• Gang Sung Jin <Gang_member pathlink.com> Apr 30 2004
• Ilya Minkov <minkov cs.tum.edu> Apr 30 2004

Gang Sung Jin <Gang_member pathlink.com> writes:

test long Arithmetic result is:
in D, 155744 ms
but C,  33388 ms

why so much different? int and double type Arithmetic result is similrar.
thanks.

test code D.ver is:
//from /* All code Copyright 2004 Christopher W. Cowell-Shah */
import std.math;
import std.date;

int main()
{
long longMin = 10000000000; // 10B
long longMax = 11000000000; // 11B

longArithmetic(longMin, longMax);

return 0;
}

double longArithmetic(long longMin, long longMax)
{
d_time startTime = getUTCtime();

long longResult = longMin;
long i = longMin;
while (i < longMax)
{
longResult -= i++;
longResult += i++;
longResult *= i++;
longResult /= i++;
}

d_time stopTime = getUTCtime();
double elapsedTime = (stopTime - startTime) / (TicksPerSecond / (double)
1000.0);
printf("Long arithmetic elapsed time: %1.0f ms with longMax %lld\n",
elapsedTime, longMax);
printf(" i: %lld\n", i);
printf(" longResult: %lld\n", longResult);
return elapsedTime;
}

C code is:
double longArithmetic(long long longMin, long long longMax)
{
double elapsedTime;
clock_t stopTime;
clock_t startTime = clock();

long long longResult = longMin;
long long i = longMin;
while (i < longMax)
{
longResult -= i++;
longResult += i++;
longResult *= i++;
longResult /= i++;
}

stopTime = clock();
elapsedTime = (stopTime - startTime) / (CLOCKS_PER_SEC / (double) 1000.0);
printf("Long arithmetic elapsed time: %1.0f ms with longMax %I64d\n",
elapsedTime, longMax);
printf(" i: %I64d\n", i);
printf(" longResult: %I64d\n", longResult);
return elapsedTime;
}

Ilya Minkov <minkov cs.tum.edu> writes:

Against what compiler did you test? It has already been reported that 
DigitalMars compilers back-end are somewhat weakish in performance with 
respect to 64-bit integers.

I you should get roughly the same result comparing DigitalMars D 
compiler against DigitalMars C compiler.

Now there is also a new Gnu D compiler, giving you the ability to 
compare against GCC. However, if it scores good - wonderful. If it 
doesn't - no reason to worry, it has only been released a few weeks ago 
and has a chance to improve.

-eye

Gang Sung Jin schrieb:
 test long Arithmetic result is:
 in D, 155744 ms
 but C,  33388 ms
 
 why so much different? int and double type Arithmetic result is similrar.
 thanks.
 
 test code D.ver is:
 //from /* All code Copyright 2004 Christopher W. Cowell-Shah */
 import std.math;
 import std.date;
 
 int main()
 {
 long longMin = 10000000000; // 10B
 long longMax = 11000000000; // 11B
 
 longArithmetic(longMin, longMax);
 
 return 0;
 }
 
 double longArithmetic(long longMin, long longMax)
 {
 d_time startTime = getUTCtime();
 
 long longResult = longMin;
 long i = longMin;
 while (i < longMax)
 {
 longResult -= i++;
 longResult += i++;
 longResult *= i++;
 longResult /= i++;
 }
 
 d_time stopTime = getUTCtime();
 double elapsedTime = (stopTime - startTime) / (TicksPerSecond / (double)
 1000.0);
 printf("Long arithmetic elapsed time: %1.0f ms with longMax %lld\n",
 elapsedTime, longMax);
 printf(" i: %lld\n", i);
 printf(" longResult: %lld\n", longResult);
 return elapsedTime;
 }
 
 C code is:
 double longArithmetic(long long longMin, long long longMax)
 {
 double elapsedTime;
 clock_t stopTime;
 clock_t startTime = clock();
 
 long long longResult = longMin;
 long long i = longMin;
 while (i < longMax)
 {
 longResult -= i++;
 longResult += i++;
 longResult *= i++;
 longResult /= i++;
 }
 
 stopTime = clock();
 elapsedTime = (stopTime - startTime) / (CLOCKS_PER_SEC / (double) 1000.0);
 printf("Long arithmetic elapsed time: %1.0f ms with longMax %I64d\n",
 elapsedTime, longMax);
 printf(" i: %I64d\n", i);
 printf(" longResult: %I64d\n", longResult);
 return elapsedTime;
 }