Android内存之VSS/RSS/PSS/USS

短命女 2022-04-12 10:27 528阅读 0赞

[http://hubingforever.blog.163.com/blog/static/17104057920114411313717/][http_hubingforever.blog.163.com_blog_static_17104057920114411313717]

VSS 表示所有能够访问的内存地址大小。不能够用来表示一个进程真正使用的多少内存。

RSS表示进程能访问的内存。也不能用来表示一个进程真正使用了多少内存。因为RSS的大小包含了可能被多个进程使用的(包括本进程)shared libraries的大小。

PSS。 如果一个shared library有30 pages, 并被三个process使用。 那么这个shared library为每个进程添加10pages 的PSS. 所以所有进程的PSS总和就是当前系统使用的内存大小。 但是，当一个process 退出时，它的PSS的大小并不是将要还给系统的free内存的大小, 因为PSS中包含的shared library贡献的内存会被其他使用Shared Library的进程均分掉。

USS是最准确的确定一个process内存使用情况的数值。当一个process 被kill时， 它还给系统的内存数目就是USS的值。

USS == Private Dirty

<table style="line-height:21px; color:rgb(0,0,0); font-family:song,Verdana; font-size:12px; text-align:left; background-color:rgb(245,245,245); border-collapse:collapse; word-wrap:break-word"> 
 <tbody> 
 <tr style="line-height:normal; font-size:12px"> 
 <td align="center" style="font-family:song,Verdana"> 
 <table style="line-height:21px; border-collapse:collapse; word-wrap:break-word"> 
 <tbody> 
 <tr style="line-height:normal; font-size:12px"> 
 <td style="font-family:song,Verdana"> 
 <div style="line-height:25px; margin:15px"> 
 <div> 
 
 <h2 style="line-height:37px">Terms</h2> 
 
 <ul style="margin:0px; padding:0px; list-style-type:none"> 
 <li style="line-height:23px; clear:both; padding-left:10px"> 
 <table> 
 <tbody> 
 <tr> 
 <td> VSS </td> 
 <td> Virtual Set Size: how much virtual memory associated withprocess </td> 
 </tr> 
 <tr> 
 <td> RSS </td> 
 <td> Resident Set Size: how much physical pages allocated for theprocess. Pages shared between processes are counted multiple times </td> 
 </tr> 
 <tr> 
 <td> PSS </td> 
 <td> Proportional Set Size. Take the RSS number but evenlydistribute shared pages among the sharing processes. For example,if three processes are sharing 3MB, each process gets 1MB in PSS. </td> 
 </tr> 
 <tr> 
 <td> USS </td> 
 <td> Also known as Private Dirty, which is basically the amount ofRAM inside the process that can not be paged to disk (it is notbacked by the same data on disk), and is not shared with any otherprocesses. </td> 
 </tr> 
 </tbody> 
 </table> </li> 
 <li style="line-height:23px; clear:both; padding-left:10px">VSS&nbsp;- Virtual Set Size 虚拟耗用内存（包含共享库占用的内存）</li> 
 <li style="line-height:23px; clear:both; padding-left:10px">RSS&nbsp;- Resident Set Size 实际使用物理内存（包含共享库占用的内存）</li> 
 <li style="line-height:23px; clear:both; padding-left:10px">PSS&nbsp;- Proportional Set Size 实际使用的物理内存（比例分配共享库占用的内存）</li> 
 <li style="line-height:23px; clear:both; padding-left:10px">USS&nbsp;- Unique Set Size 进程独自占用的物理内存（不包含共享库占用的内存）</li> 
 </ul> 
 一般来说内存占用大小有如下规律：VSS &gt;= RSS &gt;= PSS &gt;= USS 
 
 <h2 style="line-height:37px"><a style="color:rgb(245,149,19)" rel="nofollow"></a>Overview</h2> 
 The aim of this post is to provide information that will assist in interpreting memory reports from various tools so the true memory usage for Linux processes and the system can be determined. 
 Android has a tool called&nbsp;procrank&nbsp;(/system/xbin/procrank), which lists out the memory usage of Linux processes in order from highest to lowest usage. The sizes reported per process are VSS, RSS, PSS, and USS. 
 For the sake of simplicity in this description, memory will be expressed in terms of pages, rather than bytes. Linux systems like ours manage memory in 4096 byte pages at the lowest level. 
 VSS&nbsp;(reported as VSZ from ps) is&nbsp;the total accessible address space of a process.&nbsp;This size also includes memory that may not be resident in RAM like mallocs that have been allocated but not written to. VSS is of very little use for determing real memory usage of a process. 
 
 RSS&nbsp;is the&nbsp;total memory actually held in RAM for a process.&nbsp;RSS can be misleading, because it reports the total all of the shared libraries that the process uses, even though a shared library is only loaded into memory once regardless of how many processes use it. RSS is not an accurate representation of the memory usage for a single process. 
 
 PSS&nbsp;differs from RSS in that it reports the proportional size of its shared libraries, i.e. if three processes all use a shared library that has 30 pages, that library will only contribute 10 pages to the PSS that is reported for each of the three processes. PSS is a very useful number because when the PSS for all processes in the system are summed together, that is a good representation for the total memory usage in the system. When a process is killed, the shared libraries that contributed to its PSS will be proportionally distributed to the PSS totals for the remaining processes still using that library. In this way PSS can be slightly misleading, because when a process is killed, PSS does not accurately represent the memory returned to the overall system. 
 
 USS&nbsp;is&nbsp;the total private memory for a process, i.e. that memory that is completely unique to that process.&nbsp;USS is an extremely useful number because it indicates the true incremental cost of running a particular process. When a process is killed, the USS is the total memory that is actually returned to the system. USS is the best number to watch when initially suspicious of&nbsp;memory leaks&nbsp;in a process. 
 
 For systems that have Python available, there is also a nice tool called&nbsp;smem&nbsp;that will report memory statistics including all of these categories. 
 # procrank procrank PID&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Vss &nbsp; &nbsp; &nbsp; &nbsp;Rss &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; Pss &nbsp; &nbsp; &nbsp; &nbsp; Uss&nbsp;cmdline 481&nbsp;&nbsp; 31536K&nbsp;&nbsp; 30936K&nbsp;&nbsp; 14337K&nbsp;&nbsp;&nbsp; 9956K&nbsp;system_server 475&nbsp;&nbsp; 26128K&nbsp;&nbsp; 26128K&nbsp;&nbsp; 10046K&nbsp;&nbsp;&nbsp; 5992K&nbsp;zygote 526&nbsp;&nbsp; 25108K&nbsp;&nbsp; 25108K&nbsp;&nbsp;&nbsp; 9225K&nbsp;&nbsp;&nbsp; 5384K&nbsp;android.process.acore 523&nbsp;&nbsp; 22388K&nbsp;&nbsp; 22388K&nbsp;&nbsp;&nbsp; 7166K&nbsp;&nbsp;&nbsp; 3432K&nbsp;com.android.phone 574&nbsp;&nbsp; 21632K&nbsp;&nbsp; 21632K&nbsp;&nbsp;&nbsp; 6109K&nbsp;&nbsp;&nbsp; 2468K&nbsp;com.android.settings 521&nbsp;&nbsp; 20816K&nbsp;&nbsp; 20816K&nbsp;&nbsp;&nbsp; 6050K&nbsp;&nbsp;&nbsp; 2776K&nbsp;jp.co.omronsoft.openwnn 474&nbsp;&nbsp;&nbsp; 3304K&nbsp;&nbsp;&nbsp; 3304K&nbsp;&nbsp;&nbsp; 1097K&nbsp;&nbsp;&nbsp;&nbsp; 624K&nbsp;/system/bin/mediaserver 37&nbsp;&nbsp;&nbsp;&nbsp; 304K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 304K&nbsp;&nbsp;&nbsp;&nbsp; 289K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;288K&nbsp;/sbin/adbd 29&nbsp;&nbsp;&nbsp;&nbsp; 720K&nbsp;&nbsp;&nbsp; &nbsp; 720K&nbsp;&nbsp;&nbsp;&nbsp; 261K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;212K&nbsp;/system/bin/rild 601&nbsp;&nbsp;&nbsp;&nbsp; 412K&nbsp;&nbsp;&nbsp;&nbsp; 412K&nbsp;&nbsp;&nbsp;&nbsp; 225K&nbsp;&nbsp;&nbsp;&nbsp; 216K&nbsp;procrank &nbsp;&nbsp; 1&nbsp;&nbsp;&nbsp;&nbsp; 204K&nbsp;&nbsp;&nbsp;&nbsp; 204K&nbsp;&nbsp;&nbsp;&nbsp; 185K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;184K&nbsp;/init 35&nbsp;&nbsp;&nbsp;&nbsp; 388K&nbsp;&nbsp;&nbsp;&nbsp; 388K&nbsp;&nbsp;&nbsp;&nbsp; 182K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;172K&nbsp;/system/bin/qemud 284&nbsp;&nbsp;&nbsp;&nbsp; 384K&nbsp;&nbsp;&nbsp;&nbsp; 384K&nbsp;&nbsp;&nbsp;&nbsp; 160K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;148K&nbsp;top 27&nbsp;&nbsp;&nbsp;&nbsp; 376K&nbsp;&nbsp;&nbsp;&nbsp; 376K&nbsp;&nbsp;&nbsp;&nbsp; 148K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;136K&nbsp;/system/bin/vold 261&nbsp;&nbsp;&nbsp;&nbsp; 332K&nbsp;&nbsp;&nbsp;&nbsp; 332K&nbsp;&nbsp;&nbsp;&nbsp; 123K&nbsp;&nbsp;&nbsp;&nbsp;112K&nbsp;logcat 33&nbsp;&nbsp;&nbsp;&nbsp; 396K&nbsp;&nbsp;&nbsp;&nbsp; 396K&nbsp;&nbsp;&nbsp;&nbsp; 105K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;80K&nbsp;/system/bin/keystore 32&nbsp;&nbsp;&nbsp;&nbsp; 316K&nbsp;&nbsp;&nbsp;&nbsp; 316K&nbsp;&nbsp;&nbsp;&nbsp; 100K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;88K&nbsp;/system/bin/installd 269&nbsp;&nbsp;&nbsp;&nbsp; 328K&nbsp;&nbsp;&nbsp;&nbsp; 328K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 95K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;72K&nbsp;/system/bin/sh 26&nbsp;&nbsp;&nbsp;&nbsp; 280K&nbsp;&nbsp;&nbsp;&nbsp; 280K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 93K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;84K&nbsp;/system/bin/servicemanager 45&nbsp;&nbsp;&nbsp;&nbsp; 304K&nbsp;&nbsp;&nbsp;&nbsp; 304K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 91K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;80K&nbsp;/system/bin/qemu-props 34&nbsp;&nbsp;&nbsp;&nbsp; 324K&nbsp;&nbsp;&nbsp;&nbsp; 324K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 91K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;68K&nbsp;/system/bin/sh 260&nbsp;&nbsp;&nbsp;&nbsp; 324K&nbsp;&nbsp;&nbsp;&nbsp; 324K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 91K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;68K&nbsp;/system/bin/sh 600&nbsp;&nbsp;&nbsp;&nbsp; 324K&nbsp;&nbsp;&nbsp;&nbsp; 324K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 91K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;68K&nbsp;/system/bin/sh 25&nbsp;&nbsp;&nbsp;&nbsp; 308K&nbsp;&nbsp;&nbsp;&nbsp; 308K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 88K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;68K&nbsp;/system/bin/sh 28&nbsp;&nbsp;&nbsp;&nbsp; 232K&nbsp;&nbsp;&nbsp;&nbsp; 232K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 67K&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;60K&nbsp;/system/bin/debuggerd # 
 </div> 
 </div> </td> 
 </tr> 
 </tbody> 
 </table> </td> 
 </tr> 
 </tbody> 
</table>

[http_hubingforever.blog.163.com_blog_static_17104057920114411313717]: http://hubingforever.blog.163.com/blog/static/17104057920114411313717/