Fixed Point C++ support designed for x86 and x86_64 architectures. The goal is to enable approximate computing by applying reduced precision computation.
fixed_point_t<I,F>
can be used as data type to represent a number with
I
bits for the integer part (including sign)F
bits for the fractional part
I
and F
must be greater than zero
I + F
must be a valid size for a standard integer data type
{ 8, 16, 32, 64, 128}