Skip to content

Equidistant sampling of uniform FP values #1270

Description

@cmpute

I just noticed a paper recently, which reports that the implementation of drawing uniform floating point numbers in rand has some imperfections. Specifically, according to the paper rng.gen_range(-f64::MAX..=f64::MAX) will only generate negative floating point numbers.

After a quick search over the issues I didn't find any issued raised with regard to this paper. I just want to post it in case it's not mentioned here.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    A-newPropose usage of a new algorithm

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions