www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - float has too much precision

reply Faux Amis <faux amis.com> writes:
I'm dumbfounded, why does the following code write '35' on DMD32 D 
Compiler v2.091.0-dirty?

module magic;

float magic( float f )
{
     return f + 35f - f;
}

void main()
{
     import std.stdio;
     writeln( magic(1_000_000_000f) );
}
Apr 21 2020
parent reply Steven Schveighoffer <schveiguy gmail.com> writes:
On 4/21/20 3:47 PM, Faux Amis wrote:

 I'm dumbfounded, why does the following code write '35' on DMD32 D 
 Compiler v2.091.0-dirty?
 
 module magic;
 
 float magic( float f )
 {
      return f + 35f - f;
 }
 
 void main()
 {
      import std.stdio;
      writeln( magic(1_000_000_000f) );
 }


On run.dlang.io, it prints 64. Also on my mac.

Possibly it's working because intermediate floating point calculations 
are generally done at max precision. On your system, that might be 
80-bit reals.

Also possible that some optimization is figuring out that it can just 
return 35f?

Try instead:

float magic( float f)
{
    float result = f + 35f;
    return result - f;
}

Is it worth worrying about? floating point is supposed to be inexact and 
subject to variance on different machines.

-Steve
Apr 21 2020
parent Faux Amis <faux amis.com> writes:
On 2020-04-21 22:10, Steven Schveighoffer wrote:

 On 4/21/20 3:47 PM, Faux Amis wrote:

 I'm dumbfounded, why does the following code write '35' on DMD32 D 
 Compiler v2.091.0-dirty?

 module magic;

 float magic( float f )
 {
      return f + 35f - f;
 }

 void main()
 {
      import std.stdio;
      writeln( magic(1_000_000_000f) );
 }


 
 On run.dlang.io, it prints 64. Also on my mac.
 
 Possibly it's working because intermediate floating point calculations 
 are generally done at max precision. On your system, that might be 
 80-bit reals.
 
 Also possible that some optimization is figuring out that it can just 
 return 35f?
 
 Try instead:
 
 float magic( float f)
 {
     float result = f + 35f;
     return result - f;
 }
 
 Is it worth worrying about? floating point is supposed to be inexact and 
 subject to variance on different machines.
 
 -Steve


No, it doesn't matter. I just wanted to understand why it happened.
It seems splitting it up did the trick. So intermediate it is.

Thanks!
Apr 21 2020