The float representations of a constant and a variable are different?

Question

I executed the following C++ code and get the answer "Not the same? Weird.....". Originally I expected to get "Of course, they are the same!!!". I was using Dev-C++ 5.11.

#include <iostream>
using namespace std;

int main() {
    float p;
    p = 0.3;
    if ((p-0.3)==0)
        cout << "Of course, they are the same !!!\n";
    else
        cout << "Not the same? Weird.....\n";
    return 0;
}

Note that 0.3 is not exactly representable in neither single nor double precision IEEE 754 format. You can try it, for example, here: https://baseconvert.com/ieee-754-floating-point. — Daniel Langr, Jun 17 '22 at 10:15

score 1 · Answer 1 · answered Jun 17 '22 at 16:40

1

As Mat points out, the problem here is that 0.3 is a double. And (double)(float)0.3 != 0.3. You lose precision when casting 0.3 to float. Casting it back does not magically undo the rounding.

answered Jun 17 '22 at 16:40

MSalters

173,980
10
155
350

1

And that's why this question isn't really a duplicate of the one it was closed against. – Mark Ransom Jun 17 '22 at 17:47

The float representations of a constant and a variable are different?

1 Answers1