FPU: Introduction to Designing a Floating Point Unit based on IEEE 754
Introduction
A Floating Point Unit (FPU) for a compute device is a module that operates on floating point values. It allows a computer to compute non-integer values.
The Main Arithmetic for Floating Point Values include
- Addition $(x + y)$ and Subtraction $(x - y)$
- Multiplication $(x * y)$ and Division $(x / y)$
- Square Root $( \sqrt{x} )$
- Exponents $(x ^ y)$ and Logarithms $ \log_{b}{x} $
- Fused Multiply Add $(x * y + z)$
- Convert From Int $(x_{int} \rightarrow x_{float})$
Format
Sign | Biased Exponent | Significand |