boost::multiprecision::uint128_t sizeof is 16

Question

Per this link,

boost::multiprecision::uint128_t sizeof is 24

uint128_t should be 24.

I tried it on Linux Ubuntu 16.04 x64, and it says 16.

Any reason?

Answer 1

cpp_int_base uses some auxiliary variables

   data_type   m_data;

   unsigned    m_limbs;
   bool        m_sign, m_internal;