John Hauser's SoftFloat

[no-zzz-un]

Is this project related to the following c software implementation of ieee 754?

http://www.jhauser.us/arithmetic/SoftFloat.html

[opless]

(From a project watcher)

Not directly. I'm porting jhauser's implementation to C# though, just the 128bit implementation, however.

If I'm successful I may post it to GitHub.

-S

On 12 May 2013, at 15:34, noun [email protected] wrote:

Is this project related to the following c software implementation of ieee 754?

http://www.jhauser.us/arithmetic/SoftFloat.html

— Reply to this email directly or view it on GitHubhttps://github.com/CodesInChaos/SoftFloat/issues/1 .

[no-zzz-un]

Are you only porting quadruple precision because performance considerations are negligible?

Sounds like an interesting project in any case - hope to see it on github eventually.

[opless]

No. The extra precision will be helpful though. Not to mention FPU independence.

On 12 May 2013, at 18:05, noun [email protected] wrote:

Are you only porting quadruple precision because performance considerations are negligible?

Sounds like an interesting project in any case - hope to see it on github eventually.

— Reply to this email directly or view it on GitHubhttps://github.com/CodesInChaos/SoftFloat/issues/1#issuecomment-17781193 .

[CodesInChaos]

No, this is unrelated to John Hauser's SoftFloat. I needed high performance deterministic floats for a now abandoned project, so this project is on hold as well.

Some key differences:

1. This project is quite incomplete
2. While it uses the same format as IEEE floats it doesn't follow IEEE arithmetic exactly.
3. Main goal of this code is performance, correct rounding is less of an issue.

[Piedone]

@opless did you finish with the implementation?

[opless]

Honestly, I can't remember. I do remember moving to a segmented 64 bit solution to my issue.

But that project is on hiatus.

[Piedone]

I see, thanks anyway.