#define signbit(x) (sizeof(x) == sizeof(float) ? __float32_signbit(x) : __float64_signbit(x))
signbit macro determines whether the sign of x is negative. signbit returns a non-zero value if and only if x is negative.