获取 Android 中的内存使用情况

声明:本页面是StackOverFlow热门问题的中英对照翻译,遵循CC BY-SA 4.0协议,如果您需要使用它,必须同样遵循CC BY-SA许可,注明原文地址和作者信息,同时你必须将它归于原作者(不是我):StackOverFlow 原文地址: http://stackoverflow.com/questions/3118234/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me): StackOverFlow

提示:将鼠标放在中文语句上可以显示对应的英文。显示中英文
时间:2020-08-20 08:47:36  来源:igfitidea点击:

Get Memory Usage in Android

androidcpu-usage

提问by Badal

Is there any API by which we can get CPU or Memory usage of android?

是否有任何 API 可以让我们获得 android 的 CPU 或内存使用情况?

I have tried one code as below:

我已经尝试了一个代码如下:

package com.infostretch.mainactivity;

import java.io.BufferedReader;
import java.io.FileInputStream;
import java.io.IOException;
import java.io.InputStreamReader;

public class CPULoad 
{
    long total = 0;
    long idle = 0;

    float usage = 0;

    public CPULoad()
    {
        readUsage();
    }

    public float getUsage()
    {
        readUsage();
        return usage;
    }

    private void readUsage()
    {
        try
        {
            BufferedReader reader = new BufferedReader(new InputStreamReader(new FileInputStream("/proc/stat")), 1000);
            String load = reader.readLine();
            reader.close();

            String[] toks = load.split(" ");

            long currTotal = Long.parseLong(toks[2]) + Long.parseLong(toks[3]) + Long.parseLong(toks[4]);
            long currIdle = Long.parseLong(toks[5]);

            this.usage = (currTotal - total) * 100.0f / (currTotal - total + currIdle - idle);
            this.total = currTotal;
            this.idle = currIdle;
        }
        catch(IOException ex)
        {
            ex.printStackTrace();
        }
    }
}

Is this the correct way to do it?

这是正确的方法吗?

回答by szcoder

I use this function to calculate cpu usage. Hope it can help you.

我使用这个函数来计算 CPU 使用率。希望它可以帮助你。

private float readUsage() {
    try {
        RandomAccessFile reader = new RandomAccessFile("/proc/stat", "r");
        String load = reader.readLine();

        String[] toks = load.split(" +");  // Split on one or more spaces

        long idle1 = Long.parseLong(toks[4]);
        long cpu1 = Long.parseLong(toks[2]) + Long.parseLong(toks[3]) + Long.parseLong(toks[5])
              + Long.parseLong(toks[6]) + Long.parseLong(toks[7]) + Long.parseLong(toks[8]);

        try {
            Thread.sleep(360);
        } catch (Exception e) {}

        reader.seek(0);
        load = reader.readLine();
        reader.close();

        toks = load.split(" +");

        long idle2 = Long.parseLong(toks[4]);
        long cpu2 = Long.parseLong(toks[2]) + Long.parseLong(toks[3]) + Long.parseLong(toks[5])
            + Long.parseLong(toks[6]) + Long.parseLong(toks[7]) + Long.parseLong(toks[8]);

        return (float)(cpu2 - cpu1) / ((cpu2 + idle2) - (cpu1 + idle1));

    } catch (IOException ex) {
        ex.printStackTrace();
    }

    return 0;
} 

回答by rcl

An easy way to check the CPU usage is to use the adb tool w/ top. I.e.:

检查 CPU 使用率的一种简单方法是使用带有 top 的 adb 工具。IE:

adb shell top -m 10

adb shell top -m 10

回答by slash33

Based on the previous answers and personnal experience, here is the code I use to monitor CPU use. The code of this class is written in pure Java.

根据之前的回答和个人经验,这里是我用来监控 CPU 使用情况的代码。这个类的代码是用纯Java编写的。

import java.io.IOException;
import java.io.RandomAccessFile;

/**
 * Utilities available only on Linux Operating System.
 * 
 * <p>
 * A typical use is to assign a thread to CPU monitoring:
 * </p>
 * 
 * <pre>
 * &#064;Override
 * public void run() {
 *  while (CpuUtil.monitorCpu) {
 * 
 *      LinuxUtils linuxUtils = new LinuxUtils();
 * 
 *      int pid = android.os.Process.myPid();
 *      String cpuStat1 = linuxUtils.readSystemStat();
 *      String pidStat1 = linuxUtils.readProcessStat(pid);
 * 
 *      try {
 *          Thread.sleep(CPU_WINDOW);
 *      } catch (Exception e) {
 *      }
 * 
 *      String cpuStat2 = linuxUtils.readSystemStat();
 *      String pidStat2 = linuxUtils.readProcessStat(pid);
 * 
 *      float cpu = linuxUtils.getSystemCpuUsage(cpuStat1, cpuStat2);
 *      if (cpu &gt;= 0.0f) {
 *          _printLine(mOutput, &quot;total&quot;, Float.toString(cpu));
 *      }
 * 
 *      String[] toks = cpuStat1.split(&quot; &quot;);
 *      long cpu1 = linuxUtils.getSystemUptime(toks);
 * 
 *      toks = cpuStat2.split(&quot; &quot;);
 *      long cpu2 = linuxUtils.getSystemUptime(toks);
 * 
 *      cpu = linuxUtils.getProcessCpuUsage(pidStat1, pidStat2, cpu2 - cpu1);
 *      if (cpu &gt;= 0.0f) {
 *          _printLine(mOutput, &quot;&quot; + pid, Float.toString(cpu));
 *      }
 * 
 *      try {
 *          synchronized (this) {
 *              wait(CPU_REFRESH_RATE);
 *          }
 *      } catch (InterruptedException e) {
 *          e.printStackTrace();
 *          return;
 *      }
 *  }
 * 
 *  Log.i(&quot;THREAD CPU&quot;, &quot;Finishing&quot;);
 * }
 * </pre>
 */
public final class LinuxUtils {

    // Warning: there appears to be an issue with the column index with android linux:
    // it was observed that on most present devices there are actually
    // two spaces between the 'cpu' of the first column and the value of 
    // the next column with data. The thing is the index of the idle 
    // column should have been 4 and the first column with data should have index 1. 
    // The indexes defined below are coping with the double space situation.
    // If your file contains only one space then use index 1 and 4 instead of 2 and 5.
    // A better way to deal with this problem may be to use a split method 
    // not preserving blanks or compute an offset and add it to the indexes 1 and 4.

    private static final int FIRST_SYS_CPU_COLUMN_INDEX = 2;

    private static final int IDLE_SYS_CPU_COLUMN_INDEX = 5;

    /** Return the first line of /proc/stat or null if failed. */
    public String readSystemStat() {

        RandomAccessFile reader = null;
        String load = null;

        try {
            reader = new RandomAccessFile("/proc/stat", "r");
            load = reader.readLine();
        } catch (IOException ex) {
            ex.printStackTrace();
        } finally {
            Streams.close(reader);
        }

        return load;
    }

    /**
     * Compute and return the total CPU usage, in percent.
     * 
     * @param start
     *            first content of /proc/stat. Not null.
     * @param end
     *            second content of /proc/stat. Not null.
     * @return 12.7 for a CPU usage of 12.7% or -1 if the value is not
     *         available.
     * @see {@link #readSystemStat()}
     */
    public float getSystemCpuUsage(String start, String end) {
        String[] stat = start.split("\s");
        long idle1 = getSystemIdleTime(stat);
        long up1 = getSystemUptime(stat);

        stat = end.split("\s");
        long idle2 = getSystemIdleTime(stat);
        long up2 = getSystemUptime(stat);

        // don't know how it is possible but we should care about zero and
        // negative values.
        float cpu = -1f;
        if (idle1 >= 0 && up1 >= 0 && idle2 >= 0 && up2 >= 0) {
            if ((up2 + idle2) > (up1 + idle1) && up2 >= up1) {
                cpu = (up2 - up1) / (float) ((up2 + idle2) - (up1 + idle1));
                cpu *= 100.0f;
            }
        }

        return cpu;
    }

    /**
     * Return the sum of uptimes read from /proc/stat.
     * 
     * @param stat
     *            see {@link #readSystemStat()}
     */
    public long getSystemUptime(String[] stat) {
        /*
         * (from man/5/proc) /proc/stat kernel/system statistics. Varies with
         * architecture. Common entries include: cpu 3357 0 4313 1362393
         * 
         * The amount of time, measured in units of USER_HZ (1/100ths of a
         * second on most architectures, use sysconf(_SC_CLK_TCK) to obtain the
         * right value), that the system spent in user mode, user mode with low
         * priority (nice), system mode, and the idle task, respectively. The
         * last value should be USER_HZ times the second entry in the uptime
         * pseudo-file.
         * 
         * In Linux 2.6 this line includes three additional columns: iowait -
         * time waiting for I/O to complete (since 2.5.41); irq - time servicing
         * interrupts (since 2.6.0-test4); softirq - time servicing softirqs
         * (since 2.6.0-test4).
         * 
         * Since Linux 2.6.11, there is an eighth column, steal - stolen time,
         * which is the time spent in other operating systems when running in a
         * virtualized environment
         * 
         * Since Linux 2.6.24, there is a ninth column, guest, which is the time
         * spent running a virtual CPU for guest operating systems under the
         * control of the Linux kernel.
         */

        // with the following algorithm, we should cope with all versions and
        // probably new ones.
        long l = 0L;

        for (int i = FIRST_SYS_CPU_COLUMN_INDEX; i < stat.length; i++) {
            if (i != IDLE_SYS_CPU_COLUMN_INDEX ) { // bypass any idle mode. There is currently only one.
                try {
                    l += Long.parseLong(stat[i]);
                } catch (NumberFormatException ex) {
                    ex.printStackTrace();
                    return -1L;
                }
            }
        }

        return l;
    }

    /**
     * Return the sum of idle times read from /proc/stat.
     * 
     * @param stat
     *            see {@link #readSystemStat()}
     */
    public long getSystemIdleTime(String[] stat) {
        try {
            return Long.parseLong(stat[IDLE_SYS_CPU_COLUMN_INDEX]);
        } catch (NumberFormatException ex) {
            ex.printStackTrace();
        }

        return -1L;
    }

    /** Return the first line of /proc/pid/stat or null if failed. */
    public String readProcessStat(int pid) {

        RandomAccessFile reader = null;
        String line = null;

        try {
            reader = new RandomAccessFile("/proc/" + pid + "/stat", "r");
            line = reader.readLine();
        } catch (IOException ex) {
            ex.printStackTrace();
        } finally {
            Streams.close(reader);
        }

        return line;
    }

    /**
     * Compute and return the CPU usage for a process, in percent.
     * 
     * <p>
     * The parameters {@code totalCpuTime} is to be the one for the same period
     * of time delimited by {@code statStart} and {@code statEnd}.
     * </p>
     * 
     * @param start
     *            first content of /proc/pid/stat. Not null.
     * @param end
     *            second content of /proc/pid/stat. Not null.
     * @return the CPU use in percent or -1f if the stats are inverted or on
     *         error
     * @param uptime
     *            sum of user and kernel times for the entire system for the
     *            same period of time.
     * @return 12.7 for a cpu usage of 12.7% or -1 if the value is not available
     *         or an error occurred.
     * @see {@link #readProcessStat(int)}
     */
    public float getProcessCpuUsage(String start, String end, long uptime) {

        String[] stat = start.split("\s");
        long up1 = getProcessUptime(stat);

        stat = end.split("\s");
        long up2 = getProcessUptime(stat);

        float ret = -1f;
        if (up1 >= 0 && up2 >= up1 && uptime > 0.) {
            ret = 100.f * (up2 - up1) / (float) uptime;
        }

        return ret;
    }

    /**
     * Decode the fields of the file {@code /proc/pid/stat} and return (utime +
     * stime)
     * 
     * @param stat
     *            obtained with {@link #readProcessStat(int)}
     */
    public long getProcessUptime(String[] stat) {
        return Long.parseLong(stat[14]) + Long.parseLong(stat[15]);
    }

    /**
     * Decode the fields of the file {@code /proc/pid/stat} and return (cutime +
     * cstime)
     * 
     * @param stat
     *            obtained with {@link #readProcessStat(int)}
     */
    public long getProcessIdleTime(String[] stat) {
        return Long.parseLong(stat[16]) + Long.parseLong(stat[17]);
    }

    /**
     * Return the total CPU usage, in percent.
     * <p>
     * The call is blocking for the time specified by elapse.
     * </p>
     * 
     * @param elapse
     *            the time in milliseconds between reads.
     * @return 12.7 for a CPU usage of 12.7% or -1 if the value is not
     *         available.
     */
    public float syncGetSystemCpuUsage(long elapse) {

        String stat1 = readSystemStat();
        if (stat1 == null) {
            return -1.f;
        }

        try {
            Thread.sleep(elapse);
        } catch (Exception e) {
        }

        String stat2 = readSystemStat();
        if (stat2 == null) {
            return -1.f;
        }

        return getSystemCpuUsage(stat1, stat2);
    }

    /**
     * Return the CPU usage of a process, in percent.
     * <p>
     * The call is blocking for the time specified by elapse.
     * </p>
     * 
     * @param pid
     * @param elapse
     *            the time in milliseconds between reads.
     * @return 6.32 for a CPU usage of 6.32% or -1 if the value is not
     *         available.
     */
    public float syncGetProcessCpuUsage(int pid, long elapse) {

        String pidStat1 = readProcessStat(pid);
        String totalStat1 = readSystemStat();
        if (pidStat1 == null || totalStat1 == null) {
            return -1.f;
        }

        try {
            Thread.sleep(elapse);
        } catch (Exception e) {
            e.printStackTrace();
            return -1.f;
        }

        String pidStat2 = readProcessStat(pid);
        String totalStat2 = readSystemStat();
        if (pidStat2 == null || totalStat2 == null) {
            return -1.f;
        }

        String[] toks = totalStat1.split("\s");
        long cpu1 = getSystemUptime(toks);

        toks = totalStat2.split("\s");
        long cpu2 = getSystemUptime(toks);

        return getProcessCpuUsage(pidStat1, pidStat2, cpu2 - cpu1);
    }

}

There are several ways of exploiting this class. You can call either syncGetSystemCpuUsageor syncGetProcessCpuUsagebut each is blocking the calling thread. Since a common issue is to monitor the total CPU usage and the CPU use of the current process at the same time, I have designed a class computing both of them. That class contains a dedicated thread. The output management is implementation specific and you need to code your own.

有几种方法可以利用这个类。您可以调用syncGetSystemCpuUsagesyncGetProcessCpuUsage但每个都阻塞调用线程。由于一个常见的问题是同时监控总 CPU 使用率和当前进程的 CPU 使用率,因此我设计了一个计算它们的类。该类包含一个专用线程。输出管理是特定于实现的,您需要自己编写代码。

The class can be customized by a few means. The constant CPU_WINDOWdefines the depth of a read, i.e. the number of milliseconds between readings and computing of the corresponding CPU load. CPU_REFRESH_RATEis the time between each CPU load measurement. Do not set CPU_REFRESH_RATEto 0 because it will suspend the thread after the first read.

可以通过几种方式自定义该类。该常量CPU_WINDOW定义了读取的深度,即读取和计算相应 CPU 负载之间的毫秒数。CPU_REFRESH_RATE是每次 CPU 负载测量之间的时间。不要设置CPU_REFRESH_RATE为 0,因为它会在第一次读取后挂起线程。

import java.io.File;
import java.io.FileNotFoundException;
import java.io.FileOutputStream;
import java.io.OutputStream;

import android.app.Application;
import android.os.Handler;
import android.os.HandlerThread;
import android.util.Log;

import my.app.LinuxUtils;
import my.app.Streams;
import my.app.TestReport;
import my.app.Utils;

public final class CpuUtil {

    private static final int CPU_WINDOW = 1000;

    private static final int CPU_REFRESH_RATE = 100; // Warning: anything but > 0

    private static HandlerThread handlerThread;

    private static TestReport output;

    static {
        output = new TestReport();
        output.setDateFormat(Utils.getDateFormat(Utils.DATE_FORMAT_ENGLISH));
    }

    private static boolean monitorCpu;

    /**
     * Construct the class singleton. This method should be called in
     * {@link Application#onCreate()}
     * 
     * @param dir
     *            the parent directory
     * @param append
     *            mode
     */
    public static void setOutput(File dir, boolean append) {
        try {
            File file = new File(dir, "cpu.txt");
            output.setOutputStream(new FileOutputStream(file, append));
            if (!append) {
                output.println(file.getAbsolutePath());
                output.newLine(1);

                // print header
                _printLine(output, "Process", "CPU%");

                output.flush();
            }

        } catch (FileNotFoundException e) {
            e.printStackTrace();
        }
    }

    /** Start CPU monitoring */
    public static boolean startCpuMonitoring() {
        CpuUtil.monitorCpu = true;

        handlerThread = new HandlerThread("CPU monitoring"); //$NON-NLS-1$
        handlerThread.start();

        Handler handler = new Handler(handlerThread.getLooper());
        handler.post(new Runnable() {

            @Override
            public void run() {
                while (CpuUtil.monitorCpu) {

                    LinuxUtils linuxUtils = new LinuxUtils();

                    int pid = android.os.Process.myPid();
                    String cpuStat1 = linuxUtils.readSystemStat();
                    String pidStat1 = linuxUtils.readProcessStat(pid);

                    try {
                        Thread.sleep(CPU_WINDOW);
                    } catch (Exception e) {
                    }

                    String cpuStat2 = linuxUtils.readSystemStat();
                    String pidStat2 = linuxUtils.readProcessStat(pid);

                    float cpu = linuxUtils
                            .getSystemCpuUsage(cpuStat1, cpuStat2);
                    if (cpu >= 0.0f) {
                        _printLine(output, "total", Float.toString(cpu));
                    }

                    String[] toks = cpuStat1.split(" ");
                    long cpu1 = linuxUtils.getSystemUptime(toks);

                    toks = cpuStat2.split(" ");
                    long cpu2 = linuxUtils.getSystemUptime(toks);

                    cpu = linuxUtils.getProcessCpuUsage(pidStat1, pidStat2,
                            cpu2 - cpu1);
                    if (cpu >= 0.0f) {
                        _printLine(output, "" + pid, Float.toString(cpu));
                    }

                    try {
                        synchronized (this) {
                            wait(CPU_REFRESH_RATE);
                        }
                    } catch (InterruptedException e) {
                        e.printStackTrace();
                        return;
                    }
                }

                Log.i("THREAD CPU", "Finishing");
            }

        });

        return CpuUtil.monitorCpu;
    }

    /** Stop CPU monitoring */
    public static void stopCpuMonitoring() {
        if (handlerThread != null) {
            monitorCpu = false;
            handlerThread.quit();
            handlerThread = null;
        }
    }

    /** Dispose of the object and release the resources allocated for it */
    public void dispose() {

        monitorCpu = false;

        if (output != null) {
            OutputStream os = output.getOutputStream();
            if (os != null) {
                Streams.close(os);
                output.setOutputStream(null);
            }

            output = null;
        }
    }

    private static void _printLine(TestReport output, String process, String cpu) {
        output.stampln(process + ";" + cpu);
    }

}

回答by robguinness

Since the OP asked about CPU usage AND memory usage (accepted answer only shows technique to get cpu usage), I'd like to recommend the ActivityManager class and specifically the accepted answer from this question: How to get current memory usage in android?

由于 OP 询问了 CPU 使用率和内存使用率(接受的答案仅显示获取 CPU 使用率的技术),我想推荐 ActivityManager 类,特别是这个问题的公认答案:How to get current memory usage in android?

回答by Mathias Conradt

Check the Debugclass. http://developer.android.com/reference/android/os/Debug.htmli.e. Debug.getNativeHeapAllocatedSize()

检查Debug班级。http://developer.android.com/reference/android/os/Debug.htmlDebug.getNativeHeapAllocatedSize()

It has methods to get the used native heap, which is i.e. used by external bitmaps in your app. For the heap that the app is using internally, you can see that in the DDMS tool that comes with the Android SDK and is also available via Eclipse.

它具有获取使用的本机堆的方法,即由应用程序中的外部位图使用。对于应用程序在内部使用的堆,您可以在 Android SDK 附带的 DDMS 工具中看到它,也可以通过 Eclipse 获得。

The native heap + the heap as indicated in the DDMS make up the total heap that your app is allocating.

本机堆 + DDMS 中指示的堆构成了您的应用程序正在分配的总堆。

For CPU usage I'm not sure if there's anything available via API/SDK.

对于 CPU 使用率,我不确定是否有任何可用的 API/SDK。

回答by wangxigui

enter the android terminal and then you can type the following commands :dumpsys cpuinfo

进入android终端,然后你可以输入以下命令:dumpsys cpuinfo

shell@android:/ $ dumpsys cpuinfo                                              
Load: 0.8 / 0.75 / 1.15
CPU usage from 69286ms to 9283ms ago with 99% awake:
  47% 1118/com.wxg.sodproject: 12% user + 35% kernel
  1.6% 1225/android.process.media: 1% user + 0.6% kernel
  1.3% 263/mpdecision: 0.1% user + 1.2% kernel
  0.1% 32747/kworker/u:1: 0% user + 0.1% kernel
  0.1% 883/com.android.systemui: 0.1% user + 0% kernel
  0.1% 521/system_server: 0.1% user + 0% kernel / faults: 14 minor
  0.1% 1826/com.quicinc.trepn: 0.1% user + 0% kernel
  0.1% 2462/kworker/0:2: 0.1% user + 0% kernel
  0.1% 32649/kworker/0:0: 0% user + 0.1% kernel
  0% 118/mmcqd/0: 0% user + 0% kernel
  0% 179/surfaceflinger: 0% user + 0% kernel
  0% 46/kinteractiveup: 0% user + 0% kernel
  0% 141/jbd2/mmcblk0p26: 0% user + 0% kernel
  0% 239/sdcard: 0% user + 0% kernel
  0% 1171/com.xiaomi.channel:pushservice: 0% user + 0% kernel / faults: 1 minor
  0% 1207/com.xiaomi.channel: 0% user + 0% kernel / faults: 1 minor
  0% 32705/kworker/0:1: 0% user + 0% kernel
12% TOTAL: 3.2% user + 9.4% kernel + 0% iowait