The Open Group Base Specifications Issue 6
IEEE Std 1003.1, 2004 Edition
Copyright © 2001-2004 The IEEE and The Open Group, All Rights reserved.
A newer edition of this document exists here


times - get process and waited-for child process times


#include <sys/times.h>

clock_t times(struct tms *


The times() function shall fill the tms structure pointed to by buffer with time-accounting information. The tms structure is defined in <sys/times.h>.

All times are measured in terms of the number of clock ticks used.

The times of a terminated child process shall be included in the tms_cutime and tms_cstime elements of the parent when wait() or waitpid() returns the process ID of this terminated child. If a child process has not waited for its children, their times shall not be included in its times.


Upon successful completion, times() shall return the elapsed real time, in clock ticks, since an arbitrary point in the past (for example, system start-up time). This point does not change from one invocation of times() within the process to another. The return value may overflow the possible range of type clock_t. If times() fails, (clock_t)-1 shall be returned and errno set to indicate the error.


No errors are defined.

The following sections are informative.


Timing a Database Lookup

The following example defines two functions, start_clock() and end_clock(), that are used to time a lookup. It also defines variables of type clock_t and tms to measure the duration of transactions. The start_clock() function saves the beginning times given by the times() function. The end_clock() function gets the ending times and prints the difference between the two times.

#include <sys/times.h>
#include <stdio.h>
void start_clock(void);
void end_clock(char *msg);
static clock_t st_time;
static clock_t en_time;
static struct tms st_cpu;
static struct tms en_cpu;
    st_time = times(&st_cpu);

/* This example assumes that the result of each subtraction is within the range of values that can be represented in an integer type. */ void end_clock(char *msg) { en_time = times(&en_cpu);
fputs(msg,stdout); printf("Real Time: %jd, User Time %jd, System Time %jd\n", (intmax_t)(en_time - st_time), (intmax_t)(en_cpu.tms_utime - st_cpu.tms_utime), (intmax_t)(en_cpu.tms_stime - st_cpu.tms_stime)); }


Applications should use sysconf(_SC_CLK_TCK) to determine the number of clock ticks per second as it may vary from system to system.


The accuracy of the times reported is intentionally left unspecified to allow implementations flexibility in design, from uniprocessor to multi-processor networks.

The inclusion of times of child processes is recursive, so that a parent process may collect the total times of all of its descendants. But the times of a child are only added to those of its parent when its parent successfully waits on the child. Thus, it is not guaranteed that a parent process can always see the total times of all its descendants; see also the discussion of the term ``realtime'' in alarm().

If the type clock_t is defined to be a signed 32-bit integer, it overflows in somewhat more than a year if there are 60 clock ticks per second, or less than a year if there are 100. There are individual systems that run continuously for longer than that. This volume of IEEE Std 1003.1-2001 permits an implementation to make the reference point for the returned value be the start-up time of the process, rather than system start-up time.

The term ``charge'' in this context has nothing to do with billing for services. The operating system accounts for time used in this way. That information must be correct, regardless of how that information is used.




alarm(), exec(), fork(), sysconf(), time(), wait(), the Base Definitions volume of IEEE Std 1003.1-2001, <sys/times.h>


First released in Issue 1. Derived from Issue 1 of the SVID.

End of informative text.

UNIX ® is a registered Trademark of The Open Group.
POSIX ® is a registered Trademark of The IEEE.
[ Main Index | XBD | XCU | XSH | XRAT ]