libqd0v5 2.3.20+dfsg.1-1build1 (i386 binary) in ubuntu cosmic

 This package supports both a double-double datatype (approx. 32
 decimal digits) and a quad-double datatype (approx. 64 decimal
 digits). The computational library is written in C++. Both C++ and
 Fortran-90 high-level language interfaces are provided to permit one
 to convert an existing C++ or Fortran-90 program to use the library
 with only minor changes to the source code.