This started as a binding of Campary, which is released under GPLv2.
With time it became a port of Campary. It is not production ready but many of the algorithms from the Ph. D. thesis of Valentina Popescu Towards fast and certified multiple-precision libraries are implemented.
An important disclaimer is that I am not related in any form to the original authors so I take the blame for anything not working as it should.
The library implements multiprecision through the use of vectors of Floating point numbers. This is implemented in classical libraries as the double-double library or the qd library.
The idea behind Campary is to use vectors of double, error free transformations and GPU to implement a fast multiprecision library.
It is possible to construct a CamparyFloat by
x = CamparyFloat{4}(1.0)
y = CamparyFloat{4}([1.0, 2^-54, 2^-108, 0])there is a lot of syntactic sugar, so most operations work out of the box.
