WO2019136963A1

WO2019136963A1 - Method and device for reclaiming memory

Info

Publication number: WO2019136963A1
Application number: PCT/CN2018/099419
Authority: WO
Inventors: 郑华晋; 韩晓征; 陆伟
Original assignee: 华为技术有限公司
Priority date: 2018-01-10
Filing date: 2018-08-08
Publication date: 2019-07-18
Also published as: CN110018899B; CN110018899A

Abstract

Provided by the present disclosure are a method and device for reclaiming memory, belonging to the technical field of terminals. Said method comprises: determining a plurality of application processes running in the background when detecting that free memory does not satisfy a preset condition; determining a user preference for each application process in said plurality of application processes in a time period to which a current time point belongs, said user preference being used to indicate the possibility of the application process being called; determining an application process to be reclaimed in said plurality of application processes according to the user preference for each application process in said plurality of application processes; and reclaiming memory occupied by said application process to be reclaimed. The present disclosure takes into account usage habits of a user when reclaiming memory and ensures that the overall impact of the process of reclaiming memory on the operation of a terminal is as small as possible.

Description

Method and device for recycling memory

This application claims priority to Chinese Patent Application No. 201 810 022 138.X, entitled "Recovering Memory Method and Apparatus", issued on January 10, 2018, the entire contents of which are incorporated herein by reference. In the application.

Technical field

The present disclosure relates to the field of terminal technologies, and in particular, to a method and an apparatus for reclaiming memory.

Background technique

With the development of the terminal technology, the terminal can run various processes, and the process can be the system process of the operating system and the application process of the application, wherein each process occupies a certain amount of memory during the running process, and stores itself in the memory. The data. As the terminal runs, more and more processes are running, more and more memory is occupied, and free memory is less and less. When the terminal faces the situation that the free memory does not satisfy the memory requirement, the memory occupied by the process is required. Recycling is performed to ensure that the free memory of the terminal meets the memory requirements.

Taking the process of reclaiming memory by the Low Memory Killer (LMK) process as an example, the process of reclaiming memory is specifically as follows: the LMK process acquires the current oom_adj of each process, and acquires the memory occupied by each process, wherein oom_adj uses To indicate the priority of the process, the higher the om_adj value, the lower the priority of the process. Then, according to the oom_adj and occupied memory of each process, each process is scored to obtain the score of each process. After that, according to the score of each process, the process to be killed is determined. The greater the score of the process, the higher the priority is to kill the process. After the process is killed, the associated process is closed, and the memory occupied by the process is released. The memory occupied by the process is reclaimed. Among them, the value of oom_adj is fixed. For example, when the process is a background process, oom_adj is 2. When the process is a native process, oom_adj is -17, the larger the oom_adj value, the higher the score of the process, and the memory occupied by the process. The larger the score, the higher the score of the process.

In the process of implementing the present disclosure, the inventors have found that the related art has at least the following problems:

When the above solution reclaims memory, only the oom_adj of the process and the memory occupied by the process are considered, and it is easy to kill the process and affect the normal operation of the process.

Summary of the invention

The embodiments of the present disclosure provide a method and an apparatus for reclaiming memory, which can solve the problem of easy to kill the process in the related art. The technical solution is as follows:

In a first aspect, a method of reclaiming memory is provided, the method comprising:

Determining a plurality of application processes running in the background when detecting that the free memory does not satisfy the preset condition;

Determining a user preference of a time period to which each application process in the plurality of application processes belongs at a current time point, the user preference being used to indicate a possibility of an application process being invoked;

Determining, according to a user preference of each application process in the multiple application processes, an application process to be recycled in the multiple application processes;

Reclaiming the memory occupied by the application process to be recycled.

The method provided in this embodiment takes the user's usage habits into consideration when reclaiming the memory, determines the application process to be recycled according to the user preference of each application process, and then recovers the memory occupied by the application process to be recycled, and preferentially recycles the user preference. The memory occupied by the lower application process avoids reclaiming the memory occupied by the application process with high user preference, ensuring the accuracy and rationality of retrieving the memory of each application process, avoiding the killing process and the application process. The data is transferred to the preset storage space, so as to avoid affecting the normal operation of the application process, and ensuring that the overall impact of the process of reclaiming the memory on the operation of the terminal is as small as possible.

In a possible design, the determining the user preference of each of the plurality of application processes in the time period to which the current application point belongs includes:

And determining, according to the active duration of the time period of each application process in the historical operation, the user preference of each application process in the plurality of application processes at a time period to which the current time point belongs.

Based on this design, the process of ensuring the memory recovery is time-sensitive: considering that the application process preferred by the user can change over time, for example, a user prefers to use the application process of the Weibo application in the past, and now prefers to use the instant messaging application. Application process, etc. In this embodiment, the user preference for each application process is not obtained in advance, but each application is calculated in real time according to the active duration of the historical running process of the application process when the terminal reclaims the memory occupied by the application process. The user preference of the process, that is, the user preference of the self-learning application process, the calculated user preference can be dynamically changed with time, and the order of memory occupied by the application process is recycled each time. Dynamic changes, for example, when the terminal calculates the user preference last month, the application process of the instant messaging application has the highest user preference, and the probability of reclaiming the memory of the application process of the instant messaging application is the smallest, and the music playing application is calculated when the user preference is calculated this month. The application process has the highest user preference, and the probability of reclaiming the memory of the application process of the music playing application is the smallest.

In a possible design, determining, according to the active duration of the time period of each application process in the historical operation, the application time of each application process in the multiple application processes at the current time point User preference for the time period to which it belongs, including:

The active duration is taken as the user preference.

In this design, the terminal can obtain user preference simply and efficiently.

Or calculating a ratio of the active duration to the sum of the active durations, where the sum of the active durations is the sum of the active durations of the plurality of application processes in the historical operation. .

In this design, the terminal can ensure that the user's preference is within a reasonable range of values.

Or calculating a product between the active duration and the cold start duration of the application process as the user preference;

In this design, on the basis of accurately recovering the memory occupied by the application process, the cold boot duration of each application process is taken into consideration, and the overall impact on the terminal is minimized.

Or calculating a ratio between the active duration and the sum of the active durations, and calculating a product between the ratio and the cold start duration of the application process as the user preference.

In this design, on the basis of providing a way to accurately obtain the user's preference, the user's preference can be guaranteed to be within a reasonable numerical range, and the cold start time of the application process is taken into account, and the memory occupied by the application process is guaranteed to be recovered. The overall impact on the terminal is as small as possible.

In a possible design, the determining, according to the user preference of each application process in the multiple application processes, the application process to be reclaimed in the multiple application processes, including:

Retrieving a revocation priority of each application process based on a user preference of each application process in the plurality of application processes, a reclaiming priority of each application process, and a user preference of each application process Negative correlation

Determining an application process to be reclaimed among the plurality of application processes based on a recycling priority of each application process.

In a possible design, the obtaining, according to the user preference of each application process in the multiple application processes, the reclaiming priority of each application process, including:

Determining a recent active weight of each application process, the most recent active weight of each application process is used to indicate a degree of recent activeness of each application process;

Retrieving the revocation priority of each application process according to the user preference of each application process and the most recent active weight of each application process.

In a possible design, the determining the most recent active weight of each application process includes:

Obtaining the latest active duration of each application process, where the latest active duration of each application process is used to indicate the active duration of the last time each application process is running in the foreground;

And attenuating the most recent active duration of each application process to obtain the most recent active duration after the attenuation based on the time difference between the end time point of the most recent active duration of the each application process and the current time point. The most recent active duration after the decay is used as the most recent active weight.

In this design, by determining the recycling priority by combining the user preference and the recent active weight of the application process, the accuracy of the reclaimed memory can be further improved: for a plurality of application processes whose user preference is close, if an application process is the most recent If the active application process has a large active weight, the revocation priority will be lower, and the memory of the application process can be avoided, so as to avoid affecting the normal operation of the recently active application process. If an application process is an active application process that is active for a long time, if the recent active weight is small, the reclaim priority will be higher, and the memory of the application process will be reclaimed preferentially, and the memory occupied by the active application process long ago is preferentially reclaimed. On the basis of realizing the recovery of memory, the overall impact on the operation of the terminal is guaranteed to be as small as possible.

Determining a function weight corresponding to the function of each application process, the function weight of each application process is used to indicate the importance degree of the function of each application process;

Retrieving the revocation priority of each application process according to the user preference of each application process and the function weight of each application process.

In this design, by combining user preference and function weight to determine the reclaim priority, the accuracy of reclaiming memory can be further improved: for multiple application processes with similar user preferences, if the process function of an application process is less important It can be considered that the current running activity of the application process is of low importance to the terminal, and the impact of reclaiming the memory of the application process is small, and when the function weight is determined to determine the recycling priority, the recycling priority is higher, and the terminal takes precedence. Reclaim the memory of the application process to minimize the overall impact of the reclaimed memory on the terminal.

Determining power consumption weights corresponding to power consumption of each application process, the power consumption weight of each application process is used to indicate a size of power consumption of each application process;

Retrieving the revocation priority of each application process according to the user preference of each application process and the power consumption weight of each application process.

In this design, by combining the user preference with the power consumption of the application process to determine the recycling priority, the following technical effects can be achieved: for a plurality of application processes with similar user preferences, if the power consumption of an application process is greater, The greater the power consumption weight, the higher the reclaim priority, and the terminal will preferentially reclaim the memory occupied by the application process. By preferentially reclaiming the memory occupied by the application process with high power consumption, the power consumption of the terminal can be saved on the basis of realizing the function of reclaiming memory.

Determining a memory ratio corresponding to a memory ratio of each application process, where a memory ratio of each application process is used to indicate a ratio between a memory occupied by each application process and a total memory overhead;

Retrieving the revocation priority of each application process according to the user preference of each application process and the memory weight of each application process.

In this design, by determining the reclaim priority by combining the user preference with the memory ratio of the application process, the accuracy of the reclaimed memory can be further improved: for a plurality of application processes with similar user preferences, it is assumed that the memory of an application process accounts for If the memory size is larger, the memory weight will be larger, and the recycling priority will be higher. Therefore, the memory of the application process will be reclaimed preferentially, and the memory that is recovered will be guaranteed to ensure that the free memory satisfies the memory requirement as quickly as possible.

Determining a frequency of invocation of each application process, the calling frequency of each application process is used to indicate the total number of times the plurality of application processes invoke the each application process in the time period in the historical operation;

Retrieving the revocation priority of each application process according to the user preference of each application process and the calling frequency of each application process.

In a possible design, the determining the calling frequency of each application process includes:

Obtaining a call record of each application process, where the call record of each application process is used to record an event of each application process calling an application process other than each of the plurality of application processes;

Determining the calling frequency of each application process based on the call record of each application process.

In this design, by combining the user preference and the calling frequency of the application process to determine the recycling priority, it can ensure that the recycling process of the frequently called application process is lower, and the recycling priority will be lower, thereby avoiding recycling the frequently called application. Process-induced effects on the operation of other application processes.

In a possible design, before the determining that the plurality of application processes are running in the background, the method further includes:

Counting the active duration of the time period of each application process in the plurality of application processes during the historical operation.

Based on this design, the process of ensuring the memory recovery is close to the user's personal usage habits: considering that the application process preferred by different users in daily life may be different, for example, user A prefers the application process using the instant messaging application, and user B prefers to use music to play. The application process of the application, user C prefers to use the application process of the online shopping application, and the like. In this embodiment, the user preference of the application process is not uniformly set, but the terminal calculates the user preference of each application process in real time by self-learning. The user preference of each application process can reflect the user's personal usage habits. Therefore, when the memory is recycled in combination with the user preference, the personalized requirement can be satisfied. For example, the probability that the user A's terminal reclaims the application process of the instant messaging application has the smallest probability, and the user has the smallest probability. The terminal of B recycles the memory of the application process of the music playing application with the least probability of the memory and the like.

In a possible design, the recovering the memory occupied by the application process to be recycled includes:

Killing the application process to be recycled; or,

Transfer the data in the memory of the application process to be recycled to a preset storage space.

In a second aspect, there is provided an apparatus for reclaiming memory, the apparatus comprising a plurality of functional modules to implement the method of reclaiming memory in the first aspect and any of the possible designs of the first aspect.

In a third aspect, an apparatus for reclaiming memory is provided, and the apparatus for reclaiming memory includes:

Memory for storing programs;

a processor, configured to execute the program stored in the memory, when the program is executed, the processor is configured to perform the method of reclaiming memory in any of the possible aspects of the first aspect and the first aspect described above .

In a fourth aspect, a computer readable storage medium is provided, comprising instructions, when executed on a computer, causing the computer to perform the method of reclaiming memory in any of the possible aspects of the first aspect and the first aspect described above.

DRAWINGS

FIG. 1 is a schematic structural diagram of an apparatus for recycling a memory according to an embodiment of the present disclosure;

2 is a flowchart of a method for reclaiming memory provided by an embodiment of the present disclosure;

3 is a schematic diagram of a design of a method for reclaiming memory provided by an embodiment of the present disclosure;

FIG. 4 is a schematic structural diagram of an apparatus for reclaiming memory provided by an embodiment of the present disclosure.

Detailed ways

The embodiments of the present disclosure will be further described in detail below with reference to the accompanying drawings.

For ease of understanding, the nouns involved in the embodiments of the present disclosure are first explained below:

Cold boot: When the application is started, if there is no main process applied in the background, the terminal creates a process and assigns the process to the application, which is the main process of the application, and then starts the application. The way is called cold start. In the scenario where the application is started for the first time after the terminal is powered on, or in the scenario where the application is restarted after the main process of the application is killed, the application is started in a cold boot manner. When the application is cold-started, the terminal needs to read the application data from the disk, then write the application data into the memory, and then load the data from the memory to start the application based on the data, because the data is exchanged from the disk to the memory. The duration is usually longer, so the cold start is usually longer.

Hot start: When the application is started, if there is a main process associated with the application in the background, the terminal will directly start the application by using the main process. This method of starting the application is called hot start. After the terminal is powered on, the application is started, and then the application process is switched to run in the background. Then, the application is switched to the foreground running scenario, and the application is started in a warm boot mode. When the application is warm-started, the terminal can start the application by directly loading data from the memory without re-switching the application's data from disk to memory, so the warm boot time is usually shorter.

Low Memory Killer (LMK) mechanism: It is a memory protection mechanism of the Linux operating system. The LMK process detects the free memory of the terminal. When the free memory is less than the memory threshold, it acquires the current oom_adj of each process and obtains The memory occupied by each process, according to the oom_adj of each process and the occupied memory, score each process, get the score of each process (oom_score_adj), kill each process in turn according to the score of each process, Reclaim free memory occupied by the process until the free memory is not less than the memory threshold.

Kswapd process mechanism: A memory protection mechanism in the Linux operating system. The kswapd process, also known as the page swapping daemon, also detects the free memory of the terminal. When the free memory is less than the memory threshold, the kswapd process scans the stored data in the memory. Each memory page, when the number of memory pages is accessed less or the storage time is earlier, the data in the memory page is transferred to the zram partition (a kind of storage space), thereby freeing the memory until the free memory is not Less than the memory threshold.

Direct reclaim mechanism: A memory protection mechanism in the Linux operating system. When a process requests memory from the operating system and the requested memory exceeds the current free memory, the process scans the stored data in the memory. Each memory page, when the number of memory pages is accessed less frequently or stored earlier, will transfer the data in the memory page to the zram partition, thereby freeing the memory until the free memory is not less than the requested memory.

User Preference: Used to indicate the likelihood of an application process being called. The application process can be called by the user or other application process. The higher the user preference, the greater the possibility that the application is called. The lower the user preference, the less likely the application is called. The user preference can represent the user. The degree of preference for the application process. For any application process, the user preference of the application process at the current time point is determined according to the active duration of the time period of the application process in the historical operation, for example, assuming that the current time point is 9:10, belonging to The time period is from 9:00 to 10:00, and the user preference of the application process is determined according to the active duration of 9:00-10:00 in the historical operation, for example, according to the application process at yesterday's 9:00-10:00. The active duration is determined, based on the active duration of the application process from 9:00 to 10:00 last month.

Recycling priority: negatively related to user preference. It is used to represent the priority when reclaiming memory. The revocation priority of the application process is high. The higher the order of memory used by the reclaiming application process, the lower the revocation priority of the application process. The later the memory usage of the application process is recycled.

1 is a schematic structural diagram of an apparatus for reclaiming memory according to an embodiment of the present disclosure. Referring to FIG. 1, the apparatus for reclaiming memory includes: a receiver 101, a transmitter 102, a memory 103, and a processor 104. The receiver 101, The transmitter 102 and the memory 103 are respectively connected to the processor 104. The memory 103 stores a program for executing a program stored in the memory 103. When the program is executed, the processor 104 is configured to execute the following. The method of recycling memory.

In an exemplary embodiment, there is also provided a computer readable storage medium comprising instructions, such as a memory including instructions, that when executed on a computer, cause the computer to execute as in the embodiments described below The method of recycling memory. For example, the computer readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device.

The embodiment of the present disclosure introduces a user's personal usage habits, obtains the user preference of the application process in a dynamic self-learning manner, and combines the user preference to recover the memory strategy, and can embed the existing Linux operating system and Android. (Android) operating system, as a new memory protection mechanism in various operating systems. The embodiments of the present disclosure can be applied to various memory recovery scenarios, and can recover and control the memory occupied by the application process to ensure smoothness and responsiveness of the terminal operation. The following is an example of three exemplary scenarios:

Scene 1, LMK mechanism. When the LMK process detects that the free memory is less than the memory threshold, the terminal may determine the user preference of each application process running in the background by using the method provided by the embodiment of the present disclosure, and determine the application process to be recycled, which is to be recovered. The application process is provided to the LMK process, and the LMK process can kill the application process to be recycled, thereby reclaiming the memory occupied by the application process to be recycled.

In this scenario, the LMK process preferentially kills application processes with low user preference, and stops the kill process when the free memory is not less than the memory threshold, thereby reducing the probability of killing the application process frequently used by the user, thereby avoiding the killing application. The process also avoids affecting the operation of the application process that the user frequently uses, and ensures the smoothness of the terminal operation.

Scene 2, kswapd process mechanism. When the kswapd process detects that the free memory is less than the memory threshold, the terminal may determine the user preference of each application process running in the background by using the method provided by the embodiment of the present disclosure, and determine the application process to be recycled, which is to be recovered. The application process is provided to the kswapd process, and the kswapd process transfers the data in the in-memory application process to the preset storage space to recover the memory occupied by the application process.

In this scenario, the kswapd process preferentially transfers the in-memory data of the application process with lower user preference to the preset storage space, and stops the process of transferring data when the free memory is not less than the memory threshold, thus reducing the kswapd process. The probability of transferring the data in the memory of the application process frequently used by the user to the preset storage space avoids affecting the operation of the application process frequently used by the user and ensuring the smoothness of the terminal operation.

Scene 3, the direct memory recovery mechanism. When any process requests memory from the terminal, and the requested memory exceeds the free memory, the terminal may determine the user preference of each application process running in the background by using the method provided by the embodiment of the present disclosure, thereby determining the application to be recycled. The process provides the process to be reclaimed to the process of requesting the memory, and the process of requesting the memory transfers the data in the memory of the application process to be recycled to the preset storage space to recover the memory occupied by the application process.

In this scenario, the process requesting memory preferentially transfers the data in the memory of the application process with lower user preference to the preset storage space, and stops the process of transferring data when the free memory exceeds the requested memory, thereby reducing the process. The probability of transferring the data in the memory of the application process frequently used by the user to the preset storage space avoids affecting the operation of the application process frequently used by the user and ensuring the smoothness of the terminal operation.

2 is a flowchart of a method for reclaiming memory according to an embodiment of the present disclosure. As shown in FIG. 2, the execution body of the method is a terminal, and the terminal may be a mobile phone, a tablet computer, or the like. The method includes the following steps:

201. The terminal collects an active duration of each application process in each application process in a historical operation.

Active duration: Active duration refers to the length of time the application process is running in the foreground. Whenever an application process is running in the foreground, the terminal can record the active duration: when the application process starts running in the foreground, the terminal can record the current time point as the start time of the active duration, when the application process ends in the foreground. The terminal can record the current time point as the end time point of the active duration, calculate the time difference between the end time point and the start time point, and obtain the active duration that the application process is currently running in the foreground. For example, when the application process starts running in the foreground at 9:00 and ends in the foreground at 9:10, the record active time is 10 minutes.

The time period: the time period may be divided into a day, for example, the time period may include a time period corresponding to the morning, a time period corresponding to the noon, or a time period corresponding to the night, and the duration of each time period may be 8 hours.

Optionally, the time period may be determined according to a physiological condition of the user, and a time period in which a certain physiological behavior is generated may be used as a time period. For example, the lunch time period may be used as a time period, and the dinner time period may be used as a time period. The time period of the sleep period is taken as a time period and the like, and the timeliness of the user preference can be improved by the method of determining the time period. For the detailed description of the technical effects, refer to the following step 204.

The specific method of counting the active duration: the terminal can accumulate the active duration of each application process from zero every time when it starts to enter any time period, that is, the current time point is the starting point of the time period, when leaving the time period, that is, the current time When the point is the end point of the time period, the accumulated active time of each application process in the time period is ended, the active duration accumulated by each application process in the time period is stored, and the next statistical process is started, and the accumulation is started again from zero. ,And so on. In the process of accumulating the active duration in each time period, each application process may be active one or more times in each time period, and may start from zero when the application process is active for the first time in each time period. Record the active duration of the application process until the application process is paused, and when the application process is active for the second time to the Nth active time, the active time of the application process may be added to the recorded time period. Active active time.

Exemplarily, if the current time point enters the time period of 8:00-10:00, if the application process A is active for the first time and is active for 5 minutes, then the active time of the application process A is 5 minutes, if the application process A is active for the second time and is active for 10 minutes. At this time, the active duration of application process A is 5+10=15 minutes. If application process A is active for the third time and is active for 8 minutes, then application process A is active. The duration is 15+8=23 minutes, and at the current time point, it reaches 10:00, and when the next time period is 10:00-13:00, the active duration of the application process A will start counting again from 0.

Optionally, each time the terminal acquires an active duration of the application process in the foreground, the time period to which the current time point belongs is obtained, the current date is obtained as the record date, and the application process identifier of the application process is obtained, and the record is active. The correspondence between the duration, the time period, the record date, and the application process identifier, so that the active duration can be obtained according to the correspondence. Illustratively, the terminal can record "00:00-08:00 (time period), December 7 (recording date), APP1 (application process identification), 50 minutes (active duration)".

It should be noted that this step 201 can be performed by the activity management service (AMS) of the terminal. AMS is the core service of the Android operating system. In the implementation, the program code of the AMS can be improved, and the program code of the active time of each application process running in the foreground is added to the program code of the AMS, so that the AMS is loaded and When you run the program code, you can implement the function of counting active time.

202. Determine, when the terminal detects that the free memory does not meet the preset condition, determine that multiple application processes are running in the background.

To ensure the normal operation of the terminal, the terminal can detect whether the free memory meets the preset condition during the running process. When it detects that the free memory does not meet the preset condition, it determines the multiple application processes running in the background, so as to Determine the application process to be recycled, and then perform memory recycling. The preset condition and the process of detecting whether the free memory meets the preset condition may specifically include the following two designs:

The default condition is that the free memory is less than the memory threshold, and the terminal can detect whether the current free memory is less than the memory threshold in real time or periodically. The memory threshold is used to indicate the minimum memory required for the terminal to operate normally. When the memory threshold is less than the memory threshold, the terminal determines that the free memory does not meet the preset condition. Optionally, the terminal may start the LMK process or the kswapd process, and the LMK process or the kswapd process performs a process of detecting whether the free memory is less than a memory threshold.

When the LMK process detects that the free memory is less than the memory threshold, it can obtain the oom_adj of all the currently running processes, and generate a list of the to-be-killed processes according to the oom_adj of each process, and the list of the to-be-killed processes includes the LMK process that is determined to be able to be killed in the background. The running process, the terminal can determine the application process associated with each process in the to-be-killed process list, as an application process that subsequently acquires user preference and can reclaim memory.

The second preset condition may be that the requested memory exceeds the free memory. When any application process requests memory from the terminal, the memory request may be sent to the terminal, where the memory request carries the memory requested by the application process, and the terminal may request the memory according to the memory request. The memory requested by the application process is obtained to determine whether the requested memory exceeds the free memory. When the requested memory exceeds the free memory, the terminal may determine that the free memory does not meet the preset condition.

203. The terminal determines a user preference of each application process in the multiple application processes at a time period to which the current time point belongs.

After the terminal determines that multiple application processes are currently running in the background, the user preference of each application process in the multiple application processes is determined, so as to determine from multiple application processes based on the user preference of each application process. The application process to be recycled. Optionally, the specific process of determining the user preference may include the following steps 1 to 2:

Step 1: Obtain an active duration of a time period to which the current time point of each application process belongs in the historical operation.

For each application process in a plurality of application processes, the terminal may obtain an active duration of the time period to which the current time point of the application process belongs in the historical operation, and the active duration reflects the activity level of the application process in the historical operation. Since the time period is determined according to the current time point, it can be considered that the active duration matches the activity level of the application process before and after the current time point in the current running process, thereby ensuring that the user preference determined based on the active duration is highly time-sensitive. High accuracy.

The process of obtaining the active duration of the time period of the application process in the historical running may specifically include the following steps (1)-(3):

(1) Obtain the time period.

The terminal can acquire the current time point and determine the time period to which the current time period belongs.

(2) Obtain the historical period.

The terminal may determine the historical period according to the current date and the preset statistical period, where the current period is the time end point and the duration is the preset statistical period. Taking the current date as December 7, 2017 as an example, assuming that the preset statistical period is 1 day, the historical period is from December 6, 2017 to December 7, 2017, assuming that the preset statistical period is one month, then The historical period is from November 7, 2017 to December 7, 2017. The preset statistical period can be determined according to the actual requirement. The specific value of the preset statistical period is not limited in this embodiment.

(3) According to the historical period and the time period, obtain the active duration of the application process in the historical operation period.

The terminal can query the correspondence between the active duration, the time period, the recording date, and the application process identifier according to the historical period, the time period, and the application process identifier of the application process, and obtain the active corresponding to the time period, the historical period, and the application process identifier. The duration, that is, the corresponding time period, the corresponding recording date is the historical period, and corresponds to the active duration of the application process identifier, as the active duration of the application process in the historical operation. For example, suppose the historical period is from November 7, 2017 to December 7, 2017, the time period is 17:00-19:00, the application process ID is service1, and the terminal can obtain service1 from November 7, 2017 to 2017. The active time of 17:00-19:00 on December 7th.

Step 2: Determine user preference of each application process in multiple application processes according to the active duration of the time period of each application process in the plurality of application processes.

After obtaining the active duration of each application process in the historical running period (that is, the time period to which the current time point belongs), the terminal may acquire the user preference of each application process based on the active duration, so as to be based on the user preference. Determine the application process to be recycled. The specific process of obtaining user preferences can include the following designs:

Design one, the active time is taken as the user preference.

That is, the terminal can directly use the active duration of the application process in the historical running period as the user preference of the application process, thereby obtaining the user preference simply and effectively.

Design 2: Calculate the ratio between the active duration and the active duration as the user preference.

The sum of the active durations refers to the sum of the active durations of the multiple application processes currently running in the background during the historical running. For the process of obtaining the sum of the active durations, the terminal can traverse each application process in multiple application processes. The active duration of each application process is determined by the method provided in the foregoing step 203, and the active duration of the multiple application processes is obtained, and the sum of the active durations is calculated to obtain the sum of the active durations.

In this design, when the user preference of an application process is to be determined, the ratio between the active duration of the application process and the sum of the active durations is calculated, and the ratio is used as the user preference, and the user preference may also be referred to as The proportion of active time, that is, the proportion of the active duration of the application process in the active duration of multiple application processes. By combining the active duration of the application process with the active duration of multiple application processes to obtain user preference, the following technical effects can be achieved:

Ensure that the user preference is within a reasonable range of values: Considering the actual application process, the user's overall usage of the terminal on different dates may be different, and the active duration obtained when the user preference is counted on different dates will also be different. For example, the holiday can use the peak period of the terminal for the user. During the holiday period, the active time of each application process in each time period is usually longer. If the user preference is counted at this time, the active time length of each application process obtained is obtained. Long, and by calculating the ratio between the active duration of the application process and the sum of the active durations, it is guaranteed that the user's preference is not too high. For example, the test week and the workday may be the low tide period for the user to use the terminal. During the test week and the working day, the active time of each application process in each time period will be short. If the user preference is collected at this time, the obtained time is obtained. The active duration obtained by each application process is short, and by calculating the ratio between the active duration and the active duration, the user preference is not too low. That is, the user preference of the calculated application process is relatively stable, and the situation is not too high or too low due to the influence of the current date, so that the subsequent application process to be recovered based on the user preference can be guaranteed. Recycle the accuracy of the memory process used by the application process.

Design 3: Calculate the product between the active duration and the cold start duration of the application process as the user preference.

The cold start duration of an application process refers to the length of time it takes for the application process to start the corresponding application in a cold boot manner. For the method of obtaining the cold boot duration, the relationship between the cold boot duration and the application process that starts the application may be recorded in advance when any application is started in a cold boot manner, and the application may be acquired when the cold boot duration is to be obtained. The cold-start duration of the most recent record of all cold-start durations that have been recorded by the process to calculate user preference based on the cold-start duration. In addition, it is also possible to obtain the cold start duration of the last n records of all the cold boot durations recorded by the application process, and calculate the average of the n cold boot durations as the obtained cold boot duration.

In this design, when the user preference is calculated, the cold start duration of the application process is obtained, and the active duration is weighted by the cold start duration, that is, the product between the cold start duration and the active duration is calculated as the user preference, by combining The cold start time and active time to obtain user preference can achieve the following technical effects:

For application process A and application process B with the same active time, it is assumed that the cold start time of application process A is long, the cold start time of application process B is short, and the application process A is slow when cold start, if the application is recycled. The memory occupied by process A, and then the application process A is started again. The slower speed will have a greater impact on the operation of the terminal, and it will easily bring the user experience of the terminal. In this design, by weighting the active duration by using the cold start duration, the user preference of the application process A can be appropriately increased, and when the application process to be recovered is subsequently determined, the application process A is determined to be determined as the application process to be recycled. In addition, when the application process A is started again, the long-term cold running time has a large impact on the operation of the terminal.

Similarly, the application process B is faster when it is cold-started. Even if the memory occupied by the application process B is recovered, the application process B is restarted again because of the faster speed, which has less impact on the terminal operation and is less likely to cause the terminal. Caton's user experience. In this design, by weighting the active duration by using the cold boot duration, the user preference of the application process B can be appropriately attenuated, and then the application process B is preferentially used as the application process to be recycled, and the memory occupied by the application process B is preferentially recycled. That is, on the basis of accurately recovering the memory occupied by the application process, taking into account the cold boot time of each application process, and minimizing the overall impact on the terminal.

It should be noted that, in the implementation, the application may start multiple application processes during the running process, and the main process in the multiple application processes starts the application in a cold start manner or a hot start manner, and each of the main processes The child processes are used to provide functions other than launching the application. In this step, when the application process is the main process, the cold boot time of the corresponding application started by the recorded main process can be directly obtained to calculate the user preference of the main process. When the application process is a child process, the cold boot time of the main process corresponding to the child process may be determined as the cold start duration of the child process to calculate the user preference of the child process.

Design 4. Calculate the ratio between the active duration and the active duration and value, and calculate the product between the ratio and the cold start duration of the application process as the user preference.

This design is the combination of design two and design three-phase. The acquisition method of active duration and value is detailed in Design 2. The acquisition method of cold start time is detailed in Design III, and will not be described here. In this design, on the basis of providing a method for accurately obtaining user preference, the user preference can be guaranteed to be within a reasonable numerical range, and the cold start time of the application process is taken into consideration, and the memory occupied by the application process is guaranteed to be recycled to the terminal. The overall impact is as small as possible.

It should be noted that, when the time period in the above step 201 is determined according to the physiological behavior of the user, in this step, the effect that the user preference matches the physiological habits of the user can be achieved: for example, considering that the user usually has a lunch time period and During the dinner time period, the application process of the take-out ordering application is preferred, and the lunch time period and the dinner time period are set as the time period in advance to calculate the active time length in real time, and if the current time point is the lunch time period and the dinner time period, the memory is to be recovered, and the calculation is performed. The user preference of the application process of the take-out ordering application is higher, and the memory occupied by the application process of the take-out ordering application can be avoided, and the application process of the user who prefers to use the music playing application during the sleeping time period is considered, for example. The sleep time period is set to the time period in advance to calculate the active duration in real time. If the memory is to be reclaimed during the current sleep time period, the calculated user preference of the application process of the music playing application is higher, then Avoid reclaiming memory used by the app process of the music player application.

204. The terminal acquires a recycling priority of each application process based on user preference of each application process.

The terminal can obtain the recycling priority of each application process, so as to determine the application process to be recycled based on the recycling priority of each application process. The reclaiming priority of each application process is negatively related to the user preference of each application process. If the user preference of the application process is higher, the revocation priority of the application process is lower, and the user preference of the application process is lower. , the higher the recycling priority of the application process.

Further, on the basis of determining the revocation priority based on the user preference, the embodiment further provides the memory ratio of the user preference and the application process, the function of the application process, the latest active duration of the application process, and the application. The power consumption of the process and other factors together determine the way to reclaim the priority. See Design 1 to Design 4 below for details. It should be noted that the value of the recycling priority in the following various designs is inversely proportional to the level of the recycling priority. That is, the smaller the value of the recycling priority, the higher the recycling priority, the larger the value of the recycling priority, and the recycling priority. The lower.

Design 1. Determine the recycling priority by combining the user preference with the memory ratio of the application process.

The design may specifically include the following steps 1 to 3:

Step 1: The terminal determines the memory weight corresponding to the memory ratio of each application process, and the memory ratio of each application process is used to indicate the ratio between the memory occupied by each application process and the total memory overhead.

Memory ratio: For each application process in multiple application processes, the terminal can obtain the memory occupied by each application process and the current total memory cost, and calculate the memory occupied by each application process and the total memory overhead. Ratio, the ratio is taken as the memory ratio. The total memory overhead refers to the total memory occupied by the terminal, which is the sum of the memory occupied by various application processes running on the terminal. Exemplarily, if the memory occupied by the application process is 100M and the current total memory overhead is 1000M, the memory ratio of the application process is 100M/1000M=10%.

Memory weight ratio: The terminal can obtain the memory weight ratio corresponding to the memory ratio of the application process based on the correspondence between the memory ratio of the application process and the memory weight. The correspondence between the memory ratio and the memory weight ratio includes a plurality of memory ratios and corresponding memory weight ratios. The larger the memory ratio, the larger the corresponding memory weight. The correspondence between the memory ratio and the memory weight can be set by the developer in advance. After the terminal obtains the memory ratio, the memory ratio is used as an index, and the correspondence between the memory ratio and the memory weight ratio is queried, and the memory ratio weight corresponding to the memory ratio is obtained from the corresponding relationship.

Optionally, the terminal may quantize the memory ratio. For example, multiple memory ratio intervals may be set, and a correspondence between the memory ratio interval and the memory weight ratio is set. When the memory ratio of the application process is obtained, The memory ratio ratio of the memory ratio may be obtained from a plurality of memory ratio intervals, and the memory ratio of the application process's memory ratio interval is obtained based on the correspondence between the memory ratio interval and the memory ratio weight. Weights. Among them, the memory ratio interval can be set by the developer according to actual needs, for example, setting 1%-10% as a memory ratio interval.

By quantifying the memory ratio, all the memory ratios in a certain range of values can be uniformly unified with a certain memory weight, and the correspondence between the memory ratio and the memory weight is not included in the value range. All memory ratios only need to include the memory ratio interval, which simplifies the correspondence and reduces the amount of data to be stored.

Step 2: The terminal obtains the recycling priority of each application process according to the user preference of each application process and the memory weight of each application process.

The recycling priority of the application process is positively correlated with the memory weight and is negatively related to user preference. Optionally, the ratio between the user preference and the memory weight may be calculated to obtain a recycling priority. Illustratively, the recycling priority can be calculated using the following formula: Yi = Bi / Ai, where Yi represents the recycling priority, Bi represents the user preference, and Ai represents the memory weight.

In this design, the greater the memory weight of the application process, the higher the reclaim priority. Therefore, the greater the probability that the application process is determined to be reclaimed, the smaller the memory weight of the application process is, and the reclaim priority is reclaimed. The lower the probability, the less likely the subsequent process is determined to be the application process to be recycled.

By determining the reclaiming priority by combining the user preference with the memory ratio of the application process, the accuracy of the reclaimed memory can be further improved: for a plurality of application processes whose user preference is close, assuming that the memory of an application process is relatively large, The memory has a larger weight, so the application process has a higher reclaim priority, and the memory of the application process is preferentially reclaimed. It can ensure that more memory is reclaimed, and the free memory is guaranteed to meet the memory requirement as quickly as possible.

Design 2. Combine user preference and application process functions to determine the recycling priority of the application process.

The design may specifically include the following steps 1 to 2:

Step 1: The terminal determines a function weight corresponding to the function of each application process.

The function of the application process: it can include push function, network connection function, etc. The application can open one or more application processes at the same time, and each application process can have corresponding process functions according to the currently executing activities, for example, the instant communication application can open the main function. The process, the push sub-process, the networked sub-process, etc., push the sub-process to push the message, and the networked sub-process performs the network connection. For any application process, the terminal can interact with the application process to learn the process function of the application process.

The process of determining the function weight corresponding to the function of the application process: for each application process in multiple application processes, after the terminal determines the process function of each application process, the process function is used as an index to query the process function and the function weight Corresponding relationship, the function weight corresponding to the process function is obtained from the corresponding relationship. The corresponding relationship between the process function and the function weight includes multiple functions and corresponding multiple function weights, which can be determined according to experience. For example, the more important the process function is, the larger the function weight is, and the more important the process function is. , the smaller the function weight is set. Exemplarily, considering that the application process of the push function usually disturbs the user during the operation, and affects the speed at which the terminal receives the network data, a small function weight can be set for the push function to ensure that the application process of the push function is preferentially recycled. The level is higher, which in turn preferentially reclaims the memory of the application process.

Step 2: According to the user preference of each application process and the function weight of each application process, obtain the recycling priority of each application process.

Among them, the recycling priority is negatively related to the function weight and negatively related to the user preference. Further, the product between the user preference and the function weight can be calculated to obtain the recycling priority. Illustratively, the recycling priority can be calculated using the following formula: Yi = Bi x Di, where Yi represents the recycling priority, Bi represents the user preference, and Di represents the functional weight.

In this design, the smaller the function weight of the application process is, the higher the reclaim priority is. Therefore, the greater the probability that the subsequent application process is determined to be reclaimed, the higher the function weight of the application process is, the lower the reclaim priority is. The greater the probability that the subsequent application process is determined to be recycled.

By combining the user preference and the function weight to determine the recycling priority, the accuracy of the reclaimed memory can be further improved: for a plurality of application processes with similar user preferences, if the process function of an application process is less important, it can be considered as The current running activity of the application process is of low importance to the terminal, and the impact of reclaiming the memory of the application process is small. When the re-prioritization priority is determined by the function weight, the revocation priority of the application process is higher, and the terminal takes precedence. Reclaim the memory of the application process to minimize the overall impact of the reclaimed memory on the terminal.

Design 3. Combine the user preference with the most recent active duration of the application process to determine the recycling priority of the application process.

The design may specifically include the following steps 1 to 3:

Step 1: Determine the most recent active weight of each application process, and the most recent active weight of each application process is used to indicate the extent to which each application process is recently active.

The specific process for determining the most recent active weights may include the following (1)-(2).

(1) Obtain the latest active duration of each application process, and the latest active duration of each application process is used to indicate the active duration of the application process running in the foreground.

For each application process in multiple application processes, the terminal can obtain the active time duration of the active time duration recorded by the application process from the current time point to obtain the latest active duration.

(2) Attenuating the most recent active duration of each application process based on the time difference between the end time point of the most recent active duration and the current time point, and using the decayed most recent active duration as the most recent active weight.

The terminal can obtain the end time point of the most recent active time, obtain the current time point, calculate the time difference between the end time point and the current time point, and attenuate the latest active time length based on the time difference, and use the decayed latest active time length as the most recent active time. Weights.

For the specific process of attenuating the most recent active time based on the time difference, the attenuation function can be used to calculate the time difference and the most recent active time, and the calculation result is used as the most recent active weight. Wherein, the attenuation function can be Ti=w×y^(-t), Ti represents the most recent active weight, w represents the normalized most active duration, y is a constant, and t represents the time difference.

Based on the above calculation method, the time difference can reflect the distance between the latest running process of the application process and the current time point. When the time difference of an application process is small, it indicates that the application process is the most active application process, and the application can be considered as the application. The process is more important to the user at present, and the attenuation is less when the latest active duration is attenuated based on the time difference, so as to ensure that the recent active weight of the application process is large. When the time difference of the application process is large, it indicates that the application process is an active application process that is active a long time ago. It can be considered that the application process is less important to the user at present, and the attenuation is greater when the latest active time is attenuated based on the time difference. Large to ensure that the most recent active weight of the application process is small.

Step 2: According to the user preference and the recent active weight of each application process, obtain the recycling priority of each application process.

The reclaim priority of the application process is negatively correlated with the most recent active weight and is negatively correlated with user preference. Alternatively, the product between the user preference and the most recent active weight can be calculated to obtain a recycling priority. Illustratively, the recycling priority can be calculated using the following formula: Yi = Bi x Ti, where Yi represents the recycling priority, Bi represents the user preference, and Ti represents the most recent active weight.

In this design, the smaller the recent active weight of the application process, the higher the reclaim priority, so the probability that the subsequent application process is determined to be reclaimed is larger. The greater the recent active weight of the application process, the lower the revocation priority, so the probability that the subsequent application process is determined to be reclaimed is smaller.

By determining the reclaim priority in combination with the user preference and the most recent active weight of the application process, the accuracy of the reclaimed memory can be further improved: for multiple application processes with similar user preferences, if an application process is the most recent active application process The recent active weight will be larger, so the application process will have lower reclaim priority, which will avoid reclaiming the memory of the application process, so as to avoid affecting the normal operation of the recently active application process. If an application process is an active application process that is active for a long time, the active weight of the application is relatively small. Therefore, the application process has a higher revocation priority, and the memory of the application process is preferentially reclaimed. The application process that was active long ago is preferentially recycled. The occupied memory, on the basis of realizing the recovery of memory, ensures that the overall impact on the operation of the terminal is as small as possible.

Design 4. Determine the recycling priority based on user preference and power consumption of the application process.

The design may specifically include the following steps 1 to 2:

Step 1: Determine the power consumption weight corresponding to the power consumption of each application process.

Power consumption: Power consumption refers to the power consumption of the application process. For the process of obtaining power consumption, the terminal can record the power consumption of the application process in real time while the application process is running. In this step, for each application process in multiple application processes, the terminal may obtain the power consumption of the recorded application process in the historical operation, for example, the power consumption of the application process may be obtained, and the application process may be acquired. The power consumption of the last preset number of runs, and the average of the power consumption of the preset number of times is calculated.

The process of determining the power consumption weight: the terminal can use the power consumption as an index, query the correspondence between the power consumption and the power weight, and obtain the power consumption weight corresponding to the power consumption from the corresponding relationship. The corresponding relationship between power consumption and power consumption weight includes multiple power consumptions and corresponding multiple power consumption weights. The greater the power consumption, the greater the corresponding power weight, and the corresponding relationship between power consumption and power weight can be It is preset by the developer and can be stored in the terminal in advance.

Step 2: According to the user preference of each application process and the power consumption weight of each application process, obtain the recycling priority of each application process.

Among them, the recycling priority is positively correlated with the power consumption weight and negatively related to the user preference. Further, the ratio between the user preference and the power consumption weight can be calculated to obtain the recycling priority. Illustratively, the recycling priority can be calculated using the following formula: Yi = Bi / Hi, where Yi represents the recycling priority, Bi represents the user preference, and Hi represents the power consumption weight.

In this design, the greater the power consumption weight of the application process, the higher the reclaim priority, and the greater the probability that the subsequent application process is determined to be reclaimed. The smaller the power consumption weight of the application process, the lower the reclaim priority, and the smaller the probability that the subsequent application process is determined to be reclaimed.

By determining the application process to be recycled in combination with the user preference and the power consumption of the application process, the following technical effects can be achieved: for a plurality of application processes with similar user preferences, if the power consumption of an application process is greater, the work is performed. If the weight of the application is higher, the subsequent terminal will preferentially reclaim the memory occupied by the application process. By preferentially reclaiming the memory occupied by the application process with high power consumption, the power consumption of the terminal can be saved on the basis of realizing the function of reclaiming memory.

Design 5. Determine the application process to be recycled in combination with the user preference and the calling frequency of the application process. The calling frequency refers to the total number of times the application process calls the application process during the time period in the historical running.

The design may specifically include the following steps 1 to 3:

Step 1: Determine the calling frequency of each application process, and the calling frequency of each application process is used to indicate the total number of times that multiple application processes call each application process in the historical running time period.

The specific process of determining the calling frequency may include the following (1)-(2):

(1) The terminal acquires a call record of each application process.

The call record is also called the call list. The call record of each application process is used to record the events of each application process in each application process except for each application process in the application process. When each application process runs, the terminal can be in real time. The event record of each application process records the events of the application process calling other application processes. For example, for the application process 1 - application process N, in the call record of the application process 1, the application process 1 is recorded to call the event in the application process 2 - application process N.

(2) The terminal determines the calling frequency of each application process based on the call record of each application process.

For each application process in multiple application processes, the terminal can calculate the sum of the number of calls to each application process in the multiple call records as the calling frequency of each application process. For example, for the application process 1, the sum of the number of calls of the application process 2 - the application process N calling the application process 1 can be obtained according to the application process 2 - the call record of the application process N, as the calling frequency of the application process 1.

Step 2: Obtain the recycling priority of each application process based on the calling frequency of each application process.

Recycling priority is negatively correlated with the calling frequency and negatively related to user preference. Optionally, the terminal can adjust the reclaim priority based on the calling frequency. The terminal may set the first frequency threshold and the second frequency threshold. After the terminal determines the calling frequency of each application process, it may determine whether the calling frequency is higher than the first calling frequency threshold, when the calling frequency is higher than the first calling frequency threshold. The application process is considered to be an application process that is frequently called. Recycling the memory occupied by the application process will have a greater impact on the operation of other application processes, and the application process's recycling priority will be lowered to avoid recycling applications. The memory occupied by the process. Similarly, the terminal can determine whether the calling frequency of the application process is lower than the second frequency threshold. When the calling frequency is lower than the second calling frequency threshold, the application process is considered to be an application process that is rarely called, and the application process is recycled. After the memory has less impact on the operation of other application processes, the recycling priority of the application process is increased. In addition, when the calling frequency is not higher than the first frequency threshold and not lower than the second frequency threshold, the terminal may not need to adjust the recycling priority of the application process. The first call frequency threshold is used to indicate the lowest value of the calling frequency of the frequently invoked application process, and may be set according to actual requirements. The second call frequency threshold is used to indicate the highest value of the calling frequency of the application process that is rarely called, and can be set according to actual needs.

Regarding the process of lowering the recycling priority, the terminal can set a preset adjustment value, and the preset adjustment value can be a positive number, and the terminal can calculate the sum of the value of the application process recovery priority and the preset adjustment value, and adjust the sum value as a tune. Low recycling priority. Alternatively, the preset adjustment value may be a positive number greater than 1, and the terminal may calculate the product of the application process recovery priority value and the preset adjustment value, and use the product as the reduction priority after the reduction.

Regarding the process of increasing the priority of recycling, the terminal can set a preset adjustment value, and the preset adjustment value can be a positive number, and the terminal can calculate the difference between the value of the application process recovery priority and the preset adjustment value, and adjust the difference as a tone. High recycling priority. Alternatively, the preset adjustment value may be a positive number greater than 1, and the terminal may calculate a ratio of the value of the application process recovery priority to the preset adjustment value, and use the ratio as the recovery priority after the adjustment.

In this design, the smaller the calling frequency of the application process, the higher the recycling priority, and the greater the probability that the subsequent application process is determined to be recycled. The higher the calling frequency of the application process, the lower the reclaiming priority, and the smaller the probability that the subsequent application process is determined to be reclaimed. By determining the reclaiming priority by combining the user preference and the calling frequency of the application process, it is possible to ensure that the reclaiming priority of the frequently invoked application process is low, thereby avoiding the operation of other application processes caused by reclaiming the frequently invoked application process. Make an impact.

It should be noted that the above design one to design five can form a method for determining the recycling priority by any combination. For example, when design one is combined with design three, the terminal can obtain the memory weight and the recent active weight, and determine the recycling priority according to the user preference, the memory weight, and the recent active weight. For example, when design one, design two, design three and design four are combined, the terminal can obtain memory weight, function weight, recent active weight and power consumption of the application process, according to user preference, memory weight, function weight The recent active weights and power consumption of the application process determine the recycling priority. For example, the terminal can calculate the recycling priority by the following formula: Yi=(Bi×Di×Ti)/(Ai×Hi), where Yi represents the recycling priority, Bi represents the user preference, Di represents the functional weight, Ti represents the nearest The active weight, Ai represents the memory weight, and Hi represents the power weight.

205. The terminal determines, according to a recycling priority of each application process, an application process to be reclaimed in multiple application processes.

The terminal can sort the application processes according to the reclaiming priority. The application process with the highest priority is used as the application process to be reclaimed, so as to reclaim the memory occupied by the application process. After that, if the free memory still does not meet the preset condition, the application process with the second highest priority is continuously used as the application process to be recycled. Then, the higher the revocation priority of the application process, the earlier the memory occupied by the application process is reclaimed. The lower the revocation priority of the application process, the later the memory occupied by the application process is reclaimed.

It should be noted that the above steps 204-205 are only based on determining the revocation priority based on the user preference, and then determining the application process to be reclaimed based on the reclaiming priority. In the implementation, the user preference may also be directly based on the user preference. Determine the application process to be reclaimed. If the user preference of the application process is higher, the application process is avoided as the application process to be reclaimed. If the user preference of the application process is lower, the application process is preferentially considered as to be recycled. Application process.

The following technical effects can be achieved by determining the application process to be recycled based on user preference:

First, improve the accuracy of the reclaimed memory: for the high-frequency application process (that is, the application process that runs normally frequently), the active duration of the application process will be longer, and the user preference corresponding to the active duration will be higher, then Avoid the high-frequency application process as the application process to be recycled, ensure that the data of the high-frequency application process resides in the memory, try to avoid affecting the normal operation of the high-frequency application process, and reduce the cold start probability of the high-frequency application process.

Second, improve the timeliness of reclaiming memory: for an application process that does not normally run, the active duration of the application process will be shorter, and the user preference corresponding to the active duration will be smaller, then the low-frequency application process will be prioritized. As the application process to be recycled, the memory occupied by the low-frequency application process is highly recovered, and the memory occupied by the low-frequency application process can be recovered in time.

Third, the process of ensuring the memory recovery is close to the user's personal usage habits: considering that the application process preferred by different users in daily life may be different, for example, user A prefers the application process using the instant messaging application, and user B prefers to use the music playing application. The application process, user C prefers to use the application process of the online shopping application, and the like. In this embodiment, the user preference of the application process is not uniformly set, but the terminal calculates the user preference of each application process in real time by self-learning. The user preference of each application process can reflect the user's personal usage habits. Therefore, when the memory is recycled in combination with the user preference, the personalized requirement can be satisfied. For example, the probability that the user A's terminal reclaims the application process of the instant messaging application has the smallest probability, and the user has the smallest probability. The terminal of B recycles the memory of the application process of the music playing application with the least probability of the memory and the like.

Fourth, the process of ensuring the recovery of memory is time-sensitive: considering that the application process preferred by the user can change over time, for example, an application process in which a user prefers to use the Weibo application in the past, and now prefers an application using an instant messaging application. Process, etc. In this embodiment, the user preference for each application process is not obtained in advance, but each application is calculated in real time according to the active duration of the historical running process of the application process when the terminal reclaims the memory occupied by the application process. The user preference of the process, that is, the user preference of the self-learning application process, the calculated user preference can be dynamically changed with time, and the order of memory occupied by the application process is recycled each time. Dynamic changes, for example, when the terminal calculates the user preference last month, the application process of the instant messaging application has the highest user preference, and the probability of reclaiming the memory of the application process of the instant messaging application is the smallest, and the music playing application is calculated when the user preference is calculated this month. The application process has the highest user preference, and the probability of reclaiming the memory of the application process of the music playing application is the smallest.

206. The terminal reclaims memory occupied by the application process to be recycled.

After the terminal determines the application process to be recycled, the memory occupied by the application process is reclaimed. After the memory occupied by the application process is reclaimed, it is possible to detect whether the free memory meets the preset condition. When the free memory has met the preset condition, the application process to be reclaimed need not be determined again, and the process of reclaiming the memory is terminated. . When the idle memory still does not meet the preset condition, the method provided in this embodiment is used again to continue to determine the application process to be recovered until the idle memory meets the preset condition.

Exemplarily, when it is detected that the idle memory does not meet the preset condition, it is determined that there are 5 application processes running in the background. After the memory occupied by the two application processes is recovered, it is determined that the free memory has met the preset condition, then You can end the process of reclaiming memory without having to continue to reclaim the memory occupied by the remaining 3 application processes.

Optionally, in combination with determining the design of the application process to be reclaimed according to the reclaiming priority, the terminal reclaims the application process with the highest priority as the application process to be reclaimed, and after reclaiming the memory occupied by the application process, can detect the free memory again. When the preset condition is met, the process of reclaiming the memory is not required to determine the application process to be reclaimed. When the idle memory still does not meet the preset condition, the application process with the second highest priority is reclaimed as the application process to be reclaimed, so as to reclaim the memory of the application process with the second highest priority, and so on.

The present embodiment provides the following two designs for the specific process of reclaiming the memory of the application process:

Design one, kill the application process to be recycled.

After killing the application process to be reclaimed, the memory occupied by the application process is released, and the effect of reclaiming the memory occupied by the application process is achieved. Optionally, the terminal can run the LMK process, and the LMK process kills the application process to be recycled.

It should be noted that, in this embodiment, the application process is a unit for counting user preference and reclaiming memory. When the application starts multiple application processes, only one application process of multiple application processes may be killed without killing other applications. The application process, and the application can still be running. For example, a reading application can open multiple application processes such as a main process, a picture display sub-process, and a text display sub-process. When calculating the user preference, the picture display sub-process is preferentially killed due to the lower user preference, and the reading is performed at this time. The application can continue to run through the main process, and the text can continue to be displayed through the text display subprocess, only the image can not be displayed temporarily.

Design 2: Transfer the data in the memory of the application process to be recycled to the preset storage space.

The terminal can sequentially transfer the data in the memory of each application process to the preset storage space according to the order of the reclaiming priority, so that the memory occupied by the application process in the memory is freed to be occupied by the recycling application process. The effect of memory. Optionally, the data may be compressed first, and the data volume of the compressed data is reduced, and then the compressed data is transferred to the preset storage space. For example, if an application process stores 50M of data in the memory, You can compress 50M data into 20M, and then store 20M data into the preset storage space, then the memory will be free of 50M space. The preset storage space may be a zram partition, and the zram partition is also called a Compcache (compression cache), which refers to a module of the Linux kernel. The zram can replace the hard disk with a memory to provide a storage data function for the operating system, and can be used for storage. The data of the compressed application process.

Optionally, the terminal can run the kswapd process or the process that requests the memory before, and the process of the in-memory data of each application process is sequentially transferred to the preset storage space by the kswapd process or the process that previously requested the memory.

It should be noted that, referring to FIG. 3, a schematic diagram of a method for reclaiming memory provided by an embodiment of the present disclosure is provided. A developer may design a program code for reclaiming a memory process in a module by referring to the architecture of FIG. 3. The calculation factors may be divided into four modules according to the attribute, the requirement module includes a memory requirement and a free memory, and the user preference module includes an active duration, a time period, a recent active weight, and a calling frequency, and the application process attribute module includes an application process. The memory ratio, cold boot time, and process functions, and the expansion module includes the power consumption of the application process. In addition, the calculation factor of each module can be divided into five levels according to importance. The developer can set the value range of the calculation factor according to the level of the calculation factor. For example, the higher the level of the calculation factor, that is, the stronger the importance, the more can be set. The range of values of the calculation factor is higher, and the lower the level of the calculation factor, that is, the weaker the importance, the lower the range of values of the calculation factor can be set.

The method provided in this embodiment takes the user's usage habits into consideration when reclaiming the memory, determines the application process to be recycled according to the user preference of each application process, and then recovers the memory occupied by the application process. Priority is given to reclaiming memory occupied by application processes with low user preference, avoiding reclaiming memory occupied by application processes with high user preference, ensuring the accuracy and rationality of retrieving the memory of each application process, and avoiding the killing process and The application process data is transferred to the preset storage space by mistake, so as to avoid affecting the normal operation of the application process, and ensuring that the overall impact of the process of reclaiming memory on the operation of the terminal is as small as possible.

FIG. 4 is a schematic structural diagram of an apparatus for reclaiming memory according to an embodiment of the present disclosure. As shown in FIG. 4, the apparatus includes: a determining module 401 and a recycling module 402.

The determining module 401 is configured to determine, when the idle memory does not meet the preset condition, determine a plurality of application processes that are running in the background;

The determining module 401 is further configured to determine a user preference of a time period to which the application process belongs to the current time point in the plurality of application processes, where the user preference is used to indicate a possibility of the application process being invoked;

The determining module 401 is further configured to determine, according to user preference of each application process in the multiple application processes, an application process to be recycled in the multiple application processes;

The recycling module 402 is configured to recover the memory occupied by the application process to be recycled.

In a possible design, the determining module 401 is further configured to: determine, according to an active duration of the time period of each application process in the plurality of application processes, the application duration of the multiple application processes User preference for the time period to which the current time point belongs.

In a possible design, the determining module 401 is further configured to:

Using the active duration as the user preference; or,

The ratio of the active duration to the sum of the active durations is calculated as the user preference, and the sum of the active durations is the sum of the active durations of the plurality of application processes in the historical operation; or

Calculating the product of the active duration and the cold start duration of the application process as the user preference; or

The ratio between the active duration and the sum of the active durations is calculated, and the product between the ratio and the cold start duration of the application process is calculated as the user preference.

In a possible design, the determining module 401 further includes:

Obtaining a sub-module, configured to obtain, according to user preference of each application process in the multiple application processes, a reclaiming priority of each application process, and a reclaiming priority of each application process and a user of each application process Preference is negatively correlated;

Determining a sub-module for determining an application process to be reclaimed in the plurality of application processes based on a reclaiming priority of each application process.

In a possible design, the obtaining sub-module is further configured to determine a recent active weight of each application process, and a recent active weight of each application process is used to indicate a degree of recent activeness of each application process; The user preference of each application process and the most recent active weight of each application process, and the recycling priority of each application process is obtained.

In a possible design, the obtaining sub-module is further configured to obtain a recent active duration of each application process, where a recent active duration of each application process is used to indicate that each application process is last running in the foreground. The active duration; based on the time difference between the end time point of the latest active duration of each application process and the current time point, attenuating the most recent active duration of each application process to obtain the most recent active duration after the attenuation, The most recent active time after decay is used as the most recent active weight.

In a possible design, the obtaining sub-module is further configured to: determine a function weight corresponding to a function of each application process, where a function weight of each application process is used to indicate a function of each application process The degree of importance; according to the user preference of each application process and the functional weight of each application process, the recycling priority of each application process is obtained.

In a possible design, the obtaining sub-module is further configured to: determine a power consumption weight corresponding to the power consumption of each application process, and the power consumption weight of each application process is used to indicate the each application process The size of the power consumption; according to the user preference of each application process and the power consumption weight of each application process, the recycling priority of each application process is obtained.

In a possible design, the obtaining submodule is further configured to: determine a memory ratio weight corresponding to a memory ratio of each application process, and a memory ratio of each application process is used to indicate each of the The ratio of the memory occupied by the application process to the total memory cost; and the reclaiming priority of each application process is obtained according to the user preference of each application process and the memory weight of each application process.

In a possible design, the obtaining sub-module is further configured to: determine a calling frequency of each application process, and the calling frequency of each application process is used to indicate that the multiple application processes are in the historical running time period The total number of times each application process is invoked; the revocation priority of each application process is obtained according to the user preference of each application process and the calling frequency of each application process.

In a possible design, the obtaining sub-module is further configured to: obtain a call record of each application process, and the calling record of each application process is used to record that each application process calls the multiple application processes. An event of an application process other than each application process; determining a calling frequency of each application process based on the invocation record of each application process.

In a possible design, the device further includes: a statistics module, configured to count an active duration of the time period of each application process in the plurality of application processes in the historical operation.

In a possible design, the recovery module 402 is configured to kill the application process to be recycled; or transfer the data in the memory of the application process to be recycled to the preset storage space.

The unit and method processes of the examples described in the embodiments of the present application can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the solution. The skilled person can use different methods for each particular application to implement the described functionality.

In the several embodiments provided by the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the device embodiments described above are merely illustrative. For example, the division of the unit is only a logical function division, and the actual implementation may have another division manner. For example, multiple units or components may be combined or integrated into another system, or some steps may be omitted or not performed. Furthermore, the coupling or direct coupling or communication connection of the various units to each other may be through some interfaces, which may be in electrical, mechanical or other form.

The units described as separate components may or may not be physically separate, and may be located in one place or on multiple network elements.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, it may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in accordance with embodiments of the present invention are generated in whole or in part. The computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer readable storage medium or transferred from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be from a website site, computer, server or data center Transfer to another website site, computer, server, or data center by wire (eg, coaxial cable, fiber optic, digital subscriber line (DSL), or wireless (eg, infrared, wireless, microwave, etc.). The computer readable storage medium can be any available media that can be accessed by a computer or a data storage device such as a server, data center, or the like that includes one or more available media. The usable medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a DVD), or a semiconductor medium (such as a solid state disk (SSD)).

Claims

A method of recycling memory, the method comprising:

Determining a plurality of application processes running in the background when detecting that the free memory does not satisfy the preset condition;

Determining a user preference of a time period to which each application process in the plurality of application processes belongs at a current time point, the user preference being used to indicate a possibility of an application process being invoked;

Determining, according to a user preference of each application process in the multiple application processes, an application process to be recycled in the multiple application processes;

Reclaiming the memory occupied by the application process to be recycled.
The method according to claim 1, wherein the determining the user preference of each of the plurality of application processes in the time period to which the current time point belongs includes:

And determining, according to the active duration of the time period of each application process in the historical operation, the user preference of each application process in the plurality of application processes at a time period to which the current time point belongs.
The method according to claim 2, wherein the determining, according to the active duration of the time period of each application process in the historical operation, determines each application of the plurality of application processes The user preference of the time period to which the process belongs at the current time point, including:

Using the active duration as the user preference; or,

Calculating a ratio of the active duration to the sum of the active durations, where the sum of the active durations refers to a sum of active durations of the plurality of application processes in the historical operation; or ,

Calculating a product between the active duration and a cold start duration of the application process as the user preference; or

Calculating a ratio between the active duration and the sum of active durations, and calculating a product between the ratio and a cold start duration of the application process as the user preference.
The method according to any one of claims 1-3, wherein the determining, according to a user preference of each application process in the plurality of application processes, an application process to be recycled in the plurality of application processes, include:

Retrieving a revocation priority of each application process based on a user preference of each application process in the plurality of application processes, a reclaiming priority of each application process, and a user preference of each application process Negative correlation

Determining an application process to be reclaimed among the plurality of application processes based on a recycling priority of each application process.
The method according to claim 4, wherein the obtaining a revocation priority of each application process based on user preference of each application process in the plurality of application processes comprises:

Determining a recent active weight of each application process, the most recent active weight of each application process is used to indicate a degree of recent activeness of each application process;

Retrieving the revocation priority of each application process according to the user preference of each application process and the most recent active weight of each application process.
The method according to claim 5, wherein the determining the most recent active weight of each application process comprises:

Obtaining the latest active duration of each application process, where the latest active duration of each application process is used to indicate the active duration of the last time each application process is running in the foreground;

And attenuating the most recent active duration of each application process to obtain the most recent active duration after the attenuation based on the time difference between the end time point of the most recent active duration of the each application process and the current time point. The most recent active duration after the decay is used as the most recent active weight.
The method according to claim 4, wherein the obtaining a revocation priority of each application process based on user preference of each application process in the plurality of application processes comprises:

Determining a function weight corresponding to the function of each application process, the function weight of each application process is used to indicate the importance degree of the function of each application process;

Retrieving the revocation priority of each application process according to the user preference of each application process and the function weight of each application process.
The method according to claim 4, wherein the obtaining a revocation priority of each application process based on user preference of each application process in the plurality of application processes comprises:

Determining power consumption weights corresponding to power consumption of each application process, the power consumption weight of each application process is used to indicate a size of power consumption of each application process;

Retrieving the revocation priority of each application process according to the user preference of each application process and the power consumption weight of each application process.
The method according to claim 4, wherein the obtaining a revocation priority of each application process based on user preference of each application process in the plurality of application processes comprises:

Determining a memory ratio corresponding to a memory ratio of each application process, where a memory ratio of each application process is used to indicate a ratio between a memory occupied by each application process and a total memory overhead;

Retrieving the revocation priority of each application process according to the user preference of each application process and the memory weight of each application process.
The method according to claim 4, wherein the obtaining a revocation priority of each application process based on user preference of each application process in the plurality of application processes comprises:

Determining a frequency of invocation of each application process, the calling frequency of each application process is used to indicate the total number of times the plurality of application processes invoke the each application process in the time period in the historical operation;

Retrieving the revocation priority of each application process according to the user preference of each application process and the calling frequency of each application process.
The method according to claim 10, wherein the determining the calling frequency of each application process comprises:

Obtaining a call record of each application process, where the call record of each application process is used to record an event of each application process calling an application process other than each of the plurality of application processes;

Determining the calling frequency of each application process based on the call record of each application process.
The method according to any one of claims 2 to 11, wherein before the determining a plurality of application processes running in the background, the method further comprises:

Counting the active duration of the time period of each application process in the plurality of application processes during the historical operation.
The method according to any one of claims 1 to 12, wherein the reclaiming the memory occupied by the application process to be recycled includes:

Killing the application process to be recycled; or,

Transfer the data in the memory of the application process to be recycled to a preset storage space.
A device for recovering a memory, characterized in that the device comprises:

a determining module, configured to determine, when the idle memory does not meet the preset condition, determine a plurality of application processes running in the background;

The determining module is further configured to determine a user preference of a time period that each application process in the multiple application processes belongs to at a current time point, where the user preference is used to indicate a possibility of an application process being invoked;

The determining module is further configured to determine, according to user preference of each application process in the multiple application processes, an application process to be recycled in the multiple application processes;

The recycling module is configured to recover the memory occupied by the application process to be recycled.
The device according to claim 14, wherein the determining module is further configured to: determine, according to an active duration of the time period of each application process in the historical operation, The user preference of each application process in the time period to which the application process belongs at the current time point.
The device according to claim 15, wherein the determining module is further configured to:

Using the active duration as the user preference; or,

Calculating a ratio of the active duration to the sum of the active durations, where the sum of the active durations refers to a sum of active durations of the plurality of application processes in the historical operation; or ,

Calculating a product between the active duration and a cold start duration of the application process as the user preference; or

Calculating a ratio between the active duration and the sum of active durations, and calculating a product between the ratio and a cold start duration of the application process as the user preference.
The device according to any one of claims 14-16, wherein the determining module further comprises:

Obtaining a sub-module, configured to acquire, according to a user preference of each application process in the multiple application processes, a reclaim priority of each application process, where each application process has a reclaim priority and each of the The user preference of the application process is negatively correlated;

Determining a sub-module, configured to determine, according to a recycling priority of each application process, an application process to be reclaimed among the multiple application processes.
The apparatus according to claim 17, wherein the obtaining submodule is further configured to determine a most recent active weight of each application process, and a recent active weight of each application process is used to indicate the each The extent to which the application process is recently active; the revocation priority of each application process is obtained according to the user preference of each application process and the most recent active weight of each application process.
The device according to claim 18, wherein the obtaining sub-module is further configured to obtain a recent active duration of each application process, and a latest active duration of each application process is used to indicate the each The active time period in which the application process was last run in the foreground; based on the time difference between the end time point of the latest active time of each application process and the current time point, the latest active time of each application process is performed. Attenuation to obtain the most recent active duration after attenuation, and the decayed most recent active duration is taken as the most recent active weight.
The apparatus according to claim 17, wherein the obtaining submodule is further configured to: determine a function weight corresponding to a function of each application process, and a function weight of each application process is used to indicate The importance of the function of each application process; obtaining the reclaim priority of each application process according to the user preference of each application process and the functional weight of each application process.
The apparatus according to claim 17, wherein the obtaining submodule is further configured to: determine a power consumption weight corresponding to the power consumption of each application process, and a power consumption weight of each application process And a size of the power consumption of each application process is obtained, and a recycling priority of each application process is obtained according to the user preference of each application process and the power consumption weight of each application process.
The device according to claim 17, wherein the obtaining submodule is further configured to: determine a memory ratio weight corresponding to a memory ratio of each application process, and the memory of each application process The ratio is used to indicate the ratio between the memory occupied by each application process and the total memory overhead; and the user preference is calculated according to the user preference of each application process and the memory weight of each application process. Recycling priority for each application process.
The apparatus according to claim 17, wherein the obtaining submodule is further configured to: determine a calling frequency of each application process, and the calling frequency of each application process is used to indicate the multiple The application process invokes the total number of times of each application process in the time period of the historical operation; and acquires each application according to the user preference of each application process and the calling frequency of each application process. The recycling priority of the process.
The apparatus according to claim 23, wherein the obtaining submodule is further configured to: acquire a call record of each application process, and the call record of each application process is used to record each of the An application process invokes an event of an application process other than each of the plurality of application processes; and based on the call record of each application process, determines a calling frequency of each application process.
The device according to any one of claims 15 to 24, wherein the device further comprises:

The statistics module is configured to count the active duration of the time period of each application process in the plurality of application processes during the historical operation.
The device according to any one of claims 14-25, wherein the recycling module is configured to kill the application process to be recycled; or transfer the data of the application process to be recycled in memory To the default storage space.
The device for recovering a memory is characterized in that: the device for recovering memory comprises:

Memory for storing programs;

A processor for executing the program stored by the memory, the processor for performing the steps of any of claims 1-13 when the program is executed.
A computer readable storage medium, comprising instructions that, when run on a computer, cause the computer to perform the method of any of claims 1-13.