[gmx-users] extending the simulation with modified mdp option

Justin A. Lemkul jalemkul at vt.edu
Fri Jan 14 00:47:24 CET 2011



Ramachandran G wrote:
> thanks for your reply.
> Still i have the same problem.
> My aim is to extend my simulation from 6ns to 7ns.
> 
> First i prepared the tpr as
> 
> grompp_mpi -f nve.mdp -c nve6-g.tpr -o nve7-g.tpr -n crom.ndx
> 
> where 'nve6-g.tpr' is my old 'tpr' . In the 'nve.mdp'  file i have 
> changed the option of storing the 'nstxout' from 1 to 100.
> That is the only change i have done.
> 
> For simulation i do,
> 
> mpirun  -nolocal -np 8 -machinefile machine mdrun_mpi -nice 0 -v -s 
> nve7-g.tpr -o nve-g.trr -c nve-g.gro -e nve-g.edr  -cpi nve6-g.cpt -g 
> md-g.log -append -cpo nve7-g.cpt  -pd
> 
>  From the gromacs website, 
> http://www.gromacs.org/Documentation/How-tos/Extending_Simulations
> i understood that i need to use the option of 'tpbconv' also to extend 
> the simulation. It says that
> 'If your |old.cpt| is for a run that has finished, then use |tpbconv 
> -extend |after grompp <http://www.gromacs.org/grompp> and before mdrun 
> <http://www.gromacs.org/Documentation/Gromacs_Utilities/mdrun>.'
> 
> 

The easiest thing for you to do is to simply extend your run with tpbconv and 
ignore grompp.  If you just need to reduce output volume, post-process the 
trajectory with trjconv.  Otherwise, you have 6 ns of coordinates at every step, 
then 1 ns of coordinates every 100 steps.  The frame interval mismatch will 
likely cause numerous Gromacs tools to fail, anyway.

> My 'mdp' options are:
> 
> title               = production-dynamics
> ;warnings           = 10
> cpp                 = /lib/cpp
> ;DEFINE             = -DPOSRES
> DEFINE             = -DFLEXIBLE

Flexible water should not be used for MD.

http://lists.gromacs.org/pipermail/gmx-users/2008-October/037571.html

> constraints         = none
> ;constraint_algorithm = shake
> integrator         = md
> dt                 = 0.0005  ;
> nsteps             = 2000000  ;
> ;nstcomm           = 1     ; reset c.o.m motion
> nstlist            = 10
> ns_type            = grid
> nstenergy          = 1  ; print energies
> nstlog             = 1  ; print to logfile
> nstvout            = 0   ; write velocities
> nstxout            = 100  ; collect data in fs (write coords)
> ;nstxtcout         = 0   ; to print corrdinates to xtc trajectory
> energygrps         = Protein CROM HOH1 HOH2 SOL
> nstfout            = 0
> coulombtype        = shift
> fourierspacing     = 0.12
> pme_order          = 4
> vdwtype            = switch
> rvdw               = 1.0
> rvdw_switch        = 0.9
> rlist              = 1.2
> rcoulomb           = 1.0
> rcoulomb-switch    = 0.9
> pbc                = xyz
> ;dispcorr           = Ener
> continuation      = yes
> ;Berendsen temperature coupling is on
> Tcoupl         = no   ; temperature bath (yes, no)
> ;tau_t           = 0.1
> ;tc-grps         = system
> ;ref_t           = 300
> ;Berendsen Pressure coupling is on
> pcoupl          = no         ; pressure bath (yes, no)
> ;pcoupltype      = isotropic
> ;tau_p           = 0.5
> ;compressibility = 4.5e-05
> ;ref_p           = 1.0
> ;Generate velocities  is on at 300
> gen_vel         = no       ; generate initial velocities
> ;gen_temp        = 300.0     ; initial temperature
> ;gen_seed        = 173529    ; random seeD
> 

You haven't set the tinit parameter in this file.  In conjunction with the fact 
that this .mdp file is for 1 ns, it will start at time zero and run to 1 ns.  If 
you're trying to pick up from the end of a 6-ns simulation, you have two options:

1. Set tinit=6000 and leave nsteps and dt alone.  Pass the .cpt file to the -t 
flag of grompp and run it as if it were a new simulation, then concatenate the 
output.

2. Set tinit=0 and set nsteps=350000 to generate a complete 7-ns .tpr file. 
Pass this to mdrun in conjunction with your .cpt file from the end of your 6-ns run.

3. Use tpbconv and ignore grompp entirely, then post-process appropriately.

http://www.gromacs.org/Documentation/How-tos/Extending_Simulations#Version_4

-Justin

> 
> Please help.
> thank you,
> Rama
> 
> 
> 
> On Wed, Jan 12, 2011 at 3:06 PM, Justin A. Lemkul <jalemkul at vt.edu 
> <mailto:jalemkul at vt.edu>> wrote:
> 
> 
> 
>     Ramachandran G wrote:
> 
>         Hi gmx-users,
>            I tried extending one of simulation with modified mdp option.
>         First i used the command,
> 
>             grompp -f new.mdp -c old.tpr -o new.tpr
> 
>         then i used the command to extend
> 
>             tpbconv -s new.tpr -extend timetoextendby -o next.tpr
> 
> 
>         Then,
> 
>             mdrun -s next.tpr -cpi previous.cpt
> 
> 
>         I surprised to see the following simulation output,
>         WARNING: This run will generate roughly 2059394074917430016 Mb
>         of data
> 
>         starting mdrun 'GROningen MAchine for Chemical Simulation in water'
> 
>         2000000 steps, infinite ps (continuing from step 12000000,  
>         6000.0 ps).
>         step 12000000 performance: 0.3 ns/day    step 12000100
>         performance: 0.8 ns/day    step 12000200 performance: 0.8 ns/day
>            step 12000300 performance: 0.8 ns/day    
>         step 12000400 performance: 0.8 ns/day    step 12000500
>         performance: 0.8 ns/day    step 12000600 performance: 0.8 ns/day
>            step 12000700 performance: 0.8 ns/day    step 12000800
>         performance: 0.8 ns/day    
>         step 12000900 performance: 0.8 ns/day    step 12001000
>         performance: 0.8 ns/day    step 12001100 performance: 0.8 ns/day
>            --------------------------------------------------
>         Is this correct? Why it needs to generate that much huge Mb of data?
> 
>         I have extended for 1000.0ps but it shows 'infinite ps'.
>         can anybody help me if there is error made in my procedure?
> 
> 
>     It seems like you're trying to extend your run using both grompp and
>     tpbconv, which is not correct.  Use one or the other.
> 
>     If you want a real answer to this question, please post real
>     commands, not just the generic format of what should be correct.
>      Also possibly useful is the contents of the .mdp file you used.
> 
>     -Justin
> 
>         Thank you
>         Rama
> 
>          
> 
> 
> 
>     -- 
>     ========================================
> 
>     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
> 
> 
> 
> 
> 

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================



More information about the gromacs.org_gmx-users mailing list