Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Printing Floating-Point Numbers A Faster, Always Correct Method #27

Closed
lygstate opened this issue Feb 3, 2016 · 6 comments
Closed

Printing Floating-Point Numbers A Faster, Always Correct Method #27

lygstate opened this issue Feb 3, 2016 · 6 comments

Comments

@lygstate
Copy link

lygstate commented Feb 3, 2016

This is a faster and complete way to do double conversation.

http://cseweb.ucsd.edu/~lerner/papers/fp-printing-popl16.pdf

@floitschG
Copy link
Contributor

Apparently they didn't measure the right configuration of the double-conversion library. (Which they confirmed in a private email). My current understanding is, that grisu is still 1.7 times faster.

They promised to update their talk, PDF and web presence.

@lygstate
Copy link
Author

lygstate commented Feb 4, 2016

@floitschG If that's was the case, then we could replace Dragon4 with Errol3

@floitschG
Copy link
Contributor

We could, but I'm not sure it's worth the effort: dragon is only hit for very few cases, and the code is well tested.
In any case: I would like to see the new measurements first.

@kientzle
Copy link

Another option worth considering:

The Swift standard library is now using a fast algorithm that combines features of Grisu2 and Errol4. The C implementation includes detailed comments. The original pull request has some additional discussion and explanation.

@lygstate
Copy link
Author

sounds good

@floitsch
Copy link
Collaborator

floitsch commented Sep 8, 2018

Closing this issue. Ryu is superseding both Errol as well as Grisu.

@floitsch floitsch closed this as completed Sep 8, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants