How can I see which CPU core a thread is running i

Question

In Linux, supposing a thread's pid is [pid], from the directory /proc/[pid] we can get many useful information. For example, these proc files, /proc/[pid]/status,/proc/[pid]/stat and /proc/[pid]/schedstat are all useful. But how can I get the CPU core number that a thread is running in? If a thread is in sleep state, how can I know which core it will run after it is scheduled again?

BTW, is there a way to dump the process(thread) list of running and sleeping tasks for each CPU core?

Answer 1

The answer below is no longer accurate as of 2014

Tasks don't sleep in any particular core. And the scheduler won't know ahead of time which core it will run a thread on because that will depend on future usage of those cores.

To get the information you want, look in /proc/<pid>/task/<tid>/status. The third field will be an 'R' if the thread is running. The sixth from the last field will be the core the thread is currently running on, or the core it last ran on (or was migrated to) if it's not currently running.

31466 (bc) S 31348 31466 31348 34819 31466 4202496 2557 0 0 0 5006 16 0 0 20 0 1 0 10196934 121827328 1091 18446744073709551615 4194304 4271839 140737264235072 140737264232056 217976807456 0 0 0 137912326 18446744071581662243 0 0 17 3 0 0 0 0 0

Not currently running. Last ran on core 3.

31466 (bc) R 31348 31466 31348 34819 31466 4202496 2557 0 0 0 3818 12 0 0 20 0 1 0 10196934 121827328 1091 18446744073709551615 4194304 4271839 140737264235072 140737264231824 4235516 0 0 0 2 0 0 0 17 2 0 0 0 0 0

Currently running on core 2.

To see what the rest of the fields mean, have a look at the Linux kernel source -- specifically the do_task_stat function in fs/proc/array.c or Documentation/filesystems/stat.txt.

Note that all of this information may be obsolete by the time you get it. It was true at some point between when you made the open call on the file in proc and when that call returned.

Answer 2

The "top" command may help towards this, it does not have CPU-grouped list of threads but rather you can see the list of threads (probably for a single process) and which CPU cores the threads are running on by

top -H -p {PROC_ID}

then pressing f to go into field selection, j to enable the CPU core column, and Enter to display.

Answer 3

You can also use ps, something like this:

ps -mo pid,tid,%cpu,psr -p `pgrep BINARY-NAME`

Answer 4

Accepted answer is not accurate. Here are the ways to find out which CPU is running the thread (or was the last one to run) at the moment of inquiry:

1. Directly read /proc/<pid>/task/<tid>/stat. Before doing so, make sure format didn't change with latest kernel. Documentation is not always up to date, but at least you can try https://www.kernel.org/doc/Documentation/filesystems/proc.txt. As of this writing, it will be the 14th value from the end.
2. Use ps. Either give it -F switch, or use output modifiers and add code PSR.
3. Use top with Last Used Cpu column (hitting f gets you to column selection)
4. Use htop with PROCESSOR column (hitting F2 gets you to setup screen)

Answer 5

This can be done with top command. The default top command output does not show these details. To view this detail you will have to press f key while on top command interface and then press j(press Enter key after you pressed j). Now the output will show you details regarding a process and which processor its running. A sample output is shown below.

top - 04:24:03 up 96 days, 13:41,  1 user,  load average: 0.11, 0.14, 0.15
Tasks: 173 total,   1 running, 172 sleeping,   0 stopped,   0 zombie
Cpu(s):  7.1%us,  0.2%sy,  0.0%ni, 88.4%id,  0.1%wa,  0.0%hi,  0.0%si,  4.2%st
Mem:   1011048k total,   950984k used,    60064k free,     9320k buffers
Swap:   524284k total,   113160k used,   411124k free,    96420k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  P COMMAND
12426 nginx     20   0  345m  47m  29m S 77.6  4.8  40:24.92 7 php-fpm
 6685 mysql     20   0 3633m  34m 2932 S  4.3  3.5  63:12.91 4 mysqld
19014 root      20   0 15084 1188  856 R  1.3  0.1   0:01.20 4 top
    9 root      20   0     0    0    0 S  1.0  0.0 129:42.53 1 rcu_sched
 6349 memcache  20   0  355m  12m  224 S  0.3  1.2   9:34.82 6 memcached
    1 root      20   0 19404  212   36 S  0.0  0.0   0:20.64 3 init
    2 root      20   0     0    0    0 S  0.0  0.0   0:30.02 4 kthreadd
    3 root      20   0     0    0    0 S  0.0  0.0   0:12.45 0 ksoftirqd/0

The P column in the output shows the processor core number where the process is currently being executed. Monitoring this for a few minutes will make you understand that a pid is switching processor cores in between. You can also verify whether your pid for which you have set affinity is running on that particular core only

top f navigation screen ^{( a live system example )} :

Fields Management for window 1:Def, whose current sort field is forest view
   Navigate with Up/Dn, Right selects for move then <Enter> or Left commits,
   'd' or <Space> toggles display, 's' sets sort.  Use 'q' or <Esc> to end!

* PID     = Process Id
* USER    = Effective User Name
* PR      = Priority
* NI      = Nice Value
* VIRT    = Virtual Image (KiB)
* RES     = Resident Size (KiB)
* SHR     = Shared Memory (KiB)
* S       = Process Status
* %CPU    = CPU Usage
* %MEM    = Memory Usage (RES)
* TIME+   = CPU Time, hundredths
* COMMAND = Command Name/Line
  PPID    = Parent Process pid
  UID     = Effective User Id
  RUID    = Real User Id
  RUSER   = Real User Name
  SUID    = Saved User Id
  SUSER   = Saved User Name
  GID     = Group Id
  GROUP   = Group Name
  PGRP    = Process Group Id
  TTY     = Controlling Tty
  TPGID   = Tty Process Grp Id
  SID     = Session Id
  nTH     = Number of Threads
* P       = Last Used Cpu (SMP)
  TIME    = CPU Time
  SWAP    = Swapped Size (KiB)
  CODE    = Code Size (KiB)
  DATA    = Data+Stack (KiB)
  nMaj    = Major Page Faults
  nMin    = Minor Page Faults
  nDRT    = Dirty Pages Count
  WCHAN   = Sleeping in Function
  Flags   = Task Flags <sched.h>
  CGROUPS = Control Groups
  SUPGIDS = Supp Groups IDs
  SUPGRPS = Supp Groups Names
  TGID    = Thread Group Id
  ENVIRON = Environment vars
  vMj     = Major Faults delta
  vMn     = Minor Faults delta
  USED    = Res+Swap Size (KiB)
  nsIPC   = IPC namespace Inode
  nsMNT   = MNT namespace Inode
  nsNET   = NET namespace Inode
  nsPID   = PID namespace Inode
  nsUSER  = USER namespace Inode
  nsUTS   = UTS namespace Inode

Answer 6

The threads are not necessary to bound one particular Core (if you did not pin it). Therefore to see the continuous switching of the core you can use (a modified answer of Dmitry):

watch -tdn0.5 ps -mo pid,tid,%cpu,psr -p \`pgrep BINARY-NAME\`

For example:

watch -tdn0.5 ps -mo pid,tid,%cpu,psr -p \`pgrep firefox\`