Implemented according to the IEEE standard 754 for normalized double-precision floating-point numbers. The IP is optimized for performance and HW resources usage and it supports the following operations:
- Arithmetic: Add, Subtract, Multiply, Divide
- Comparison: Greater or equal, Smaller or equal, Greater, Smaller, Equal, Not equal
- Conversion: Integer to double float, Unsigned integer to double float, Double float to integer, Double float to unsigned integer
The supported rounding mode is round-to-nearest-even
Dedicated output flag interface to highlight the occurrence of the following events:
- An input operand is Not a Number
- An input operand is ± infinite
- The result is ± Underflow
- The result is ± Overflow
- The result is Not a Number
- The result is Zero
- The result is positive or negative.
- A division by zero
It has a simple user interface and can be used also as a HW accelerator for processor’s custom instructions
For flexibility, all the operations are configurable to be active or not within the IP.