I’m using the Catalyst V2 interface for insitu visualization. This works great if i only use one MPI process.
On multiple MPI processes, every rank passes its part of the grid to catalyst. E.g. for a structured rectangle [-1,1] x [-1, 1] :
catalyst: channels: example_grid: type: "mesh" data: coordsets: coords: type: uniform dims: i: 100 j: 200 origin: x: -1 y: -1 z: 0 spacing: dx: 0.01 dy: 0.01 dz: 0 topologies: mesh: type: "uniform" coordset: "coords" fields: example_field: association: "element" topology: "mesh" volume_dependent: "false" values: [...]
On rank 1 the same but with:
In that setting ParaView obtains the correct data, but does not consider the parts of the grid as one object. E.g.
RescaleTransferFunctionToDataRange scales the Data Range on each part separately.
What am i missing? I can’t find a way to pass the adjacency information to ParaView. I’ve found
adjsets in the conduit mesh blueprint, but that seems to be ignored by paraview.
I really appreciate your help!