问题
#include <stdio.h>
#include <stdlib.h>
#define answer 3.141593
void main(int argc, char **argv) {
float a = (argc - 2)?: strtod(argv[1], 0);
printf("double = %lf ,float = %f", a-answer , a-answer);
}
when I run it like that:
./a.out 3.141593
the output is
double = -0.000000 ,float = -0.000000
Why does it -0.00000 ? how can I make it output 0.000000 ?
How can I make a == answer ?
How comes there is -0 value if it uses 2's complement?
回答1:
Floating point numbers doesn't use 2's complement. They have sign, exponent and mantissa and your numbers have just zero with sign, or more probably, you have some number like -1.0e-15, which is printed as -0.0000. try %e instead of %f. The small difference is made by inability to store numbers with infinte precission in finite precission data type (some rounding occured) and when you change double to float, additional rounding have to occur. (take in mind, that 3.141593 is infinite periodic number in binary representation, this it really depends on in which type is this number stored)
来源：https://stackoverflow.com/questions/15175218/what-is-0-0000-in-c-when-using-floats-and-double