In C++ there are (sometimes) subnormal doubles that are finite, not nan
and not zero.
How can I round them to nearest non subnormal double representation?
For example: std::isnormal(DBL_MIN/2.0) = false
In C++ there are (sometimes) subnormal doubles that are finite, not nan
and not zero.
How can I round them to nearest non subnormal double representation?
For example: std::isnormal(DBL_MIN/2.0) = false