Cellmax is not saved in the restart file if changed by user about sparta

[aklothakis]

When i change the value of maxcellpersurf manually in the grid.cpp file the code runs and restarts fine but when set in the input script it doesn't

[stanmoore1]

I can't seem to reproduce this. Are you sure your global cellmax command is before the read_restart command?

[stanmoore1]

in.sphere_start.txt
in.sphere_restart.txt

Here are the two input files I used to test. If I don't set the cellmax then I get this error:

Max cells overlapping any surf = 228
ERROR: Too many cells overlap one surf - set global cellmax (../grid_surf.cpp:284)

So I used global cellmax 300.

[stanmoore1]

Just to clarify, the global cellmax command isn't stored by the restart file, so it has to be set again, before the read_restart command.

[stanmoore1]

@aklothakis I'm going to close this issue since I can't reproduce it, but please let us know if you have more info on this.

[aklothakis]

Hello there and thank you for your fast response. In the Manual page 321 says that the read_restart command stores global settings. If you check the read_restart.cpp and write_restart.cpp code you can see that indeed the global settings are saved apart from the cellmax setting.

In order to be able to reproduce the issue try restarting the code without the global command (global commands shouldn't be in the restart script because are saved in the restart file) in your restart script. You will see that all the global settings are saved but you will get a message that the cellmax value is not set. But the cellmax value is set in your input script.

This means that the cellmax value is not saved by the write_restart nor read by the read_restart.

[stanmoore1]

OK I see what you are saying now. The surfmax variable is stored in the restart file but cellmax is not. There are probably other global commands not included as well, like mem/limit and particle/reorder.

[aklothakis]

Ok i see, never used the mem/limit or particle/reorder but i will have a look at them too. In the pool request i have changed the appropriate lines in the code to include the cellmax. Tested it for a several timesteps and it seems to be working. I haven't tested it extensively though.

If you have any specific tests in mind let me know and i will perform them.

Does it sound nice to do a list of which global commands are not included?

I have attached the "modified" restart to reproduce the issue.

Thanks again.

in.sphere_restart_v2.txt

[stanmoore1]

Here are the global commands that are missing from the restart file:

• surfs
• surfgrid
• cellmax
• splitmax
• surftally
• surfpush
• particle/reorder
• mem/limit

[aklothakis]

Thank you i will have have a look at them and i will submit a pull request when i have something ready.
Meanwhile tested the cellmax for my case for several thousand timesteps now with 200 cores and is still working.....

[stanmoore1]

Thanks for reporting and working on this. I can help out as well if needed.

[stanmoore1]

Here are the global commands that are missing from the restart file:

• surfs
• surfgrid
• cellmax
• splitmax
• surftally
• surfpush
• particle/reorder
• mem/limit

We are planning to fix this in #303.