SE250:lab-1:dcho040

How long does it take to execute an addition (+) operation in C?

Information

• Variable sizes

Int   : -2,147,483,648    to 2,147,483,647
Long  : -2,147,483,648    to 2,147,483,647
Short : -32,768           to 32,767
float : (+-)1.175494e-38  to (+-)3.402823e+38
double: (+-)2.225074e-308 to (+-)1.797693e+308

• Clock function

• In the file 'time.h'
• Use type 'clock_t'

Expectation

• The time of running time would be effected by the range of each type

Expected order by the speed

• 1: short
• 2: int and long
• 3: float
• 4: double

Testing codes

Sample1

• Codes

#define _CRT_SECURE_NO_DEPRECATE
#include <stdio.h>
#include <time.h>
int main(void)
{
   int a;
   int b=1;
   clock_t t0 = clock();
   for (a=1; a<1000; a++) {
       b = b+a;
   }
   printf("%d\n", b);
   printf("%d", clock() - t0);
   return 0;
}

• Results

b = 499501
time = 0

• discussion

The time is too small -> more calculations are needed

Sample2

• Codes

#define _CRT_SECURE_NO_DEPRECATE
#include <stdio.h>
#include <time.h>
int main(void)
{
   int a;
   int b=1;
   clock_t t0 = clock();
   for (a=1; a<1000000000; a++) {
       b = b+a;
   }
   printf("%d\n", b);
   printf("%d", clock() - t0);
   return 0;
}

• Results

b = -1243309311
Time = 3475

• Discussion

B = -1243309311 because B is bigger than the bigger than 2,147,483,647(maximum int value) -> All calculation should be calculated with short size (the smallest variable: -32,768 to 32,767)

Sample3

• Codes

#define _CRT_SECURE_NO_DEPRECATE
#include <stdio.h>
#include <time.h>
int main(void)
{
   int a, b;
   int c=0;
   clock_t t0 = clock();
   for (a=1; a<32767; a++) {
       for(b=1; b<32767; b++) { 
          c = c+a-b;
       }
   }
   printf("c needs to be 0 = %d\n", c);
   printf("time = %d", clock() - t0);
   return 0;
}

• Results

c = 0
Time = 3892

• discussion

While calculate the answer, 'c' goes over 32767, yet this looks fine.

Sample4(final)

• Codes

#define _CRT_SECURE_NO_DEPRECATE
#include <stdio.h>
#include <time.h>
int main(void)
{
   int a, b, c=1, d=0;
   clock_t t0 = clock();
   for (a=1; a<32767; a++) {
       for(b=1; b<32767; b++) { 
          d=d+c;
       }
       c=c*-1;
   }
   printf("c needs to be 32767 = %d\n", d);
   printf("time = %d", clock() - t0);
   return 0;
}

• Results

c = 32767
Time = 2627

• Discussion

• This example includes ‘*’, ‘+’, and ’-‘ which increase the processing time, yet this lab suggests to use addition (+) operation.

Testing

Vista(Visual Studio 2008)

Type   Int   long  Short Float Double
--------------------------------------
Test1  2346  2334  3011  9074  9397
Test2  2463  2339  3251  9295  9122
Test3  2447  2343  3086  9029  9054
Test4  2337  2373  2981  9056  9053
Test5  2346  2336  2991  9026  9057
--------------------------------------
AVG    2388  2345  3064  9096  9137

• Discussion

• The results are not always same.

• The speed of a programme is may affected by the condition of the CPU and other running programmes.

• ‘int’ and ‘long’ are similar, and are the fastest.
• ’short’ is little slower than ‘int’ and ‘long’
• ’float’ and ‘double’ are similar, and about 3 times slower than ‘short’

• Data types, ‘float’ and ‘double’, have different range, but floating point data may types show same processing speed.

Vista(CYGWIN)

Type   Int   long  Short Float Double
--------------------------------------
Test1  3588  3541  6615  5102  5085
Test2  3540  3603  6723  5132  5132
Test3  3557  3603  6614  5070  5085
Test4  3541  3525  6693  5070  5101
Test5  3572  3556  6661  5116  5085
--------------------------------------
AVG    3560  3566  6661  5098  5098

• Discussion

• ‘int’ and ‘long’ are similar, and are the fastest.
• ’float’ and ‘double’ are similar, and are slower than ‘int’ and ‘long’

• The difference between ‘float’ and ‘double’, and ‘int’ and ‘long’ are smaller than in Visual Studio 2005.

• ’Short’ is slower than ‘float and ‘double which is unexpected.

• The type, ’short’, may uses difference processing method than ‘int’, and ‘long’

Conclusion

• The speed of a programme is may changeable by condition of the CPU and other running programmes.
• Same range data types may show similar processing speed (e.g. int and long).
• Floating point data types may show similar processing speed (e.g. float and double).
• This report suggests some rules with the time, but these rules could be different in other conditions. (e.g. 64bit windows, Linux, different cpus, etc)