AD_LANGEVIN thermostat for QM/MM with fixed_atoms

[Mayank Dodia]

Hi,

I was trying to test AD_LANGEVIN thermostat settings for a QM/MM system with fixed MM atoms. So instead of using REGION MASSIVE for the thermostat settings I am using REGION GLOBAL. When I tested this setting with cp2k regtest files in ../cp2k/tests/QMMM/SE/regtest-force-mixing/ (Lysozyme_small_NVT.inp and tyrosine_NVT.inp), I get the error message below for REGION GLOBAL:

 *******************************************************************************

 *   ___                                                                       *

 *  /   \                                                                      *

 * [ABORT]                                                                     *

 *  \___/   QM/MM QM atoms must be fully contained in the same image of the QM *

 *    |                box - No wrapping of coordinates is allowed!            *

 *  O/|                                                                        *

 * /| |                                                                        *

 * / \                                                        qmmm_force.F:141 *

 *******************************************************************************

which is does appear for REGION MASSIVE. Similarly if I change the settings to THERMOSTAT CSVR and REGION GLOBAL I do not receive this error. I have checked this with cp2k v. 7.0 (git:d15a9c53f) and 8.0 (git:91d9f7f). Does anyone have any ideas on how to resolve this issue or if I am using incorrect settings?

Best Regards,

Mayank 

[Mayank Dodia]

Hello all,

As a test I tried a simple MM water box simulation (files attached here). When the thermostat settings are:

     &THERMOSTAT
       REGION MASSIVE
!       REGION GLOBAL
       TYPE  AD_LANGEVIN
       &AD_LANGEVIN
         TIMECON_LANGEVIN     100
         TIMECON_NH     100
       &END AD_LANGEVIN
     &END THERMOSTAT

(MM.inp)

I can run the simulation. However when I change the REGION MASSIVE to REGION GLOBAL I receive the error:

*******************************************************************************
 *   ___                                                                       *
 *  /   \                                                                      *
 * [ABORT]                                                                     *

 *  \___/            Potential energy is an abnormal value (NaN/Inf).          *
 *    |                                                                        *

 *  O/|                                                                        *
 * /| |                                                                        *

 * / \                                                 force_env_methods.F:353 *
 *******************************************************************************

 ===== Routine Calling Stack =====

            3 velocity_verlet
            2 qs_mol_dyn_low
            1 CP2K

If I use the Nose-Hoover thermostat (MM_v2.inp), the run executed without any errors. So is not possible to set REGION GLOBAL, and hence use fixed atoms with AD_LANGEVIN thermostat?

Best Regards,

Mayank

[Noam Bernstein]

It’s quite possible that AD_LANGEVIN doesn’t handle constrained atoms correctly. I certainly didn’t have them in mind when I coded it up. Can you confirm that it works OK with REGION GLOBAL if there are no fixed atoms?

    Noam

[Mayank Dodia]

Hi Noam,

The original files I attached earlier contained no constraints (fixed atoms or intramolecular).

What I have noticed is that for AD_LANGEVIN thermostat with:

1) REGION GLOBAL: Does not work in QM, MM  or QM/MM system (with or without constraints). For pure QM, no constraints (H2O-32.inp) I get a cholesky decomposition error which is gone once I change the thermostat settings.

2) REGION MASSIVE: Fully functional.

Since we need to use a global thermostat for fixed atoms, so effectively AD_LANGEVIN doesn't seems to work with constraints and/or REGION GLOBAL thermostat setting.

Best,

Mayank

[Noam Bernstein]

GitHub pull request 930 fixes this issue. If a developer could take a look and merge or let me know what else needs to be done, I’d appreciate it.

    Noam

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/1218f75b-78cc-4e63-ae2d-44ab79ae931f%40googlegroups.com.

[Ole Schütt]

I just merged it. Thanks for the patch!

-Ole

>> [1].

>
> --
> You received this message because you are subscribed to the Google
> Groups "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to cp2k+uns...@googlegroups.com.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/cp2k/CAD4xyBJaqbr8h9GX2oxhBBWMg-MT%3Dqtc-x%3Dsg-KP0S59i0cP-Q%40mail.gmail.com
> [2].
>
>
> Links:
> ------
> [1]
> https://groups.google.com/d/msgid/cp2k/1218f75b-78cc-4e63-ae2d-44ab79ae931f%40googlegroups.com?utm_medium=email&utm_source=footer
> [2]
> https://groups.google.com/d/msgid/cp2k/CAD4xyBJaqbr8h9GX2oxhBBWMg-MT%3Dqtc-x%3Dsg-KP0S59i0cP-Q%40mail.gmail.com?utm_medium=email&utm_source=footer