CODEWITHSUNDEEP
c++doubledecimal
slozano95
slozano95

Reputation: 278

Float addition not including decimal c++

Im trying to do an elipsoid and to do so I need the following operation

float result = this->eValues[0]+this->eValues[2]-this->eValues[1]

in this case 

eValues[0]=-1.86265e-09;
eValues[1]=8.999999;
eValues[2]=8.999999;

It supposed it should return -1.86265e-09 but the result is 0

Thanks

Upvotes: 0

Views: 188

Answers (1)

asimes
asimes

Reputation: 5894

This is a problem of floating point precision, it is made more clear by the order that the additions is carried out:

#include <iostream>

int main() {
        float a = -1.86265e-09;
        float b = 8.999999;
        float c = 8.999999;
        std::cout << ((a + b) - c) << std::endl; // 0
        std::cout << (a + (b - c)) << std::endl; // -1.86265e-09
        return 0;
}

The reason why the first output produces 0 is because -1.86265e-09 is very small relative to 8.999999. The difference in the exponent of how floating point numbers is stored is too great for the float to produce the correct arithmetic answer

Upvotes: 7

Related Questions