Timing

The header file time.h defines a number of library functions which can be used to assess how much CPU time a C program consumes during execution. The simplest such function is called clock(). A call to this function, with no argument, will return the amount of CPU time used so far by the calling program. The time is returned in a special data type, clock_t, defined in time.h. This time must be divided by CLOCKS_PER_SEC, also defined in time.h, in order to covert it into seconds. The ability to measure how much CPU time a given code consumes is useful in scientific programming: e.g., because it allows the effectiveness of the various available compiler optimization flags to be determined. Optimization usually (but not always!) speeds up the execution of a program. However, over aggressive optimization can often slow a program down again.

The program listed below illustrates the simple use of the clock() function. The program compares the CPU time required to raise a double to the fourth power via a direct calculation and via a call to the pow() function. Actually, both operations are performed a million times and the elapsed CPU time is then divided by a million.

/* timing.c */
/* 
  Program to test operation of clock() function
*/

#include <time.h>
#include <math.h>
#define N_LOOP 1000000

int main()
{
  int i;
  double a = 11234567890123456.0, b;
  clock_t time_1, time_2;

  time_1 = clock();
  for (i = 0; i < N_LOOP; i++) b = a * a * a * a;
  time_2 = clock();
  printf ("CPU time needed to evaluate a*a*a*a:    %f microsecs\n",
          (double) (time_2 - time_1) / (double) CLOCKS_PER_SEC);

  time_1 = clock();
  for (i = 0; i < N_LOOP; i++) b = pow(a, 4.);
  time_2 = clock();
  printf ("CPU time needed to evaluate pow(a, 4.): %f microsecs\n",
          (double) (time_2 - time_1) / (double) CLOCKS_PER_SEC);

  return 0;
}

The typical output from this program is as follows:

 
CPU time needed to evaluate a*a*a*a:    0.190000 microsecs
CPU time needed to evaluate pow(a, 4.): 1.150000 microsecs
%

Clearly, evaluating a fourth power using the pow() function is a lot more expensive than the direct calculation. Hence, as has already been mentioned, the pow() function should not be used to raise floating point quantities to small integer powers.