M-Hwcosim uses fixed point data types internally, while it consumes and produces double precision floating point values to external entities. All data samples passing through a port are fixed point numbers. Each sample has a preset data width and an implicit binary point position that are fixed at the compilation time. Data conversions (from double precision to fixed point) happen on the boundary of M-Hwcosim. In the current implementation, quantization of the input data is handled by rounding, and overflow is handled by saturation.