Re: ABF simulation fails with "FATAL ERROR: Low global exclusion count!"

From: Maria Pikoula (maria.pikoula_at_eng.ox.ac.uk)
Date: Fri Feb 28 2014 - 12:50:27 CST

Thanks for the tips Jerome,

I attach here the highest values output from the ABF gradient file (is that too large?) and also the final colvar trajectory information which looks normal, although that was about 1000 steps before the crash.

Cheers,

Maria

----------------
Gradient:

 3.75 -17.4257
 3.85 -13.5601
 3.95 -9.01215
 4.05 -5.62778
 4.15 -3.51482
 4.25 -3.10521
 4.35 -2.65639
 4.45 -2.27599
 4.55 -1.53575
 4.65 -1.29962
 4.75 -1.0401
 4.85 -0.847205
 4.95 -1.05582
 5.05 -0.707276
 5.15 -0.86581

Colvars Trajectory, final output:

     7979250 9.21938587001597e+00 -1.19191023819836e-01
     7979500 8.99085146295547e+00 -9.03306597326653e-02
     7979750 9.35263382270071e+00 -1.42858663471416e-01
     7980000 8.99325331654785e+00 -9.05210165761842e-02

On 28 Feb 2014, at 17:58, Jérôme Hénin wrote:

> Hi Maria,
>
> If you inspect the ABF gradient file and look for high values, you will see if ABF applies very large forces. That would most probably be due to a bad coordinate definition.
> For more information on instantaneous forces, you could enable the outputAppliedforce option in the colvar.
>
> Cheers,
> Jerome
>
>
> ----- Original Message -----
>> Dear Thanassis,
>>
>> Thanks for your comments and input. I reset the margin to 0 and I am still
>> getting the same issue. I don't think the problem is a non-equilibrated
>> system either. Do you think I should just restart every time it happens
>> then?
>>
>> Cheers,
>>
>> Maria
>>
>> On 25 Feb 2014, at 17:15, Athanassios Stavrakoudis wrote:
>>
>>>
>>> Hi,
>>>
>>> Setting margin to 30 is a bad idea.
>>> Better keep it close to zero.
>>> It might be a poorly equilibrated system.
>>> Anyway you can restart your simulation from the point it stopped.
>>> Afterwards you can merge your dcd files.
>>>
>>> Best,
>>> Thanassis
>>>
>>>
>>>> Dear all,
>>>>
>>>> In my simulation of two GlcNS(6S) restrained close to each other using
>>>> ABF, the simulation fails at at some point, usually after 10 or more
>>>> nanoseconds. I get the following error:
>>>>
>>>> FATAL ERROR: Low global exclusion count! (9601 vs 9603) System unstable
>>>> or pairlistdist or cutoff too small.
>>>>
>>>> Increasing the pairlistdist hasn't helped. Any ideas what may be causing
>>>> this? Below I copy the configuration file and the colvars input file.
>>>>
>>>> Regards,
>>>>
>>>> Maria
>>>>
>>>> Config. File:
>>>> #############################################################
>>>> ## SIMULATION PARAMETERS ##
>>>> #############################################################
>>>> coordinates eqlb.coor
>>>> margin 30;
>>>> numsteps 25000000
>>>> set temp 300
>>>> set outputname abf_3.7to10
>>>>
>>>> firsttimestep 0
>>>>
>>>> binaryoutput no
>>>> binaryrestart no
>>>>
>>>>
>>>> velocities eqlb.vel
>>>>
>>>> paraTypeCharmm off
>>>>
>>>> amber on
>>>> parmfile ../monomers_WAT.prmtop
>>>>
>>>> # Force-Field Parameters
>>>> exclude scaled1-4
>>>> 1-4scaling 1.0
>>>> cutoff 12.
>>>> switching off
>>>> pairlistdist 13.5
>>>>
>>>>
>>>> # Integrator Parameters
>>>> timestep 2
>>>> rigidBonds all
>>>> nonbondedFreq 1
>>>> fullElectFrequency 2
>>>> stepspercycle 10
>>>>
>>>>
>>>> # Constant Temperature Control Using Berendsen Method (Coupling)
>>>>
>>>> tCouple on
>>>> tCoupleTemp $temp
>>>> tCoupleFile temp_coupling.pdb
>>>> tCoupleCol B
>>>>
>>>> wrapAll off
>>>>
>>>> extendedSystem eqlb.xsc
>>>>
>>>> # PME (for full-system periodic electrostatics)
>>>> PME yes
>>>> PMEGridSpacing 1.0
>>>>
>>>> # Constant Pressure Control (variable volume)
>>>> useGroupPressure yes
>>>> useFlexibleCell no
>>>> useConstantArea no
>>>>
>>>> BerendsenPressure on
>>>> BerendsenPressureTarget 1.01325 ;# in bar -> 1 atm
>>>> BerendsenPressureCompressibility 4.57E-5
>>>> BerendsenPressureRelaxationTime 100
>>>> BerendsenPressureFreq 10
>>>>
>>>>
>>>> # Output
>>>> outputName $outputname
>>>>
>>>>
>>>> restartfreq 500
>>>> dcdfreq 250
>>>> xstFreq 250
>>>> outputEnergies 100
>>>> outputPressure 100
>>>>
>>>> colvars on
>>>> colvarsConfig Distance_3.7to10.in
>>>>
>>>> ##############################################
>>>>
>>>> Colvars file:
>>>>
>>>> ##############################################
>>>>
>>>> colvarsTrajFrequency 250
>>>> colvarsRestartFrequency 20000
>>>>
>>>> colvar {
>>>> name AtomDistance
>>>> outputAppliedForce on
>>>> width 0.1
>>>>
>>>> lowerboundary 3.7
>>>> upperboundary 10.0
>>>>
>>>> lowerwallconstant 10.0
>>>> upperwallconstant 10.0
>>>>
>>>> distance {
>>>> group1 {
>>>> atomnumbers { 8 16 }
>>>> }
>>>> group2 {
>>>> atomnumbers { 42 50 }
>>>> }
>>>> }
>>>> }
>>>>
>>>>
>>>> abf {
>>>> colvars AtomDistance
>>>> fullSamples 1000
>>>> hideJacobian
>>>> }
>>>> ##############################################
>>>>
>>>> Maria Pikoula
>>>> Doctoral Candidate in Centre for Doctoral Training in Healthcare
>>>> Innovation
>>>> Institute of Biomedical Engineering
>>>> Department of Engineering Science
>>>> University of Oxford
>>>> maria.pikoula_at_eng.ox.ac.uk
>>>>
>>>
>>>
>>>
>>> --
>>> Athanassios Stavrakoudis
>>> University of Ioannina web: http://stavrakoudis.econ.uoi.gr/
>>> Department of Economics e-mail: mailto:astavrak_at_uoi.gr
>>> GR-45110, Ioannina tel: +30 265 100 5935
>>> GREECE fax: +30 265 100 5092
>>
>> Maria Pikoula
>> Doctoral Candidate in Centre for Doctoral Training in Healthcare Innovation
>> Institute of Biomedical Engineering
>> Department of Engineering Science
>> University of Oxford
>> maria.pikoula_at_eng.ox.ac.uk
>>
>>

Maria Pikoula
Doctoral Candidate in Centre for Doctoral Training in Healthcare Innovation
Institute of Biomedical Engineering
Department of Engineering Science
University of Oxford
maria.pikoula_at_eng.ox.ac.uk

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:22:11 CST