Double precision constant for Not a Number.
NAN_DBL is considered unordered, which means all comparisons involving it are false except for not equal (operator !=).
As a consequence, NAN_DBL == NAN_DBL is false but NAN_DBL != NAN_DBL is true.
Additionally, there are multiple bit representations for Not a Number, so if you must test if your value
is NAN_DBL, use isnan().