Numerical Analysis
Machine epsilon refers to a parameter of computational precision. For
floating point numbers,
eps is the distance from 1.0 to the next largest floating point number. For numbers
x smaller than
eps, floating point math yields the perhaps surprising result
1.0 +x = 1.0 (0 < x < eps)
The machine precision can be determined using an
algorithm such as
eps = 1.0;
while (1.0+eps > 1.0)
eps = eps /2;
end
eps = eps *2;
For
IEEE double precision floating point numbers, eps is around 2.2e-16.