libunifloat
===========
A C library for fast high-precision floating-point computations.
ABOUT:
You can use this library instead of a standart libm mathematical library for
fast high-precision calculations. You do not have to change your source code
since it's compatible with libm.
You can use sin(), cos(), tan(), etc. functions from the library to minimize
computation error w/o the change of your source code or rewrite your algorithms
with the help of sin_UF(), cos_UF(), tan_UF(), etc. functions to perform
computations with extended precision.
The library was originally created in 2006-2007 as a part of bachelor's degree
work at MIPT for testing various implementations of the libm library on different
hardware platforms (see AUTHORS).
The library is currently used as part of the official LSB Core test suite for
testing precision and behaviour of math functions.
INSTALL:
mkdir build
cd build
cmake .. -DCMAKE_INSTALL_PREFIX=/usr
make
make install
USAGE:
You have to add two compilation flags to your project:
-fno-builtin do not use built-in math functions
-lunifloat link with libunifloat library and call its functions (sin, cos, ...)
Do not forget to remove -lm compilation flag from your build scripts.