I doubt a full port of gsl would be possible since some of it is intrinsically bound to certain precision levels. For example some of the special functions and quadrature routines make use of precomputed parameters/rules. I imagine a subset of the ODE integrators and nonlinear solvers would be amenable to multiple precision rewrites using mpfr.
> ------------------------------ > > Message: 2 > Date: Thu, 7 Jun 2012 16:52:59 +0200 > From: PAU ROLDAN <[email protected]> > To: [email protected] > Subject: [Help-gsl] Multiple precision GSL? > Message-ID: > <CALHR=Et88fkKRa=kpcpxjhrvoqj+vl1mys6zaya8pujmz+h...@mail.gmail.com> > Content-Type: text/plain; charset=ISO-8859-1 > > Hi, > > are you aware of any multiple-precision implementation of the GSL library, > or > something close to it? I have been browsing the web but nothing seems to be > available. > > The MPFR website is listing > > "The MPGSL (Multiple Precision GSL), a collection of routines for > numerical > computing. This is a partial rewrite of the GSL using MPFR, by Marco > Maggi." > > but this project seems to be dead. > > The closest thing I found is the mpmath python library > http://code.google.com/p/mpmath/ > > In particular, I would like to port some of my (double-precision) code using > numerical integration, ODE, and root-finding GSL routines into > multiple-precision.
