This is a fast C implementation of posit based on Berkeley SoftFloat. Release candidate 0.4.1rc supports posit8_t, posit16_t, posit32_t and new flexible type posit_2_t that enables 2 to 32-bit posits(es=2). It includes support of a C++ wrapper that overrides operators. A python wrapper (SoftPosit-Python) has also been provided.
It contains the complete functionalities from the draft Posit Standard such as add, subtract, divide, multiply, fused-multiply-add, quire fused-dot-product, square root, etc.!