Welcome to GNSTLIB’s documentation!¶
GNSTLIB is a numerical library written in C++11 for fast and accurate computation of special functions in double precision floating-point arithmetic. GNSTLIB can be used as stand-alone C++ library and provides wrappers for the major programming languages used in scientific computing, such as Fortran, C and Python. GNSTLIB will provide support to R and Julia in future releases.
GNSTLIB implements vectorized versions of all special functions. These routines include an option to trigger OpenMP for parallelization, thus taking advantage of multi-core processors. Furthermore, several algorithms for multi-evaluation of univariate special functions via analytic continuation in parallel are implemented. These routines are suitable for the evaluation of large samples with values within a small range.
GNSTLIB is a work in progress. Future releases will expand the available set of routines with univariate and multivariate statistical distributions.