JP7479919B2 - Information processing device - Google Patents

Description

The present invention relates to an information processing device.

In information processing devices, a process is carried out to cool a processor that has become hot in order to prevent the device from becoming too hot and to protect the device. Examples of processes for cooling a processor include using a cooling fan and reducing the frequency of the clock signal supplied to the processor.

Patent document 1 describes, for example, an information processing device that controls the processing load of an electronic circuit based on the temperature measured by a temperature monitor so that the temperature does not exceed an allowable temperature. Patent document 2 describes, for example, a terminal temperature control method that drives a control unit at a first drive frequency when the temperature of the terminal corresponds to a first set temperature.

JP 2012-243274 A JP 2013-33469 A

In the case of information processing devices that do not have a cooling fan to cool the processor, a processor that has become too hot is cooled by lowering the frequency of the clock signal. With this type of cooling method, in order to cool a processor that has already become too hot, the frequency of the clock signal is significantly lowered. If the frequency of the clock signal is significantly lowered, the amount of processing that the processor can process per unit time is significantly reduced. As a result, there is a risk that the processing power that the processor can provide will be insufficient for the processing power required by the application program being executed. If the processing power is insufficient, issues such as poor operability, such as slow response to operations and non-smooth movement of displayed images, may occur.

One aspect of the disclosed technology is to provide an information processing device that can suppress heat generation from the processor by reducing processing power while suppressing degradation of operability.

One aspect of the disclosed technology is exemplified by an information processing device as follows. The information processing device includes a processor, a supply unit that supplies a clock signal to the processor, and a limiting unit that reduces the processing power of the processor to a first processing power when the temperature of the processor reaches a limit temperature. The limiting unit measures a load fluctuation pattern of the processor, and reduces the processing power to a second processing power higher than the first processing power when a difference between a reference pattern indicating the load fluctuation pattern of the processor until the limit temperature is reached, which is stored in advance in a storage unit, and the measured load fluctuation pattern becomes equal to or less than a threshold value.

The disclosed technology can reduce heat generation from the processor by lowering processing power while still preventing a decrease in operability.

FIG. 1 is a diagram illustrating an example of a hardware configuration of a smartphone according to an embodiment. FIG. 2 is a diagram illustrating a reduction interval for implementing performance restrictions in an embodiment. FIG. 3 is a diagram illustrating an example of dictionary data according to the embodiment. FIG. 4 is a diagram illustrating an example of the relationship between the frequency of a clock signal supplied to a CPU and the current flowing through the CPU in the embodiment. FIG. 5 illustrates an example of a predicted heat generation amount table used for approximating a predicted heat generation amount in a small interval within a prediction interval in the embodiment. FIG. 6 is a first diagram illustrating governor parameters in the embodiment. FIG. 7 is a second diagram illustrating governor parameters in the embodiment. FIG. 8 is a third diagram illustrating governor parameters in the embodiment. FIG. 9 is a diagram illustrating a determination as to whether or not the temperature of the CPU reaches the limit temperature in the embodiment. FIG. 10 is a diagram illustrating an example of a processing flow of the smartphone according to the embodiment. FIG. 11 is a diagram illustrating an example of a process flow of dictionary data generation processing according to the embodiment. FIG. 12 is a diagram illustrating an example of a processing flow of performance restriction in the embodiment. FIG. 13 is a diagram illustrating an example of a change in CPU temperature depending on whether or not performance restrictions are implemented.

<Embodiment>
The configurations of the embodiments described below are merely examples, and the disclosed technology is not limited to the configurations of the embodiments. An information processing device according to the embodiments has, for example, the following configuration.
The information processing device according to this embodiment includes:
A processor;
a supply unit for supplying a clock signal to the processor;
a limiting unit that reduces a processing capability of the processor to a first processing capability when the temperature of the processor reaches a limit temperature;
The limiting portion is
measuring a load variation pattern of the processor;
When the difference between a reference pattern indicating the load fluctuation pattern of the processor until the limit temperature, which is pre-stored in a memory unit, is reached and the measured load fluctuation pattern becomes equal to or smaller than a threshold value, the processing capacity is reduced to a second processing capacity higher than the first processing capacity.

The information processing device is, for example, an information processing device that does not have a cooling fan for cooling a processor. Examples of such information processing devices include smartphones, feature phones, tablet personal computers, and notebook personal computers. The information processing device may also be a desktop personal computer that is quiet because it does not have a cooling fan. The information processing device may also be a small device used in the Internet of Things (IoT).

When the temperature of the processor reaches the limit temperature, the information processing device cools the processor by reducing the processing power of the processor to a first processing power. Here, in order to cool a processor that has already reached the limit temperature, the processing power of the processor is significantly reduced. If the processing power of the processor is significantly reduced, there is a risk that the operability of the information processing device will decrease.

The information processing device determines whether the temperature of the processor will reach the limit temperature based on the load fluctuation pattern of the processor. If the information processing device determines that the temperature will reach the limit temperature, it reduces the processing power of the processor to a second processing power higher than the first processing power. Because the temperature of the processor has not yet reached the limit temperature, the processor can be cooled by reducing the processing power of the processor to the second processing power without reducing it to the first processing power. Furthermore, because the second processing power is higher than the first processing power, the deterioration of operability is suppressed more than when the processing power is reduced to the first processing power. Here, reducing the processing power to the second processing power may include, for example, reducing the number of cores to be driven when the processor has a multi-core configuration.

The information processing device may have the following features: The supply unit supplies a clock signal of a frequency selected from a predetermined frequency range to the processor, and reducing the processing power to the first processing power includes reducing the upper limit of the predetermined frequency range, and reducing the processing power to the second processing power includes reducing the proportion of the frequencies in the predetermined frequency range to which a clock signal of the upper limit frequency is supplied. Even when the processing power is reduced to the second processing power, the upper limit of the predetermined frequency range does not change. In other words, even when the processing power of the processor is reduced to the second processing power, a clock signal of the upper limit frequency of the predetermined frequency range may be supplied to the processor in accordance with the processing load. Therefore, even when high-load processing is required, the information processing device suppresses deterioration in operability.

The information processing device may have the following features. The storage unit stores the reference pattern in association with information identifying a process, and the limiting unit identifies a process being executed by the processor, and when the difference between the reference pattern corresponding to the identified process and the measured load fluctuation pattern becomes equal to or less than a threshold, the processing capacity is reduced to a second processing capacity that is higher than the first processing capacity. Here, the process may be an application program unit, or may be a window unit within the application program. For example, in an application program that plays a video, the load on the processor is considered to be different between a window that plays the video and a window that sets the application program. Therefore, it is also useful to store the reference pattern in association with a window rather than an application program. By having such features, the information processing device can reduce the processing capacity to the second processing capacity according to the characteristics of the load fluctuation of the process.

The information processing device may have the following features. When a load fluctuation pattern is not stored in the storage unit, the limiting unit measures a load fluctuation pattern until the temperature of the processor reaches the limit temperature, and stores the measured load fluctuation pattern until the limit temperature is reached as the reference pattern in the storage unit. By providing such features, a load fluctuation pattern that reflects a user's operating habits and characteristics of each information processing device (such as processes executed in the background) can be stored in the storage unit.

Below, an embodiment in which the above-mentioned information processing device is applied to a smartphone will be further described with reference to the drawings. FIG. 1 is a diagram showing an example of the hardware configuration of a smartphone according to an embodiment. The smartphone 100 illustrated in FIG. 1 includes a Central Processing Unit (CPU) 101, a memory unit 102, a thermistor 103, a governor control unit 104, a temperature control unit 105, a frequency control unit 106, and a governor temperature control unit 107. Due to space limitations within the housing, the smartphone 100 does not include a cooling fan for cooling the CPU 101.

The CPU 101 is also called a microprocessor unit (MPU) or a processor.
The CPU 101 is not limited to a single processor, and may have a multi-processor configuration. Also, a single CPU 101 connected by a single socket may have a multi-core configuration. At least a part of the processing executed by the CPU 101 may be executed by a processor other than the CPU 101, for example, a dedicated processor such as a digital signal processor (DSP), a graphics processing unit (GPU), a numerical calculation processor, a vector processor, or an image processing processor. Also, at least a part of the processing executed by the CPU 101 may be executed by an integrated circuit (IC) or other digital circuit. Also, at least a part of the CPU 101 may include an analog circuit. The integrated circuit includes a large scale integrated circuit (LSI), an application specific integrated circuit (ASIC), and a programmable logic device (PLD). The PLD includes, for example, a Field-Programmable Gate Array (FPGA). The CPU 101 may be a combination of a processor and an integrated circuit. The combination is called, for example, a microcontroller unit (MCU), a system-on-a-chip (SoC), a system LSI, a chipset, etc. In the smartphone 100, the CPU 101 executes an application program stored in the storage unit 102 to control peripheral devices. This allows the smartphone 100 to execute processing that matches a predetermined purpose.

The amount of heat generated by the CPU 101 varies depending on the frequency of the clock signal supplied to it. Therefore, if a high-frequency clock signal is supplied for a long period of time, the CPU 101 becomes hot. The CPU 101 is an example of a "processor."

The memory unit 102 includes a main memory unit and an auxiliary memory unit. The main memory unit is exemplified as a memory unit that is directly accessed by the CPU 101. The main memory unit includes a Random Access Memory (RAM) and a Read Only Memory (ROM).

The auxiliary memory stores various programs and various data in a readable and writable manner on a recording medium. The auxiliary memory stores an operating system (OS), various programs, various tables, and the like. The auxiliary memory also stores application programs (referred to as "apps" in the figure) installed on the smartphone 100.

Examples of auxiliary memory units include erasable programmable ROM (EPROM), solid state drive (SSD), and hard disk drive (HDD). Memory unit 102 is an example of a "memory unit."

Thermistor 103 measures the temperature of CPU 101. Temperature control unit 105 acquires the temperature of CPU 101 measured by thermistor 103. If the acquired temperature is equal to or higher than a preset limit temperature, temperature control unit 105 imposes a temperature restriction on governor control unit 104, lowering the upper limit of the frequency of the clock signal supplied to CPU 101 to a first frequency that is lower than the maximum frequency at which CPU 101 can operate.

The governor control unit 104 controls the frequency of the clock signal supplied by the frequency control unit 106 to the CPU 101. The governor control unit 104 controls the frequency of the clock signal supplied by the frequency control unit 106, for example, by applying a governor parameter to the frequency control unit 106. The governor parameter is, for example, a parameter that associates the load factor of the CPU 101, the frequency of the clock signal to be supplied, and the maintenance time for maintaining the supply of the clock signal of that frequency. The governor parameter will be described in detail later. When a temperature limit is applied, the governor control unit 104 determines the frequency of the clock signal to be supplied to the CPU 101 within a range with the first frequency as the upper limit. The governor control unit 104 is an example of a "limiting unit".

The frequency control unit 106 determines the frequency of the clock signal according to the governor parameters applied by the governor control unit 104, and supplies the clock signal of the determined frequency to the CPU 101. The CPU 101 operates based on the clock signal supplied from the frequency control unit 106. The frequency control unit 106 is an example of a "supply unit".

Here, in the temperature restriction by the governor control unit 104, the upper limit of the frequency of the clock signal supplied to the CPU 101 is lowered to the first frequency, which may cause a significant decrease in the performance of the smartphone 100. Therefore, in this embodiment, the governor temperature control unit 107 determines whether the temperature of the CPU 101 reaches the limit temperature. When the governor temperature control unit 107 determines that the limit temperature is reached, it imposes a performance restriction on the CPU 101 that is more lenient than the temperature restriction. Since the performance restriction is more lenient than the temperature restriction, a significant decrease in the performance of the CPU 101 is suppressed. In addition, the performance restriction executed by the governor temperature control unit 107 suppresses the amount of heat generated by the CPU 101. By implementing such a performance restriction, it is possible to prevent the temperature of the CPU 101 from reaching the limit temperature or to delay the timing at which the temperature of the CPU 101 reaches the limit temperature.

FIG. 2 is a diagram illustrating a reduction interval in which performance restrictions are implemented in an embodiment. In FIG. 2, the vertical axis illustrates the temperature of CPU 101, and the horizontal axis illustrates time. Governor temperature control unit 107 determines whether the temperature of CPU 101 will reach the limit temperature in the prediction interval. If governor temperature control unit 107 determines that the limit temperature will be reached, it implements performance restrictions on CPU 101 in the reduction interval. If the temperature of CPU 101 reaches the limit temperature even after performance restrictions are implemented, temperature restrictions are implemented by governor control unit 104.

The governor temperature control unit 107 monitors the application programs being executed by the CPU 101 and the load fluctuation pattern of the CPU 101. In making a judgment in the prediction interval, the governor temperature control unit 107 may refer to dictionary data, described below, based on the monitored load fluctuation pattern and judge whether the temperature of the CPU 101 will reach the limit temperature.

Performance restrictions include allowing a maximum frequency clock signal to be supplied to CPU 101 while shortening the supply time during which the maximum frequency clock signal is supplied. For example, shortening the supply time may include quickly lowering the frequency of the clock signal in response to a reduction in the load on CPU 101, even if the frequency of the clock signal transitions to the maximum frequency in response to the load. Such control is performed, for example, using a governor table that associates the load on CPU 101 with the frequency of the clock signal.

If dictionary data corresponding to an application program being executed by CPU 101 does not yet exist, governor temperature control unit 107 generates dictionary data corresponding to the application program. When governor temperature control unit 107 receives a notification from temperature control unit 105 that a temperature limit has been executed, governor temperature control unit 107 associates the load fluctuation pattern of CPU 101 and the reduced governor parameters in the prediction interval and stores them as dictionary data.

Here, the prediction interval may be different for each application. For example, the prediction interval may be set longer for an application program in which the load on the CPU 101 fluctuates frequently, and shorter for an application program in which the load on the CPU 101 fluctuates little. The reduced governor parameter is a parameter used when imposing a performance restriction on the CPU 101 that is more lenient than the temperature restriction. The governor temperature control unit 107 may store dictionary data for applications that have received a notification that a temperature restriction has been imposed a predetermined number of times or more. The governor temperature control unit 107 is an example of a "restriction unit".

Figure 3 is a diagram showing an example of dictionary data in an embodiment. Dictionary data 1071 illustrated in Figure 3 includes the following items: "Application name", "Length of prediction interval", "Feature data", and "Governor parameters". "Application name" stores the name of an application program. "Length of prediction interval" stores the length of the prediction interval corresponding to an application program. The length of the prediction interval is, for example, the length of the time until the temperature limit is reached for each application program minus the reduction interval. "Feature data" stores, for example, data indicating the load fluctuation pattern of CPU 101.

Data indicating the load fluctuation pattern of the CPU 101 may be, for example, a time series change in a pair of the predicted heat generation amount of the CPU 101 and the temperature of the CPU 101 measured by the thermistor 103. An example of the time series change may be a list of pairs of the predicted heat generation amount of the CPU 101 and the temperature of the CPU 101 measured by the thermistor 103 in each of the sub-intervals, with the prediction interval divided into a number of sub-intervals. The dictionary data 1071 is stored, for example, in a storage unit in the governor temperature control unit 107. The dictionary data 1071 may also be stored, for example, in the storage unit 102. The dictionary data 1071 is an example of a "reference pattern".

The predicted heat generation amount can be estimated, for example, from the frequency of the clock signal supplied to the CPU 101. FIG. 4 is a diagram illustrating the relationship between the frequency of the clock signal supplied to the CPU and the current flowing through the CPU in an embodiment. When a large current flows through the CPU 101, the power consumption of the CPU 101 increases. Therefore, FIG. 4 can be said to illustrate the relationship between the frequency of the clock signal supplied to the CPU and the power consumption of the CPU 101. With reference to FIG. 4, it can be seen that the higher the frequency of the clock signal supplied to the CPU 101, the higher the power consumption of the CPU 101. And, the amount of heat generated by the CPU 101 increases in response to the increased power consumption. Therefore, the predicted heat generation amount can be roughly calculated based on the frequency distribution.

5 is an example of a predicted heat generation amount table used to estimate the predicted heat generation amount in a small section in a prediction interval in an embodiment. The predicted heat generation amount table 1072 includes the items of "frequency", "weight", "frequency distribution", and "predicted heat generation amount". "Frequency" lists the frequency of the clock signal supplied to the CPU 101 in the small section. The unit of frequency is, for example, "MHz". FIG. 5 illustrates that clock signals of frequencies 100 MHz, 200 Mhz, and 300 Mhz are supplied in the small section. "Weight" stores information indicating the weight used to estimate the heat generation amount for each frequency. For example, the higher the frequency, the greater the weight. "Frequency distribution" stores information indicating the time when the clock signal of each frequency is supplied as a ratio with the entire small section set to 1. "Predicted heat generation amount" stores an approximate value of the predicted heat generation amount calculated based on "weight" and "frequency distribution". In the example of Figure 5, the "weight" for the frequency 200 MHz is "2" and the "frequency distribution" is "0.3", so as a result of these multiplications, the predicted heat generation amount is calculated as "0.6". The heat generation amount in the prediction interval is the sum of the predicted heat generation amounts at each frequency, so in the example of Figure 5, the predicted heat generation amount in the small interval is calculated as "2.3".

Returning to FIG. 3, the "Governor Parameter" stores information indicating a reduced governor parameter applied when performance restriction is performed. FIGS. 6 to 8 are diagrams illustrating governor parameters in the embodiment. FIG. 6 illustrates a governor parameter applied when neither temperature restriction nor performance restriction is performed. FIGS. 7 and 8 illustrate a reduced governor parameter used for performance restriction. The governor parameters and reduced governor parameters illustrated in FIGS. 6 to 8 are illustrated as a table that associates the items of "frequency,""target load factor," and "maintenance time." The "frequency" stores the frequency (MHz) of the clock signal supplied to the CPU 101 by the frequency control unit 106. The "target load factor" stores a value that determines the frequency so as to be close to the load factor (%) of the CPU 101. The "maintenance time" stores the minimum time (ms) for which the clock signal of the changed frequency is continuously supplied.

For example, in the governor parameters 1073 illustrated in FIG. 6, when the frequency of the clock signal supplied to the CPU 101 is 100 MHz and the load factor of the CPU 101 exceeds 40%, the frequency of the clock signal supplied to the CPU 101 is changed to 200 MHz. In the example of the governor parameters 1073, for example, when the frequency of the clock signal is changed to "300 MHz", a clock signal of 300 MHz is supplied to the CPU 101 for 200 ms continuously. After 200 ms have elapsed, the frequency of the clock signal supplied to the CPU 101 is changed to "200 MHz", which is one step lower in the governor parameters 1073.

The reduced governor parameter 1074 illustrated in FIG. 7 is set to have a higher target load factor at frequencies of 200 MHz or higher and a shorter maintenance time at frequencies of 300 MHz or higher than the governor parameter 1073. Therefore, the reduced governor parameter 1074 is a governor parameter that makes it difficult to transition to high frequencies and makes it easy to transition to low frequencies.

The reduction governor parameter 1075 illustrated in FIG. 8 is a governor parameter that makes it even more difficult to transition to a higher frequency and easier to transition to a lower frequency than the reduction governor parameter 1074. Here, since none of the reduction governor parameters lowers the upper limit of the frequency supplied by the frequency control unit 106, the frequency control unit 106 can supply to the CPU 101 a clock signal of the maximum frequency at which the CPU 101 can operate, depending on the load on the CPU 101. The reduction governor parameters 1074 and 1075 can be said to be governor parameters that have different frequency fluctuation patterns of the clock signal supplied to the CPU 101.

The ease of transitioning to a higher frequency in the governor parameters and reduced governor parameters can be controlled by the target load rate. Increasing the target load rate makes it more difficult to transition to a higher frequency. On the other hand, lowering the target load rate makes it easier to transition to a higher frequency. Also, the ease of transitioning to a lower frequency in the reduced governor parameters can be controlled by the maintenance time. Increasing the maintenance time makes it more difficult to transition to a lower frequency. On the other hand, shortening the maintenance time makes it easier to transition to a lower frequency.

FIG. 9 is a diagram showing a schematic diagram of a determination as to whether or not the temperature of the CPU reaches the limit temperature in an embodiment. In FIG. 9, the vertical axis illustrates an example of feature data, and the horizontal axis illustrates an example of time. Here, a predicted heat generation amount is adopted as the feature data. For each small section that divides the prediction interval, the governor temperature control unit 107 calculates the absolute value of the difference between the predicted heat generation amount calculated based on the frequency of the clock signal supplied to the CPU 101 and the predicted heat generation amount in the dictionary data 1071. The governor temperature control unit 107 calculates a dissimilarity by adding up the absolute values of the differences calculated for each small section. The larger the dissimilarity value, the greater the deviation between the actual measurement data and the dictionary data 1071. In other words, the governor temperature control unit 107 determines that the greater the dissimilarity, the greater the deviation between the actual measurement data and the dictionary data 1071.
The governor temperature control unit 107 determines that the temperature of the CPU 101 is unlikely to reach the limit temperature. In other words, the magnitude of the dissimilarity indicates the low probability that the CPU 101 will reach the temperature limit. In this embodiment, the dictionary data 1071 stores the temperature measured by the thermistor 103 together with the predicted heat generation amount as feature data. Thus, the governor temperature control unit 107 calculates the dissimilarity between the temperature of the CPU 101 actually measured by the thermistor 103 and the temperature stored in the dictionary data 1071, as with the predicted heat generation amount. The governor temperature control unit 107 may determine that the temperature of the CPU 101 will reach the limit temperature when the dissimilarity of at least one of the predicted heat generation amount and the temperature is equal to or less than a threshold value. The governor temperature control unit 107 may also determine that the temperature of the CPU 101 will reach the limit temperature when both the dissimilarity of the predicted heat generation amount and the dissimilarity of the temperature are equal to or less than a threshold value.

<Processing flow of the smartphone 100>
Fig. 10 is a diagram showing an example of a processing flow of the smartphone according to the embodiment. The processing illustrated in Fig. 10 is repeatedly executed while the smartphone 100 is activated. Hereinafter, an example of the processing flow of the smartphone 100 will be described with reference to Fig. 10.

At J1, the governor temperature control unit 107 acquires the name of the application program being executed by the CPU 101. At J2, the governor temperature control unit 107 determines whether or not dictionary data 1071 corresponding to the name of the application program acquired at J1 has been generated. If dictionary data 1071 has been generated (YES at J2), the process proceeds to J3. If dictionary data 1071 has not been generated (NO at J2), the process proceeds to J5.

In J3, the governor temperature control unit 107 determines whether the temperature of the CPU 101 will reach the limit temperature based on the actual measurement data of the CPU 101 and the dictionary data 1071. If it is determined that the limit temperature will be reached (YES in J3), the process proceeds to J4. If it is determined that the limit temperature will not be reached (NO in J3), the process ends.

In J4, the governor temperature control unit 107 imposes performance restrictions on the CPU 101. Details of the performance restrictions are described with reference to FIG. 12.

In J5, the governor temperature control unit 107 generates dictionary data 1071. Details of the process of generating the dictionary data 1071 will be described with reference to FIG. 11.

<Generation process of dictionary data 1071>
Fig. 11 is a diagram showing an example of a process flow of dictionary data generation processing in an embodiment. The process illustrated in Fig. 11 is executed, for example, in J5 in Fig. 10. Hereinafter, an example of a process flow of dictionary data 1071 generation processing will be described with reference to Fig. 11.

In J51, the governor temperature control unit 107 monitors the load fluctuation pattern of the CPU 101. In J52, the governor temperature control unit 107 determines whether or not a temperature limit has been implemented by the governor control unit 104. If it has been implemented (YES in J52), the process proceeds to J53. If it has not been implemented (NO in J52), the process proceeds to J51.

In J53, the governor temperature control unit 107 stores the load fluctuation pattern for the period corresponding to the prediction interval until the temperature limit is executed. In storing the load fluctuation pattern, the period corresponding to the prediction interval may be divided into small intervals, and feature data indicating the load fluctuation pattern in each of the divided small intervals may be stored. The area for storing the load fluctuation pattern may be a memory unit within the governor temperature control unit 107, or may be the memory unit 102.

In J54, the governor temperature control unit 107 determines whether the load change pattern in J53 has been saved a threshold number of times or more. In other words, the governor temperature control unit 107 determines whether a threshold number of load change patterns or more have been saved. If the number of saved times is equal to or greater than the threshold number (YES in J54), processing proceeds to J55. If the number of saved times is less than the threshold number (NO in J54), processing proceeds to J51.

In J55, the governor temperature control unit 107 determines the feature data to be stored in the dictionary data 1071. The governor temperature control unit 107 calculates the average value of the feature data stored a threshold number of times for each small section. The governor temperature control unit 107 determines the calculated average value as the feature data for that small section.

In J56, the governor temperature control unit 107 determines the reduced governor parameters to be applied based on the characteristics of the load fluctuation pattern in the prediction interval. For example, when the load rate of the CPU 101 repeatedly rises and falls in a long cycle in the prediction interval, the governor temperature control unit 107 may adopt reduced governor parameters that make it difficult to transition to a high frequency and that make it easy to transition from a high frequency to a low frequency. For example, when the load rate of the CPU 101 repeatedly rises and falls in a short cycle in the prediction interval, the governor temperature control unit 107 may adopt reduced governor parameters that make it easy to transition to a high frequency and that make it easy to transition from a high frequency to a low frequency. Here, the length of the cycle may be determined based on an appropriately determined threshold value.

In J57, the governor temperature control unit 107 generates dictionary data 1071 by associating the feature data determined in J55 and the reduced governor parameters determined in J56 with the names of the application programs obtained in J1 of FIG. 10.

<Performance limit processing flow>
Fig. 12 is a diagram illustrating an example of a processing flow of performance restriction in the embodiment. The processing illustrated in Fig. 12 is executed, for example, at J3 and J4 in Fig. 10. Hereinafter, an example of the processing flow of performance restriction will be described with reference to Fig. 12.

In J31, the governor temperature control unit 107 calculates the dissimilarity between the measured load fluctuation pattern of the CPU 101 and the feature data stored in the dictionary data 1071. In J32, the governor temperature control unit 107 determines whether the dissimilarity calculated in J31 is equal to or less than a threshold value. If it is equal to or less than the threshold value (YES in J32), the process proceeds to J41. If it is greater than the threshold value (NO in J32), the process ends. The processes of J31 and J32 are an example of the process executed in J3 in FIG. 10.

In J41, the governor temperature control unit 107 instructs the governor control unit 104 to apply the reduced governor parameters associated with the name of the application program acquired in J1 of FIG. 10 in the dictionary data 1071. The governor control unit 104 applies the instructed reduced governor parameters to the frequency control unit 106. Applying the reduced governor parameters to the frequency control unit 106 is an example of "reducing to the second processing capacity."

<Effects of the embodiment>
In the embodiment, before implementing a temperature restriction that reduces the upper limit of the frequency of the clock signal supplied from the frequency control unit 106 to the CPU 101 to a first frequency, it is determined whether or not a temperature restriction will occur in the prediction interval. Then, when it is determined that the temperature restriction will occur, a performance restriction that is a more lenient restriction than the temperature restriction is implemented. In the performance restriction, the upper limit of the frequency of the clock signal supplied from the frequency control unit 106 to the CPU 101 is not reduced. Therefore, even in a state in which the performance restriction is implemented, a clock signal of the maximum frequency may be supplied to the CPU 101 when a high-load process is executed. Therefore, according to the present embodiment, the deterioration of the operability of the smartphone 100 is suppressed.

In this embodiment, the smartphone 100 suppresses the amount of heat generated by the CPU 101 by implementing a performance restriction in the reduction section. FIG. 13 is a diagram illustrating a change in the temperature of the CPU depending on whether or not a performance restriction is implemented. In FIG. 13, the vertical axis illustrates the temperature of the CPU 101, and the horizontal axis illustrates time. The solid line graph illustrates the transition of the temperature of the CPU 101 when a performance restriction is implemented in the reduction section. The dotted line graph illustrates the transition of the temperature of the CPU 101 when a performance restriction is not implemented. As can be understood by referring to FIG. 13, the rise in the temperature of the CPU 101 in the reduction section is suppressed by implementing a performance restriction. As a result, the smartphone 100 can reduce the frequency with which a temperature restriction occurs. In other words, the smartphone 100 can reduce the frequency with which operability is reduced due to a temperature restriction.

In the embodiment, the smartphone 100 quantifies the probability that the CPU 101 will reach the temperature limit as a dissimilarity. When the smartphone 100 determines to change the governor parameters to be applied to the frequency control unit 106 based on this numerical value, the smartphone 100 selects the reduced governor parameters according to the dictionary data 1071. Here, in the dictionary data 1071, suitable governor parameters are selected based on the frequency fluctuation frequency pattern of the clock signal and the frequency fluctuation width of the clock signal. The selected reduced governor parameters are then registered in association with the name of the application program. In other words, in the dictionary data 1071, the combination of the frequency fluctuation frequency pattern of the clock signal and the frequency fluctuation width of the clock signal can be indicated by the name of the application program. Therefore, the smartphone 100 can select suitable reduced governor parameters based on the frequency fluctuation frequency pattern of the clock signal and the frequency fluctuation width of the clock signal by following the dictionary data 1071. Moreover, by selecting the reduced governor parameters in this way, the smartphone 100 can reduce the amount of calculation required for selecting the reduced governor parameters. In turn, the smartphone 100 can improve the responsiveness of the application of the reduced governor parameters to changes in the temperature of the CPU 101.

In the embodiment, the dictionary data 1071 stores the reduction governor parameters in association with the application programs, but the dictionary data 1071 is not limited to this configuration. For example, the dictionary data 1071 may associate the reduction governor parameters with a combination of a frequency fluctuation frequency pattern of a clock signal and a frequency fluctuation width of the clock signal. In the embodiment, the "length of the prediction interval" in the dictionary data 1071 is set for each application program, but the "length of the prediction interval" may be a predetermined constant value.

In the embodiment, the load fluctuation pattern of the application program is stored as dictionary data 1071, and the dictionary data 1071 is referenced to determine whether or not a temperature limit will occur. Therefore, according to the present embodiment, the smartphone 100 can determine whether or not a temperature limit will occur according to the characteristics of each of the multiple installed application programs.

<Modification>
In the embodiment described above, the predicted heat generation amount and temperature are used as the feature data stored in the dictionary data 1071. However, the feature data may be other data. For example, as described above, the heat generation amount of the CPU 101 varies depending on the frequency of the clock signal supplied. Therefore, the frequency of the clock signal supplied to the CPU 101 may be used as the feature data.

In the embodiment described above, the CPU 101 is cooled by controlling the frequency of the clock signal supplied to the CPU 101. However, the cooling of the CPU 101 is not limited to this method, and for example, when the CPU 101 has a multi-core configuration, the number of cores to be driven may be reduced.

The embodiments and variations disclosed above can be combined with each other.

100: Smartphone 101: CPU
102: Memory unit 103: Thermistor 104: Governor control unit 105: Temperature control unit 106: Frequency control unit 107: Governor temperature control unit 1071: Dictionary data 1072: Predicted heat generation amount table 1073: Governor parameters 1074, 1075: Reduced governor parameters

Claims (4)

A processor;
a supply unit for supplying a clock signal to the processor;
a limiting unit that reduces a processing capability of the processor to a first processing capability when the temperature of the processor reaches a limit temperature;
The limiting portion is
measuring a load variation pattern of the processor;
when a difference between a reference pattern indicating a load fluctuation pattern of the processor until the limit temperature is reached, which is stored in advance in a storage unit, and the measured load fluctuation pattern becomes equal to or smaller than a threshold value, the processing capability of the processor is reduced to a second processing capability higher than the first processing capability before reducing the processing capability of the processor to the first processing capability;
Information processing device. the supply unit supplies a clock signal having a frequency selected from a predetermined frequency range to the processor;
reducing the processing capability to the first processing capability includes limiting a frequency of a clock signal supplied to the processor to a first frequency or less that is set lower than an upper limit frequency of the predetermined frequency range;
The step of reducing the processing capability to the second processing capability includes reducing a ratio of a clock signal having the upper limit frequency of the predetermined frequency range to be supplied.
The information processing device according to claim 1 . the storage unit stores the reference pattern in association with information for identifying a process;
The limiting portion is
Identifying a process being executed by said processor;
when a difference between a reference pattern corresponding to the identified process and the measured load fluctuation pattern becomes equal to or smaller than a threshold, reducing the processing power of the processor to the second processing power before reducing the processing power to the first processing power .
3. The information processing device according to claim 1 or 2. The limiting portion is
If the load change pattern is not stored in the storage unit, the load change pattern is measured until the temperature of the processor reaches the limit temperature;
The load change pattern measured until the limit temperature is reached is stored in the storage unit as the reference pattern.
The information processing device according to claim 1 .

Images