外贸网站建设价格,江苏建设准考证打印在哪个网站,文化体育局网站建设,手机网站判断代码【Android性能优化】Android CPU占用率检测原理和优化方向
CPU相关知识 CPU占用的基本计算公式 (1 - 空闲态运行时间/总运行时间) * 100% Hz、Tick、Jiffies#xff1a; Hz#xff1a;Linux核心每隔固定周期会发出timer interrupt (IRQ 0)#xff0c;HZ是用来定义每一秒有…【Android性能优化】Android CPU占用率检测原理和优化方向
CPU相关知识 CPU占用的基本计算公式 (1 - 空闲态运行时间/总运行时间) * 100% Hz、Tick、Jiffies HzLinux核心每隔固定周期会发出timer interrupt (IRQ 0)HZ是用来定义每一秒有几次timer interrupts。举例来说HZ为1000代表每秒有1000次timer interrupts。 通过getconf CLK_TC命令可以查看当前系统的Hz。以某某项目为例Hz为100。
130|console:/ # getconf CLK_TCK
100TickTick 1/Hz即多久发出一次timer interrupt。 以上述情况为例则为10ms发生一次。
JiffiesJiffies是Linux的核心变数。用于记录系统启动后发生timer interrupt的次数timer interrupt每发生一次jiffies增加一次。Jiffies可以认为是Linux下CPU的单位时间
user、nice、system、idle、iowait、irq、softirq kernel/Documentation/filesystems/proc.txt 1.8 Miscellaneous kernel statistics in /proc/stat
-------------------------------------------------Various pieces of information about kernel activity are available in the
/proc/stat file. All of the numbers reported in this file are aggregates
since the system first booted. For a quick look, simply cat the file: cat /proc/statcpu 2255 34 2290 22625563 6290 127 456 0 0 0cpu0 1132 34 1441 11311718 3675 127 438 0 0 0cpu1 1123 0 849 11313845 2614 0 18 0 0 0intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]ctxt 1990473btime 1062191376processes 2915procs_running 1procs_blocked 0softirq 183433 0 21755 12 39 1137 231 21459 2263The very first cpu line aggregates the numbers in all of the other cpuN lines. These numbers identify the amount of time the CPU has spent performing different kinds of work. Time units are in USER_HZ (typically hundredths of a second). The meanings of the columns are as follows, from left to right:- user: normal processes executing in user mode
- nice: niced processes executing in user mode
- system: processes executing in kernel mode
- idle: twiddling thumbs
- iowait: In a word, iowait stands for waiting for I/O to complete. But thereare several problems:1. Cpu will not wait for I/O to complete, iowait is the time that a task is waiting for I/O to complete. When cpu goes into idle state for outstanding task io, another task will be scheduled on this CPU.2. In a multi-core CPU, the task waiting for I/O to complete is not running on any CPU, so the iowait of each CPU is difficult to calculate.3. The value of iowait field in /proc/stat will decrease in certainconditions.So, the iowait is not reliable by reading from /proc/stat.
- irq: servicing interrupts
- softirq: servicing softirqs
- steal: involuntary wait
- guest: running a normal guest
- guest_nice: running a niced guest
第一行数字是其他所有CPUN行中的数字总和这些数字表示CPU执行不同工作花费的时间时间单位为USER_HZ一般为10ms从左到右各列的含义如下
标题含义user在用户模式下执行的正常进程niceniced进程在用户模式下执行system在内核模式下执行的进程idle空闲iowait总而言之iowait代表等待I/O完成。但是存在以下几个问题1.Cpu不会等待I/O完成iowait是任务等待I/O完成的时间。当cpu为未完成的任务io进入空闲状态时将在该cpu上调度另一个任务。2.在多核CPU中等待I/O完成的任务不在任何CPU上运行因此每个CPU的iowait很难计算。3./proc/stat中iowait字段的值将在一定程度上减少条件因此从/proc/stat读取iowait是不可靠的。irq服务中断softirq服务软中断steal非自愿等待时间另一个解释是其它系统所花的时间guest执行时间为客户操作系统下的虚拟CPU控制guest_nice低优先级程序所占用的用户态的cpu时间
Cpu花费时间总和cpuTime user nice system idle iowait irq softirq steal guest guest_nice
计算整机CPU占用率
根据上面的公式我们可以采取两个时间点计算CPU总时间差值和CPU空闲时间IDLE的差值以此得出CPU使用率 采样两个足够短的时间间隔的cpu数据分别记作t1、t2其中t1、t2的结构均为 (user、nice、system、idle、iowait、irq、softirq、stealstolen、guest、guest_nice)的10元组;当然这里依据Linux内核的不同有些数据可能没有就不必计入 计算t1、t2总的cpu时间片totalCPUTime a) 把第一次的所有cpu10元组数据求和得到totalCPUTime1; b) 把第二次的所有cpu10元组数据求和得到totalCPUTime2; 计算空闲时间idle cpu空闲时间对应第四列的数据 a获得第一次的idle数据记为idle1 b获得第二次的idle数据记为idle2 计算cpu使用率 totalCPUrate 1 - ((totalCPUTime2-idle2)-(totalCPUTime1-idle1))totalCPUTime2-totalCPUTime1x100% 计算某个进程CPU占用率 kernel/msm-5.4/Documentation/filesystems/proc.txt The stat filecontains details information about the process itself. Its fields are explained in Table 1-4.Table 1-4: Contents of the stat files (as of 2.6.30-rc7)
..............................................................................Field Contentpid process idtcomm filename of the executablestate state (R is running, S is sleeping, D is sleeping in anuninterruptible wait, Z is zombie, T is traced or stopped)...cmaj_flt number of major faults with childsutime user mode jiffies utime指进程在用户态的运行时间stime kernel mode jiffies ;stime指进程在内核态的运行时间cutime user mode jiffies with childs utime指所有子进程在用户态的运行时间总和cstime kernel mode jiffies with childs cstime指所有子进程在核心态的运行时间总和priority priority levelnice nice levelnum_threads number of threadsit_real_value (obsolete, always 0)start_time time the process started after system bootvsize virtual memory sizerss resident set memory sizersslim current limit in bytes on the rssstart_code address above which program text can run
...这里只关注 utime、 stime、 cutime、 cstime它们的总和就是该进程的CPU时间。取两个时间点进行采样并取差值就可以得到这段时间内的该进程所占用的CPU时间片。
scheduledTime utime stime cutime sctime;if (lastCpuTime lastScheduledTime)cpuUsage (scheduledTime - lastScheduledTime) * 100. / (cpuTime - lastCpuTime);
lastScheduledTime scheduledTime;
lastCpuTime cpuTime;优化的基本思路和方向
确定CPU占用高的进程使用top命令查看确定该进程CPU占用高的线程top -p PID查看线程状态和线程CPU结合logcat命令和trace工具确定该线程处于什么工作逻辑中以及相关方法耗时实在找不到就打断点加日志控制变量法 频繁报错 — 解决报错过度绘制 — 重新布局绘制解码/编码 — 排查具体算法逻辑频繁调用 — 确定各个调用者并协调修改…
