[gmx-users] Different box volumes from gmxdump and g_energy
Johnny Lu
johnny.lu128 at gmail.com
Wed Dec 3 19:14:22 CET 2014
The error in pressure of the nvt simulation (about -1.4 bar) is not
explained by the error in pressure of the npt simulation.
I wonder if the error in pressure of the NVT simulation is caused by that
in the NPT simulation I used a berendsen barostat with a very low coupling
time of 0.2 ps?
I guess the worse case is that I get the NPT frames with volumes around
(but not equal to) the average volume I get from g_energy, and then run NVT
simulations starting from each of these NPT frames. And hopefully I get a
NVT simulation with the correct pressure... I don't know what went wrong. I
did convert the cpt file to pdb file, and then look at them in vmd to check
for vacuum bubbles (and i don't see such bubbles in the water molecules).
Below is my estimation of the errors:
Somehow, the pressure of the nvt14d simulation is about -1.4 bar, instead
of 1 bar. This error in pressure is not explained by the err in volume from
the g_energy of npt13c, and the isothermal compressibility of water.
[= g_energy of npt13c =]
Statistics over 214381681 steps [ 0.0000 through 428763.3600 ps ], 5
data sets
All statistics are over 17108015 points
Energy Average Err.Est. RMSD Tot-Drift
-------------------------------------------------------------------------------
Total Energy -307350 6.1 nan -15.4944
(kJ/mol)
Temperature 299.851 0.0018 nan
-0.00756116 (K)
Pressure 1.01127 5.7e-05 nan
0.000126851 (bar)
Volume 298.385 0.0043 nan 0.00672558
(nm^3)
Density 1015.42 0.015 nan -0.0229017
(kg/m^3)
[= g_energy of nvt14d =]
Last energy frame read 4715311 time 94306.219
Statistics over 22153111 steps [ 50000.0000 through 94306.2200 ps ],
1 data sets
All statistics are over 2215312 points
Energy Average Err.Est. RMSD Tot-Drift
-------------------------------------------------------------------------------
Pressure -1.40405 0.39 146.262 3.11836
(bar)
gcq#98: "Jesus Can't Save You, Though It's Nice to Think He Tried"
(Black Crowes)
[= Approximate Maximal Error in Pressure (95% confidence, or two
standard deviation) =]
Assume the error estimation in averaged volume of npt13c is correct.
4.5e-5 = - (dV/dP)_T / V
(Delta P)= - (Delta V)/ V / 4.5e-5
(Delta P)= - (0.0043*2)/ (298.385 - 0.0043*2) / 4.5e-5 = -0.6 bar
Thus, there is 5% chance that the NPT simulation has a pressure as
low as 1.0 - 0.6 = 0.4 bar.
Yet, the actual pressure of the NVT simulation is -1.40405 +- 0.39,
and 0.4 bar is not within two standard deviations of this value.
On Wed, Dec 3, 2014 at 1:08 PM, Johnny Lu <johnny.lu128 at gmail.com> wrote:
> yes.
>
> I get an xvg file from the edr file with g_energy. Then from the xvg file,
> i selected the observable (volume) corresponding to the timestep at which
> the .cpt file is written.
> Then I compared this value of volume, with the value of volume calculated
> from the box vectors in the .cpt file.
>
> On Wed, Dec 3, 2014 at 1:04 PM, Mark Abraham <mark.j.abraham at gmail.com>
> wrote:
>
>> A .cpt file has one frame in it. An .edr file has values and statistics
>> from many frames. Are you comparing things that are comparable?
>>
>> Mark
>>
>> On Wed, Dec 3, 2014 at 6:56 PM, Johnny Lu <johnny.lu128 at gmail.com> wrote:
>>
>> > g_energy seems more correct than gmxdump in box size.
>> >
>> >
>> > [= npt13c_step169777880. gmxdump =]
>> > box (3x3):
>> > box[ 0]={ 7.50577e+00, 0.00000e+00, 0.00000e+00}
>> > box[ 1]={ 0.00000e+00, 7.50577e+00, 0.00000e+00}
>> > box[ 2]={ 3.75289e+00, 3.75289e+00, 5.30739e+00}
>> >
>> > volume from box vector:
>> > 7.50577 * 7.50577 * 5.30739 = 299.000218803
>> >
>> > [= npt13c. g_energy =]
>> > time(ps) volume
>> > ------------- ----------
>> > 339555.760000 298.385254
>> >
>> > [= commands for starting a NVT simulation, from the npt frame. =]
>> > ../../sofware/gromacs-4.6.7/bin/grompp -f nvt14d.mdp -c
>> npt13c.tpr -t
>> > npt13c_step169777880.cpt -o nvt14d.tpr
>> >
>> > [= gmxdump_nvt14d_step62571050. gmxdump =]
>> > box (3x3):
>> > box[ 0]={ 7.50062e+00, 0.00000e+00, 0.00000e+00}
>> > box[ 1]={ 0.00000e+00, 7.50062e+00, 0.00000e+00}
>> > box[ 2]={ 3.75032e+00, 3.75032e+00, 5.30375e+00}
>> >
>> > volume from box vector:
>> > 7.50062 * 7.50062 * 5.30375 = 298.385264414 (the same value
>> in
>> > g_energy of npt13c)
>> >
>> > [= gmxdump_nvt14d_step6247910. gmxdump =]
>> > box (3x3):
>> > box[ 0]={ 7.50062e+00, 0.00000e+00, 0.00000e+00}
>> > box[ 1]={ 0.00000e+00, 7.50062e+00, 0.00000e+00}
>> > box[ 2]={ 3.75032e+00, 3.75032e+00, 5.30375e+00}
>> >
>> > volume from box vector:
>> > 7.50062 * 7.50062 * 5.30375 = 298.385264414
>> >
>> > On Wed, Dec 3, 2014 at 11:34 AM, Johnny Lu <johnny.lu128 at gmail.com>
>> wrote:
>> >
>> > > Hi.
>> > >
>> > > I get two different volumes with g_energy and gmxdump.
>> > >
>> > > Which one is true?
>> > >
>> > > For now, I'm trying to get a NVT simulation with correct pressure,
>> after
>> > a
>> > > 300 ns NPT equilibration.
>> > > The 400 ns NPT equilibration has the following statistics, according
>> to
>> > > g_energy of gromacs 4.6.7, and the equilibration seems to be
>> sufficient.
>> > >
>> > > Last energy frame read 21438168 time 428763.375
>> > >
>> > > Statistics over 214381681 steps [ 0.0000 through 428763.3600 ps
>> ],
>> > > 5 data sets
>> > > All statistics are over 17108015 points
>> > >
>> > > Energy Average Err.Est. RMSD
>> > Tot-Drift
>> > >
>> > >
>> >
>> -------------------------------------------------------------------------------
>> > > Total Energy -307350 6.1 nan
>> > > -15.4944 (kJ/mol)
>> > > Temperature 299.851 0.0018 nan
>> > > -0.00756116 (K)
>> > > Pressure 1.01127 5.7e-05 nan
>> > > 0.000126851 (bar)
>> > > Volume 298.385 0.0043 nan
>> > > 0.00672558 (nm^3)
>> > > Density 1015.42 0.015 nan
>> > > -0.0229017 (kg/m^3)
>> > >
>> > > gcq#328: "I used to be blond and stupid, but now I dyed it
>> black"
>> > > (Miss Li)
>> > >
>> > > The nan in RMSD is caused by that I did one restart of the
>> > > simulation.
>> > >
>> > > Getting the correct volume is very important for me because the
>> > isothermal
>> > > compressibility of water is 4.5e-5, and a small change in volume can
>> lead
>> > > to a relatively large change in pressure.
>> > >
>> > > gromacs5.0.1. g_energy_d
>> > > time(ps) box-x box-y box-z volume
>> > > ------------- -------- -------- -------- ----------
>> > > 125142.100000 7.500611 7.500611 5.303739 298.383881
>> > >
>> > > gromacs4.6.7. g_energy
>> > >
>> > > time(ps) volume
>> > > ------------- ----------
>> > > 125142.100000 298.383881
>> > >
>> > > gromacs4.6.7. gmxdump
>> > > box (3x3):
>> > > box[ 0]={ 7.50577e+00, 0.00000e+00, 0.00000e+00}
>> > > box[ 1]={ 0.00000e+00, 7.50577e+00, 0.00000e+00}
>> > > box[ 2]={ 3.75289e+00, 3.75289e+00, 5.30739e+00}
>> > >
>> > > from the box vectors, volume is |z dot (x cross y)|
>> > > 7.50577 * 7.50577 * 5.30739 = 299.000218803
>> > >
>> > > gromacs5.0.1. gmxdump_d
>> > > box (3x3):
>> > > box[ 0]={ 7.50577e+00, 0.00000e+00, 0.00000e+00}
>> > > box[ 1]={ 0.00000e+00, 7.50577e+00, 0.00000e+00}
>> > > box[ 2]={ 3.75289e+00, 3.75289e+00, 5.30739e+00}
>> > >
>> > > example command for running gmx_dump:
>> > > gmxdump_d -s ../npt13c.tpr -f ../npt13c_step62571050.cpt &>
>> > > gmxdump5.0.1d_step62571050
>> > >
>> > > example command for running g_energy:
>> > > g_energy_d -f ../npt13c.edr npt13c_volume_gromacs5.0.1_d.xvg
>> > >
>> > > since the timestep is 0.002 ps, the time(ps) corresponding to
>> > > the check point file is:
>> > >
>> > > 62571050 * 0.002 = 125142.1 ps
>> > >
>> > > All the files (tpr, cpt, edr) are generated by mdrun of gromacs4.6.7
>> > > (mixed precision). It seems that it is ok to use the g_energy and
>> gmxdump
>> > > of gromacs 5.0.1 on these files.
>> > >
>> > > Size of the .edr file is 5 gb.
>> > >
>> > > Thank you again.
>> > >
>> > --
>> > Gromacs Users mailing list
>> >
>> > * Please search the archive at
>> > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>> > posting!
>> >
>> > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>> >
>> > * For (un)subscribe requests visit
>> > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>> > send a mail to gmx-users-request at gromacs.org.
>> >
>> --
>> Gromacs Users mailing list
>>
>> * Please search the archive at
>> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>> posting!
>>
>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>> * For (un)subscribe requests visit
>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>> send a mail to gmx-users-request at gromacs.org.
>>
>
>
More information about the gromacs.org_gmx-users
mailing list