[gmx-users] MTTK with double precision GROMACS

[Michael Shirts]

MTTK with constraints has a tendency to fail with numerical stability
issues.  It is perfectly good as long as it doesn't fail.  We will
retire it in 5.1 or so.

In the meantime, things that can reduce failures due to numerical instability.

- Apply constraints only to h-bonds
- make tau_p longer (5 ps).
- increase the tolerance on the shake constraints to 10^-10 or 10^-12.

However, MTTK does not currently have semiisotropic scaling -- it was
too nasty to do with constraints.

If you want to semiisotropic scaling, I would probably do md +
Parrinello-Rahman.  Currently MTTK really only has an advantage with
certain algorithms that require velocity Verlet.


On Mon, Dec 23, 2013 at 1:43 PM, khaos <spirit.khaos at gmail.com> wrote:
> Dear all,
>
> I'm trying to run the DPPC bilayer simulation using the double precision
> GROMACS with the MTTK pressure coupling method.
> However, I've got an error like this:
>
> -------------------------------------------------------
> Program mdrun_d, VERSION 4.6.5
> Source code file:
> /users/khaos/prog/src/gromacs-4.6.5/src/mdlib/iteratedconstraints.c, line:
> 226
>
> Fatal error:
> Could not converge NPT constraints
>
> For more information and tips for troubleshooting, please check the GROMACS
> website at http://www.gromacs.org/Documentation/Errors
> -------------------------------------------------------
>
> What can I do to solve this problem?
> Additionally, as I know, the MTTK pressure coupling method only supports
> isotropic scaling.
> But I want to use the MTTK method with semi-isotropic scaling because my
> system is membrane system.
> Is there any trick for using the MTTK method with semi-isotropic scaling?
>
> Here is the input parameters that I used.
> =============================
> integrator                   = md-vv
> dt                              = 0.002
> nsteps                        = 500000
> nstlog                        = 100
> nstxout                       = 500
> nstenergy                   = 500
> nstlist                        = 10
> rlist                           = 1.2
> coulombtype               = pme
> rcoulomb                    = 1.2
> cutoff-scheme             = group
> vdw-type                    = switch
> rvdw_switch                = 1.0
> rvdw                          = 1.2
> tcoupl                        = nose-hoover
> tc-grps                       = SYSTEM
> tau-t                          = 0.1
> nsttcouple                   = 1
> ref-t                          = 323
> pcoupl                        = MTTK
> pcoupltype                  = isotropic
> tau-p                         = 1.0
> compressibility            = 4.5e-5 0.0
> ref-p                         = 1.0
> constraints                 = all-bonds
> constraint_algorithm    = shake
> =============================
>
> Best,
> Jumin
>
>
> --
> View this message in context: http://gromacs.5086.x6.nabble.com/MTTK-with-double-precision-GROMACS-tp5013492.html
> Sent from the GROMACS Users Forum mailing list archive at Nabble.com.
> --
> Gromacs Users mailing list
>
> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-request at gromacs.org.