AKMC saddle search error

[pshinyeong]

Hello,

I found an error while following through this tutorial : https://docs.quantumatk.com/tutorials/boron_diffusion_in_si/boron_diffusion_in_si.html

I attached a python file which was used for the Adaptive Kinetic Monte Carlo (AKMC) calculation and the output file which is the log file. The saddle search stopped after 69th saddle search and apparently, there were no new state discovered after this.

This is the error message shown in the log file:

  Traceback (most recent call last):
    File "zipdir/NL/ComputerScienceUtilities/ParallelTools/DynamicTaskScheduler.py", line 427, in __runInParallelModeAsDelegator
    File "zipdir/NL/Dynamics/AdaptiveKineticMonteCarlo.py", line 1313, in schedulerCode
    File "zipdir/NL/Dynamics/AdaptiveKineticMonteCarlo.py", line 840, in search
    File "zipdir/NL/ComputerScienceUtilities/ParallelTools/DynamicTaskScheduler.py", line 793, in waitAnyTaskFinished
    File "zipdir/NL/ComputerScienceUtilities/ParallelTools/DynamicTaskScheduler.py", line 1120, in __monitorTasks
    File "Request.pyx", line 47, in mpi4py.MPI.Request.Test (src/mpi4py.MPI.c:50413)
  mpi4py.MPI.Exception: Unknown error class, error stack:
  PMPI_Test(189)........................: MPI_Test(request=0x1529336f4f70, flag=0x7ffca46265c8, status=0x1) failed
  MPIR_Test_impl(67)....................: fail failed
  MPIDU_Complete_posted_with_error(1710): Process failed

Can anyone guide me through this error and explain why this error message showed up? Thank you!

[Anders Blom]

If you are trying to learn the AKMC method, I strongly recommend the other tutorial https://docs.quantumatk.com/tutorials/akmc_pt_on_pt_100_surface/akmc_pt_on_pt_100_surface.html which uses forcefield and thus run a lot faster. The B in Si is at best a demo case, not fully developed, and extremely time-consuming.