Floating Point Utility

[MathMan]

Hello all,

The ever reappearing questions on floating point numbers, their decimal representation and why converted values do not look as expected inspired me to develop a small utility.

Every finite binary floating point can be converted to an exact finite decimal representation and the tool does exactly that for Single & Double types.

It converts to a fixed point decimal representation, is reasonably fast - in the range of FloatToStr - and supports a small memory model build.

It should run on 32/64 big/little target architectures, but so far is only tested on x86-64 in 64 bit mode.

Beside the conversion unit there is

- a command-line unit test build on consoleapprunner
- a command-line table generation unit (in case the required tables get lost)
- a command-line benchmark utility

The tool targets my own amusement and maybe educational use.

Feedback in any form is appreciated.

Cheers,
MathMan

[schuler]

git repo?

[MathMan]

Hm,

Have been thinking about getting acquaintant with git from time to time. In the past it has always turned out - nah.

But due to your nudge I'll think it over again.

[schuler]

easier to inspect code, submit issues, PRs, look at the readme and even give it a star!