FP→FP Floating Point Converter
| Library: | Floating Point Arithmetic |
Behavior
This component converts a floating-point value between independently selected input and output formats. The letters drawn beside the ports identify the format: M for 8-bit minifloat, H for binary16, F for binary32, and D for binary64.
Narrowing conversion rounds to the output precision. A finite value may overflow to infinity or underflow to a subnormal value or zero. The conversion is numeric rather than bit-preserving, so a NaN payload may be normalized even when both sizes are the same.
The ERR output is 1 when the input is NaN and 0 otherwise. An input containing unknown or error bits is converted to NaN. Overflow to infinity does not set ERR.
Pins
- West edge (input, bit width matches Input float size)
- The floating-point value to convert.
- East edge (output, bit width matches Output float size)
- The value represented in the selected output format.
- South edge (output, bit width 1)
- ERR: 1 if the input is NaN; otherwise 0.
Attributes
- Input float size
- The input format: 8, 16, 32, or 64 bits.
- Output float size
- The output format: 8, 16, 32, or 64 bits.
Poke Tool Behavior
None.
Text Tool Behavior
None.