[CIG-SHORT] Restart

Brad Aagaard baagaard at usgs.gov
Sun Sep 16 10:47:36 PDT 2012


Birendra,

There is currently no way to "restart" PyLith. We have not implemented 
checkpointing (which is what you want). This is something we will 
consider more seriously once we start working on earthquake cycle modeling.

Regards,
Brad


On 9/16/12 10:40 AM, Birendra jha wrote:
> Hi
> 1. Is it possible to restart a pylith run from where it stopped? I only
> have displacements at every point (vtk output of the
> problem.formulation.output.domain). Can I use something like following,
> but displacements in place of stress:
>
> [pylithapp.timedependent.materials.upper_crust]
>
> db_initial_stress = spatialdata.spatialdb.SimpleDB
>
> db_initial_stress.iohandler.filename = spatialdb/initial_stress.spatialdb
>
> 2. Is there something like db_initial_displacement for each material region?
> 3. Right now I have displacements for entire domain together. Can I
> get nodal displacements by material region (to be used above for restart)?
> Following doesn't work because
> elasticisotropic3d.outputmatelastic.vertex_data_fields does not exist:
>
> [pylithapp.timedependent.materials.elastic.output]
>
> output_freq = time_step
>
> time_step = 1.0*yr
>
> vertex_data_fields = [displacements]
>
> Any suggestions on how to restart?
>
> Thanks and regards
> Birendra
>



More information about the CIG-SHORT mailing list