Boost::multiprecision docs [1] mentions that the type parameter of number::convert_to method should be "fundamental", pointing to the list of standard types ('A number can be converted to any fundamental (built-in) type)'.
This is different to number::convert_to reference [2], which states that "Type T may be any arithmetic type".
Can [unsigned] __int128 be treated as "fundamental" on compilers/platforms which support it?