[Paraview-developers] Plugin client/server

[Utkarsh Ayachit]

Doh! Yes indeed :). Thanks for catching.

 bounds [0] = bmin[0];
 bounds [2] = bmin[1];
 bounds [4] = bmin[2];
 bounds [1] = bmax[0];
 bounds [3] = bmax[1];
 bounds [5] = bmax[2];

Utkarsh

On Wed, Jul 27, 2011 at 10:45 AM, David Cole <david.cole at kitware.com> wrote:
> I think your indices may be a bit incorrect, but the spirit of the
> code looks right.
>
> Double check all of those array indices: 0,1,2 (not 0,1,3) and 1,3,5
> (not 1,3,4)...
>
>
> On Wed, Jul 27, 2011 at 10:35 AM, Utkarsh Ayachit
> <utkarsh.ayachit at kitware.com> wrote:
>> An alternative to what John suggests:
>>
>> double bounds[6];
>> input->GetBounds(bounds);
>>
>> vtkMultiProcessController* controller =
>> vtkMultiProcessController::GetGlobalController();
>> if (controller && controller->GetNumberOfProcesses() > 1)
>>  {
>>  double bmin[3], bmn[3] = {bounds[0], bounds[2], bounds[4]};
>>  double bmax[3], bmx[3] = {bounds[1], bounds[3], bounds[5]};
>>  controller->AllReduce(bmn, bmin, 3,  vtkCommunicator::MIN_OP);
>>  controller->AllReduce(bmx, bmax, 3, vtkCommunicator::MAX_OP);
>>  bounds [0] = bmin[0];
>>  bounds [2] = bmin[1];
>>  bounds [4] = bmin[3];
>>  bounds [1] = bmax[0];
>>  bounds [3] = bmax[1];
>>  bounds [4] = bmax[3];
>>  }
>>
>> Utkarsh
>>
>> On Wed, Jul 27, 2011 at 1:11 AM, Biddiscombe, John A. <biddisco at cscs.ch> wrote:
>>> I believe that the client does collect boxes from all nodes and generate the
>>> information you need, but on the servers ... if my filters require the
>>> global bounds, I use this.
>>>
>>>
>>>
>>> #ifdef VTK_USE_MPI
>>>
>>>   vtkMPICommunicator *communicator =
>>> vtkMPICommunicator::SafeDownCast(this->Controller->GetCommunicator());
>>>
>>>   MPI_Comm mpiComm = MPI_COMM_NULL;
>>>
>>>   if (communicator) {
>>>
>>>     mpiComm = *(communicator->GetMPIComm()->GetHandle());
>>>
>>>   }
>>>
>>>
>>>
>>> ...
>>>
>>>
>>>
>>>   double bounds[6];
>>>
>>>   input->GetBounds(bounds);
>>>
>>>   double bmin[3], bmn[3] = {bounds[0], bounds[2], bounds[4]};
>>>
>>>   double bmax[3], bmx[3] = {bounds[1], bounds[3], bounds[5]};
>>>
>>>   MPI_Allreduce(bmn, bmin, 3, MPI_DOUBLE, MPI_MIN, mpiComm);
>>>
>>>   MPI_Allreduce(bmx, bmax, 3, MPI_DOUBLE, MPI_MAX, mpiComm);
>>>
>>>
>>>
>>> Cut’n’pasted with some extra code removed. Apologies if I deleted the wrong
>>> lines
>>>
>>>
>>>
>>> JB
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> From: paraview-developers-bounces at paraview.org
>>> [mailto:paraview-developers-bounces at paraview.org] On Behalf Of Scott, W Alan
>>> Sent: 27 July 2011 03:56
>>> To: paraview-developers at paraview.org
>>> Subject: [Paraview-developers] Plugin client/server
>>>
>>>
>>>
>>> I have written a filter that runs client/ multiple remote server.  This
>>> filter works correctly, from what I can see.  (It just replicates a dataset,
>>> in order to create a larger dataset in memory for testing and benching.)
>>>
>>>
>>>
>>> I am trying to convert this filter to a plugin.  It works local server, but
>>> fails client/ remote server.  I believe that my problem is that I am trying
>>> to calculate the geometric bounds of the dataset, and each of the processors
>>> is returning a different bounding box – representing the bounding box of
>>> this server’s data.  What I want is the bounding box of all server’s data.
>>> I am using a vtkDataSet function called GetBounds.
>>>
>>>
>>>
>>> My questions are as follows:
>>>
>>> ·         Is there an example of how to make plugins that work for remote
>>> servers, that calculate the bounds of all of the data, for all processors?
>>>
>>> ·         If not, is there anywhere that I should look for ldeas?
>>>
>>> ·         If not, is there someone I should talk to?
>>>
>>>
>>>
>>> Thanks,
>>>
>>>
>>>
>>> Alan
>>>
>>>
>>>
>>>
>>>
>>> --------------------------------------------------------
>>>
>>> W. Alan Scott
>>>
>>> ParaView Support Manager
>>>
>>>
>>>
>>> GAITS
>>>
>>> Sandia National Laboratories, MS 0822
>>>
>>> Org 9326 - Building 880 A1-C
>>>
>>> (505) 284-0932   FAX (505) 845-0833
>>>
>>> ---------------------------------------------------------
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> 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
>>
>