RFR 8014076: Arrays parallel and serial sorting improvements

Chris Hegarty chris.hegarty at oracle.com
Wed May 8 08:57:44 UTC 2013


Doug,

David raises a good question here. Is this implementation detail still 
correct:
   "The algorithm requires a working space equal to the size of the
original array."

I did get clarifications on a number of the minor spec/implementation 
details, but I don't think we discuss this one.

  (I believe lambda's version of Arrays was not sync'ed with TL after 
integration of parallelSort, and before Doug did his work. Or at least 
there was some confusion about the latest version of this file).

-Chris.

On 08/05/2013 01:36, David Holmes wrote:
> Hi Chris,
>
> Didn't Doug's changes also remove the constraint:
>
> "The algorithm requires a working space equal to the size of the
> original array."
>
> ?
>
> David
>
> On 8/05/2013 12:51 AM, Chris Hegarty wrote:
>> Doug has made some updates to the java.util.Arrays sorting code to
>> provide stable sorting. There have also been some changes to the
>> original Parallel Array Sorting ( MIN_ARRAY_SORT_GRAN is public again ).
>>
>> Right now a copy of this work is sitting in the lambda repo. This issue
>> proposed to integrate this work into jdk8.
>>
>> http://cr.openjdk.java.net/~chegar/8014076/ver.00/specdiff/java/util/Arrays.html
>>
>>
>> http://cr.openjdk.java.net/~chegar/8014076/ver.00/webrev/
>>
>> Thanks,
>> -Chris.



More information about the core-libs-dev mailing list