[Paraview] The use of paraview's parallel

[David E DeMarle]

To test if paraview is using the entire set of nodes:

* connect the client to the pvserver mpi job,

* open up views->Memory inspector.
You should then see as many processors as their are on the system, plus the
client.

Next create a Sources->Sphere and apply the Filters->Process Id scalars
filter. There should be as many colors on the sphere as there are
processors in the server. Increase the resolution on the sphere source if
you have more than 8 nodes.

* next open up your data file and use the the memory inspector and process
id scalars filters to see how the reader chose to distribute the data. If
it is all loaded on process 1, apply the D3 filter to redistribute it
amongst all ranks.

hope that helps



David E DeMarle
Kitware, Inc.
R&D Engineer
21 Corporate Drive
Clifton Park, NY 12065-8662
Phone: 518-881-4909

On Wed, Nov 5, 2014 at 10:20 PM, 周恒众 <zhz1993622 at 163.com> wrote:

>         I'm a fresh man in paraview.I use the paraview by the way of compiling.I
> want to use paraview to realize the rendering of large data .
>         But now I find a problem .when I use paraview in paralleling ,it
> seems that  the first node's pressure is very big .It should use more than
> the size of the data ,but the others node only use a little memory .So in
> this way ,it dosen't share the datas to all the node,I use many computers
> ,but it can't mange large datas .The first node should read all of the
> large data.So I suspect that what the use of  paraview's parallel .
>         I read some news of your paraview ,you say that it can deal with
> large datas .I want to konw how to do it .I have compiled the paraview's
> source code with cmake,mpi ,qt,python cuda .And I have used the mpi to
> realise the parallel.In the client of mpi ,I can see many nodes , and I
> use the nfs to share files ,all the nodes can read files in a computer.Then
> i want to konw what the problems of my configuration .Who can help me
> .Thank you.
>
>
>
>                                                   zhou
>
>
>
>
> _______________________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Please keep messages on-topic and check the ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://public.kitware.com/mailman/listinfo/paraview
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview/attachments/20141106/451c37d9/attachment.html>