Re: [Paraview] [Non-DoD Source] Re: Using 64 or more cores on a single node when running client/server (UNCLASSIFIED)

CLASSIFICATION: UNCLASSIFIED A little bit more background information, we were trying to run it on Wai (large SMP machine) and we ran into issue getting more than 64 core for our setup. Let’s see if a simple MPI hello world will show where the limit is coming from. thanks -simon

Re: [Paraview] [Non-DoD Source] Re: Using 64 or more cores on a single node when running client/server (UNCLASSIFIED)

CLASSIFICATION: UNCLASSIFIED thank you Utkarsh. that's the confirmation that we are looking. We will test the MPI. thanks -simon -Original Message- From: ParaView [mailto:paraview-boun...@paraview.org] On Behalf Of Utkarsh Ayachit Sent: Wednesday, September 13, 2017 12:12 PM To:

Re: [Paraview] [EXT] Re: Select points based on coordinates

Yifei, Glad that helped. If memory becomes critical, you could write a Programmable Filter that created array Loc by shallow copying array coords – this would sharply reduce the memory footprint of the Loc array. Next time I’ll ask you for help! Dennis From: Yifei Ma

Re: [Paraview] [EXT] Re: Select points based on coordinates

Hi Dennis, Calculator filter did the trick. I created Loc using vector array coords. Then I have no problem finding the points using Find Data by specifying certain Loc. Thanks for the help. Best, Yifei On Wed, Sep 13, 2017 at 10:03 AM, Dennis Conklin < dennis_conk...@goodyear.com> wrote: >

Re: [Paraview] [EXT] Re: Select points based on coordinates

Yifei, I don’t know anything about vtu files – I don’t see any array of “Position”. However, perhaps you can use vector array coords – this is Paraviews internal array of point positions. You could use this to create Loc after you read your file into Paraview – your filesize would not

Re: [Paraview] Using 64 or more cores on a single node when running client/server

If you're running on the DSRCs I'm guessing that the limit is coming from the node selection option (i.e. "#PBS -l select=8:ncpus=36:mpiprocs=8"). mpiprocs here is the number of MPI processes per node while ncpus is the number of cores per node (in this case only nodes that have 36 cores per node

Re: [Paraview] Select points based on coordinates

Hi Dennis, Sorry that I wasn't clear about the data file. The "Loc" DataArray is not included in the original data. I simply copy the "Position" DataArray and re-named it "Loc" in order to find data based on its location. This is a workaround and it works just fine for small data files. However,

Re: [Paraview] Catalyst distributed graph example

vtkMutableDirectedGraph has an API that is hopefully pretty intuitive with AddVertex, AddEdge, etc. You can also shallow-copy from the mutable graph into a DAG object, which performs a check, if it is important to verify you

Re: [Paraview] [EXTERNAL] Select points based on coordinates

On Tue, Sep 12, 2017 at 6:10 PM, Scott, W Alan wrote: > This is on the todo list. > https://gitlab.kitware.com/paraview/paraview/issues/15530 A fix is available in master (https://gitlab.kitware.com/paraview/paraview/merge_requests/1724).

Re: [Paraview] Using 64 or more cores on a single node when running client/server

> It should be related to your mpi environment: may be oversubscribing (more > than one mpi process by core) is not the default behavior. I am tempted to second that. There's nothing in ParaView that checks how many cores your node has, as a result if there's a cap, it's coming from the MPI

Re: [Paraview] Catalyst distributed graph example

Hi Matthieu, There isn't a Catalyst example for graphs but it shouldn't be too hard to make. I'm not very familiar with the vtkGraph and derived classes but if I were to to try doing that I'd just see how the graph object is generated in some filter that generates one (look in the RequestData()

[Paraview] Catalyst distributed graph example

Hi, I would like to use Catalyst to visualize a distributed directed graph. Each node of the graph is characterized by an id (not necessarily contiguous, as some nodes are added and removed during the execution) and a list of connections (ids of "friends"). The nodes can be assigned

Re: [Paraview] Using 64 or more cores on a single node when running client/server

Yes, it works. Paraview 5.4 with OpenMPI 1.7.3 konsole─┬─bash───mpirun───80*[pvserver] ├─bash └─{konsole} Node: 20 cores with hyperthreading. It should be related to your mpi environment: may be oversubscribing (more than one mpi process by core) is not the default behavior.

Re: [Paraview] Two issues with time-dependent data

Expanding issue 1: I have created the simple programmable source below, which creates a periodic animation to demonstrate the problem of double display of time steps in periodic data. It is a line that moves according to a sine-function with 3 as cycle duration. When the animation ist set to

Re: [Paraview] Select points based on coordinates

Yifei, I don't understand the issue. I DL'ed your file and proceeded as follows Edit/Settings/General/Auto Convert Properties checked (On) Find Data Find POINTS in test_data.vtu Loc(0) is >= 0.009 Run Selection Query Extract Selection This correctly extracted the 1 point met that requirement

[Paraview] Paraview MIP viewer

Hi everyone, I am struggling with the paraview MIP viewer (Maximum Intensity Projection Viewer) right now. I know the basic settings in paraview doesn’t have the mip viewer designed in it, so I compiled it from the source code, and added the blend mode in the volume rendering, to allow the