A C++ fixed point math library providing a Number class that can be used by all arithmetic and relational operators in place of using the floating point data types (float, double or long double). This is to provide the exact precision required by financial applications.

Linux x86 64bit

gcc 4.8.1

Required compile flags -std=gnu++11.

make

output will be build/libfixed.a

Include the file include/fixed/Number.h

Simply instantiate Numbers and use them in normal arithmetic and relational expressions.

• Number (40) Produces a Number with a value of 40.
• Number (-5, 2, 3) Produces a Number with the a value of -5.002.
• Number (5, 2, 3, Number::Sign::NEGATIVE) Produces a Number with the value of -5.002.
• Number (0, 2, 2, Number::Sign::NEGATIVE) Produces a Number with the value of -0.02.

Can also instantiate with a string, though this is more expensive than the integer based constructor.

• Number ("1.0234") produces a number with a value of 1.0234

Also there are constructors that accept floating point values, but these are less accurate than the integer based versions. These are provided for completeness and convenience, though the integer based constructor is encouraged.

For the full details see include/fixed/Number.h