[gmx-users] Fwd: Problem in applying walls in the z direction

[Stephen Chan]

Hi Justin,

Just above the fatal error message, there is some additional information:

Initializing Domain Decomposition on 112 ranks
Dynamic load balancing: auto
Will sort the charge groups at every domain (re)decomposition
Initial maximum inter charge-group distances:
     two-body bonded interactions: 10.867 nm, Exclusion, atoms 30544 30545
   multi-body bonded interactions: 0.488 nm, CMAP Dih., atoms 1075 1091
Minimum cell size due to bonded interactions: 11.954 nm
Maximum distance for 5 constraints, at 120 deg. angles, all-trans: 0.222 nm
Estimated maximum distance required for P-LINCS: 0.222 nm
Guess for relative PME load: 0.23
Will use 84 particle-particle and 28 PME only ranks
This is a guess, check the performance at the end of the log file
Using 28 separate PME ranks, as guessed by mdrun
Scaling the initial minimum size with 1/0.8 (option -dds) = 1.25
Optimizing the DD grid for 84 cells with a minimum initial size of 14.942 nm
The maximum allowed number of cells is: X 0 Y 0 Z 2

Do I need to adjust other parameters in the mdp accordingly??

Stephen


On 07/02/16 22:02, Justin Lemkul wrote:
>
>
> On 7/2/16 3:50 PM, Stephen Chan wrote:
>> Dear all,
>>
>> I'm running a simulation of a bilayer system and I would like to 
>> apply wall
>> condition to both xy planes (i.e. nwall = 2). On top of the normal MD 
>> settings,
>> I added a couple of lines to the mdp file:
>> pbc             = xy
>> nwall           = 2
>> wall-type       = 10-4
>> wall-density    = 5  5
>> wall-atomtype   = OT OT
>> ewald-geometry  = 3dc
>>
>> integrator              = md
>> dt                      = 0.002
>> nsteps                  = 50000000
>> nstlog                  = 10000
>> nstxout                 = 0
>> nstvout                 = 0
>> nstfout                 = 0
>> nstcalcenergy           = 10000
>> nstenergy               = 10000
>> nstxout-compressed      = 10000
>> ;
>> cutoff-scheme           = Verlet
>> nstlist                 = 20
>> rlist                   = 1.0
>> coulombtype             = pme
>> rcoulomb                = 1.0
>> vdwtype                 = Cut-off
>> vdw-modifier            = Force-switch
>> rvdw_switch             = 0.9
>> rvdw                    = 1.0
>> ;
>> tcoupl                  = Nose-Hoover
>> tc_grps                 = PROT   MEMB   SOL_ION
>> tau_t                   = 1.0    1.0    1.0
>> ref_t                   = 310 310 310
>> ;
>> pcoupl                  = Parrinello-Rahman
>> pcoupltype              = semiisotropic
>> tau_p                   = 5.0
>> compressibility         = 0.0     4.5e-5
>> ref_p                   = 1.0     1.0
>> ;
>> constraints             = h-bonds
>> constraint_algorithm    = LINCS
>> continuation            = yes
>> ;
>> nstcomm                 = 100
>> comm_mode               = linear
>> comm_grps               = PROT   MEMB   SOL_ION
>> ;
>> refcoord_scaling        = com
>>
>> When running grompp, I managed to generate the tpr file from my input
>> gro file. However, when running mdrun, I got the following error:
>>
>> Fatal error:
>> There is no domain decomposition for 84 ranks that is compatible with
>> the given box and a minimum cell size of 14.9422 nm
>> Change the number of ranks or mdrun option -rdd or -dds
>> Look in the log file for details on the domain decomposition
>>
>> Any help would be appreciated :)
>>
>
> The error message tells you where to start:
>
> "Look in the log file for details on the domain decomposition"
>
> What do you find there?  You have bonded interactions at very long 
> distances, and the .log file tells you exactly which atoms are engaged 
> in the limiting interactions.
>
> -Justin
>