[gmx-users] Domain decomposition configuration
Mark Abraham
Mark.Abraham at anu.edu.au
Thu Feb 10 03:16:38 CET 2011
On 10/02/2011 12:45 PM, Justin A. Lemkul wrote:
>
>
> Mark Abraham wrote:
>> On 10/02/2011 8:10 AM, Justin A. Lemkul wrote:
>>>
>>>
>>> Denny Frost wrote:
>>>> Is tpbconv with the "pbc" option the best way to make the molecules
>>>> whole again?
>>>>
>>>
>>> The only way, as far as I'm aware (aside from editconf's crude
>>> approach). Do this before you concatenate your systems.
>>
>> And you will then have potential issues with clashing atoms, so need
>> to re-equilibrate. Unfortunately, in general there is no box
>> definition where all molecules are whole and inside the box, and it's
>> certainly not worthwhile looking for one.
>>
>
> Not worthwhile? In the case of the OP's problem, there's no way out
> unless you make the molecules whole.
Sure, whole molecules are necessary for the OP's approach, but my point
is that it is not worthwhile looking for a box with all of the molecules
whole *and* inside the (pre-concatenation) box (e.g. by trying new box
centers), because in general there need not be one.
Mark
> You certainly can't have it both ways (both whole and in the box, but
> take, for instance, atoms A and B in the same molecule, which are
> bonded. The normal situation across PBC:
>
> | |
> -B A-
> | |
>
> is fine. DD can re-establish the proper bond across PBC. But if this
> system is concatenated to one with water (*):
>
> | |******|
> -B A--------
> | |******|
>
> the bond is physically unrealistic and DD fails. Even if you *could*
> get the simulation to run (i.e. in serial or mdrun -pd), it would
> instantly crash due either to LINCS failure or some such problem. So
> making the molecules whole prior to concatenation gives:
>
> | |******|
> | A-B*****|
> | |******|
>
> which works. Of course, you do have to worry about clashes, so maybe
> concatenation is not the best approach, but rather a run through
> genbox would be more suitable.
>
> -Justin
>
>> Mark
>>
>>>
>>> -Justin
>>>
>>>> On Wed, Feb 9, 2011 at 2:06 PM, Justin A. Lemkul <jalemkul at vt.edu
>>>> <mailto:jalemkul at vt.edu>> wrote:
>>>>
>>>>
>>>>
>>>> Denny Frost wrote:
>>>>
>>>> This run is actually a combination of two 5x5x5 nm boxes,
>>>> one if
>>>> which was previously run in DD, and the other is water. Since
>>>> the length of that bond is almost 5 nm, is it possible that
>>>> the
>>>> pbc's are not being recognized? There is no way I have a bond
>>>> that long from my previous run.
>>>>
>>>>
>>>> I'll venture a guess that there were broken molecules in the
>>>> system
>>>> you concatenated? That would gel with a bond that stretches
>>>> across
>>>> a 5-nm box. You have to deal with whole molecules in the input
>>>> configuration.
>>>>
>>>> -Justin
>>>>
>>>> On Wed, Feb 9, 2011 at 1:56 PM, Justin A. Lemkul
>>>> <jalemkul at vt.edu <mailto:jalemkul at vt.edu>
>>>> <mailto:jalemkul at vt.edu <mailto:jalemkul at vt.edu>>> wrote:
>>>>
>>>>
>>>>
>>>> Denny Frost wrote:
>>>>
>>>> I'm using version 4.5.3
>>>>
>>>> Here's the output from the log file from DD
>>>> initiation to
>>>> the error:
>>>>
>>>> Initializing Domain Decomposition on 8 nodes
>>>> Dynamic load balancing: auto
>>>> Will sort the charge groups at every domain
>>>> (re)decomposition
>>>> Initial maximum inter charge-group distances:
>>>> two-body bonded interactions: 4.893 nm, Bond, atoms
>>>> 8994 8996
>>>> multi-body bonded interactions: 4.893 nm, Angle, atoms
>>>> 8994 8997
>>>> Minimum cell size due to bonded interactions: 5.382 nm
>>>>
>>>>
>>>> Bonded interactions should normally not occur over such
>>>> a length.
>>>> The information printed here points to the culprits.
>>>> What are
>>>> these atoms, and why are they bonded if they are so far
>>>> away?
>>>>
>>>> -Justin
>>>>
>>>> Using 0 separate PME nodes
>>>> Scaling the initial minimum size with 1/0.8 (option
>>>> -dds)
>>>> = 1.25
>>>> Optimizing the DD grid for 8 cells with a minimum
>>>> initial
>>>> size
>>>> of 6.728 nm
>>>> The maximum allowed number of cells is: X 0 Y 0 Z 1
>>>>
>>>> -------------------------------------------------------
>>>> Program mdrun_mpi, VERSION 4.5.3
>>>> Source code file: domdec.c, line: 6428
>>>>
>>>> Fatal error:
>>>> There is no domain decomposition for 8 nodes that is
>>>> compatible
>>>> with the given box and a minimum cell size of
>>>> 6.72787 nm
>>>> Change the number of nodes or mdrun option -rdd or -dds
>>>> Look in the log file for details on the domain
>>>> decomposition
>>>> For more information and tips for troubleshooting,
>>>> please
>>>> check
>>>> the GROMACS
>>>> website at http://www.gromacs.org/Documentation/Errors
>>>>
>>>> And here is my mdp file
>>>>
>>>> title = BMIM+PF6
>>>> cpp = /lib/cpp
>>>> constraints = hbonds
>>>> integrator = md
>>>> dt = 0.002 ; ps !
>>>> nsteps = 75000 ; total 150 ps
>>>> nstcomm = 10
>>>> nstxout = 50000
>>>> nstvout = 50000
>>>> nstfout = 0
>>>> nstlog = 5000
>>>> nstenergy = 5000
>>>> nstxtcout = 25000
>>>> nstlist = 10
>>>> ns_type = grid
>>>> pbc = xyz
>>>> coulombtype = PME
>>>> vdwtype = Cut-off
>>>> rlist = 1.2
>>>> rcoulomb = 1.2
>>>> rvdw = 1.2
>>>> fourierspacing = 0.12
>>>> pme_order = 4
>>>> ewald_rtol = 1e-5
>>>> ; Berendsen temperature coupling is on in two groups
>>>> Tcoupl = berendsen
>>>> tc_grps = BMI PF6 SOL
>>>> tau_t = 0.2 0.2 0.2
>>>> ref_t = 300 300 300
>>>> nsttcouple = 1
>>>> ; Energy monitoring
>>>> energygrps = BMI PF6 SOL
>>>> ; Isotropic pressure coupling is now on
>>>> Pcoupl = berendsen
>>>> pcoupltype = isotropic
>>>> ;pc-grps = BMI PFF
>>>> tau_p = 2.0
>>>> ref_p = 1.0
>>>> compressibility = 4.5e-5
>>>>
>>>> ; Generate velocites is off at 300 K.
>>>> gen_vel = yes
>>>> gen_temp = 300.0
>>>> gen_seed = 100000
>>>>
>>>>
>>>> On Wed, Feb 9, 2011 at 1:39 PM, Justin A. Lemkul
>>>> <jalemkul at vt.edu <mailto:jalemkul at vt.edu>
>>>> <mailto:jalemkul at vt.edu <mailto:jalemkul at vt.edu>>
>>>> <mailto:jalemkul at vt.edu <mailto:jalemkul at vt.edu>
>>>> <mailto:jalemkul at vt.edu <mailto:jalemkul at vt.edu>>>> wrote:
>>>>
>>>>
>>>>
>>>> Denny Frost wrote:
>>>>
>>>> I am trying to start a run using domain
>>>> decomposition on a
>>>> 5x5x10 nm box with about 26,000 atoms in it.
>>>> I've
>>>> tried
>>>> running
>>>> 8-16 pp nodes, but gromacs always throws an
>>>> error
>>>> saying that
>>>> there is no domain decomposition compatible with
>>>> this box
>>>> and a
>>>> minimum cell size of 6.728 nm. I've tried
>>>> many values
>>>> for -dds
>>>> and a few dd vectors, but with no luck. Does
>>>> anyone know
>>>> to get
>>>> domain decomposition working on a rectangular
>>>> system like
>>>> this?
>>>>
>>>>
>>>> Not without significantly more information.
>>>> Please post:
>>>>
>>>> 1. Your Gromacs version
>>>> 2. Any DD-related information that is printed to
>>>> either the
>>>> log file
>>>> or stdout
>>>> 3. Your .mdp file
>>>>
>>>> -Justin
>>>>
>>>> -- ========================================
>>>>
>>>> Justin A. Lemkul
>>>> Ph.D. Candidate
>>>> ICTAS Doctoral Scholar
>>>> MILES-IGERT Trainee
>>>> Department of Biochemistry
>>>> Virginia Tech
>>>> Blacksburg, VA
>>>> jalemkul[at]vt.edu <http://vt.edu> <http://vt.edu>
>>>> <http://vt.edu> | (540)
>>>>
>>>> 231-9080
>>>>
>>>>
>>>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>>>
>>>> ========================================
>>>> -- gmx-users mailing list
>>>> gmx-users at gromacs.org
>>>> <mailto:gmx-users at gromacs.org>
>>>> <mailto:gmx-users at gromacs.org <mailto:gmx-users at gromacs.org>>
>>>> <mailto:gmx-users at gromacs.org
>>>> <mailto:gmx-users at gromacs.org> <mailto:gmx-users at gromacs.org
>>>> <mailto:gmx-users at gromacs.org>>>
>>>>
>>>>
>>>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>>>> Please search the archive at
>>>>
>>>> http://www.gromacs.org/Support/Mailing_Lists/Search before
>>>> posting!
>>>> Please don't post (un)subscribe requests to the
>>>> list.
>>>> Use the www
>>>> interface or send it to
>>>> gmx-users-request at gromacs.org
>>>> <mailto:gmx-users-request at gromacs.org>
>>>> <mailto:gmx-users-request at gromacs.org
>>>> <mailto:gmx-users-request at gromacs.org>>
>>>> <mailto:gmx-users-request at gromacs.org
>>>> <mailto:gmx-users-request at gromacs.org>
>>>> <mailto:gmx-users-request at gromacs.org
>>>> <mailto:gmx-users-request at gromacs.org>>>.
>>>>
>>>> Can't post? Read
>>>> http://www.gromacs.org/Support/Mailing_Lists
>>>>
>>>>
>>>>
>>>> -- ========================================
>>>>
>>>> Justin A. Lemkul
>>>> Ph.D. Candidate
>>>> ICTAS Doctoral Scholar
>>>> MILES-IGERT Trainee
>>>> Department of Biochemistry
>>>> Virginia Tech
>>>> Blacksburg, VA
>>>> jalemkul[at]vt.edu <http://vt.edu> <http://vt.edu> | (540)
>>>> 231-9080
>>>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>>>
>>>> ========================================
>>>> -- gmx-users mailing list gmx-users at gromacs.org
>>>> <mailto:gmx-users at gromacs.org>
>>>> <mailto:gmx-users at gromacs.org <mailto:gmx-users at gromacs.org>>
>>>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>>>> Please search the archive at
>>>> http://www.gromacs.org/Support/Mailing_Lists/Search before
>>>> posting!
>>>> Please don't post (un)subscribe requests to the list.
>>>> Use the www
>>>> interface or send it to gmx-users-request at gromacs.org
>>>> <mailto:gmx-users-request at gromacs.org>
>>>> <mailto:gmx-users-request at gromacs.org
>>>> <mailto:gmx-users-request at gromacs.org>>.
>>>> Can't post? Read
>>>> http://www.gromacs.org/Support/Mailing_Lists
>>>>
>>>>
>>>>
>>>> -- ========================================
>>>>
>>>> Justin A. Lemkul
>>>> Ph.D. Candidate
>>>> ICTAS Doctoral Scholar
>>>> MILES-IGERT Trainee
>>>> Department of Biochemistry
>>>> Virginia Tech
>>>> Blacksburg, VA
>>>> jalemkul[at]vt.edu <http://vt.edu> | (540) 231-9080
>>>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>>>
>>>> ========================================
>>>> -- gmx-users mailing list gmx-users at gromacs.org
>>>> <mailto:gmx-users at gromacs.org>
>>>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>>>> Please search the archive at
>>>> http://www.gromacs.org/Support/Mailing_Lists/Search before
>>>> posting!
>>>> Please don't post (un)subscribe requests to the list. Use the www
>>>> interface or send it to gmx-users-request at gromacs.org
>>>> <mailto:gmx-users-request at gromacs.org>.
>>>> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>>
>>>>
>>>
>>
>
More information about the gromacs.org_gmx-users
mailing list