Re: AW: restarting Gaussian accelerated MD fails

From: Vlad Cojocaru (vlad.cojocaru_at_mpi-muenster.mpg.de)
Date: Fri Jan 25 2019 - 04:48:50 CST

well, I want to start from the last step of the previous run which is
25000000 (firstTimeStep is also set to this) ... and original
accelerated MD is fine with that ...
but sure, its something I should test just to be sure ...

Thanks, Vlad

On 1/25/19 11:30 AM, Utesch, Tillmann wrote:
>
> Dear Vlad,
>
>
> maybe you should set accelMDFirstStep to 0.
>
> ------------------------------------------------------------------------
> *Von:* owner-namd-l_at_ks.uiuc.edu <owner-namd-l_at_ks.uiuc.edu> im Auftrag
> von Vlad Cojocaru <vlad.cojocaru_at_mpi-muenster.mpg.de>
> *Gesendet:* Freitag, 25. Januar 2019 11:11
> *An:* Pang Andrew; Miao, Yinglong
> *Cc:* David Hardy; NAMD
> *Betreff:* Re: namd-l: restarting Gaussian accelerated MD fails
> Thanks Yinglong, Andrew for the explanation ...
>
> The restart problem is still there. The restart file is not read
> properly (namd log says at first time step that dV = 0, iE = 0, k0 =
> 0) and the simulation crashes soon after that although NAMD does
> recognize the restart file as input ("accelMDG USING RESTART FILE
> oct4_hoxb1_mur_01_box_gamd010102.gamd" ) ... The restart file also
> seems correct.
>
> Curiously that it appears that nobody else has encountered this issue
> ....
>
> I guess I will wait for David's feed back ... He has all files needed
> to reproduce the issue ..
> I provide below again the relevant input lines and the restart file
> ...  Maybe somebody sees something obvious that I am missing ....
>
> Vlad
>
> ---- NAMD input -----
> #*** Accelerated MD
> **********************************************************************************
> accelMD                                       on
> accelMDG                                      on
> accelMDdihe                                   on
> accelMDdual                                   on
> accelMDFirstStep                        25000000
> accelMDLastStep                         50000000
> accelMDOutFreq                              2500
>
> accelMDGiE                                     1
> accelMDGcMDPrepSteps                           0
> accelMDGcMDSteps                               0
> accelMDGEquiPrepSteps                          0
> accelMDGEquiSteps                              0
> accelMDGSigma0P                                6
> accelMDGSigma0D                                6
> accelMDGRestart                               on
> accelMDGRestartFile               test.gamd
>
>
> ---------- GAMD restart  (test.gamd) --------
> # NAMD accelMDG restart file
> # D/T Vn Vmax Vmin Vavg sigmaV M2 E k
> D 10000000 4779.3504030500444 4469.4517421030359 4633.215337071334
> 49.785203930094141 2.47856653036106071e+10 4779.3504030500444
> 0.0008247009977261319
> T 10000000 -974348.98058993137 -984746.7871256324 -976797.22733897553
> 856.08499832075404 7.32881524349845312e+12 -974348.98058993137
> 2.8627220105560853e-06
>
>
>
>
> On 1/24/19 2:02 AM, Pang Andrew wrote:
>>
>> Dear Vlad,
>>
>> Dr Miao is right. The second column of the restart file shows the
>> number of steps used to obtain the statistics instead of the number
>> of simulation steps.
>>
>> Sincerely,
>>
>> Andrew
>>
>> On Tue, Jan 22, 2019, 3:10 PM Miao, Yinglong <miao_at_ku.edu
>> <mailto:miao_at_ku.edu> wrote:
>>
>>
>>
>> On Jan 22, 2019, at 2:56 AM, Vlad Cojocaru
>> <vlad.cojocaru_at_mpi-muenster.mpg.de
>> <mailto:vlad.cojocaru_at_mpi-muenster.mpg.de>> wrote:
>>
>> I also have a question regarding the k0 values. During the
>> gAMD equilibration, these values are updated at each time
>> step whereas during a production run , k0 is maintained
>> constant throughout the simulation ...  That is correct, right ?
>>
>> Dear Vlad,
>>
>> Yes, k0 works in the way as you described.
>>
>> Hope David can help with the simulation restart problem, and
>> Andrew on the step number in the restart file (Should it be the
>> number used to calculate average and standard deviation of the
>> boost potential?)
>>
>> Best wishes,
>>
>> Yinglong
>>
>> Yinglong Miao, Ph.D.
>> Assistant Professor
>> Center for Computational Biology and
>> Department of Molecular Biosciences
>> University of Kansas
>> http://miao.compbio.ku.edu
>> <https://urldefense.proofpoint.com/v2/url?u=http-3A__miao.compbio.ku.edu&d=DwMD-g&c=OCIEmEwdEq_aNlsP4fF3gFqSN-E3mlr2t9JcDdfOZag&r=jUfnSyKkfkyVRBIUzlG1GSGGZAZGcznwr8YliSSCjPc&m=L3Ljvh2mjX-areRBOG_9pALiW6yI70tfzsC3aBusJuI&s=nYQO2xqZD_XwMl6QyTyYJ4LeVOCcUGBBcAOzMLm2WqI&e=>
>>
>
> --
> Vlad Cojocaru, PD (Habil.), Ph.D.
> -----------------------------------------------
> Project Group Leader
> Department of Cell and Developmental Biology
> Max Planck Institute for Molecular Biomedicine
> Röntgenstrasse 20, 48149 Münster, Germany
> -----------------------------------------------
> Tel: +49-251-70365-324; Fax: +49-251-70365-399
> Email: vlad.cojocaru[at]mpi-muenster.mpg.de
> http://www.mpi-muenster.mpg.de/43241/cojocaru

-- 
Vlad Cojocaru, PD (Habil.), Ph.D.
-----------------------------------------------
Project Group Leader
Department of Cell and Developmental Biology
Max Planck Institute for Molecular Biomedicine
Röntgenstrasse 20, 48149 Münster, Germany
-----------------------------------------------
Tel: +49-251-70365-324; Fax: +49-251-70365-399
Email: vlad.cojocaru[at]mpi-muenster.mpg.de
http://www.mpi-muenster.mpg.de/43241/cojocaru

This archive was generated by hypermail 2.1.6 : Fri Oct 18 2019 - 23:20:09 CDT