[Paraview-developers] MPI Parallel Sorting

[Biddiscombe, John A.]

Utkarsh

I look forward to seeing the details. I would like to do some parallel sort stuff on particle data to enable ghost particles to be divvied up nicely. Anything that can be reused outside of the spreadsheet sorting would be very useful.

JB

-----Original Message-----
From: Utkarsh Ayachit [mailto:utkarsh.ayachit at kitware.com] 
Sent: 15 September 2010 19:18
To: Biddiscombe, John A.
Cc: Moreland, Kenneth; ParaView Developers; Greenfield, John A; Thompson, David C
Subject: Re: [Paraview-developers] MPI Parallel Sorting

That sounds interesting too. BTW, sorting in spreadsheet was recently
committed to git-master by Sebastien based on the "Parallel Merge Sort
with Load Balancing" paper. He will be posting the details soon.

Utkarsh

On Wed, Sep 15, 2010 at 4:27 AM, Biddiscombe, John A. <biddisco at cscs.ch> wrote:
> This one looks interesting
> http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.86.8588
>
>
> comes with code.
>
> A Novel Parallel Sorting Algorithm for Contemporary Architectures
> http://gauss.cs.ucsb.edu/~viral/PAPERS/psort/html/psort.html
>
> JB
>
> -----Original Message-----
> From: paraview-developers-bounces at paraview.org [mailto:paraview-developers-bounces at paraview.org] On Behalf Of Utkarsh Ayachit
> Sent: 24 August 2010 15:01
> To: Moreland, Kenneth
> Cc: ParaView Developers; Greenfield, John A; Thompson, David C
> Subject: Re: [Paraview-developers] MPI Parallel Sorting
>
> Actually, Joe Hennessey from JHU sent me a reference:
> http://portal.acm.org/citation.cfm?id=608798
>
> We are taking a look at this one.  It seems to do what we want.
>
> Utkarsh
>
> On Mon, Aug 23, 2010 at 7:34 PM, Moreland, Kenneth <kmorel at sandia.gov> wrote:
>> No it does not.  The spreadsheet on the client carefully loads only the
>> portion of the data that is visible.  If you were to pull all the data to
>> the client, you would run out of memory and kill it.
>>
>> In short, that won't work.
>>
>> -Ken
>>
>>
>> On 8/23/10 5:16 PM, "Scott, W Alan" <wascott at sandia.gov> wrote:
>>
>> Greenfield has a really good point - doesn't the data for the whole
>> spreadsheet have to be passed to the client for display?  Thus, couldn't you
>> just sort on the client?
>>
>> Alan
>>
>> -----Original Message-----
>> From: paraview-developers-bounces at paraview.org
>> [mailto:paraview-developers-bounces at paraview.org] On Behalf Of Utkarsh
>> Ayachit
>> Sent: Monday, August 09, 2010 8:48 AM
>> To: ParaView Developers; Thompson, David C
>> Subject: [Paraview-developers] MPI Parallel Sorting
>>
>> Folks,
>>
>> I am looking at the long standing feature request to add support for
>> sorting in the spreadsheet view. Our use case is that we have a large
>> array is distributed among N processes.
>> We want to end up with that array sorted but still distributed among
>> all processes. Most implementations I am looking at end up creating
>> one large sorted array on one of nodes. Does anyone know any approach
>> that avoids this generation of a large array on the root node?
>>
>> Utkarsh
>> _______________________________________________
>> Paraview-developers mailing list
>> Paraview-developers at paraview.org
>> http://public.kitware.com/mailman/listinfo/paraview-developers
>>
>>
>> _______________________________________________
>> Paraview-developers mailing list
>> Paraview-developers at paraview.org
>> http://public.kitware.com/mailman/listinfo/paraview-developers
>>
>>
>>
>>
>>    ****      Kenneth Moreland
>>     ***      Sandia National Laboratories
>> ***********
>> *** *** ***  email: kmorel at sandia.gov
>> **  ***  **  phone: (505) 844-8919
>>     ***      web:   http://www.cs.unm.edu/~kmorel
>>
>>
> _______________________________________________
> Paraview-developers mailing list
> Paraview-developers at paraview.org
> http://public.kitware.com/mailman/listinfo/paraview-developers
>