Expand description
Software floating point types.
Manipulating floats in any kind of tool is very complicated. Parsing can be tricky, and floating point hardware is at best untrustworthy from a cross-target reproducibility perspective. This module provides helpers to make those operations just a little bit easier.
Structs§
- A “brain float”; an IEEE547 32-bit floating point value with a truncated mantissa.
- Returned by some
Digital
functions ifilex
does not have compiled support for a particular parsing operation; ifExotic
is returned, no diagnostics will be emitted (although it almost certainly indicates a bug). - An IEEE547-formatted 16-bit floating point value.
- An IEEE547-formatted 32-bit floating point value.
- An IEEE547-formatted 64-bit floating point value.
- An IEEE547-formatted 128-bit floating point value.
- An error returned by a soft float type’s parsing functions.
Enums§
- A category of floating point value, returned by e.g.
Fp64::classify()
.