We recommend using Visual Studio 2017

difftime, _difftime32, _difftime64


The new home for Visual Studio documentation is Visual Studio 2017 Documentation on docs.microsoft.com.

The latest version of this topic can be found at difftime, _difftime32, _difftime64.

Finds the difference between two times.

double difftime(   
   time_t timer1,  
   time_t timer0   
double _difftime32(   
   __time32_t timer1,  
   __time32_t timer0   
double _difftime64(   
   __time64_t timer1,  
   __time64_t timer0   


Ending time.

Beginning time.

difftime returns the elapsed time in seconds, from timer0 to timer1. The value returned is a double precision floating-point number. The return value may be 0, indicating an error.

The difftime function computes the difference between the two supplied time values timer0 and timer1.

The time value supplied must fit within the range of time_t. time_t is a 64-bit value. Thus, the end of the range was extended from 23:59:59 January 18, 2038, UTC to 23:59:59, December 31, 3000. The lower range of time_t is still midnight, January 1, 1970.

difftime is an inline function that evaluates to either _difftime32 or _difftime64 depending on whether _USE_32BIT_TIME_T is defined. _difftime32 and _difftime64 can be used directly to force the use of a particular size of the time type.

These functions validate their parameters. If either of the parameters is zero or negative, the invalid parameter handler is invoked, as described in Parameter Validation. If execution is allowed to continue, these functions return 0 and set errno to EINVAL.

RoutineRequired header

For additional compatibility information, see Compatibility in the Introduction.

// crt_difftime.c  
// This program calculates the amount of time  
// needed to do a floating-point multiply 100 million times.  
#include <stdio.h>  
#include <stdlib.h>  
#include <time.h>  
#include <float.h>  
double RangedRand( float range_min, float range_max)  
   // Generate random numbers in the half-closed interval  
   // [range_min, range_max). In other words,  
   // range_min <= random number < range_max  
   return ((double)rand() / (RAND_MAX + 1) * (range_max - range_min)  
            + range_min);  
int main( void )  
   time_t   start, finish;  
   long     loop;  
   double   result, elapsed_time;  
   double   arNums[3];  
   // Seed the random-number generator with the current time so that  
   // the numbers will be different every time we run.  
   srand( (unsigned)time( NULL ) );  
   arNums[0] = RangedRand(1, FLT_MAX);  
   arNums[1] = RangedRand(1, FLT_MAX);  
   arNums[2] = RangedRand(1, FLT_MAX);  
   printf( "Using floating point numbers %.5e %.5e %.5e\n", arNums[0], arNums[1], arNums[2] );  
   printf( "Multiplying 2 numbers 100 million times...\n" );  
   time( &start );  
   for( loop = 0; loop < 100000000; loop++ )  
      result = arNums[loop%3] * arNums[(loop+1)%3];   
   time( &finish );  
   elapsed_time = difftime( finish, start );  
   printf( "\nProgram takes %6.0f seconds.\n", elapsed_time );  

Using random floating point numbers 1.04749e+038 2.01482e+038 1.72737e+038Multiplying 2 floating point numbers 100 million times...Program takes      3 seconds.Multiplying 2 floating point numbers 500 million times...  
Program takes      5 seconds.  


Floating-Point Support
Time Management
time, _time32, _time64