time man page

 

 

TIME(7)                    Linux Programmer's Manual                   TIME(7)

 

NAME

       time - overview of time and timers

 

DESCRIPTION

   Real time and process time

       Real  time  is  defined  as time measured from some fixed point, either

       from a standard point in the past (see the description of the Epoch and

       calendar  time below), or from some point (e.g., the start) in the life

       of a process (elapsed time).

 

       Process time is defined as the amount of CPU time used  by  a  process.

       This  is  sometimes  divided into user and system components.  User CPU

       time is the time spent executing code in user mode.  System CPU time is

       the  time spent by the kernel executing in system mode on behalf of the

       process (e.g., executing system calls).  The  time(1)  command  can  be

       used  to determine the amount of CPU time consumed during the execution

       of a program.  A program can determine the amount of CPU  time  it  has

       consumed using times(2), getrusage(2), or clock(3).

 

   The Hardware Clock

       Most computers have a (battery-powered) hardware clock which the kernel

       reads at boot time in order to initialize the software clock.  For fur‐

       ther details, see rtc(4) and hwclock(8).

 

   The Software Clock, HZ, and Jiffies

       The  accuracy  of  various  system  calls  that  set  timeouts,  (e.g.,

       select(2), sigtimedwait(2)) and measure CPU time  (e.g.,  getrusage(2))

       is  limited by the resolution of the software clock, a clock maintained

       by the kernel which measures time in jiffies.  The size of a  jiffy  is

       determined by the value of the kernel constant HZ.

 

       The  value  of HZ varies across kernel versions and hardware platforms.

       On i386 the situation is as follows: on kernels  up  to  and  including

       2.4.x,  HZ was 100, giving a jiffy value of 0.01 seconds; starting with

       2.6.0, HZ was raised to 1000, giving a jiffy of 0.001  seconds.   Since

       kernel 2.6.13, the HZ value is a kernel configuration parameter and can

       be 100, 250 (the default) or 1000, yielding a jiffies value of, respec‐

       tively,  0.01, 0.004, or 0.001 seconds.  Since kernel 2.6.20, a further

       frequency is available: 300, a number that divides evenly for the  com‐

       mon video frame rates (PAL, 25 HZ; NTSC, 30 HZ).

 

       The  times(2)  system  call is a special case.  It reports times with a

       granularity defined by the kernel constant USER_HZ.  Userspace applica‐

 

       tions    can    determine    the   value   of   this   constant   using

       sysconf(_SC_CLK_TCK).

 

   High-Resolution Timers

       Before Linux 2.6.21, the accuracy of timer and sleep system calls  (see

       below) was also limited by the size of the jiffy.

 

       Since  Linux  2.6.21,  Linux  supports  high-resolution  timers (HRTs),

       optionally configurable via CONFIG_HIGH_RES_TIMERS.  On a  system  that

       supports  HRTs,  the  accuracy  of  sleep  and timer system calls is no

       longer constrained by the jiffy, but instead can be as accurate as  the

       hardware  allows  (microsecond accuracy is typical of modern hardware).

       You can determine  whether  high-resolution  timers  are  supported  by

       checking  the resolution returned by a call to clock_getres(2) or look‐

       ing at the "resolution" entries in /proc/timer_list.

 

       HRTs are not supported on all hardware architectures.  (Support is pro‐

       vided on x86, arm, and powerpc, among others.)

 

   The Epoch

       Unix  systems  represent  time  in  seconds  since  the Epoch, which is

       defined as 0:00:00 UTC on the morning of 1 January 1970.

 

       A program can determine the calendar time using gettimeofday(2),  which

       returns  time (in seconds and microseconds) that have elapsed since the

       Epoch; time(2) provides similar information, but only with accuracy  to

       the  nearest  second.   The system time can be changed using settimeof‐

       day(2).

 

   Broken-down time

       Certain library functions use a structure of type tm to represent  bro‐

       ken-down time, which stores time value separated out into distinct com‐

       ponents (year, month, day, hour, minute, second, etc.).  This structure

       is  described  in ctime(3), which also describes functions that convert

       between calendar time and broken-down time.  Functions  for  converting

       between  broken-down  time  and printable string representations of the

       time are described in ctime(3), strftime(3), and strptime(3).

 

 

   Sleeping and Setting Timers

       Various system calls and functions allow a program  to  sleep  (suspend

       execution)   for   a   specified  period  of  time;  see  nanosleep(2),

       clock_nanosleep(2), and sleep(3).

 

       Various system calls allow a process to set a  timer  that  expires  at

       some  point  in  the  future, and optionally at repeated intervals; see

       alarm(2), getitimer(2), timerfd_create(2), and timer_create(2).

 

SEE ALSO

       date(1),    time(1),    adjtimex(2),    alarm(2),     clock_gettime(2),

       clock_nanosleep(2),  getitimer(2), getrlimit(2), getrusage(2), gettime‐

       ofday(2), nanosleep(2), stat(2), time(2), timer_create(2), timerfd_cre‐

       ate(2),   times(2),   utime(2),   adjtime(3),  clock(3),  clock_getcpu‐

       clockid(3), ctime(3), pthread_getcpuclockid(3), sleep(3),  strftime(3),

       strptime(3), timeradd(3), usleep(3), rtc(4), hwclock(8)

 

COLOPHON

       This  page  is  part of release 3.23 of the Linux man-pages project.  A

       description of the project, and information about reporting  bugs,  can

       be found at http://www.kernel.org/doc/man-pages/.

 

Linux                             2009-02-10                           TIME(7)