...Thank you for that. What he is really looking for, for me, is a C/C++ solution to rounding floating point numbers(doubles) to a specific number of decimal places with the most accuracy that can be had...


Ray,

Fast Rounding of Floating Point Numbers in C/C++ on Wintel Platform
- http://ldesoras.free.fr/doc/articles/rounding_en.pdf -

and ROUND 4 C++ example ex http://cis.stvincent.edu/swd/basic/arithmetic.html

quote

/* Given:   Value      a floating point number
           NumPlaces  a positive integer giving the number of
                      decimal places to which to round the answer.
   Task:    To round Value to NumPlaces decimal places.  If NumPlaces
           is not postive, Value is returned unchanged as the answer.
   Return:  The rounded number in the function name.
*/   
float Round4(float Value, int NumPlaces)
   {
   int k, Temp;
   float Factor;
   
   Factor = 1;
   for (k = 0; k < NumPlaces; k++)
     Factor = Factor * 10;
     
   Temp = Value * Factor + 0.5;
   return Temp / Factor;
   }

unquote

Hope, it helps.

- M.

Quote:


Except, this is using FOR to loop while POW is a function ..., I understand from your previously mentioned:

quote
double MyRound(double Value, int NumPlaces)
{
   double Factor;
   long Temp;
   
   Factor = pow(10.0, NumPlaces);
   Temp = (Value * Factor) + 0.5;
   return Temp / Factor;
}
unquote

that it is talking LONG TEMP while INT TEMP is different ...

- M.

The Pow version was one of the many I tried and I figured it would be the easiest to understand which is why I posted it.

-Ray

Both of us have seen it, and the examples have been compiled and tried.

The problem is not one of unknown technique. The techniques are generally known and implemented. By now, Ray has gathered all of the reliable methods. All of them have specific pros and cons within the context of the application. We are in the process of weighing those pros and cons.

To a large degree, I think we may be dealing as much with Ray's admirable desire for mathmatical absolutes as with the numbers themselves.