← Back to team overview

yade-dev team mailing list archive

Re: [Bug 729079] Re: Performance optimization of InsertionSortCollider

 

Yes, you are right.

You have put in this code the idea of insertionsortcollider, which is
in the trunk now?

Anton




On Tue, Mar 22, 2011 at 2:05 PM, Sergei D. <sj2001@xxxxxxxxx> wrote:
> 22.03.2011 14:58, Bruno Chareyre wrote:
>>
>> but parallelizing a serial algo the non-intrusive way is
>> exactly what OpenMP is supposed to do
>
> No. OpenMP is supposed to run a loops in a few threads. Not anymore.
> Parallelizing of a _algo_ is a problem of a developers.
> Proof in the attached file. Compile it and run (may be several times) by
>
> g++ -fopenmp -lgomp insertion-sort-test.cpp -o insertion-sort-test
> ./insertion-sort-test
>
> You should see few "insertion sort failed" and negative values on wrong
> places in resulting sorted vector.
>
>
> --
>
> Best regards,
> Sergei D.
>
>
> _______________________________________________
> Mailing list: https://launchpad.net/~yade-dev
> Post to     : yade-dev@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~yade-dev
> More help   : https://help.launchpad.net/ListHelp
>
>



Follow ups

References