Re: cell basis vector restart

From: Jeffrey J. Potoff (jpotoff_at_chem1.eng.wayne.edu)
Date: Tue Mar 02 2010 - 13:09:37 CST

Patch grid error is a common problem, and if you search the NAMD mailing
list archives you should find a number of posts on this.

What is going on is your system, as it equilibrates, is shrinking to
achieve the correct density. If the simulation box shrinks too much,
NAMD spits the patch grid error. The patch grid is how NAMD splits the
box into sections so that calculations may be distributed efficiently to
the various processors.

You can fix this by starting your system closer to the final (correct)
density, or running a series of short simulations, where each successive
simulation uses the final configuration of the last calculation.
Regards,
Jeff

Mark M Huntress wrote:
> Now I am getting this error, and I am not totally sure what it means:
>
> Periodic cell has become too small for original patch grid!
> Possible solutions are to restart from a recent checkpoint,
> increase margin, or disable useFlexibleCell for liquid simulation.
> FATAL ERROR: Periodic cell has become too small for original patch grid!
> Possible solutions are to restart from a recent checkpoint,
> increase margin, or disable useFlexibleCell for liquid simulation.
>
> This is the .namd file from the first job where I heat the system and the box grows:
>
> amber yes
> parmfile kfretBlysA.prmtop
>
> # Le coordinate sono lette dal file pdb.
> # In questo modo si eleina la
> # dipendenza da ambercoord etc....
>
> coordinates kfretBlysAmin.coor
> scnb 2.0
>
> # output
> set output kfretBlysAheat
> outputname $output
> dcdfile ${output}.dcd
> xstFile ${output}.xst
> dcdfreq 2000
> xstFreq 2000
>
> binaryoutput no
> binaryrestart no
> outputEnergies 200
> restartfreq 10000 #makes restart files (xsc, coor, vel) every x steps
>
>
> # Basic dynamics
> exclude scaled1-4 #neglect 1-2, 1-3 estatic? interactions, weaken 1-4
> 1-4scaling 0.8333 #between 0-1 0: neglect 1-4 estatic interactions
> COMmotion no #??
> dielectric 1.0 # 1.0 is default?
>
> # Simulation space partitioning
> switching on
> switchdist 10 #starts smoothly lowering vdw curve at this distance
> cutoff 12 #vdw ignored after this distance
> pairlistdist 14 #only looks for possible close atoms in this distance
>
> # Multiple timestepping
> firsttimestep 0
> timestep 1
> stepspercycle 20 #frequency with which atoms within cutoff are looked for
> nonbondedFreq 1 #full nonbonded interactions calculated only every x steps
> fullElectFrequency 1 #full estatics calculated only every x steps
> rigidBonds water # no vibration, change in bond length
>
> # Temperature control
>
> # set temperature 298
> temperature 0 # initial temperature
> reassignfreq 1000
> reassigntemp 0
> reassignincr 10
> reassignhold 300
>
> cellBasisVector1 -62.0900006295 0.0 0.0
> cellBasisVector2 0.0 -59.0930003524 0.0
> cellBasisVector3 0.0 0.0 -68.6139990091
> cellOrigin 32.4652252197 30.272726059 35.9197502136
> wrapAll on
> dcdUnitCell yes
> useConstantArea no # CAMBIATO
>
>
> PME yes
> PMEGridSizeX 65
> PMEGridSizeY 60
> PMEGridSizeZ 70
>
> # additional simulation parameters--fixed atoms
> fixedAtoms on
> fixedAtomsForces on
> fixedAtomsCol B
> fixedAtomsFile kfretBlysAfix.pdb
>
>
> #constraints on
> #consref kfretBlysAharmref.pdb
> #conskfile kfretBlysAharm.pdb
> #conskcol B
>
>
> # Scripting
>
> minimize 1000
> #reinitvels $temperature
> #run 10000
> #reinitvels $temperature
> #run 10000
> #reinitvels $temperature
> #run 10000
> #reinitvels $temperature
> #run 10000
> #reinitvels $temperature
> #run 10000
>
> ## I primi 50 ps sono stati eseguiti nell'insieme NPT
> ## riassegnando le velocit¨¤ ogni 10 ps
>
> ## Seguono 5 ns di simulazione NPT
> # su cui farai le analisi
> run 40000
>
>
>
>
> Do I need to put in a keyword to keep the volume constant, or something?
>
>
>
>
>
> This is the .namd file from the next job where I got the error:
>
> # input
> amber yes
> parmfile ../kfretBlysA.prmtop
>
> # Le coordinate sono lette dal file pdb.
> # In questo modo si eleina la
> # dipendenza da ambercoord etc....
>
> #coordinates klechHmin.coor
>
> set inputname kfretBlysAheat
> Coordinates $inputname.coor
> Velocities $inputname.vel
> extendedSystem $inputname.xsc
>
>
> scnb 2.0
>
> # output
> set output kfretBlysArun
> outputname $output
> dcdfile ${output}.dcd
> xstFile ${output}.xst
> dcdfreq 2000
> xstFreq 2000
>
> binaryoutput no
> binaryrestart no
> outputEnergies 200
> restartfreq 10000 #makes restart files (xsc, coor, vel) every x steps
>
>
> # Basic dynamics
> exclude scaled1-4 #neglect 1-2, 1-3 estatic? interactions, weaken 1-4
> 1-4scaling 0.8333 #between 0-1 0: neglect 1-4 estatic interactions
> COMmotion no # ??
> dielectric 1.0 # 1.0 is default?
>
> # Simulation space partitioning
> switching on
> switchdist 10 #starts smoothly phasing out vdw curve at this distance
> cutoff 12 #vdw ignored after this distance
> pairlistdist 14 #only looks for possible close atoms in this distance
>
> # Multiple timestepping
> firsttimestep 0
> timestep 1
> stepspercycle 20 #frequency with which atoms within cutoff are looked for
> nonbondedFreq 1 ##full nonbonded interactions calculated only every x steps
> fullElectFrequency 2 #full estatics calculated only every x steps
> rigidBonds water # no vibration, change in bond length
>
> # Temperature control
>
> set temperature 298
>
> # Langevin Dynamics
> langevin on; # do langevin dynamics
> langevinDamping 5; # damping coefficient (gamma) of 5/ps
> langevinTemp $temperature; # bath temperature
> langevinHydrogen no; # don't couple langevin bath to hydrogens
> seed 12345
>
> # Pressure control
> langevinPiston on
> langevinPistonTarget 1.01325; # in bar -> 1.01325 bar = 1 atm
> langevinPistonPeriod 200
> langevinPistonDecay 100
> langevinPistonTemp $temperature
> useFlexibleCell no # CAMBIATO (cell dimensions vary independently. useful for membranes)
> useGroupPressure yes # must be set to yes if rigidbonds is set
>
> # PBC
>
> if (0) {
> cellBasisVector1 -63.3060009479 0.0 0.0
> cellBasisVector2 0.0 -59.0939998031 0.0
> cellBasisVector3 0.0 0.0 -72.3179966211
> cellOrigin 32.7434005737 30.3597278595 37.5696868896
> }
> wrapAll on
> dcdUnitCell yes
> useConstantArea no # CAMBIATO
>
>
> PME yes #shoud PME part be commented out on restart? email says no.
> PMEGridSpacing 1.0 #automatically selects appropriate PME dimensions
> #PMEGridSizeX 65
> #PMEGridSizeY 60
> #PMEGridSizeZ 75
>
> # additional simulation parameters--fixed atoms
> fixedAtoms on
> fixedAtomsForces on
> fixedAtomsCol B
> fixedAtomsFile kfretBlysAfix.pdb
>
> run 900000
>
>
>
> Sorry there is so much info here, I would appreciate help though.
>
> Thanks
> Mark
>
>
> ________________________________________
> From: owner-namd-l_at_ks.uiuc.edu [owner-namd-l_at_ks.uiuc.edu] On Behalf Of Mark M Huntress [markmh_at_bgsu.edu]
> Sent: Tuesday, March 02, 2010 10:51 AM
> To: ÁÖÐÇÓê; namd-l_at_ks.uiuc.edu
> Subject: RE: namd-l: cell basis vector restart
>
> ah, I was wondering if I should do that, thank you, but the dimensions listed in the xsc file are exactly the same as the ones I started with. But if I measure, in VMD, the dimensions of the .coor output file I am using as the starting coordinates for this run, the box dimensions are bigger than those listed in the xsc file. Is that OK?
> Thanks
> Mark
> ________________________________________
> From: ÁÖÐÇÓê [linxingyu1986_at_gmail.com]
> Sent: Monday, March 01, 2010 10:08 PM
> To: Mark M Huntress
> Subject: Re: namd-l: cell basis vector restart
>
> read the box vector from the last xsc file when you restart your simualtion
>
> On Tue, Mar 2, 2010 at 6:47 AM, Mark M Huntress <markmh_at_bgsu.edu<mailto:markmh_at_bgsu.edu>> wrote:
> Hi, I just did a slow heat, for which I set my cell basis vectors equal to the solvent box dimensions. After the slow heat, the box has increased in size. Do I need to respecify the cell basis vectors to the larger values when restarting?
> Thanks
>
>
>
>
> --
> Xingyu Lin
>
> Currently:
> National Institute of Biological Sciences, Beijing
> Dr. Niu Huang's Lab
> http://www.nibs.ac.cn/
> Tel: 80726688-8572
> Address: Beijing ZhongGuanCun Life Science Park, NO.7 Science Park East Road, NIBS B427
> Postcode: 102206
>
>
>

-- 
======================================================================
Jeffrey J. Potoff                         jpotoff_at_chem1.eng.wayne.edu
Associate Professor                       Wayne State University		
Department of Chemical Engineering and Materials Science
5050 Anthony Wayne Dr                     Phone:(313)577-9357		
Detroit, MI 48202                         Fax:  (313)578-5815
http://potoff1.eng.wayne.edu
======================================================================

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:55:31 CST