[gmx-users] question about energygrps

Mon May 13 09:27:00 CEST 2013

Huyjin,

If you look in your log file from the rerun you will see that when running using the GPU, energy groups are ignored or unused for energy reporting. Thus for this to work you need to run it on the CPUs.

This is what I get in my output md.log when trying the same thing...
>> NOTE: With GPUs, reporting energy group contributions is not supported

Best regards,
Jesper

On May 12, 2013, at 11:06 AM, Hyunjin Kim <hyunjink at andrew.cmu.edu> wrote:

>> 
>> 
>> On 5/12/13 1:53 PM, Hyunjin Kim wrote:
>>>> 
>>>> 
>>>> On 5/12/13 1:31 PM, Hyunjin Kim wrote:
>>>>>> 
>>>>>> 
>>>>>> On 5/12/13 1:14 PM, Hyunjin Kim wrote:
>>>>>>>> 
>>>>>>>> 
>>>>>>>> On 5/12/13 12:54 PM, Hyunjin Kim wrote:
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>>> On 5/12/13 2:25 AM, Hyunjin Kim wrote:
>>>>>>>>>>> Dear,
>>>>>>>>>>> 
>>>>>>>>>>> I want to calculate LJ and Electrostatic energies between two
>>>>>>>>>>> groups
>>>>>>>>>>> defined in index.ndx during rerun with original trajectory.
>>>>>>>>>>> 
>>>>>>>>>>> The following is what I tried:
>>>>>>>>>>> 
>>>>>>>>>>> 1. insert "energygrps r_1 r_25" in the test.mdp file.
>>>>>>>>>>> 
>>>>>>>>>>> 2. grompp -f test.mdp -c x.gro -o x.tpr -p x.top -n index.ndx
>>>>>>>>>>> 
>>>>>>>>>>> 3. mdrun -nt 2 -rerun o.trr -s x.tpr -o x.trr -c x.gro -g x.log
>>>>>>>>>>> -e
>>>>>>>>>>> x.ene
>>>>>>>>>>> 
>>>>>>>>>>> The simulation runs properly, but if I collect energy using
>>>>>>>>>>> g_energy,
>>>>>>>>>>> then
>>>>>>>>>>> values of all energy terms are same, although I choose different
>>>>>>>>>>> groups.
>>>>>>>>>> 
>>>>>>>>>> Which specific groups did you choose?
>>>>>>>>> 
>>>>>>>>> Thanks a lot. I chose residues 1 and 25.
>>>>>>>>> 
>>>>>>>> 
>>>>>>>> I understood that from the original post.  What I was asking about
>>>>>>>> was
>>>>>>>> the
>>>>>>>> exact
>>>>>>>> groups you chose from the g_energy selection menu.
>>>>>>>> 
>>>>>>> 
>>>>>>> When I use g_energy, I chose 4 5 6 7 8 9 10 from the below list.
>>>>>>> Thanks.
>>>>>>> 
>>>>>>>     1  Bond             2  Angle            3  Ryckaert-Bell.   4
>>>>>>> LJ-14
>>>>>>>     5  Coulomb-14       6  LJ-(SR)          7  Disper.-corr.    8
>>>>>>> Coulomb-(SR)
>>>>>>>     9  Coul.-recip.    10  Potential       11  Kinetic-En.     12
>>>>>>> Total-Energy
>>>>>> 
>>>>>> Dispersion correct, PME mesh (Coul-recip), and total potential are
>>>>>> not
>>>>>> decomposable.  The other terms you selected should be.  The fact the
>>>>>> these
>>>>>> options do not appear in the .edr file indicate either (1) you did
>>>>>> not
>>>>>> rerun
>>>>>> with the right .tpr file or (2) you analyzed the wrong .edr file.
>>>>>> The
>>>>>> approach
>>>>>> you have suggested is correct and should result in the desired energy
>>>>>> groups
>>>>>> being present.  What version of Gromacs are you using?
>>>>> 
>>>>> I use 4.6.1 with gpu.
>>>>> 
>>>> 
>>>> Are you doing the rerun on GPU, as well, or on CPU?  What happens if
>>>> you
>>>> run in
>>>> serial (-nt 1)?  I don't believe reruns work properly in parallel, but
>>>> maybe
>>>> that's outdated information.  Does a rerun on the .xtc file produce the
>>>> desired
>>>> output?
>>>> 
>>> 
>>> I am about test on rerun with -nt 1 or CPU. But, previously, I only
>>> tested
>>> rerun with -nt 2 and GPU and looked to produce proper output. However,
>>> what you mean "desired output" seems nonbonded interactions between two
>> 
>> The proper output is an .edr file with the decomposed energy terms you
>> want, so
>> I don't understand what "looked to produce proper output" means.  Any
>> other
>> output (trajectory, coordinate, log, etc) should be identical to the
>> previous
>> run, since you're doing a rerun, not a new simulation.
>> 
>> -Justin
> 
> I see. Then no I have not obtained the proper result yet. I will test what
> you suggested.
> Thanks. Hyunjin.
> 
>> 
>>> groups I set, which I have not. I will test rerun in CPU and even -nt 1
>>> and see whether I got results properly.
>>> Thanks a lot for your support.
>>> 
>>> Hyunjin.
>>> 
>>> 
>>>> -Justin
>>>> 
>>>> --
>>>> ========================================
>>>> 
>>>> Justin A. Lemkul, Ph.D.
>>>> Research Scientist
>>>> Department of Biochemistry
>>>> Virginia Tech
>>>> Blacksburg, VA
>>>> jalemkul[at]vt.edu | (540) 231-9080
>>>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>>> 
>>>> ========================================
>>>> --
>>>> gmx-users mailing list    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.
>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>> 
>>> 
>>> 
>> 
>> --
>> ========================================
>> 
>> Justin A. Lemkul, Ph.D.
>> Research Scientist
>> Department of Biochemistry
>> Virginia Tech
>> Blacksburg, VA
>> jalemkul[at]vt.edu | (540) 231-9080
>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>> 
>> ========================================
>> --
>> gmx-users mailing list    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.
>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>> 
> 
> 
> -- 
> gmx-users mailing list    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.
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists