I revised the header file gmpxx.h to replace the gmp floating point class (mpf_class) with an alternative class (mpfr_class) in which all computations are done with correct rounding. I implemented all the necessary routines to allow the new mpfr_class to mix freely with gmp integers (mpz_class) and rational numbers (mpq_class) in algebraic expressions, and to output them via C++ iostreams. I also added all the functions (sin, cos, exp, log, etc.) that are part of mpfr but not gmp. As in gmpxx.h, algebraic expressions are represented using inlined templates which resolve to direct function calls to the appropriate mpfr or gmp function without creating copies or temporary variables whenever possible.

Last updated 2008/11/16 to incorporate new features of MPFR 2.3.*

gmpfrxx.zip contains the following files:

• README
• COPYING
• example.cpp
• gmpfrxx.h
• gmpfrxx.cpp
• mpfr_mul_d.h
• mpfr_mul_d.c
• Makefile

previous release of gmpfrxx (using mpfr-2.2.1):   gmpfrxx-2007-08-14.zip