Error reading coordinates from restart file in QM/MM force mixing calculations
157 views
Skip to first unread message
Pedro Maximiano
Nov 29, 2021, 9:31:59 AM11/29/21
to cp2k
Dear CP2K users and developers
I'm trying to do QM/MM simulations of a condensation reaction between two orthosilicic acid molecules, using the force mixing method to model the first solvation layer at the QM level.
I have noticed that, when starting the QM/MM simulations from a restart file of a previous fully-MM run or when restarting the QM/MM simulation itself, the coordinates of the restart file are not read (even though RESTART_POS .TRUE. is activated in the &EXT_RESTART section), yielding the following error message:
*******************************************************************************
* ___ *
* / \ *
* [ABORT] *
* \___/ No atomic coordinates have been found! *
* | *
* O/| *
* /| | *
* / \ topology.F:595 *
*******************************************************************************
This issue happens only if force mixing is enabled. However, if I copy the coordinates from the &COORD section of the restart file into an external file and then load that file in the section &FORCE_EVAL / &SUBSYS / &TOPOLOGY the simulation runs normally.
On the other hand, if I load the same coordinates as a pdb file and not a coordinate file in CP2K format, the simulation runs but the atoms appear almost frozen in the trajectory. It is not just a visualization artifact, since the fluctuations in temperature and energies change completely and the RDF of water oxygen atoms around the Si atoms is completely wrong. This is true even when the QM/MM simulation is restarted in this way.
The calculations were done with the xTB model for the QM region, although the same issues occurred with PM6 and PM6-FM. The coordinates issue also occurred when starting a DFT GPW simulation from a restart file. I have tried both the 7.1 and 8.2 versions of the code.
Are these issues related to a bug or am I missing something?
Best regards
Pedro Maximiano
I'm trying to do QM/MM simulations of a condensation reaction between two orthosilicic acid molecules, using the force mixing method to model the first solvation layer at the QM level.
I have noticed that, when starting the QM/MM simulations from a restart file of a previous fully-MM run or when restarting the QM/MM simulation itself, the coordinates of the restart file are not read (even though RESTART_POS .TRUE. is activated in the &EXT_RESTART section), yielding the following error message:
*******************************************************************************
* ___ *
* / \ *
* [ABORT] *
* \___/ No atomic coordinates have been found! *
* | *
* O/| *
* /| | *
* / \ topology.F:595 *
*******************************************************************************
This issue happens only if force mixing is enabled. However, if I copy the coordinates from the &COORD section of the restart file into an external file and then load that file in the section &FORCE_EVAL / &SUBSYS / &TOPOLOGY the simulation runs normally.
On the other hand, if I load the same coordinates as a pdb file and not a coordinate file in CP2K format, the simulation runs but the atoms appear almost frozen in the trajectory. It is not just a visualization artifact, since the fluctuations in temperature and energies change completely and the RDF of water oxygen atoms around the Si atoms is completely wrong. This is true even when the QM/MM simulation is restarted in this way.
The calculations were done with the xTB model for the QM region, although the same issues occurred with PM6 and PM6-FM. The coordinates issue also occurred when starting a DFT GPW simulation from a restart file. I have tried both the 7.1 and 8.2 versions of the code.
Are these issues related to a bug or am I missing something?
Best regards
Pedro Maximiano
Reply all
Reply to author
Forward
0 new messages