Saturday, June 20, 2009

How do I Find Out Linux CPU Utilization?

Whenever a Linux system CPU is occupied by a process, it is unavailable for processing other requests. Rest of pending requests must wait till CPU is free. This becomes a bottleneck in the system. Following command will help you to identify CPU utilization, so that you can troubleshoot CPU related performance problems.

Finding CPU utilization is one of the important tasks. Linux comes with various utilities to report CPU utilization. With these commands, you will be able to find out:

* CPU utilization
* Display the utilization of each CPU individually (SMP cpu)
* Find out your system's average CPU utilization since the last reboot etc
* Determine which process is eating the CPU(s)

top - display top CPU processes


top [-] [d delay] [p pid] [q] [c] [C] [S] [s] [i] [n iter] [b]


top provides an ongoing look at processor activity in real time. It displays a listing of the most CPU-intensive tasks on the system, and can provide an interactive interface for manipulating processes. It can sort the tasks by CPU usage, memory usage and runtime. can be better configured than the standard top from the procps suite. Most features can either be selected by an interactive command or by specifying the feature in the personal or system-wide configuration file. See below for more information.


Specifies the delay between screen updates. You can change this with the sinteractive command.
Monitor only processes with given process id. This flag can be given up to twenty times. This option is neither available interactively nor can it be put into the configuration file.

This causes top to refresh without any delay. If the caller has superuser privileges, top runs with the highest possible priority.
Specifies cumulative mode, where each process is listed with the CPU time that it as well as its dead children has spent. This is like the -S flag to ps(1). See the discussion below of the S interactive command.
Tells top to run in secure mode. This disables the potentially dangerous of the interactive commands (see below). A secure top is a nifty thing to leave running on a spare terminal.
Start top ignoring any idle or zombie processes. See the interactive command i below.
display total CPU states instead of individual CPUs. This option only affects SMP systems.
display command line instead of the command name only. The default behavior has been changed as this seems to be more useful.
Show all threads.
Number of iterations. Update the display this number of times and then exit.
Batch mode. Useful for sending output from top to other programs or to a file. In this mode, top will not accept command line input. It runs until it produces the number of iterations requested with the n option or until killed. Output is plain text suitable for display on a dumb terminal.


top displays a variety of information about the processor state. The display is updated every 5 seconds by default, but you can change that with the d command-line option or thes interactive command.
This line displays the time the system has been up, and the three load averages for the system. The load averages are the average number of process ready to run during the last 1, 5 and 15 minutes. This line is just like the output of uptime(1). The uptime display may be toggled by the interactive l command.
The total number of processes running at the time of the last update. This is also broken down into the number of tasks which are running, sleeping, stopped, or undead. The processes and states display may be toggled by the t interactive command.
"CPU states"
Shows the percentage of CPU time in user mode, system mode, niced tasks, iowait and idle. (Niced tasks are only those whose nice value is positive.) Time spent in niced tasks will also be counted in system and user time, so the total will be more than 100%. The processes and states display may be toggled by the t interactive command.
Statistics on memory usage, including total available memory, free memory, used memory, shared memory, and memory used for buffers. The display of memory information may be toggled by the m interactive command.
Statistics on swap space, including total swap space, available swap space, and used swap space. This and Mem are just like the output of free(1).
The process ID of each task.
The parent process ID each task.
The user ID of the task's owner.
The user name of the task's owner.
The priority of the task.
The nice value of the task. Negative nice values are higher priority.
The size of the task's code plus data plus stack space, in kilobytes, is shown here.
The code size of the task. This gives strange values for kernel processes and is broken for ELF processes.
Data + Stack size. This is broken for ELF processes.
Text resident size.
Size of the swapped out part of the task.
Size of pages marked dirty.
Last used processor. (That this changes from time to time is not a bug; Linux intentionally uses weak affinity. Also notice that the very act of running top may break weak affinity and cause more processes to change current CPU more often because of the extra demand for CPU time.)
The total amount of physical memory used by the task, in kilobytes, is shown here. For ELF processes used library pages are counted here, for a.out processes not.
The amount of shared memory used by the task is shown in this column.
The state of the task is shown here. The state is either S for sleeping, D for uninterruptible sleep, R for running, Z for zombies, or T for stopped or traced. These states are modified by trailing < for a process with negative nice value, N for a process with positive nice value, W for a swapped out process (this does not work correctly for kernel processes).
depending on the availability of either /boot/psdatabase or the kernel link map/boot/ this shows the address or the name of the kernel function the task currently is sleeping in.
Total CPU time the task has used since it started. If cumulative mode is on, this also includes the CPU time used by the process's children which have died. You can set cumulative mode with the S command line option or toggle it with the interactive command S. The header line will then be changed to CTIME.
The task's share of the CPU time since the last screen update, expressed as a percentage of total CPU time per processor.
The task's share of the physical memory.
The task's command name, which will be truncated if it is too long to be displayed on one line. Tasks in memory will have a full command line, but swapped-out tasks will only have the name of the program in parentheses (for example, "(getty)").
"A , WP"
these fields from the kmem top are not supported.


Several single-key commands are recognized while top is running. Some are disabled if thes option has been given on the command line.
Immediately updates the display.
Erases and redraws the screen.
h or ?
Displays a help screen giving a brief summary of commands, and the status of secure and cumulative modes.
Kill a process. You will be prompted for the PID of the task, and the signal to send to it. For a normal kill, send signal 15. For a sure, but rather abrupt, kill, send signal 9. The default signal, as with kill(1), is 15, SIGTERM. This command is not available in secure mode.
Ignore idle and zombie processes. This is a toggle switch.
Toggle between Solaris (CPU percentage divided by total number of CPUs) and Irix (CPU percentage calculated solely by amount of time) views. This is a toggle switch that affects only SMP systems.
n or #
Change the number of processes to show. You will be prompted to enter the number. This overrides automatic determination of the number of processes to show, which is based on window size measurement. If 0 is specified, then top will show as many processes as will fit on the screen; this is the default.
Re-nice a process. You will be prompted for the PID of the task, and the value to nice it to. Entering a positve value will cause a process to be niced to negative values, and lose priority. If root is running top, a negative value can be entered, causing a process to get a higher than normal priority. The default renice value is 10. This command is not available in secure mode.
This toggles cumulative mode, the equivalent of ps -S, i.e., that CPU times will include a process's defunct children. For some programs, such as compilers, which work by forking into many separate tasks, normal mode will make them appear less demanding than they actually are. For others, however, such as shells and init, this behavior is correct. In any case, try cumulative mode for an alternative view of CPU use.
Change the delay between updates. You will be prompted to enter the delay time, in seconds, between updates. Fractional values are recognized down to microseconds. Entering 0 causes continuous updates. The default value is 5 seconds. Note that low values cause nearly unreadably fast displays, and greatly raise the load. This command is not available in secure mode.
f or F
Add fields to display or remove fields from the display. See below for more information.
o or O
Change order of displayed fields. See below for more information.
toggle display of load average and uptime information.
toggle display of memory information.
toggle display of processes and CPU states information.
toggle display of command name or full command line.
sort tasks by pid (numerically).
sort tasks by age (newest first).
sort tasks by CPU usage (default).
sort tasks by resident memory usage.
sort tasks by time / cumulative time.
Write current setup to ~/.toprc. This is the recommended way to write a top configuration file.

The Field and Order Screens

After pressing f, F, o or O you will be shown a screen specifying the field order on the top line and short descriptions of the field contents. The field order string uses the following syntax: If the letter in the filed string corresponding to a field is upper case, the field will be displayed. This is furthermore indicated by an asterisk in front of the field description. The order of the fields corresponds to the order of the letters in the string.
From the field select screen you can toggle the display of a field by pressing the corresponding letter.
From the order screen you may move a field to the left by pressing the corresponding upper case letter resp. to the right by pressing the lower case one.

Configuration Files

The command top reads it's default configuration from two files, /etc/toprc and ~/.toprc. The global configuration file may be used to restrict the usage of top to the secure mode for non-non-privileged users. If this is desired, the file should contain a 's' to specify secure mode and a digit d (2<=d<=9) for the default delay (in seconds) on a single line. The personal configuration file contains two lines. The first line contains lower and upper letters to specify which fields in what order are to be displayed. The letters correspond to the letters in the Fields or Order screens from top. As this is not very instructive, it is recommended to select fields and order in a running top process and to save this using theW interactive command.

The second line is more interesting (and important). It contains information on the other options. Most important, if you have saved a configuration in secure mode, you will not get an insecure top without removing the lower 's' from the second line of your ~/.toprc.

A digit specifies the delay time between updates, a capital 'S' cumulative mode, a lower 'i' no-idle mode, a capital 'I' Irix view. As in interactive mode, a lower 'm', 'l', and 't' suppresses the display of memory, uptime resp. process and CPU state information. Currently changing the default sorting order (by CPU usage) is not supported.