式可根據由監視器模組114所提供之一或多個信號或提供於各種處理器核心之各別控制輸入上的一或多個信號來操作,以按比例調整各別處理器核心之效能。就此而言,監視器模組114可提供事件、程序、應用程式、資源狀態條件、耗用時間、溫度等之一或多個指示符。Management logic 260 includes one or more executable instructions for terminating operational performance scaling procedures on one or more of the respective processor cores, and selectively identifying, carrying Entering and executing a more appropriate replacement program for managing or controlling the performance of one or more of the available cores. Management logic 260 is configured to perform such functions during the execution phase or when the mobile device is powered and used by an operator of the device. The replacement program can be found in the program store 296 of the embedded file system 290 and is defined by a specific combination of the performance scaling algorithm 297 and the set of parameters 298. As illustrated in FIG. 2, the program storage 296 can be configured in a hierarchical manner within the core storage 294 and the DVFS storage 292. When executed by one or more of the core processors in the digital signal processor or the core 210 in the RF transceiver 268, the replacement program can be based on one or more signals or provided by the monitor module 114 One or more signals on respective control inputs of various processor cores operate to scale the performance of the respective processor cores. In this regard, the monitor module 114 can provide one or more indicators of events, programs, applications, resource status conditions, elapsed time, temperature, and the like.

DVFS介面邏輯或介面邏輯270包括用於呈現外部輸入、管理外部輸入且與外部輸入交互以觀測、組態或以其他方式更新儲存於嵌入式檔案系統290中之資訊的一或多個可執行指令。在一實施例中,介面邏輯270可結合經由USB埠142接收之製造商輸入來操作。此等輸入可包括待自程式儲存器296刪除或待添加至程式儲存器296之一或多個程式。或者,該等輸入可包括對程式儲存器296中之程式中之一或多者的編輯或改變。此外,該等輸入可識別對起動邏輯250及管理邏輯260中之一者或兩者的一或多個改變或整體替換。藉由實例,該等輸入可包括對管理邏輯260之改變,該改變指示行動裝置100在所接收之信號功率下降到低於所識別之臨限值時暫時中止RF收發器268中之所有效能按比例調整。藉由其他實例,該等輸入可包括對管理邏輯260之改變,該改變指示行動裝置100在視訊編碼解碼器134處於作用中時應用所要程式。DVFS interface logic or interface logic 270 includes one or more executable instructions for presenting external inputs, managing external inputs, and interacting with external inputs to observe, configure, or otherwise update information stored in embedded file system 290. . In an embodiment, interface logic 270 can operate in conjunction with manufacturer input received via USB port 142. Such inputs may include one or more programs to be deleted from program storage 296 or to be added to program storage 296. Alternatively, the inputs may include edits or changes to one or more of the programs in the program store 296. Moreover, the inputs may identify one or more changes or overall replacements to one or both of the startup logic 250 and the management logic 260. By way of example, the inputs may include a change to management logic 260 that instructs mobile device 100 to temporarily suspend all performance in RF transceiver 268 when the received signal power drops below the identified threshold. Proportional adjustment. By way of other examples, the inputs may include changes to the management logic 260 that instruct the mobile device 100 to apply the desired program while the video codec 134 is active.

介面邏輯270使得製造商能夠在定義之操作條件下以可控制方式組態並調整終端使用者在行動裝置100上之體驗。當記憶體112為快閃記憶體時,可編輯、替換或以其他方式修改起動邏輯250、管理邏輯260、介面邏輯270、應用程式儲存器280中之應用程式或嵌入式檔案系統290中之資訊中的一或多者。在一些實施例中,介面邏輯270可准許行動裝置100之終端使用者或操作者搜尋、定位、修改或替換起動邏輯250、管理邏輯260、應用程式儲存器280中之應用程式及嵌入式檔案系統290中之資訊。操作者可使用所得介面以進行在行動裝置100之下一起動後便將實施的改變。或者,操作者可使用所得介面以進行在執行階段期間實施之改變。The interface logic 270 enables the manufacturer to configure and adjust the end user's experience on the mobile device 100 in a controlled manner under defined operating conditions. When the memory 112 is a flash memory, the startup logic 250, the management logic 260, the interface logic 270, the application in the application storage 280, or the information in the embedded file system 290 can be modified, replaced, or otherwise modified. One or more of them. In some embodiments, the interface logic 270 can permit an end user or operator of the mobile device 100 to search, locate, modify, or replace the startup logic 250, the management logic 260, the application in the application storage 280, and the embedded file system. Information in 290. The operator can use the resulting interface to make changes that will be implemented after moving together under the mobile device 100. Alternatively, the operator can use the resulting interface to make changes that are implemented during the execution phase.

圖3A及圖3B包括說明使用狀況之實施例之示意圖。在圖3A之表300中說明第一實例使用狀況。該使用狀況包括能夠在行動裝置100上執行之任務之集合。該等任務包括web瀏覽(亦即,當行動裝置100處於耦接至網際網路之傳輸器的通信範圍中時執行瀏覽器應用程式)、音訊/視訊播放、音訊播放、執行電子郵件應用程式及執行一或多個其他應用程式。在圖3B之表350中說明第二實例使用狀況。替代第一使用狀況之電子郵件應用程式,此第二使用狀況包括簡訊服務(SMS)應用程式。3A and 3B include schematic diagrams illustrating an embodiment of a use condition. The first example use case is illustrated in the table 300 of Figure 3A. This usage condition includes a collection of tasks that can be performed on the mobile device 100. Such tasks include web browsing (ie, executing a browser application when the mobile device 100 is in communication range coupled to the Internet's transmitter), audio/video playback, audio playback, executing an email application, and Execute one or more other applications. The second example use case is illustrated in the table 350 of Figure 3B. In place of the first use status email application, this second usage status includes a Short Message Service (SMS) application.

一或多個使用狀況可呈現行動裝置100之處理資源上之可識別工作負載係可能的。若可識別行動裝置100上之工作負載,則可應用合適組態之工作負載管理器以如可能所需要而調整行動裝置100上之操作條件,從而確保效能、節省電池電力或應用經設計以達成效能與電力消耗之間的所要取捨之特定效能按比例調整演算法及相應參數。One or more usage conditions may present an identifiable workload on the processing resources of the mobile device 100. If the workload on the mobile device 100 can be identified, a suitably configured workload manager can be applied to adjust the operating conditions on the mobile device 100 as needed to ensure performance, save battery power, or the application is designed to achieve The specific performance of the desired trade-off between performance and power consumption scales the algorithm and corresponding parameters.

工作負載預測器之效益或準確度的量測將輸入提供至控制系統(諸如,負載管理器),該控制系統可用以識別行動裝置100上之工作負載。當啟用時,負載管理器即時地操作以識別行動裝置100上之當前負載,且在特定情況下藉由調整行動裝置100上之一或多個操作參數而對經識別工作負載作出反應。對於包括已知用以達成操作者體驗與電力消耗之間的所要平衡之多個效能按比例調整演算法的行動裝置100而言,經識別工作負載的識別可用以選擇一選擇效能按比例調整演算法與一或多個參數之一適當組合且將該組合應用於行動裝置100上。The measurement of the benefit or accuracy of the workload predictor provides input to a control system, such as a load manager, that can be used to identify the workload on the mobile device 100. When enabled, the load manager operates on the fly to identify the current load on the mobile device 100 and, in certain instances, reacts to the identified workload by adjusting one or more operational parameters on the mobile device 100. For a mobile device 100 that includes a plurality of performance scaling algorithms known to achieve a desired balance between operator experience and power consumption, the identification of the identified workload can be used to select a selection performance scaling algorithm. The method is suitably combined with one of one or more parameters and applied to the mobile device 100.

圖4為說明系統400之一實施例的示意圖,該系統400用於針對圖1之行動裝置100的當前使用來量測與負載管理器415相關聯之工作負載預測器420的效益。系統400提供工作負載預測器之效益的即時定量分析以準確地識別在行動裝置100上操作之DSP 410之處理核心上的工作負載。4 is a schematic diagram illustrating one embodiment of a system 400 for measuring the benefits of the workload predictor 420 associated with the load manager 415 for the current use of the mobile device 100 of FIG. System 400 provides an instant quantitative analysis of the benefits of the workload predictor to accurately identify the workload on the processing core of DSP 410 operating on mobile device 100.

如圖4中所指示,該系統包括數位信號處理器410,該數位信號處理器410耦接至感測器405以接收行動裝置100上之當前操作狀態的一或多個指示。除耦接至感測器405之外,數位信號處理器410亦經由匯流排111耦接至記憶體元件112且耦接至偵測器480。如以上所指示,記憶體元件112可經組態而具有用於執行所要功能之多個程式或可執行陳述式(亦即,邏輯)之集合。此等功能可包括隨時間儲存誤差信號以用於經由一或多個交叉相關器進行比較,從而量化即時誤差信號與針對各別經識別使用狀況的一或多個先前記錄之誤差信號的相似性。As indicated in FIG. 4, the system includes a digital signal processor 410 coupled to the sensor 405 to receive one or more indications of the current operational state on the mobile device 100. In addition to being coupled to the sensor 405 , the digital signal processor 410 is also coupled to the memory component 112 via the bus bar 111 and coupled to the detector 480 . As indicated above, the memory component 112 can be configured with a collection of programs or executable statements (i.e., logic) for performing the desired functions. Such functions may include storing an error signal over time for comparison via one or more cross-correlators to quantify the similarity of the immediate error signal to one or more previously recorded error signals for respective identified usage conditions .

如以上所建議,負載管理器415可經配置以回應於誤差信號而如行動裝置100上可能所需要來調整一或多個操作參數。亦如以上所建議,若先前記錄之誤差信號與經識別使用狀況相關聯,且若即時誤差信號接近先前記錄之誤差信號,則負載管理器415可將先前經最佳化之參數集合及/或效能按比例調整演算法與參數集合之組合應用於行動裝置100上。As suggested above, the load manager 415 can be configured to adjust one or more operational parameters as may be needed on the mobile device 100 in response to the error signal. As also suggested above, if the previously recorded error signal is associated with the identified usage condition, and if the immediate error signal is close to the previously recorded error signal, the load manager 415 can prioritize the previously optimized parameters and/or The combination of the performance scaling algorithm and the set of parameters is applied to the mobile device 100.

感測器405或感測器之集合(圖中未示)可提供當前如何組態行動裝置100之許多指示。此等指示可由經組態以提供實體參數(諸如,電壓、電流、溫度等)之量測的實體裝置提供,或(諸如)在軟體堆疊之狀況下,此等指示可表示一或多個暫存器或其他儲存位置之內容。此外,感測器405可基於行動裝置100上之各資源的操作狀態提供一或多個額外指示符。儘管將感測器405說明為在數位信號處理器410外部,但應理解,數位信號處理器410可包括額外感測器,諸如,量測處理器核心222上之一或多個條件的感測器223、量測處理器核心224上之一或多個條件的感測器225,及量測處理器核心230上之一或多個條件的感測器231。如圖4中所展示,感測器223、感測器225及感測器231將指示行動裝置100上之實際工作負載的實際量測提供至負載管理器415。感測器405、感測器223、感測器225及感測器231中之一或多者,該感測器。A collection of sensors 405 or sensors (not shown) can provide a number of indications of how the mobile device 100 is currently configured. Such indications may be provided by a physical device configured to provide measurements of physical parameters such as voltage, current, temperature, etc., or such as in the case of a software stack, such indications may represent one or more The contents of the storage or other storage location. Moreover, sensor 405 can provide one or more additional indicators based on the operational status of each resource on mobile device 100. Although the sensor 405 is illustrated as being external to the digital signal processor 410, it should be understood that the digital signal processor 410 can include additional sensors, such as sensing one or more conditions on the processor core 222. The 223, the sensor 225 that measures one or more conditions on the processor core 224, and the sensor 231 that measures one or more conditions on the processor core 230. As shown in FIG. 4, sensor 223, sensor 225, and sensor 231 provide actual measurements indicative of the actual workload on mobile device 100 to load manager 415. One or more of the sensor 405, the sensor 223, the sensor 225, and the sensor 231, the sensor.

然而,感測器405或額外感測器可關於行動裝置100而配置、組態或散佈,數位信號處理器410為具有N個核心處理器之多核心處理器。亦即,數位信號處理器410包括第一核心222、第二核心224及第N核心230。如所已知,第一核心222、第二核心224及第N核心230中之每一者可用於支援專用應用程式或程式。或者,可跨越可用核心中之兩者或兩者以上而散佈一或多個應用程式或程式以用於處理。However, the sensor 405 or additional sensor can be configured, configured, or interspersed with respect to the mobile device 100, which is a multi-core processor with N core processors. That is, the digital signal processor 410 includes a first core 222, a second core 224, and an Nth core 230. As is known, each of the first core 222, the second core 224, and the Nth core 230 can be used to support a dedicated application or program. Alternatively, one or more applications or programs may be distributed for processing across two or more of the available cores.

第一核心222、第二核心224至第N核心230可整合於單一積體電路晶粒上,或其可整合或耦接於多電路封裝中之單獨晶粒上。設計者可經由一或多個共用快取記憶體耦接第一核心222、第二核心224至第N核心230,且其可經由諸如匯流排、環狀、網狀或縱橫式拓撲之網路拓撲來實施訊息或指令傳遞。The first core 222, the second core 224 to the Nth core 230 may be integrated on a single integrated circuit die, or they may be integrated or coupled to separate dies in a multi-circuit package. The designer can couple the first core 222, the second core 224 to the Nth core 230 via one or more shared cache memories, and can pass through a network such as a bus, ring, mesh or crossbar topology Topology to implement message or instruction delivery.

負載管理器415包括工作負載預測器420、誤差產生器430及控制器440。在可選實施例中,負載管理器415亦可包括效能按比例調整演算法之集合及參數之集合以用於跨越該等核心中之一或多者應用DVFS方案。工作負載預測器420接收一或多個輸入且在一或多個輸出端處產生經預測工作負載之指示。第一輸出端將經預測之工作負載的指示(亦即,時變信號)耦接至控制器440。第二輸出端將表示經預測工作負載的信號耦接至誤差產生器430之負輸入端。誤差產生器430在誤差產生器430之正輸入端處自感測器223、感測器225或感測器231接收行動裝置100上之工作負載的實際量測。作為回應,誤差產生器430在連接435上將誤差信號(亦即,經量測或實際工作負載與經預測工作負載的差)傳達至控制器440,傳達至一或多個交叉相關器(例如,交叉相關器450、交叉相關器452及交叉相關器458),且傳達至記憶體112中之時間序列儲存器460。The load manager 415 includes a workload predictor 420, an error generator 430, and a controller 440. In an alternative embodiment, load manager 415 can also include a set of performance scaling algorithms and a set of parameters for applying a DVFS scheme across one or more of the cores. The workload predictor 420 receives one or more inputs and produces an indication of the predicted workload at one or more outputs. The first output couples an indication of the predicted workload (ie, a time varying signal) to the controller 440. The second output couples a signal representative of the predicted workload to a negative input of the error generator 430. The error generator 430 receives the actual measurement of the workload on the mobile device 100 from the sensor 223, the sensor 225, or the sensor 231 at the positive input of the error generator 430. In response, error generator 430 communicates an error signal (i.e., the difference between the measured or actual workload and the predicted workload) to controller 440 over connection 435, to one or more cross-correlators (e.g., The cross correlator 450, the cross correlator 452, and the cross correlator 458) are communicated to the time series storage 460 in the memory 112.

在替代實施例中,可藉由負載管理器實施一個以上工作負載預測器。舉例而言,第一工作負載預測器將預測使用狀況X以及Y但不預測使用狀況Z係可能的。當使用狀況Z與行動裝置100之重要使用狀況相關聯或以其他方式定義行動裝置100之重要使用狀況時,不同於第一工作負載預測器且能夠準確地預測使用狀況Z之第二工作負載預測器對負載管理器將為有用的。In an alternate embodiment, more than one workload predictor can be implemented by the load manager. For example, the first workload predictor will predict usage states X and Y but not predict usage status Z. When the usage condition Z is associated with or otherwise defines an important usage condition of the mobile device 100, the second workload prediction different from the first workload predictor and capable of accurately predicting the usage status Z The loader will be useful for the load manager.

控制器440包括產生用於選擇性地調整應用於核心222、核心224及核心230之參數中的一或多者之一或多個控制輸入的邏輯。控制器440可進一步包括產生用於選擇性地調整待與選擇DCFS演算法297一起應用之參數集合之一或多個控制輸入的邏輯,該選擇DCVS演算法297經應用以調整選擇核心處理器之頻率及電壓。如將在以下所解釋,該邏輯可對來自感測器405、感測器223、感測器225、感測器231之輸入、連接435上之誤差信號或其函數及/或來自N個交叉相關器之結果作出回應。感測器405、感測器223、感測器225、感測器231及(可能)額外感測器可提供軟體產生之量測(諸如,處理器閒置),或可提供對電壓、電流、溫度等之直接或間接量測的指示。Controller 440 includes logic to generate one or more control inputs for selectively adjusting one or more of the parameters applied to core 222, core 224, and core 230. The controller 440 can further include logic to generate one or more control inputs for selectively adjusting a set of parameters to be applied with the selected DCFS algorithm 297, the selected DCVS algorithm 297 being applied to adjust the selection of the core processor Frequency and voltage. As will be explained below, the logic can be for input from sensor 405, sensor 223, sensor 225, sensor 231, error signal on connection 435, or a function thereof, and/or from N intersections. The result of the correlator responds. Sensor 405, sensor 223, sensor 225, sensor 231, and (possibly) additional sensors may provide software generated measurements (such as processor idle), or may provide voltage, current, An indication of direct or indirect measurement of temperature, etc.

N個交叉相關器中之每一者可包括藉由行動裝置100在已知使用狀態下之操作所產生的誤差值之時間序列或將能夠存取該時間序列。在一些實施例中,效能按比例調整演算法與參數集合之經最佳化組合與該已知使用狀態相關聯,該經最佳化組合係先前識別為在節省電力方面最有效,同時針對經識別使用狀況維持行動裝置100上之所要效能等級。記憶體112包括時間序列儲存器460,該時間序列儲存器460用於在行動裝置100上保持針對效能按比例調整演算法與參數集合之可用組合中之每一者的誤差序列。或者,N個交叉相關器中之每一者可具有用於儲存各別誤差序列之內部緩衝器。在替代實施例(圖中未示)中,神經網路可替換N個交叉相關器。該神經網路將使用不同於交叉相關之技術來識別當前使用狀況。Each of the N cross-correlators may include a time series of error values generated by operation of the mobile device 100 in a known state of use or will be able to access the time series. In some embodiments, an optimized combination of a performance scaling algorithm and a set of parameters is associated with the known usage state, the optimized combination previously identified as being most effective in conserving power, while targeting The usage status is determined to maintain the desired level of performance on the mobile device 100. The memory 112 includes a time series store 460 for maintaining an error sequence for each of the available combinations of performance scaling algorithms and parameter sets on the mobile device 100. Alternatively, each of the N cross-correlators may have an internal buffer for storing individual error sequences. In an alternative embodiment (not shown), the neural network may replace N cross-correlators. The neural network will use a different technique than cross-correlation to identify current usage.

N個交叉相關器中之每一者產生對由工作負載預測器所產生之誤差歷程記錄已知信號之成功程度的各別量測。將交叉相關器之各別輸出傳達至偵測器480,該偵測器480經組態以選擇或識別最佳地匹配即時誤差信號之交叉相關器。舉例而言,偵測器480可包括識別當輸出在-1與1之間變化時具有最接近一之正規化輸出的交叉相關器的電路元件及/或邏輯。交叉相關為估計兩個系列相關之程度的標準方法。用於執行兩個時變信號之交叉相關分析的系統及方法為熟知的且為本發明之理解起見無需在本文中詳述。Each of the N cross-correlators produces a separate measure of the degree of success of the known signal for the error history generated by the workload predictor. The respective outputs of the cross correlators are communicated to a detector 480 that is configured to select or identify a cross correlator that best matches the immediate error signal. For example, detector 480 can include circuit elements and/or logic that identify a cross-correlator that has the closest normalized output when the output changes between -1 and 1. Cross-correlation is a standard method for estimating the extent to which two series are related. Systems and methods for performing cross-correlation analysis of two time-varying signals are well known and need not be detailed herein for the purposes of the present invention.

來自N個交叉相關器之結果的比較識別哪一先前已知之使用狀況為行動裝置100之當前操作狀態或使用的最佳適合匹配(fit match)。因此,行動裝置100當前可在未知組態或使用狀況下操作,且仍選擇並應用效能按比例調整演算法與參數集合之有效組合以節省電力,同時仍維持所要效能等級。偵測器480將控制信號傳達至負載管理器415中之控制器440。控制器440將控制輸入轉遞至各種核心以管理此等資源之操作。另外,控制器440經組態以在必要時選擇並應用效能按比例調整演算法297及合適參數集合298,從而調整行動裝置100上之處理資源的操作。A comparison of the results from the N cross-correlators identifies which previously known usage condition is the current operational state of the mobile device 100 or the best fit match used. Thus, the mobile device 100 can currently operate under unknown configuration or usage conditions, and still select and apply an effective combination of performance scaling algorithms and parameter sets to conserve power while still maintaining the desired performance level. The detector 480 communicates control signals to the controller 440 in the load manager 415. Controller 440 forwards control inputs to various cores to manage the operation of such resources. Additionally, controller 440 is configured to select and apply performance scaling algorithm 297 and appropriate parameter sets 298 as necessary to adjust the operation of processing resources on mobile device 100.

圖5為說明用於判定在圖4之行動裝置上操作之工作負載預測器的效益之方法500之實施例的流程圖。大體而言,將用於一使用狀況之工作負載預測器之效益或準確度判定為誤差信號之平方在整體使用狀況期間或在所要時間週期內的總和。使誤差項成平方,此係因為誤差信號在任何特定時間瞬時可為負或正。FIG. 5 is a flow chart illustrating an embodiment of a method 500 for determining the benefit of a workload predictor operating on the mobile device of FIG. In general, the benefit or accuracy of a workload predictor for a usage condition is determined as the sum of the square of the error signal during the overall usage condition or during the desired time period. The error term is squared because the error signal can be negative or positive at any particular time instant.

方法500以區塊502開始,在區塊502中,產生行動裝置100上之工作負載的即時量測。此量測可包括每秒處理之指令之數目、每瓦特處理之指令之數目,或任何數目個替代效能量測。在區塊504中,行動裝置100使用工作負載預測器產生對行動裝置100上之工作負載的估計。此後,在區塊506中,行動裝置100比較工作負載之即時量測與工作負載之估計以產生誤差信號。如在區塊508中所指示,隨時間記錄該誤差信號。此後,在區塊510中,產生誤差信號之函數以作為工作負載預測器之效益或準確度的量測。The method 500 begins with block 502, in which an instant measurement of the workload on the mobile device 100 is generated. This measurement may include the number of instructions processed per second, the number of instructions processed per watt, or any number of alternative energy measurements. In block 504, the mobile device 100 uses the workload predictor to generate an estimate of the workload on the mobile device 100. Thereafter, in block 506, the mobile device 100 compares the immediate measurement of the workload with the estimate of the workload to generate an error signal. As indicated in block 508, the error signal is recorded over time. Thereafter, in block 510, a function of the error signal is generated to measure the benefit or accuracy of the workload predictor.

在一些實施例中,將誤差之平方的總和計算為工作負載預測器之效益的量測且轉遞該量測。因此,誤差之平方之總和值愈小,則工作負載預測器在預測當前使用狀況之行動裝置100上之經量測工作負載方面愈準確。In some embodiments, the sum of the squares of the errors is calculated as a measure of the benefit of the workload predictor and the measurement is forwarded. Therefore, the smaller the sum of the squares of the errors, the more accurate the workload predictor is in measuring the measured workload on the mobile device 100 in the current state of use.

圖6為說明可在圖4之行動裝置上執行的方法600之一替代實施例的流程圖。方法600以區塊602開始,在區塊602中,產生行動裝置100上之實際工作負載的即時量測。此量測可包括每秒處理之指令之數目、消耗每瓦特功率所處理之指令之數目,或任何數目個替代效能量測。在區塊604中,行動裝置100使用工作負載預測器產生對行動裝置100上之工作負載的估計。此後,在區塊606中,行動裝置100比較工作負載之即時量測與工作負載之估計以產生誤差信號。在圖4中所說明之實施例中,負載管理器415計算實際或經量測工作負載與經預測工作負載的差。如在區塊608中所指示,計算該誤差信號之平方。此後,在區塊610中,將誤差信號之平方轉遞至控制器,該控制器回應於誤差信號之平方的量值而調整行動裝置100。在一些實施例(圖中未示)中,誤差信號之平方可藉由控制器進行低通濾波或以其他方式處理以維持負載管理器中之穩定性。6 is a flow chart illustrating an alternate embodiment of a method 600 that may be performed on the mobile device of FIG. The method 600 begins with block 602, in which an instant measurement of the actual workload on the mobile device 100 is generated. This measurement may include the number of instructions processed per second, the number of instructions processed per watt of power consumed, or any number of alternative energy measurements. In block 604, the mobile device 100 uses the workload predictor to generate an estimate of the workload on the mobile device 100. Thereafter, in block 606, the mobile device 100 compares the immediate measurement of the workload with the estimate of the workload to generate an error signal. In the embodiment illustrated in FIG. 4, load manager 415 calculates the difference between the actual or measured workload and the predicted workload. As indicated in block 608, the square of the error signal is calculated. Thereafter, in block 610, the square of the error signal is forwarded to the controller, which adjusts the mobile device 100 in response to the magnitude of the square of the error signal. In some embodiments (not shown), the square of the error signal can be low pass filtered or otherwise processed by the controller to maintain stability in the load manager.

圖7為說明可在圖4之行動裝置上執行的方法700之一替代實施例的流程圖。方法700以區塊702開始,在區塊702中,產生行動裝置100上之實際工作負載的即時量測。此量測可包括每秒處理之指令之數目、消耗每瓦特功率所處理之指令之數目,或任何數目個替代效能量測。在區塊704中,行動裝置100使用工作負載預測器產生對行動裝置100上之工作負載的估計。此後,在區塊706中,行動裝置100比較工作負載之即時量測與工作負載之估計以產生誤差信號。在圖4中所說明之實施例中,負載管理器415計算實際或經量測工作負載與經預測工作負載的差。如在區塊708中所指示,隨時間記錄該誤差信號。此後,在區塊710中,產生誤差信號之函數作為工作負載預測器預測行動裝置100上之工作負載之準確度的量測。在區塊712中,隨時間所記錄之誤差信號與一使用狀況相關聯。當先前已使該使用狀況與針對該使用狀況而最佳化之效能按比例調整演算法與參數集合之組合相關聯時,負載管理器415可以可控制方式選擇並應用該效能按比例調整演算法及該參數集合以調整行動裝置100之操作。FIG. 7 is a flow diagram illustrating an alternate embodiment of a method 700 that may be performed on the mobile device of FIG. The method 700 begins with block 702, in which an instant measurement of the actual workload on the mobile device 100 is generated. This measurement may include the number of instructions processed per second, the number of instructions processed per watt of power consumed, or any number of alternative energy measurements. In block 704, the mobile device 100 uses the workload predictor to generate an estimate of the workload on the mobile device 100. Thereafter, in block 706, the mobile device 100 compares the immediate measurement of the workload with the estimate of the workload to generate an error signal. In the embodiment illustrated in FIG. 4, load manager 415 calculates the difference between the actual or measured workload and the predicted workload. As indicated in block 708, the error signal is recorded over time. Thereafter, in block 710, a function of the error signal is generated as a measure of the accuracy of the workload predictor predicting the workload on the mobile device 100. In block 712, the error signal recorded over time is associated with a usage condition. The load manager 415 can control select and apply the performance scaling algorithm when the usage condition has previously been associated with a combination of performance scaling and parameter sets optimized for the usage condition. And the set of parameters to adjust the operation of the mobile device 100.

圖8為說明用於管理圖4之行動裝置上的工作負載之方法的流程圖。方法800以區塊802開始,在區塊802中,產生行動裝置100上之實際工作負載的即時量測。此量測可包括每秒處理之指令之數目、消耗每瓦特功率所處理之指令之數目,或任何數目個替代效能量測。在區塊804中,行動裝置100使用工作負載預測器產生對行動裝置100上之工作負載的估計。此後,在區塊806中,行動裝置100比較工作負載之即時量測與工作負載之估計以產生誤差信號。在圖4中所說明之實施例中,負載管理器415計算實際或經量測工作負載與經預測工作負載的差。8 is a flow chart illustrating a method for managing a workload on a mobile device of FIG. The method 800 begins with block 802, in which an instant measurement of the actual workload on the mobile device 100 is generated. This measurement may include the number of instructions processed per second, the number of instructions processed per watt of power consumed, or any number of alternative energy measurements. In block 804, the mobile device 100 uses the workload predictor to generate an estimate of the workload on the mobile device 100. Thereafter, in block 806, the mobile device 100 compares the immediate measurement of the workload with the estimate of the workload to generate an error signal. In the embodiment illustrated in FIG. 4, load manager 415 calculates the difference between the actual or measured workload and the predicted workload.

如在區塊808中所指示,隨時間記錄該誤差信號。此後,在區塊810中,產生誤差信號之函數作為工作負載預測器預測行動裝置100上之工作負載之準確度的量測。在區塊812中,比較即時誤差信號之交叉相關的量測與先前記錄之誤差信號以判定當前使用狀況是否類似於已知使用狀況。當交叉相關之量測指示該誤差信號類似於已知使用狀況且該使用狀況已與效能按比例調整演算法與參數集合之經最佳化組合相關聯時,可藉由經合適組態之負載管理器415來選擇性地應用該組合,如在區塊814中所指示。The error signal is recorded over time as indicated in block 808. Thereafter, in block 810, a function of the error signal is generated as a measure of the accuracy of the workload predictor predicting the workload on the mobile device 100. In block 812, the cross-correlation of the immediate error signal is compared to the previously recorded error signal to determine if the current usage condition is similar to the known usage condition. When the cross-correlation measurement indicates that the error signal is similar to the known usage condition and the usage condition has been associated with an optimized combination of the performance scaling algorithm and the parameter set, the load can be configured by a suitable configuration. Manager 415 to selectively apply the combination, as indicated in block 814.

圖9為說明用於量測在圖1之行動裝置100上操作的工作負載預測器之效益的系統之一替代實施例的示意圖。一電腦程式產品可涵蓋以上結合圖5至圖8中說明之方法所描述之功能性。跨越圖9中所展示之各種邏輯模組而散佈適合於在電腦程式產品上或在行動裝置100上之記憶體112內儲存的可執行指令之邏輯模組或集合。大體上以與以上結合圖2所描述之系統200相同的方式組態系統900。如在圖9中所指示,系統900包括用於實施用於量化行動裝置100上之工作負載預測器的效益之線上或裝置上系統的額外軟體或邏輯模組。該線上系統包括誤差產生邏輯930,該誤差產生邏輯930評估如由工作負載邏輯910所提供的經量測工作負載與如藉由工作負載估計器邏輯920所判定的工作負載估計之間的差,且將該差轉遞至工作負載效益邏輯970及轉遞至時間序列儲存器460。9 is a schematic diagram illustrating an alternate embodiment of a system for measuring the benefits of a workload predictor operating on the mobile device 100 of FIG. A computer program product may encompass the functionality described above in connection with the methods illustrated in Figures 5-8. Logic modules or sets of executable instructions suitable for storage in a computer program product or in memory 112 on mobile device 100 are distributed across the various logic modules shown in FIG. System 900 is generally configured in the same manner as system 200 described above in connection with FIG. As indicated in FIG. 9, system 900 includes additional software or logic modules for implementing an on-line or on-device system for quantifying the benefits of the workload predictor on mobile device 100. The on-line system includes error generation logic 930 that evaluates the difference between the measured workload as provided by workload logic 910 and the workload estimate as determined by workload estimator logic 920, The difference is forwarded to the workload benefit logic 970 and forwarded to the time series storage 460.

如以上所指示,控制邏輯950可接收藉由工作負載效益邏輯970所判定的工作負載預測器之效益或準確度的量測,且將該量測應用於一或多個用於判定對行動裝置上之操作參數之所要校正的函數。此為工作負載預測器輸出之應用的實例,該工作負載預測器輸出經傳達至控制器以用於調整行動裝置100上之一或多個操作參數。如以上所描述,工作負載效益邏輯970可計算誤差信號之量值之平方在所要時間週期上的總和。誤差信號之量值之平方的總和愈接近零,則工作負載預測器在預測行動裝置100上之工作負載方面愈有效或愈準確。As indicated above, control logic 950 can receive a measure of the benefit or accuracy of the workload predictor as determined by workload benefit logic 970 and apply the measure to one or more for determining the pair of mobile devices The function to be corrected for the operating parameters above. This is an example of an application of a workload predictor output that is communicated to a controller for adjusting one or more operational parameters on the mobile device 100. As described above, workload benefit logic 970 can calculate the sum of the magnitudes of the magnitudes of the error signals over a desired time period. The closer the sum of the squares of the magnitudes of the error signals is to zero, the more efficient or accurate the workload predictor is in predicting the workload on the mobile device 100.

儘管描述為單獨程式或模組,但工作負載邏輯910、工作負載估計器邏輯920、誤差產生邏輯930、交叉相關邏輯940、控制邏輯950、時間序列管理邏輯960及工作負載效益邏輯970中之每一者可彼此組合及/或與先前引入之邏輯模組組合,該等先前引入之邏輯模組儲存於記憶體112內或儲存於可由數位信號處理器110存取之其他資料儲存器中。Although described as a separate program or module, workload logic 910, workload estimator logic 920, error generation logic 930, cross-correlation logic 940, control logic 950, time series management logic 960, and workload benefit logic 970 are each One may be combined with each other and/or with a previously introduced logic module that is stored in memory 112 or stored in other data stores accessible by digital signal processor 110.

行動裝置100可經組態以監視誤差項達所要持續時間以開發有關終端使用者實際上如何使用裝置之知識庫。系統900使用交叉相關技術來匹配誤差項時間序列,以判定對於裝置之當前使用而言,效能按比例調整演算法與儲存於行動裝置100上之參數集合之可用組合中的哪一組合對於節省電力最為適當。因為誤差項為由於饋送通過工作負載預測器之系統輸入而產生的經預測工作負載及裝置上之工作負載之實際量測的函數,所以誤差項之量值為工作負載預測器預測裝置上之未來工作負載之效益的定量量測。因此,隨著誤差信號之平方的積分隨時間接近於零,工作負載預測器在預測裝置上之未來使用/事件方面更為準確。The mobile device 100 can be configured to monitor the error term for a desired duration to develop a knowledge base about how the end user actually uses the device. System 900 uses cross-correlation techniques to match the error term time series to determine which combination of the performance scaling algorithm and the set of parameters stored on the mobile device 100 for the current use of the device is to conserve power Most appropriate. Since the error term is a function of the predicted workload generated by the system input through the workload predictor and the actual measurement of the workload on the device, the magnitude of the error term is the future on the workload predictor prediction device. Quantitative measurement of the benefits of the workload. Thus, as the integral of the square of the error signal approaches zero over time, the workload predictor is more accurate in predicting future use/events on the device.

舉例而言,工作負載邏輯910包括用於產生行動裝置100上之工作負載之即時量測的一或多個可執行指令。經量測工作負載將通常與行動裝置100上所關注之選擇處理器核心相關聯。然而,工作負載邏輯910並不限於量測單一處理器核心上之工作負載。大體而言,工作負載邏輯910經由數位信號處理器110結合監視器模組114或結合行動裝置100上之一或多個感測器起作用,以判定行動裝置100上之當前工作負載。如以上所描述,該等感測器可皆為量測電流、電壓、溫度等之實體感測器,或該等感測器可(諸如)藉由監視緩衝器或軟體堆疊之內容而以軟體啟用。For example, workload logic 910 includes one or more executable instructions for generating an instant measurement of a workload on mobile device 100. The measured workload will typically be associated with a selected processor core of interest on the mobile device 100. However, workload logic 910 is not limited to measuring workloads on a single processor core. In general, the workload logic 910 functions in conjunction with the monitor module 114 or in conjunction with one or more sensors on the mobile device 100 via the digital signal processor 110 to determine the current workload on the mobile device 100. As described above, the sensors can all be physical sensors that measure current, voltage, temperature, etc., or the sensors can be software-like, such as by monitoring the contents of the buffer or software stack. Enabled.

工作負載估計器邏輯920應用演算法以產生行動裝置100上之工作負載的估計。將經估計工作負載及經量測工作負載轉遞至誤差產生邏輯930,該誤差產生邏輯930依據經量測工作負載及經估計工作負載產生誤差信號。如圖9中所說明,可使經量測工作負載或經估計工作負載中之一者數位反向(亦即,乘以-1)且添加至剩餘信號。或者,可自其他信號減去經量測工作負載及經估計工作負載中之一者以產生誤差信號。此外,亦如以上所描述,可在選擇時間週期上使誤差信號成平方且對其進行評估。The workload estimator logic 920 applies an algorithm to generate an estimate of the workload on the mobile device 100. The estimated workload and the measured workload are forwarded to error generation logic 930, which generates an error signal based on the measured workload and the estimated workload. As illustrated in Figure 9, one of the measured workload or the estimated workload may be reversed (i.e., multiplied by -1) and added to the remaining signals. Alternatively, one of the measured workload and the estimated workload may be subtracted from other signals to generate an error signal. Moreover, as also described above, the error signal can be squared and evaluated over a selected time period.

時間序列管理邏輯960結合時間序列儲存器460起作用以隨時間記錄誤差信號。當已知時,控制輸入或識別行動裝置100之當前狀態之其他資料的當前集合經識別,且該當前集合之指示與所儲存之時間序列相關聯。Time series management logic 960 acts in conjunction with time series storage 460 to record error signals over time. When known, the current set of other materials that control input or identify the current state of the mobile device 100 are identified, and the indication of the current set is associated with the stored time series.

控制邏輯950包括用於初始化、載入或以其他方式管理實現線上系統之各種邏輯模組的一或多個可執行指令。舉例而言,控制邏輯950產生一或多個控制輸入,該一或多個控制輸入用於選擇性地調整參數集合298中之參數中的一或多者及/或經應用以調整選擇核心處理器之頻率及電壓的效能按比例調整演算法297。控制邏輯950可對以下各者作出回應:來自監視器模組114之輸入、由誤差產生邏輯930所產生之誤差信號之平方的量值及/或來自交叉相關邏輯940之結果或自相關器結果儲存器470擷取之結果。Control logic 950 includes one or more executable instructions for initializing, loading, or otherwise managing various logic modules implementing an online system. For example, control logic 950 generates one or more control inputs for selectively adjusting one or more of the parameters in parameter set 298 and/or applied to adjust selection core processing. The frequency and voltage performance of the device is scaled to algorithm 297. Control logic 950 can respond to the input from monitor module 114, the magnitude of the square of the error signal produced by error generation logic 930, and/or the result from cross-correlation logic 940 or the autocorrelator result. The result of the memory 470 is retrieved.

交叉相關邏輯940包括用於執行兩個時間序列之交叉相關分析以定量地判定即時誤差跡線與先前儲存之誤差跡線之間的最佳匹配之一或多個可執行指令。來自交叉相關邏輯940之結果可儲存於相關器結果儲存器470中以供控制邏輯950及或交叉相關邏輯940擷取及分析。以此方式,控制邏輯950可儲存藉由評估誤差信號而特徵化行動裝置上之操作之當前狀態的任何數目個指示符。當針對即時誤差項識別到相對緊密之匹配且該相對緊密之匹配已與可用效能按比例調整演算法與參數集合之最有效組合相關聯時,控制邏輯950可識別如藉由誤差信號所識別之操作狀態或使如藉由誤差信號所識別之操作狀態與該組合相關聯。此後,每當行動裝置之操作者以與由誤差信號及(可能)一或多個額外指示符所識別之方式相同或類似之方式使用該裝置時,控制邏輯950即可應用可用效能按比例調整演算法與參數集合之最有效組合以節省行動裝置100上之電力。此外,可產生並監視誤差信號以提供行動裝置正在可用之最有效模式下操作同時維持所要效能等級的即時驗證。Cross-correlation logic 940 includes one or more executable instructions for performing a cross-correlation analysis of two time series to quantitatively determine the best match between the immediate error trace and the previously stored error trace. The results from cross-correlation logic 940 can be stored in correlator result store 470 for control logic 950 and or cross-correlation logic 940 to retrieve and analyze. In this manner, control logic 950 can store any number of indicators that characterize the current state of operation on the mobile device by evaluating the error signal. When a relatively tight match is identified for an immediate error term and the relatively close match has been associated with the most efficient combination of available performance scaling algorithms and parameter sets, control logic 950 can identify the identification as identified by the error signal The operational state or associated operational state as identified by the error signal is associated with the combination. Thereafter, whenever the operator of the mobile device uses the device in the same or similar manner as identified by the error signal and (possibly) one or more additional indicators, the control logic 950 can apply the available performance scaled The most efficient combination of algorithms and parameter sets to save power on the mobile device 100. In addition, error signals can be generated and monitored to provide instant verification that the mobile device is operating in the most efficient mode available while maintaining the desired performance level.

如圖9中所展示,當以軟體來實施由行動裝置100所使用之邏輯時,應注意,工作負載邏輯910、工作負載估計器邏輯920、誤差產生邏輯930、交叉相關邏輯940、控制邏輯950、時間序列管理邏輯960及工作負載效益邏輯970中之一或多者可儲存於任何電腦可讀媒體上以供任何電腦相關系統或方法使用或結合任何電腦相關系統或方法使用。As shown in FIG. 9, when the logic used by the mobile device 100 is implemented in software, it should be noted that the workload logic 910, the workload estimator logic 920, the error generation logic 930, the cross-correlation logic 940, the control logic 950 One or more of time series management logic 960 and workload benefit logic 970 can be stored on any computer readable medium for use by or in connection with any computer related system or method.

在以硬體來實施工作負載邏輯910、工作負載估計器邏輯920、誤差產生邏輯930、交叉相關邏輯940、控制邏輯950、時間序列管理邏輯960及工作負載效益邏輯970中之一或多者的替代實施例中,可藉由各自為此項技術中已知之以下技術中之任一者或組合來實施各種功能:具有用於對資料信號實施邏輯功能之邏輯閘的離散邏輯電路、具有適當組合邏輯閘之特殊應用積體電路(ASIC)、可程式化閘陣列(PGA)、場可程式化閘陣列(FPGA)等。One or more of workload logic 910, workload estimator logic 920, error generation logic 930, cross-correlation logic 940, control logic 950, time series management logic 960, and workload benefit logic 970 are implemented in hardware. In alternative embodiments, various functions may be implemented by any one or combination of the following techniques known in the art: discrete logic circuits having logic gates for performing logic functions on data signals, with appropriate combinations Logic Gate Special Application Integrated Circuit (ASIC), Programmable Gate Array (PGA), Field Programmable Gate Array (FPGA), etc.

本說明書中所描述之程序或程序流程中的特啶步驟本質上先於其他步驟以使本發明如所描述地起作用。然而,若所描述之步驟的次序或序列不變更本發明之功能性,則本發明不限於此次序。亦即,應認識到,在不脫離本發明之範疇及精神的情況下,一些步驟可在其他步驟之前、在其他區塊之後或與其他區塊並列(實質上同時)地執行。在一些例子中,在不脫離本發明之情況下,可省略或不執行特定步驟。另外,諸如「此後」、「接著」、「接下來」等的詞並不意欲限制步驟之次序。此等詞僅用以指導讀者閱讀例示性方法之描述。The procedures in the procedures or procedures described in this specification are essentially preceded by other steps to make the invention function as described. However, the present invention is not limited to this order unless the order or sequence of steps described does not change the functionality of the present invention. That is, it will be appreciated that some steps may be performed before other steps, after other blocks, or juxtaposed (substantially simultaneously) with other blocks, without departing from the scope and spirit of the invention. In some instances, specific steps may be omitted or not performed without departing from the invention. In addition, words such as "after", "continued", "next", etc. are not intended to limit the order of the steps. These terms are only used to guide the reader in reading the description of the exemplary methods.

一般熟習程式化技術者能夠在無過多困難或實驗的情況下撰寫電腦程式碼或識別用以實施所揭示之程序的適當硬體。因此,為達成對如何製作及使用本發明之適當理解,未必將可執行指令或詳述硬體裝置之特定集合的揭示內容視為必要的。在以上描述中且結合諸圖(其可說明各種程序流程)較詳細地解釋了所主張之行動裝置100及程序的本發明功能性。Those who are familiar with stylized techniques can write computer code or identify appropriate hardware for implementing the disclosed program without undue difficulty or experimentation. Thus, the disclosure of specific sequences of executable instructions or detailed hardware devices is not necessarily deemed necessary to achieve a proper understanding of how to make and use the invention. The functionality of the present invention of the claimed mobile device 100 and program is explained in greater detail in the above description in conjunction with the figures (which may illustrate various program flows).

在一或多個例示性態樣中,所描述之功能可以硬體、軟體、韌體或其任何組合來實施。若以軟體實施,則可將該等功能作為一或多個指令或程式碼而儲存於電腦可讀媒體上或在電腦可讀媒體上傳輸。電腦可讀媒體包括電腦儲存媒體及通信媒體兩者,通信媒體包括促進電腦程式自一處至另一處之傳送的任何媒體。儲存媒體可為可由電腦存取之任何可用媒體。藉由實例且非限制,此等電腦可讀媒體可包含RAM、ROM、EEPROM、CD-ROM或其他光碟儲存器、磁碟儲存器或其他磁性儲存裝置,或可用以攜載或儲存呈指令或資料結構形式之所要程式碼且可藉由電腦存取的任何其他媒體。In one or more exemplary aspects, the functions described can be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functions may be stored as one or more instructions or code on a computer readable medium or transmitted on a computer readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transmission of the computer program from one location to another. The storage medium can be any available media that can be accessed by a computer. By way of example and not limitation, such computer-readable media can comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage device, or can be used to carry or store instructions or Any other medium in the form of the data structure and accessible by the computer.

又,將任何連接恰當地稱為電腦可讀媒體。舉例而言,若使用同軸電纜、光纖纜線、雙絞線、數位用戶線(「DSL」)或諸如紅外線、無線電及微波之無線技術而自網站、伺服器或其他遠端源傳輸軟體,則同軸電纜、光纖纜線、雙絞線、DSL或諸如紅外線、無線電及微波之無線技術包括於媒體之定義中。Also, any connection is properly termed a computer-readable medium. For example, if you use a coaxial cable, fiber optic cable, twisted pair cable, digital subscriber line ("DSL"), or wireless technology such as infrared, radio, and microwave to transmit software from a website, server, or other remote source, then Coaxial cables, fiber optic cables, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of the media.

如本文中所使用,磁碟及光碟包括緊密光碟(「CD」)、雷射光碟、光碟、數位影音光碟(「DVD」)、軟性磁碟及藍光光碟,其中磁碟通常以磁性方式再生資料,而光碟藉由雷射以光學方式再生資料。上述各物之組合亦應包括於電腦可讀媒體之範疇內。As used herein, magnetic disks and optical disks include compact discs ("CDs"), laser compact discs, optical discs, digital audio and video discs ("DVDs"), flexible magnetic discs and Blu-ray discs, where the magnetic discs are usually magnetically regenerated. The optical disc optically reproduces data by laser. Combinations of the above should also be included in the context of computer readable media.

儘管已詳細說明並描述了選定態樣,但應理解,在不脫離如藉由以下申請專利範圍所界定之本發明之範疇的情況下,可在其中進行各種取代及變更。While the invention has been described and illustrated in detail, it is understood that various modifications and changes can be made therein without departing from the scope of the invention as defined by the appended claims.

100...行動裝置100. . . Mobile device

102...系統單晶片102. . . System single chip

110...數位信號處理器110. . . Digital signal processor

111...匯流排111. . . Busbar

112...記憶體/記憶體元件112. . . Memory/memory component

114...監視器模組114. . . Monitor module

126...類比信號處理器126. . . Analog signal processor

128...顯示控制器128. . . Display controller

130...觸控螢幕控制器130. . . Touch screen controller

132...觸控螢幕顯示器132. . . Touch screen display

134...視訊編碼解碼器/視訊編碼其134. . . Video codec/video coding

136...視訊放大器136. . . Video amplifier

138...視訊埠138. . . Video

140...通用串列匯流排(「USB」)控制器140. . . Universal Serial Bus ("USB") Controller

142...通用串列匯流排(USB)埠142. . . Universal Serial Bus (USB)埠

146...用戶識別模組(SIM)卡146. . . Subscriber Identity Module (SIM) card

148...數位相機148. . . Digital camera

150...立體聲音訊編碼解碼器(CODEC)150. . . Stereo audio codec (CODEC)

152...音訊放大器152. . . Audio amplifier

154...第一立體聲揚聲器154. . . First stereo speaker

156...第二立體聲揚聲器156. . . Second stereo speaker

158...麥克風放大器158. . . Microphone amplifier

160...麥克風160. . . microphone

162...頻率調變(「FM」)無線電調諧器162. . . Frequency Modulation ("FM") Radio Tuner

164...頻率調變(FM)天線164. . . Frequency modulation (FM) antenna

166...立體聲頭戴式耳機166. . . Stereo headset

168...射頻(「RF」)收發器168. . . Radio Frequency ("RF") Transceiver

170...射頻(RF)開關170. . . Radio frequency (RF) switch

172...射頻(RF)天線172. . . Radio frequency (RF) antenna

174...小鍵盤174. . . Keypad

176...具有麥克風之單聲道耳機176. . . Mono headset with microphone

178...振動器裝置/振動器178. . . Vibrator device / vibrator

180...電源供應器180. . . Power Supplier

200...系統200. . . system

210...核心處理器/核心210. . . Core processor/core

222...第一核心/處理器核心222. . . First core/processor core

223...感測器223. . . Sensor

224...第二核心/處理器核心224. . . Second core/processor core

225...感測器225. . . Sensor

230...第N核心/處理器核心230. . . Nth core/processor core

231...感測器231. . . Sensor

250...起動邏輯250. . . Start logic

260...管理邏輯260. . . Management logic

270...動態電壓及頻率按比例調整(DVFS)介面邏輯270. . . Dynamic voltage and frequency scaling (DVFS) interface logic

280...應用程式儲存器280. . . Application storage

290...嵌入式檔案系統290. . . Embedded file system

292...動態電壓及頻率按比例調整(DVFS)儲存器292. . . Dynamic voltage and frequency scaling (DVFS) storage

294...核心儲存器294. . . Core storage

296...程式儲存器296. . . Program storage

297...效能按比例調整演算法/DCVS演算法297. . . Performance Proportional Algorithm/DCVS Algorithm

298...參數/參數集合298. . . Parameter/parameter set

300...表300. . . table

350...表350. . . table

400...用於量測工作負載預測器之效益的系統400. . . System for measuring the benefits of a workload predictor

405...感測器405. . . Sensor

410...數位信號處理器(DSP)410. . . Digital signal processor (DSP)

415...負載管理器415. . . Load manager

420...工作負載預測器420. . . Workload predictor

430...誤差產生器430. . . Error generator

435...連接435. . . connection

440...控制器440. . . Controller

450...交叉相關器450. . . Cross correlator

452...交叉相關器452. . . Cross correlator

458...交叉相關器458. . . Cross correlator

460...時間序列儲存器460. . . Time series storage

470...相關器結果儲存器470. . . Correlator result store

480...偵測器480. . . Detector

500...用於判定在行動裝置上操作之工作負載預測器之效益的方法500. . . Method for determining the benefit of a workload predictor operating on a mobile device

600...可在行動裝置上執行之方法600. . . Method that can be performed on a mobile device

700...可在行動裝置上執行之方法700. . . Method that can be performed on a mobile device

800...用於管理行動裝置上之工作負載之方法800. . . Method for managing workloads on mobile devices

900...用於量測在行動裝置100上操作之工作負載預測器之效益的系統900. . . System for measuring the benefit of a workload predictor operating on mobile device 100

910...工作負載邏輯910. . . Workload logic

920...工作負載估計器邏輯920. . . Workload estimator logic

930...誤差產生邏輯930. . . Error generation logic

940...交叉相關邏輯940. . . Cross correlation logic

950...控制邏輯950. . . Control logic

960...時間序列管理邏輯960. . . Time series management logic

970...工作負載效益邏輯970. . . Workload benefit logic

圖1為說明行動裝置之實施例的示意圖。1 is a schematic diagram illustrating an embodiment of a mobile device.

圖2為說明圖1之行動裝置之特徵的示意圖。FIG. 2 is a schematic view showing the features of the mobile device of FIG. 1. FIG.

圖3A及圖3B包括說明使用狀況之實施例之示意圖。3A and 3B include schematic diagrams illustrating an embodiment of a use condition.

圖4為說明可由圖1之行動裝置實施的用於管理一工作負載之系統之實施例的示意圖。4 is a schematic diagram illustrating an embodiment of a system for managing a workload that may be implemented by the mobile device of FIG. 1.

圖5為說明用於判定在圖4之行動裝置上操作的工作負載預測器之效益的方法之實施例的流程圖。5 is a flow chart illustrating an embodiment of a method for determining the benefit of a workload predictor operating on the mobile device of FIG.

圖6為說明可在圖4之行動裝置上執行的方法之一替代實施例的流程圖。6 is a flow chart illustrating an alternate embodiment of a method that may be performed on the mobile device of FIG.

圖7為說明可在圖4之行動裝置上執行的方法之另一替代實施例的流程圖。7 is a flow chart illustrating another alternate embodiment of a method that can be performed on the mobile device of FIG.

圖8為說明用於管理圖4之行動裝置上之工作負載的方法的流程圖。8 is a flow chart illustrating a method for managing a workload on a mobile device of FIG.

圖9為說明用於量測在圖1之行動裝置上操作的工作負載預測器之效益的系統之一替代實施例的示意圖。9 is a schematic diagram illustrating an alternate embodiment of a system for measuring the benefits of a workload predictor operating on the mobile device of FIG. 1.

100...行動裝置100. . . Mobile device

102...系統單晶片102. . . System single chip

110...數位信號處理器110. . . Digital signal processor

112...記憶體/記憶體元件112. . . Memory/memory component

114...監視器模組114. . . Monitor module

126...類比信號處理器126. . . Analog signal processor

128...顯示控制器128. . . Display controller

130...觸控螢幕控制器130. . . Touch screen controller

132...觸控螢幕顯示器132. . . Touch screen display

134...視訊編碼解碼器/視訊編碼器134. . . Video codec/video encoder

136...視訊放大器136. . . Video amplifier

138...視訊埠138. . . Video

140...通用串列匯流排(「USB」)控制器140. . . Universal Serial Bus ("USB") Controller

142...通用串列匯流排(USB)埠142. . . Universal Serial Bus (USB)埠

146...用戶識別模組(SIM)卡146. . . Subscriber Identity Module (SIM) card

148...數位相機148. . . Digital camera

150...立體聲音訊編碼解碼器(CODEC)150. . . Stereo audio codec (CODEC)

152...音訊放大器152. . . Audio amplifier

154...第一立體聲揚聲器154. . . First stereo speaker

156...第二立體聲揚聲器156. . . Second stereo speaker

158...麥克風放大器158. . . Microphone amplifier

160...麥克風160. . . microphone

162...頻率調變(「FM」)無線電調諧器162. . . Frequency Modulation ("FM") Radio Tuner

164...頻率調變(FM)天線164. . . Frequency modulation (FM) antenna

166...立體聲頭戴式耳機166. . . Stereo headset

168...射頻(「RF」)收發器168. . . Radio Frequency ("RF") Transceiver

170...射頻(RF)開關170. . . Radio frequency (RF) switch

172...射頻(RF)天線172. . . Radio frequency (RF) antenna

174...小鍵盤174. . . Keypad

176...具有麥克風之單聲道耳機176. . . Mono headset with microphone

178...振動器裝置/振動器178. . . Vibrator device / vibrator

180...電源供應器180. . . Power Supplier

Claims (20)

一種用於量測一行動裝置上之一工作負載預測器之效益的方法,該方法包含:產生該行動裝置上之工作負載的一即時量測;使用一工作負載預測器產生該行動裝置上之該工作負載的一估計;比較該工作負載之該即時量測與該行動裝置上之該工作負載之該估計以產生一誤差信號;隨時間記錄該誤差信號;及產生該誤差信號之一函數作為該工作負載預測器之該效益的一量測。A method for measuring the benefit of a workload predictor on a mobile device, the method comprising: generating an instant measurement of a workload on the mobile device; generating a mobile device using a workload predictor An estimate of the workload; comparing the immediate measurement of the workload with the estimate of the workload on the mobile device to generate an error signal; recording the error signal over time; and generating a function of the error signal as A measure of the benefit of the workload predictor. 如請求項1之方法,其中使用該誤差信號之一函數包含計算該誤差信號之平方在一選擇週期上之一積分。The method of claim 1, wherein using one of the error signals comprises calculating an integral of the square of the error signal over a selection period. 如請求項1之方法,其進一步包含:將該誤差信號施加至一控制器,該控制器回應於該誤差信號而調整至少一參數。The method of claim 1, further comprising: applying the error signal to a controller, the controller adjusting at least one parameter in response to the error signal. 如請求項1之方法,其進一步包含:使該誤差信號與一作用中效能按比例調整演算法與至少一參數之組合相關聯。The method of claim 1, further comprising: correlating the error signal with an active performance scaling algorithm and a combination of at least one parameter. 如請求項1之方法,其進一步包含:使用該誤差信號之交叉相關的一量測以判定該行動裝置之一當前操作狀態是否類似於一已知使用狀況。The method of claim 1, further comprising: using a cross-correlation of the error signal to determine whether a current operational state of one of the mobile devices is similar to a known usage condition. 如請求項5之方法,其進一步包含:在該行動裝置經如此組態時應用一效能按比例調整演算法與一或多個參數之一組合。The method of claim 5, further comprising: applying a performance scaling algorithm in combination with one of the one or more parameters when the mobile device is so configured. 如請求項1之方法,其中該行動裝置包含:一蜂巢式電話、一傳呼器、一個人數位助理、一閱讀器、一衛星電話、一導航裝置或一具有一無線連接之電腦中之一者。The method of claim 1, wherein the mobile device comprises: a cellular phone, a pager, a number of assistants, a reader, a satellite phone, a navigation device or a computer having a wireless connection. 一種用於在一行動裝置上執行之負載管理器,其包含:一工作負載預測器,其經組態以產生在該行動裝置上操作之一處理器核心上之工作負載的一估計;一感測器,其經組態以產生在該行動裝置上操作之該處理器核心上之該工作負載的一量測;一誤差產生器,其經組態以接收該工作負載之該估計及該工作負載之該量測且產生一誤差信號;及一控制器,其經配置以接收該誤差信號且調整該處理器核心之操作。A load manager for execution on a mobile device, comprising: a workload predictor configured to generate an estimate of a workload on a processor core operating on the mobile device; a detector configured to generate a measurement of the workload on the processor core operating on the mobile device; an error generator configured to receive the estimate of the workload and the operation The measurement of the load produces an error signal; and a controller configured to receive the error signal and adjust the operation of the processor core. 如請求項8之負載管理器,其中該工作負載預測器係對一效能按比例調整演算法作出回應。The load manager of claim 8, wherein the workload predictor is responsive to a performance scaling algorithm. 如請求項8之負載管理器,其中該誤差信號之平方之量值用以判定該工作負載預測器預測該工作負載之準確度。The load manager of claim 8, wherein the magnitude of the square of the error signal is used to determine the accuracy of the workload predictor predicting the workload. 如請求項8之負載管理器,其進一步包含:一時間序列儲存器,其對該誤差信號作出回應。The load manager of claim 8, further comprising: a time series store that responds to the error signal. 如請求項11之負載管理器,其進一步包含:一交叉相關器,其耦接至該時間序列儲存器,該交叉相關器經組態以產生該誤差信號與相關聯於該行動裝置之一經識別使用之一誤差信號的相似性的一量測。The load manager of claim 11, further comprising: a cross correlator coupled to the time series storage, the cross correlator configured to generate the error signal and associated with one of the mobile devices identified A measure of the similarity of one of the error signals is used. 一種用於管理一行動裝置上之一負載的系統,該系統包含:用於量測在該行動裝置上操作之一處理器核心之一工作負載的構件;用於產生在該行動裝置上操作之該處理器核心的該工作負載之一估計的構件;用於產生對該工作負載及該工作負載之該估計作出回應之一誤差信號的構件;用於隨時間儲存該誤差信號的構件;用於使一隨時間所儲存之誤差信號與該行動裝置之一當前使用狀態相關聯的構件;及用於回應於該隨時間所儲存之誤差信號而選擇性地調整該行動裝置的構件。A system for managing a load on a mobile device, the system comprising: means for measuring a workload of one of the processor cores operating on the mobile device; for generating a operation on the mobile device An estimated component of the workload of the processor core; means for generating an error signal responsive to the workload and the estimate of the workload; means for storing the error signal over time; Means for correlating an error signal stored over time with a current state of use of one of the mobile devices; and means for selectively adjusting the mobile device in response to the error signal stored over time. 如請求項13之系統,其中用於產生該工作負載之一估計的該構件包含一工作負載預測器。The system of claim 13, wherein the means for generating an estimate of the workload comprises a workload predictor. 如請求項13之系統,其中用於選擇性地調整之該構件係對一交叉相關器及一偵測器作出回應。The system of claim 13 wherein the means for selectively adjusting is responsive to a cross correlator and a detector. 一種電腦程式產品,其包含一電腦可用媒體,該電腦可用媒體具有體現於其中之一電腦可讀程式碼,該電腦可讀程式碼經調適以經執行而實施用於識別在一行動裝置上操作之一工作負載預測器之效益的一方法,該電腦程式產品包含:經組態以產生該行動裝置上之工作負載之一即時量測的邏輯;經組態以產生該行動裝置上之該工作負載之一估計的邏輯;經組態以比較該工作負載之該即時量測與該行動裝置上之該工作負載之該估計以產生一誤差信號的邏輯;經組態以隨時間記錄該誤差信號之邏輯;經組態以計算該誤差信號之一函數的邏輯;及經組態以回應於該誤差信號之該函數而判定該工作負載預測器之該效益的邏輯。A computer program product comprising a computer usable medium having a computer readable program code embodied therein, the computer readable program code being adapted to be executed for identification to operate on a mobile device A method of benefiting a workload predictor, the computer program product comprising: logic configured to generate an immediate measurement of a workload on the mobile device; configured to generate the work on the mobile device Logic for estimating one of the loads; logic configured to compare the estimate of the workload with the estimate of the workload on the mobile device to generate an error signal; configured to record the error signal over time Logic; logic configured to calculate a function of the error signal; and logic configured to determine the benefit of the workload predictor in response to the function of the error signal. 如請求項16之系統,其進一步包含:經組態以回應於該誤差信號而調整至少一參數的邏輯。The system of claim 16, further comprising: logic configured to adjust the at least one parameter in response to the error signal. 如請求項16之系統,其進一步包含:經組態以使該誤差信號與一效能按比例調整演算法與至少一參數之一組合相關聯的邏輯。The system of claim 16, further comprising: logic configured to associate the error signal with a performance scaling algorithm in combination with one of the at least one parameter. 如請求項16之系統,其進一步包含:經組態以產生交叉相關之一量測以便判定該行動裝置之一當前操作狀態是否類似於一已知使用狀況的邏輯,已針對該已知使用狀況定義了一效能按比例調整演算法與一或多個參數之一適當組合。The system of claim 16, further comprising: logic configured to generate a cross-correlation to determine whether a current operational state of one of the mobile devices is similar to a known usage condition, for which the known usage condition has been A performance scaling algorithm is defined that is suitably combined with one of one or more parameters. 如請求項19之系統,其進一步包含:經組態以儲存一效能按比例調整演算法與一或多個參數之最類似組合之一指示的邏輯;及經組態以在該行動裝置經如此組態時應用該組合之邏輯。The system of claim 19, further comprising: logic configured to store one of a most similar combination of a performance scaling algorithm and one or more parameters; and configured to act on the mobile device The logic of the combination is applied during configuration.
TW100136777A 2011-10-11 2011-10-11 Systems and methods for measuring the effectiveness of a workload predictor on a mobile device TW201316252A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW100136777A TW201316252A (en) 2011-10-11 2011-10-11 Systems and methods for measuring the effectiveness of a workload predictor on a mobile device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW100136777A TW201316252A (en) 2011-10-11 2011-10-11 Systems and methods for measuring the effectiveness of a workload predictor on a mobile device

Publications (1)

Publication Number Publication Date
TW201316252A true TW201316252A (en) 2013-04-16

Family

ID=48803054

Family Applications (1)

Application Number Title Priority Date Filing Date
TW100136777A TW201316252A (en) 2011-10-11 2011-10-11 Systems and methods for measuring the effectiveness of a workload predictor on a mobile device

Country Status (1)

Country Link
TW (1) TW201316252A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI550398B (en) * 2015-12-28 2016-09-21 英業達股份有限公司 System for determining physical location of logic cpu and method thereof
US11216055B2 (en) 2018-02-09 2022-01-04 Samsung Electronics Co., Ltd. Mobile device including context hub and operation method thereof

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI550398B (en) * 2015-12-28 2016-09-21 英業達股份有限公司 System for determining physical location of logic cpu and method thereof
US11216055B2 (en) 2018-02-09 2022-01-04 Samsung Electronics Co., Ltd. Mobile device including context hub and operation method thereof
US11971769B2 (en) 2018-02-09 2024-04-30 Samsung Electronics Co., Ltd. Mobile device including context hub and operation method thereof

Similar Documents

Publication Publication Date Title
US8478567B2 (en) Systems and methods for measuring the effectiveness of a workload predictor on a mobile device
KR101520359B1 (en) System and method for battery load management in a portable computing device
US9465423B2 (en) System and method for thermal management in a portable computing device using thermal resistance values to predict optimum power levels
US9652026B2 (en) System and method for peak dynamic power management in a portable computing device
KR101562448B1 (en) Method and system for dynamically controlling power to multiple cores in a multicore processor of a portable computing device
US8510740B2 (en) System and method of performing dynamic voltage and frequency scaling operations on a mobile device
US8595520B2 (en) System and method for determining thermal management policy from leakage current measurement
US20150026495A1 (en) System and method for idle state optimization in a multi-processor system on a chip
US8320898B2 (en) Systems and methods for optimizing the configuration of a set of performance scaling algorithms
US8850234B2 (en) Power management apparatus and method
TW201316252A (en) Systems and methods for measuring the effectiveness of a workload predictor on a mobile device