search

MolarVerse / PQ

https://molarverse.github.io/PQ/
GNU General Public License v3.0
5 stars 5 forks source link

Bug report: `qm_loop_time_limit` is not working properly #47

Closed stkroe closed 1 month ago

stkroe commented 1 month ago

Describe the bug

The new feature keyword qm_loop_time_limit does not allow running the simulation properly. If the keyword is set to a value, the simulation is stuck during the first simulation step. No trajectory and even no detailed.out of the DFTB+ calculation is written. After the loop time limit is reached, the calculation will be stopped. Also higher time limit does not make a difference. If the keyword is not set, everything works fine.

To Reproduce

A QMMD simulation was running on our cluster using DFTB+ with the PQ v.0.3.8.0. For reproducibility, an input file is attached:

input-file.zip

Expected behavior

If the keyword is set, the user specifies a loop time limit for QM calculations. If the limit is reached, the calculation will be stopped.

Screenshots

The files after the simulation is stopped due to the time limit: bug_report_qm_loop_time_limit_files The run.out file: bug_report_qm_loop_time_limit_run_out_file The slurm output: bug_report_qm_loop_time_limit_slurm_out_file

97gamjak commented 1 month ago

Hmm... really strange bug. For my test system it worked yesterday evening. On what kind of machine did you run this simulation?

I'll have a look at it later today.

stkroe commented 1 month ago

On the CPU_rune cluster. It seems that also on the local machine it does not work ... very strange. Maby I have a wrong setting.

97gamjak commented 1 month ago

It seems to be fixed now in tag 0.3.8.1 - Could you please try it yourself?

stkroe commented 1 month ago

It tried it. It works now. Thanks.