VC++ stream floating point precision error

[Val Samko]

=========== #include int main() { double z = 20.0; std::cout<

[Joseph Dempsey]

a double is only a 64bit value. You can't get a 128 bit number out of that. You're going to have to find some other way of getting your precision. Joseph Dempsey joseph_r_dempsey@yahoo.com "Software Engineering is a race between the programmers, trying to make bigger and better fool-proof software, and the universe trying to make bigger fools. So far the Universe in winning." --anonymous

[Val Samko]

double on Win32 is of course 64bit. long double on SGI and on Sun is 128bit. The question is - why does VC++ produce such a weird output? I am not trying to get 128bit number out of Win32 double, I'm just trying to output it with 50 digits after the decimal point. And when I try to do that, I get completely wrong result. Borland C++ and g++ produce correct results on Win32.

[David Crow]

Just a guess but I think it may be outputting the least number of digits required for significance.

[Val Samko]

The problem is quite different, 20 and 2000000000000000 are just very different numbers, it is not the same nubmer with a different number of digits after the decimal point .. there is no decimal point :)

[David Crow]

Perhaps you'd want this instead:

void main( void )
{
double z = 20.0123;

for (int x = 1; x < 40; x++)
    cout << setprecision(x) << z << endl;

}

[Tim Smith]

It is a bug in Microsoft's STL when the precision is greater than _MAX_SIG_DIG and the formatted number doesn't contain any decimal point. Send the bug off to Microsoft and/or Dinkumware http://www.dinkumware.com/[^]. Tim Smith I'm going to patent thought. I have yet to see any prior art.

[Val Samko]

I thought it is a bug, weird one. Thanks for the explanation.