[gmx-developers] Re: [gmx-users] Re: Re: g_velacc problem (Florian Dommert)

[Berk Hess]

On 10/13/2010 02:32 PM, Florian Dommert wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 10/13/2010 02:08 PM, Florian Dommert wrote:
>   
>> On 10/13/2010 01:01 PM, Eudes Fileti wrote:
>>     
>>> Hi Florian, thank you for your help!
>>> I informed at the file.mdp that the velocities should be saved:
>>>       
>>     
>>> ; OUTPUT CONTROL OPTIONS
>>> ; Output frequency for coords (x), velocities (v) and forces (f)
>>> nstxout                  = 10
>>> nstvout                  = 10
>>> nstfout                  = 0
>>> ;
>>>       
>>
>> Hi,
>>
>> I took a quick look at the code and realized that the problem is the -m
>> flag. I will try to fix this and report ASAP.
>>
>>
>> /Flo
>>
>>
>>     
> Hello,
>
>  there is a wrong check in line 164/165:
>
>   if (bMol)
>     bTPS = bM || ftp2bSet(efTPS,NFILE,fnm) ||
>            !ftp2bSet(efNDX,NFILE,fnm);
>
> It should be:
>
>   if (bMol || bM)
>     bTPS = ftp2bSet(efTPS,NFILE,fnm) || !ftp2bSet(efNDX,NFILE,fnm);
>
> If you change this lines in src/tools/gmx_velacc.c and recompile,
> everything should work fine.
>
> Attached there is a patch against the master branch which should fix the
> issue.
>
> /Flo
>
> P.S Perhaps a developer can commit this patch :)
>
>
>   
>>
>>     
>>> Also, I checked the trajectory using gmxcheck and I got this:
>>>       
>>     
>>> Checking file traj.trr
>>> trn version: GMX_trn_file (single precision)
>>> Reading frame       0 time    0.000   
>>> # Atoms  7178
>>> Last frame     150000 time 3000.000   
>>>       
>>
>>     
>>> Item        #frames Timestep (ps)
>>> Step        150001    0.02
>>> Time        150001    0.02
>>> Lambda      150001    0.02
>>> Coords      150001    0.02
>>> Velocities  150001    0.02
>>> Forces           0
>>> Box         150001    0.02
>>>       
>>
>>     
>>> So I believe that velocities are OK.
>>> I've been using version 4.5 and not version 4.5.1, 
>>> Do you believe that's the problem? Or it could be memory problem? 
>>> _______________________________________
>>> Eudes Eterno Fileti
>>> Centro de Ciências Naturais e Humanas
>>> Universidade Federal do ABC  CCNH
>>> Av. dos Estados, 5001
>>> Santo André - SP - Brasil
>>> CEP 09210-971
>>> http://fileti.ufabc.edu.br
>>>       
>>
>>
>>     
>>>     On 10/12/2010 11:12 PM, Eudes Fileti wrote:
>>>     >  Hi everybody, I'm trying to use g_velacc to calculate the diffusion
>>>     > coeficient for my solute.
>>>     >  For this, I performed a simple simulation to test such tool.
>>>     >
>>>     >  1000 water molecules,
>>>     >  NPT ensemble
>>>     >  positions and velocities colected every 0.01ps
>>>     >  Gromacs 4.5.
>>>     >
>>>     >  However, once I run: g_velacc -f traj -s -n index.ndx -o -m -acflen
>>>     > 1000 -nonormalize
>>>     >
>>>     > I get the following error:
>>>     >
>>>     > Group     0 (         System) has  3000 elements
>>>     > Group     1 (            SOL) has  3000 elements
>>>     > Group     2 (             OW) has  1000 elements
>>>     > Group     3 (            HW1) has  1000 elements
>>>     > Group     4 (            HW2) has  1000 elements
>>>     > Select a group: 1
>>>     > Selected 1: 'SOL'
>>>     > trn version: GMX_trn_file (single precision)
>>>     > Reading frame       0 time    0.000   Segmentation fault
>>>       
>>     
>>>     Recently I used g_velacc in gmx 4.0.7 and it worked fine. Now I just
>>>     tested the gmx4.5.1 version for you and it also worked. So it seems that
>>>     there is something wrong with your trajectory. Are you sure that
>>>     velocities were saved, which is not the case for an xtc file. You have
>>>     to explicitly define nstvout in your mdp file, that tells gromacs how
>>>     often the velocities should be written to the trr file.
>>>       
>>     
>>>     /Flo
>>>       
>>     
>>>     >
>>>     > I've tried several things. I changed flag values. I tried the -mol
>>>     > option. I tried converting a trajectory
>>>     > to another format (.trj). I tried using the version 3.3.3 of g_velacc.
>>>     > Nothing worked, always the same error.
>>>     > Could someone give me a hand?
>>>     > Grateful
>>>     > eef
>>>     > _______________________________________
>>>     > Eudes Eterno Fileti
>>>     > Centro de Ciências Naturais e Humanas
>>>     > Universidade Federal do ABC  CCNH
>>>     > Av. dos Estados, 5001
>>>     > Santo André - SP - Brasil
>>>     > CEP 09210-971
>>>     > http://fileti.ufabc.edu.br
>>>     >
>>>       
>>
>>     
>>>     - --
>>>     Florian Dommert
>>>     Dipl.-Phys.
>>>       
>>     
>>>     Institute for Computational Physics
>>>       
>>     
>>>     University Stuttgart
>>>       
>>     
>>>     Pfaffenwaldring 27
>>>     70569 Stuttgart
>>>       
>>     
>>>     Phone: +49(0)711/685-6-3613
>>>     Fax:   +49-(0)711/685-6-3658
>>>       
>>     
>>>     EMail: dommert at icp.uni-stuttgart.de
>>>     <mailto:dommert at icp.uni-stuttgart.de>
>>>     Home: http://www.icp.uni-stuttgart.de/~icp/Florian_Dommert
>>>       
>>
>>
>>
>>
>>     
> - -- 
> Florian Dommert
>   
I committed the fix.

Thanks,

Berk