Restart script with extended time

[NastasiaM]

Hi!
I have questions about the restart script.
1)Can I start all this thing from the last checkpoint of the simulation (I think I can, but I don't know exactly how)
2)Is it possible to violate a little bit the walltime limit with this thing? For example, I need the simulation to run for ~120 hours in total, and my limit is 100 hours, can I do it 2*60 hours?

Happy Halloween :)

[ax3l]

Hi,

1. by default, we restart from the most-progressed checkpoint unless otherwise requested via:

--checkpoint.restart.step <N>

https://github.com/ComputationalRadiationPhysics/picongpu/blob/master/docs/TBG_macros.cfg#L208-L214

2. yes just change the walltime in the .cfg file or maybe checkpoint more often and restart several times?

[NastasiaM]

Hi,
I tried to start from the last checkpoint and it started from the beginning. Am I doing smth wrong?
Here is .cfg file.
0004gpus.txt

[ax3l]

Are you sure your simulation already reached step 23555 and wrote at least some checkpoint? What does simOutput/output say and and ls simOutput/checkpoints/*? Otherwise, try adding --checkpoint.restart.step <N> as well to TBG_restart in your .cfg.

[ax3l]

Ah I see it, your TBG_restart should also contain a simple --checkpoint.restart to trigger the restart.

TBG_restart="--checkpoint.restart --checkpoint.restart.backend adios --checkpoint.restart.directory /bigdata/hplsim/external/mukhar40/ColloidalMelting_TF_I=6.3_10ps/simOutput/checkpoints"

[psychocoderHPC]

I will close this issue, I assume it is solved.