commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Luc Maisonobe (JIRA)" <j...@apache.org>
Subject [jira] [Resolved] (MATH-953) Replaced Double.doubleToLongBits to Double.doubleToRawLongBits where appropriate
Date Mon, 25 Mar 2013 10:23:16 GMT

     [ https://issues.apache.org/jira/browse/MATH-953?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Luc Maisonobe resolved MATH-953.
--------------------------------

       Resolution: Fixed
    Fix Version/s:     (was: 3.1.1)
                   3.2

Fixed in subversion repository as of r1460587.

Patch applied, and in fact all other use of double.ToLongBits have been replaced too. I found
no cases where we did need the NaN bit patterns to be collapsed to a canonical representation,
so it was safe to do.

Thanks for the report and for the patch.
                
> Replaced Double.doubleToLongBits to Double.doubleToRawLongBits where appropriate
> --------------------------------------------------------------------------------
>
>                 Key: MATH-953
>                 URL: https://issues.apache.org/jira/browse/MATH-953
>             Project: Commons Math
>          Issue Type: Improvement
>    Affects Versions: 3.1.1
>            Reporter: Charles Cooper
>            Priority: Minor
>              Labels: optimization, patch
>             Fix For: 3.2
>
>         Attachments: patch
>
>   Original Estimate: 0h
>  Remaining Estimate: 0h
>
> Replaced calls to Double.doubleToLongBits with Double.doubleToRawLongBits in FastMath
where appropriate (i.e. I decided it was safe and would result in a speedup). Some methods
might be significantly faster, like getExponent which I measured to be about 40% faster (do
your own tests!). Patch is attached.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message