C Date Time - JohnHau/mis GitHub Wiki

last modified July 13, 2020

In this tutorial, we show how to work with date and time in C. In the tutorial, we use C99. If you are compiling programs on Windows, we highly recommend Pelles C compiler. Windows API tutorial has a chapter dedicated to date and time.

C date time definitions We start with a few definitions. A calendar time, also called an absolute time, is a point in the time continuum, for example February 17, 2016 at 13:02:5 CET. A time interval is a continuous part of the time continuum between two calendar times, for example the hour between 13:00 and 14:00 on February 20, 2000. An elapsed time is the length of an interval, for example, 28 minutes.

An amount of time is a sum of elapsed times. The elapsed times do not need to be successive. When the work took us eleven hours, we might be working on different days. A period is the elapsed time of an interval between two events. CPU time is the amount of time for which a central processing unit (CPU) was used for processing instructions of a computer program or operating system. It is measured in clock ticks or seconds. An epoch is an instant in time chosen as the origin of a particular era. The Unix epoch is the time 00:00:00 UTC on 1 January 1970 (or 1970-01-01T00:00:00Z ISO 8601).

Simple time is a compact representation of a calendar time; it is the number of seconds of elapsed time since Unix epoch. Simple time uses the time_t data type. Broken-down time represents a human-readable calendar time. It is divided into a set of components specifying the year, month, day, and so on in the Gregorian calendar, for a specific time zone. The broken-down time uses the struct tm data type.

Wall time, also called real-world time or wall-clock time, refers to elapsed time as determined by a chronometer such as a wristwatch or wall clock. (The reference to a wall clock is how the term originally got its name.) Wall time differs from time as measured by counting microprocessor clock pulses or cycles.

The time.h contains the following function prototypes:

char *asctime(const struct tm *) — converts date and time to a string (obsolete) clock_t clock(void) — returns an approximation of the CPU time used by a process int clock_getres(clockid_t, struct timespec *) — returns the resolution of a clock int clock_gettime(clockid_t, struct timespec *) — retrieves the current time of the clock int clock_settime(clockid_t, const struct timespec *) — sets the current time of a clock specified char *ctime(const time_t *) — converts a time value to date and time string (obsolete) double difftime(time_t, time_t) — computes the difference between two calendar time values struct tm *getdate(const char *) — converts a string representation of a date or time into a broken-down time struct tm *gmtime(const time_t *) — converts a time value to a broken-down UTC time struct tm *localtime(const time_t *) — converts a time value to a broken-down local time time_t mktime(struct tm *) — converts a broken-down time into Unix time size_t strftime(char *, size_t, const char *, const struct tm *) — convert date and time to a string char *strptime(const char *, const char *, struct tm *) — converts a time string to a broken-down time time_t time(time_t *) — returns the Unix time, the number of seconds since the Unix epoch void tzset(void) — sets time zone conversion information In addition, the time.h file defines the CLOCKS_PER_SEC macro, which holds the number of processor clock ticks per second. The clock_t is the process running time type.

Unix time The Unix time is the number of seconds since the Unix epoch. The time() function returns the value of time in seconds since 0 hours, 0 minutes, 0 seconds, January 1, 1970, Coordinated Universal Time. If an error occurs, it returns -1.

time_t time(time_t *t); If t is not NULL, the return value is also stored in the memory pointed to by t.

unixtime.c #include <stdio.h> #include <time.h>

int main(void) {

time_t now = time(NULL);

if (now == -1) {
    
    puts("The time() function failed");
}

printf("%ld\n", now);

return 0;

} The example prints the Unix time.

$ ./unixtime 1455997310 At this moment, 1455997310 have passed since Unix epoch.

Broken-down time Broken-down time is a human-readable version of the calendar time. The struct tm data type is used for broken-down time. The localtime() functions converts a simple calendar time into a broken-down time. It takes the current timezone into account.

struct tm *localtime(const time_t *clock); It stores a tm structure and returns a pointer to that structure.

The following description of the struct tm was taken from FreeBSD manual.

struct tm { int tm_sec; /* seconds / int tm_min; / minutes / int tm_hour; / hours / int tm_mday; / day of the month / int tm_mon; / month / int tm_year; / year / int tm_wday; / day of the week / int tm_yday; / day in the year / int tm_isdst; / daylight saving time */ }; The members of the tm structure are:

tm_sec — the number of seconds after the minute, normally in the range 0 to 59, but can be up to 60 to allow for leap seconds. tm_min — the number of minutes after the hour, in the range 0 to 59. tm_hour — the number of hours past midnight, in the range 0 to 23. tm_mday — the day of the month, in the range 1 to 31. tm_mon — the number of months since January, in the range 0 to 11. tm_year — the number of years since 1900. tm_wday — the number of days since Sunday, in the range 0 to 6. tm_yday — the number of days since January 1, in the range 0 to 365. tm_isdst — a flag that indicates whether daylight saving time is in effect at the time described. The value is positive if daylight saving time is in effect, zero if it is not, and negative if the information is not available. brokendowntime.c #include <stdio.h> #include <time.h>

int main(void) {

time_t rawtime = time(NULL);

if (rawtime == -1) {
    
    puts("The time() function failed");
    return 1;
}

struct tm *ptm = localtime(&rawtime);

if (ptm == NULL) {
    
    puts("The localtime() function failed");
    return 1;
}

printf("The time is: %02d:%02d:%02d\n", ptm->tm_hour, 
       ptm->tm_min, ptm->tm_sec);

return 0;

} In the example, we get the Unix time and convert it into the local broken-down time.

time_t rawtime = time(NULL); We get the raw Unix time.

struct tm *ptm = localtime(&rawtime); The raw time, expressed in seconds, is converted into the brokend-down time. The functions fills a struct tm type and returns a pointer to it. The structure is statically allocated, so we do not have to free it.

printf("The time is: %02d:%02d:%02d\n", ptm->tm_hour, ptm->tm_min, ptm->tm_sec);

We use the tm_hour, tm_min, and tm_sec members to express a human-readable time format.

$ ./brokendowntime The time is: 17:00:12 This is a sample run of the brokendowntime example.

Converting broken-down time to Unix time The mktime() function converts the broken-down time, expressed as local time, in the structure pointed to by tm into the Unix time (simple calendar time).

time_t mktime(struct tm *tm); The mktime() function receives a pointer to the struct tm value and returns a time_t value.

xmas.c #include <stdio.h> #include <time.h>

int main(void) {

struct tm xmas = { 0, 0, 0, 24, 11, 116 };

time_t rawtime = mktime(&xmas);

if (rawtime == -1) {
    
     puts("The mktime() function failed");   
     return 1;
}

printf("The Unix time for Xmas is: %ld\n", rawtime);

return 0;

} In the example, we calculate a Unix time for a specific time in the future.

struct tm xmas = { 0, 0, 0, 24, 11, 116 }; This is a time value for the beginning of a XMas in 2016.

time_t rawtime = mktime(&xmas); With the mktime() function, we convert the broken-down time into the Unix time.

$ ./xmas The Unix time for Xmas is: 1482534000 The 1482534000 is the Unix time for the beginning of the XMas in 2016.

UTC time Our planet is a sphere; it revolves round its axis. The Earth rotates towards the east, so the Sun rises at different times in different locations. The Earth rotates once in about 24 hours. Therefore, the world was divided into 24 time zones. In each time zone, there is a different local time. This local time is often further modified by the daylight saving.

There is a pragmatic need for one global time. One global time helps to avoid confusion about time zones and daylight saving time. The UTC (Universal Coordinated time) was chosen to be the primary time standard. UTC is used in aviation, weather forecasts, flight plans, air traffic control clearances, and maps. Unlike local time, UTC does not change with a change of seasons.

The gmtime() converts the time_t value into the broken-down time, not adjusting to the timezone.

utctime.c #include <stdio.h> #include <time.h>

int main(void) {

time_t now = time(&now);

if (now == -1) {
    
    puts("The time() function failed");
}
    
struct tm *ptm = gmtime(&now);

if (ptm == NULL) {
    
    puts("The gmtime() function failed");
}    

printf("UTC time: %s", asctime(ptm));

return 0;

} The example prints the current UTC time.

struct tm *ptm = gmtime(&now); We get the UTC time with the gmtime() function.

printf("UTC time: %s", asctime(ptm)); We print the UTC time. The asctime() function is obsolete, we should use strftime(), which is covered next, instead.

$ ./utctime UTC time: Sat Feb 20 19:47:39 2016 $ date Sat Feb 20 20:47:40 CET 2016 For the CET time zone, there is one hour difference in time.

Formatting time The strftime() function formats date and time.

size_t strftime(char *s, size_t max, const char *format, const struct tm *tm); The strftime() function formats the broken-down time tm according to the format specification format and places the result in the character array s of size max.

The following examples present some of the format specifiers.

today.c #include <stdio.h> #include <time.h>
#include <string.h>

#define BUF_LEN 256

int main(void) {

char buf[BUF_LEN] = {0};

time_t rawtime = time(NULL);

if (rawtime == -1) {
    
    puts("The time() function failed");
    return 1;
}

struct tm *ptm = localtime(&rawtime);

if (ptm == NULL) {
 
    puts("The localtime() function failed");
    return 1;
}

strftime(buf, BUF_LEN, "%d/%m/%Y", ptm);
puts(buf);

memset(buf, 0, BUF_LEN);
strftime(buf, BUF_LEN, "%d. %m. %Y", ptm);

puts(buf);

return 0;

} The example prints today's date in a British and Slovak format.

strftime(buf, BUF_LEN, "%d/%m/%Y", ptm); This code line creates a date string in a British format. We use the %d, %m, and %Y format specifiers.

strftime(buf, BUF_LEN, "%d. %m. %Y", ptm); In this line, a Slovak date format is created.

$ ./today 20/02/2016 20. 02. 2016 This is the output of the example.

In the following example, we work with additional format specifiers.

timeformats.c #include <stdio.h> #include <time.h>
#include <string.h>

#define BUF_LEN 256

int main(void) {

char buf[BUF_LEN] = {0};

time_t rawtime = time(NULL);

if (rawtime == -1) {
    
    puts("The time() function failed");
    return 1;   
}

struct tm *ptm = localtime(&rawtime);

if (ptm == NULL) {
    
    puts("The localtime() function failed");
    return 1;
}

strftime(buf, BUF_LEN, "Today is %A", ptm);
puts(buf);

memset(buf, 0, BUF_LEN);
strftime(buf, BUF_LEN, "The month is %B", ptm);
puts(buf);

memset(buf, 0, BUF_LEN);
strftime(buf, BUF_LEN, "Today is %-d day of %B", ptm);
puts(buf);

memset(buf, 0, BUF_LEN);
strftime(buf, BUF_LEN, "Today is %-j day of %G", ptm);
puts(buf);    

memset(buf, 0, BUF_LEN);
strftime(buf, BUF_LEN, "Today is %-W week of %G", ptm);
puts(buf);        

memset(buf, 0, BUF_LEN);
strftime(buf, BUF_LEN, "The time is %T", ptm);
puts(buf);     

memset(buf, 0, BUF_LEN);
strftime(buf, BUF_LEN, "The date is %D", ptm);
puts(buf);       

memset(buf, 0, BUF_LEN);
strftime(buf, BUF_LEN, "The timezone is %Z", ptm);
puts(buf);       

return 0;

} In the code example, we determine the current time and convert it into additional information with the strftime() function.

strftime(buf, BUF_LEN, "Today is %A", ptm); With the %A specifier, we get the full weekday name.

strftime(buf, BUF_LEN, "The month is %B", ptm); With the %B specifier, we get the full month name.

strftime(buf, BUF_LEN, "Today is %-d day of %B", ptm); With the %d specifier, we get day of the month. The - character after % removes the leading zero from the value.

strftime(buf, BUF_LEN, "Today is %-j day of %G", ptm); The %j specifier gives the day of the year; the %G specifier gives the year.

strftime(buf, BUF_LEN, "Today is %-W week of %G", ptm); The %W specifier gives week number of the year.

strftime(buf, BUF_LEN, "The time is %T", ptm); The %T specifier gives the ISO 8601 time format.

strftime(buf, BUF_LEN, "The date is %D", ptm); The %D specifier gives the short date format.

strftime(buf, BUF_LEN, "The timezone is %Z", ptm); Finally, the ZT specifier gives the timezone name.

$ ./timeformats Today is Saturday The month is February Today is 20 day of February Today is 51 day of 2016 Today is 7 week of 2016 The time is 20:22:49 The date is 02/20/16 The timezone is CET On February 20, 2016 we get these results.

In the following example, we use a localized date format.

localized.c #include <stdio.h> #include <time.h>
#include <locale.h>

#define BUF_LEN 256

int main(void) {

char buf[BUF_LEN] = {0};       

if (setlocale(LC_TIME, "sk_SK.utf8") == NULL) {
    
    puts("Unable to set locale");
    return 1;
}

time_t now = time(NULL);

if (now == -1) {
    
    puts("The time() function failed");
    return 1;
}

struct tm *ptm = localtime(&now);

if (ptm == NULL) {
    
    puts("The localtime() function failed");
    return 1;
}

strftime(buf, BUF_LEN, "%c", ptm);
printf("Dnešný dátum: %s\n", buf);

return 0;

} The example prints today's date in Slovak language.

if (setlocale(LC_TIME, "sk_SK.utf8") == NULL) {

puts("Unable to set locale");
return 1;

} With the setlocale() function, we set the Slovak locale.

strftime(buf, BUF_LEN, "%c", timeinfo); We use the %c specifier, which is a national representation of date and time.

$ ./localized Dnešný dátum: So 20. február 2016, 16:25:11 CET On February 20, 2016, we get this Slovak date format in Slovak language.

Parsing time The strptime() function is used to convert a string representation of time to the tm structure (broken-down time). The strptime() function is the converse function to the strftime() function.

char *strptime(const char *s, const char *format, struct tm *tm); The function's first parameter is the time expressed in a string. The second parameter is the string format. The third parameter is a pointer to the struct tm. The function returns a pointer to the first character not processed in this function call. On failure, the function returns NULL.

parsetime.c #define _XOPEN_SOURCE 700 #include <stdio.h> #include <time.h>

int main(void) {

struct tm tm = {0};

char *s = strptime("16 Feb 2016 14:43:00", "%d %b %Y %H:%M:%S", &tm);

if (s == NULL) {
    
    printf("Cannot parse date"); 
    return 1;
}

printf("year: %d; month: %d; day: %d;\n",
        tm.tm_year + 1900, tm.tm_mon, tm.tm_mday);
printf("hour: %d; minute: %d; second: %d\n",
        tm.tm_hour, tm.tm_min, tm.tm_sec);
printf("week day: %d; year day: %d\n", tm.tm_wday, tm.tm_yday);

return 0;

} The example parses a string time and fills a struct tm; it prints the members of the structure then.

#define _XOPEN_SOURCE 700 This definition gets rid of the warning: initialization makes pointer from integer without a cast warning. It tells your compiler to include definitions for some extra functions that are defined in the X/Open and POSIX standards.

char *s = strptime("16 Feb 2016 14:43:00", "%d %b %Y %H:%M:%S", &tm); We parse a date with the strptime() function. The format specifiers must match the string parts.

printf("year: %d; month: %d; day: %d;\n", tm.tm_year + 1900, tm.tm_mon, tm.tm_mday); Here we print the tm_year, tm_mon, and tm_mday members of the struct tm, which was filled with the strptime().

$ ./parsetime year: 2016; month: 1; day: 16; hour: 14; minute: 43; second: 0 week day: 2; year day: 46 This is the output of the parsetime example.

Time arithmetic The time_t type is an arithmetic type, we can do arithmetic operations with it.

The time functions work with dates from 1900. With earlier dates, we need to refer to Julian day.

time_arithmetic.c #include <stdio.h> #include <time.h>

#define BUF_LEN 256

int main(void) {

char buf[BUF_LEN] = {0};
struct tm day = { 0, 0, 0, 20, 1, 116 };    

time_t one_day = 86400;
time_t sixty_days = one_day * 60;

time_t t1 = mktime(&day);

if (t1 == -1) {
    
    puts("The mktime() function failed");   
    return 1;
}    

time_t t2 = t1 + one_day + sixty_days;

struct tm *ptm = gmtime(&t2);    

if (ptm == NULL) {
    
    puts("The gmtime() function failed");
}  

strftime(buf, BUF_LEN, "%F", ptm);
puts(buf);

return 0;

} In the example, we add sixty days to February 20, 2016 and get the resulting date.

struct tm day = { 0, 0, 0, 20, 1, 116 }; We define the initial date; it is February 20, 2016.

time_t one_day = 86400; time_t sixty_days = one_day * 60; This is sixty days expressed in seconds.

time_t t1 = mktime(&day); The mktime() method converts the date into seconds.

time_t t2 = t1 + one_day + sixty_days; We go to the end of the day by adding one day in seconds and then add sixty days.

struct tm *ptm = gmtime(&t2);
We get the broken-down time with the gmtime() method.

strftime(buf, BUF_LEN, "%F", ptm); puts(buf); We format the date with the strftime() method and print it with the puts method.

$ ./time_arithmetic 2016-04-20 Adding 60 days to February 20, 2016 gives April 20, 2016.

Julian day A Julian day is the number of days since the beginning of the Julian Period. It is used primarily by astronomers. (It should not be confused with the Julian calendar.) The Julian Period started in 4713 BC. The Julian day number 0 is assigned to the day starting at noon on January 1, 4713 BC. The Julian Day Number (JDN) is the number of days elapsed since the beginning of this period. The Julian Date (JD) of any instant is the Julian day number for the preceding noon plus the fraction of the day since that instant. Apart from astronomy, Julian dates are often used by military and mainframe programs.

days_since_borodino.c #include <stdio.h> #include <time.h>

#define BUF_LEN 256

int getJulianDay(int day, int month, int year);

int main(void) {

time_t now = time(NULL);

if (now == -1) {
    
    puts("The time() function failed");
    return 1;
}

struct tm *ptm = localtime(&now);

if (ptm == NULL) {
 
    puts("The localtime() function failed");
    return 1;
}

int year = ptm->tm_year + 1900;
int month = ptm->tm_mon + 1;
int day = ptm->tm_mday;
    
int jtoday = getJulianDay(day, month, year);
int jborodino = getJulianDay(7, 9, 1812);

int delta = jtoday - jborodino;

printf("%d days have passed since the Borodino battle\n", delta);

return 0;

}

int getJulianDay(int day, int month, int year) {

int a = (14 - month) / 12;
int y = year + 4800 - a;
int m = month + 12*a - 3;

int jdn = 0;

if ((year > 1582) || (year == 1582 && month > 10) || 
    (year == 1582 && month == 10 && day < 15)) {
    
    jdn = day + (153*m+2)/5 + 365*y + y/4 - y/100 + y/400 - 32045;
} else {
    
    jdn = day + (153*m+2)/5 + 365*y + y/4 - 32045;
}

return jdn;

}
In the example, we calculate the number of days passed since the Borodino battle. The Borodino battle started on September 7, 1812.

time_t now = time(NULL);

if (now == -1) {

puts("The time() function failed");
return 1;

}

struct tm *ptm = localtime(&now);

if (ptm == NULL) {

puts("The localtime() function failed");
return 1;

}

int year = ptm->tm_year + 1900; int month = ptm->tm_mon + 1; int day = ptm->tm_mday; We determine today's year, month, and day.

int jtoday = getJulianDay(day, month, year); int jborodino = getJulianDay(7, 9, 1812); We calculate the Julian days for today and for the Borodino battle.

int delta = jtoday - jborodino;

printf("%d days have passed since Borodino battle\n", delta); We compute the difference in days and print it to the console.

int getJulianDay(int day, int month, int year) {

int a = (14 - month) / 12;
int y = year + 4800 - a;
int m = month + 12*a - 3;

int jdn = 0;

if ((year > 1582) || (year == 1582 && month > 10) || 
    (year == 1582 && month == 10 && day < 15)) {
    
    jdn = day + (153*m+2)/5 + 365*y + y/4 - y/100 + y/400 - 32045;
} else {
    
    jdn = day + (153*m+2)/5 + 365*y + y/4 - 32045;
}

return jdn;

}
The getJulianDay() function has the algorithm for calculating the Julian day. There are two calculations; one for Gregorian era and one for pre-Gregorian era.

$ ./days_since_borodino 74310 days have passed since the Borodino battle On February 20, 2016, 74310 days have passed since the Borodino Battle.

Time zones A time zone is a region throughout which the same standard time is used. There are 24 time zones in the world.

UTC = local time + bias The bias is the difference between UTC time and local time.

The tzset() function initiates the tzname array with the standard names of the pair of time zones: the standard time zone and the daylight saving time zone. If daylight saving time zone is not used, the second string is empty. The function also initiates the timezone variable, which contains the difference between UTC and the latest local standard time, in seconds west of UTC.

timezone.c #include <stdio.h> #include <time.h>

int main(void) {

tzset();

printf("The timezone is %s and %s\n", tzname[0], tzname[1]);

printf("The bias is %ld seconds\n", timezone); 

return 0;

} The program prints the current time zone and the bias.

tzset(); The timezone information is set with the tzset() function.

printf("The timezone is %s and %s\n", tzname[0], tzname[1]); The tzset() function fille the tzname array with timezone strings.

$ ./timezone The timezone is CET and CEST The bias is -3600 seconds On a system located in Bratislava, we get these values.

Timezone information can be also retrieved from the strftime() function.

timezone2.c #include <stdio.h> #include <time.h>

#define BUF_LEN 256

int main(void) {

char buf[BUF_LEN] = {0};    

time_t rawtime = time(NULL);

if (rawtime == -1) {
    
    puts("The time() function failed");
    return 1;
}

struct tm *ptm = localtime(&rawtime);

if (ptm == NULL) {

    puts("The localtime function failed");
    return 1;
}

strftime(buf, BUF_LEN, "Timezone: %Z  bias: %z", ptm);    
puts(buf);
  
return 0;

} In the example, we compute the simple calendar value, break it down to the tm structure, and get the timezone information with the strftime() function.

strftime(buf, BUF_LEN, "Timezone: %Z bias: %z", ptm); The %Z specifier gets the current time zone name and the %z returns the bias. The bias is specified as hours and minutes followed with two digits each with no delimiter between them. The leading plus stands for east of UTC.

$ ./timezone2 Timezone: CET bias: +0100 The timezone on the system is CET. The bias is one hour east of UTC.

Leap year A leap year is a year containing an additional day. The reason for an extra day in the calendar is the difference between the astronomical and the calendar year.

leapyear.c #include <stdio.h> #include <stdbool.h>

bool isLeapYear(int);

int main(void) {

// Assume year >= 1582 in the Gregorian calendar.
int years[] = { 2000, 2002, 2004, 2008, 2012, 2016, 2020,
    1900, 1800, 1600 };

int size = sizeof(years) / sizeof(int);

for (int i=0; i<size; i++) {

    if (isLeapYear(years[i])) {

        printf("%d is a leap year\n", years[i]);
    } else {

        printf("%d is not a leap year\n", years[i]);
    }
}

return 0;

}

bool isLeapYear(int year) {

if (year % 4 != 0) {
    
    return false;
} else if (year % 400 == 0) {
    
    return true;
} else if (year % 100 == 0) {
    
    return false;
} else {
    
    return true;
}

} We have an array of years. We check all years if they are leap years. There is no built-in function to check for a leap year; therefore, a custom isLeapYear() function is created.

// Assume year >= 1582 in the Gregorian calendar. int years[] = { 2000, 2002, 2004, 2008, 2012, 2016, 2020, 1900, 1800, 1600 }; This is an array of years that we check. The years must be in the Gregorian calendar.

for (int i=0; i<size; i++) {

if (isLeapYear(years[i])) {

    printf("%d is a leap year\n", years[i]);
} else {

    printf("%d is not a leap year\n", years[i]);
}

} With the for loop we traverse the array. We check if a year is a leap year using the isLeapYear() function.

bool isLeapYear(int year) {

if (year % 4 != 0) {
    
    return false;
} else if (year % 400 == 0) {
    
    return true;
} else if (year % 100 == 0) {
    
    return false;
} else {
    
    return true;
}

} This is the function for determining a leap year. Leap years are integer multiples of 4. A year that is an integer multiple of 100 is not a leap year, unless it is also an integer multiple of 400, in which case it is also a leap year.

$ ./leapyear 2000 is a leap year 2002 is not a leap year 2004 is a leap year 2008 is a leap year 2012 is a leap year 2016 is a leap year 2020 is a leap year 1900 is not a leap year 1800 is not a leap year 1600 is a leap year This is the output of the leapyear program.

CPU time CPU time is the amount of time for which a central processing unit was used for processing instructions of a computer program.

clock_t clock(void); The clock() function returns the processor time consumed by the program. The clock() function has a low resolution and is not consistent on platform; therefore, it should not be used in situations where high accuracy is required.

Note that on Windows, the clock() function calculates the wall clock time used by the calling process. The wall clock time is a time elapsed according to the computer's internal clock, which should match time in the outside world.

cputime.c #include <stdio.h> #include <time.h>

unsigned long long factorial(int);

int main(void) {

clock_t start = clock();

if (start == -1) {
 
    puts("The clock() function failed");
    return 1;
}
  
unsigned long long f = factorial(20);

printf("The factorial of 20 is: %llu\n", f);
   
clock_t end = clock();

if (end == -1) {
 
    puts("The clock() function failed");
    return 1;
}    

clock_t delta  = end - start;

printf("The program took %lu clicks (%lf seconds).\n", delta, 
       (double) delta/CLOCKS_PER_SEC);

return 0;

}

unsigned long long factorial(int v) {

if (v == 0) {
    
    return 1;
} else {
 
    return v * factorial(v - 1);
}    

} The program calculates a factorial of a number; we use the clock() function to get the time of the computation.

clock_t start = clock();

if (start == -1) {

puts("The clock() function failed");
return 1;

}

unsigned long long f = factorial(20);

printf("The factorial of 20 is: %llu\n", f);

clock_t end = clock(); We compute the factorial between two clock() calls.

clock_t delta = end - start; We get the difference between the start and end values.

printf("The program took %lu clicks (%lf seconds).\n", delta, (double) delta/CLOCKS_PER_SEC); Dividing the delta value by CLOCKS_PER_SEC we get an approximation of duration in seconds.

$ ./cputime The factorial of 20 is: 2432902008176640000 The program took 41 clicks (0.000041 seconds). This is a sample output of the program.

High resolution CPU time The clock_gettime() function can be used to get high resolution time clock.

int clock_gettime(clockid_t, struct timespec *) The function's first parameter specifies the clock Id; there are several types of clocks:

CLOCK_REALTIME CLOCK_MONOTONIC CLOCK_PROCESS_CPUTIME_ID CLOCK_THREAD_CPUTIME_ID The CLOCK_REALTIME is a system-wide clock that measures real (wall-clock) time. This clock is affected by discontinuous jumps in the system time caused by system administrator or by adjtime() and NTP. The CLOCK_MONOTONIC represents monotonic time since some unspecified starting point. This clock is not affected by discontinuous jumps in the system time. The CLOCK_PROCESS_CPUTIME_ID is a high-resolution per-process timer from the CPU. The CLOCK_THREAD_CPUTIME_ID is a thread-specific CPU-time clock.

The clock_gettime() function's second parameter is a timespec structure.

struct timespec { time_t tv_sec; /* seconds / long tv_nsec; / nanoseconds */ }; The timespec structure holds seconds and nanoseconds.

cputime.c #include <stdio.h> #include <time.h> #include <string.h> #include <errno.h>

unsigned long long factorial(int); double time_spec_seconds(struct timespec *);

int main(void) {

struct timespec tstart = {0,0}, tend = {0,0};
int r = clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &tstart);

if (r == -1) {
    
    printf("The clock_gettime() function failed: %s\n", strerror(errno));
    return 1;
}
  
unsigned long long f = factorial(20);

printf("The factorial of 20 is: %llu\n", f);

r = clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &tend);

if (r == -1) {
    
    printf("The clock_gettime() function failed: %s\n", strerror(errno));
    return 1;
}

double delta = time_spec_seconds(&tend) - time_spec_seconds(&tstart);

printf("The computation took %lf seconds.\n", delta);

return 0;

}

double time_spec_seconds(struct timespec* ts) {

return (double) ts->tv_sec + (double) ts->tv_nsec * 1.0e-9;

}

unsigned long long factorial(int v) {

if (v == 0) {
    
    return 1;
} else {
 
    return v * factorial(v - 1);
}    

} With the clock_gettime() we compute the CPU time of a factorial computation.

struct timespec tstart = {0,0}, tend = {0,0}; We create and initiate two timespec structures.

int r = clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &tstart); We call the clock_gettime() with the CLOCK_PROCESS_CPUTIME_ID.

double delta = time_spec_seconds(&tend) - time_spec_seconds(&tstart); We get the delta value by subtracting two timespec structures.

double time_spec_seconds(struct timespec* ts) {

return (double) ts->tv_sec + (double) ts->tv_nsec * 1.0e-9;

} The time_spec_seconds() function converts the timespec structure to seconds.

$ ./cputime2 The factorial of 20 is: 2432902008176640000 The computation took 0.000081 seconds. This is a sample output of the cputime2 program.

Sources The following sources were used to create this article:

C reference The GNU C library manual Linux manual pages wall time (real-world time or wall-clock time)