is it possible to decompose a case from an already obtained time step?
 

I am very curious about the possibility of decomposing a case from a determined time step

I ran my case with 30 nodes while after running of several hours, I think 30 nodes may not be the best choice..

So I hope to decompose my case again meanwhile I hope the results already obtained can be saved

Is this possible or I have to configure it on a new case?

Hey there,

What you can do is that you reconstruct your simulation only for the latest time step, then create a new case and initialize your fields with the results of your latest reconstructed time step. The rest is straight forward I guess... (just decompose the case with the new number of processors/nodes).

Thank you einstein-zee, let me try what you said

Another solution, without setting up a new case and initializing your previous results:
1) reconstructPar your simulation as far as you want to keep your results (or -latestTime to only reconstruct the last timestep)
2) change your decomposeParDict settings
3) decomposePar -latestTime
This should decompose your case with new settings while keeping your simulation results and avoiding the hazzle of setting up a new case and mapping the fields.

Thank you @RobertHB this is quite clear